Chapter 1 Principles and Design

Despite the varying approaches to undertaking research, handbooks on research methods continually
stress what is often called the research process. This chapter will discuss the six main building block of
the research process, the dissemination of research, writing up research, academic writing such as
dissertations, academic papers, and conferences.
This chapter considers the role and purpose of research. It draws on debates that have been taking
place in the social sciences over the last thirty years. It provides a background to understanding:
 why the research is conducted; and
 the distinctiveness of different types of research.

The aim is to encourage you to think beyond the basic questions of how you will carry out your research
and collect your data, and to consider why you are doing it and what the impact of your findings might
be. You will quickly realise that there are different ways of conducting research and consequently much
debate across and within disciplines about research and the development of knowledge.

Research and the Development of Knowledge

The concept of research as ‘adding to the knowledge’ of a particular area is a popular idea. The
production and development of knowledge are certainly the reasons that are generally given for
undertaking research; but it is clear that much research does not constitute an unproblematic progress
towards ever greater understanding. False trails are followed; progress may be patchy; and some
research is, of course, downright bad and an inadequate basis from which to draw conclusions. Another
common reason for doing research is to ‘predict’, ‘explain’, ‘describe’ or ‘prove’ something. These
concepts derive from the definition of social research as analogous to natural science.
Positivists believe that social research should be as objective as possible, operating in a similar
manner to natural science. In general, positivists advocate the use of the collection of quantifiable data.
They emphasise reliability and representativeness in research and often use statistical tests in order to
back up their conclusions. Such an approach does not, however, allow for the often unexpected
discoveries or new questions that arise in the course of undertaking research; and it is often viewed as
imposing the methods of natural science on the study of society. Critics of the positivist approach argue
that society is far more complicated than the controlled world of the laboratory where the classical
scientific research method originated. Furthermore, results from experiments are seldom so certain that
they predict or explain with any great certitude (witness the numerous controversies, not to mention
vested interests, in drug trials).
Researchers in the interpretative tradition emphasise the researcher as an active player in the
research. They are not detached ‘scientists’ observing in an objective, clinical manner. Researchers
are seen to:
‘have their own understandings, their own convictions, their own conceptual orientations; they too are
members of a particular culture at a specific historical moment.’
(Miles & Huberman, 1994, p. 8)
Interpretivists believe that social researchers should attempt to investigate how people themselves
make sense of social reality. They prefer in-depth interviews and less structured observation.
Qualitative data is seen to provide a rich source of descriptive and explanatory detail, which potentially
results in unexpected findings, generating fresh concepts and frameworks. Validity is emphasised.
Research based on qualitative data has been criticised for its potential subjectivity and thus (it can be
argued) the potentially biased nature of the findings, which are influenced heavily by the researcher’s
own interpretations of the research.
Hover over the words below for a reminder of these concepts
Activity 1: How objective can we really be?
Consider the following quote:
What we regard as obvious is much too dependent on and relative to our education, our prejudices and
our culture to be a reliable guide to what is reasonable. To many cultures, at various stages in history, it
was obvious that the earth was flat.
(Chalmers, 1982, p. 20)

1
Consider what you already ‘know’ about child trafficking and how this might influence how you went
about researching this issue.f
As the quote above reminds us ‘the facts’ seldom speak for themselves. ‘Facts’ are socially constructed
and inevitably reflect the predispositions and beliefs of the researcher who, after all, will have decided
how the facts will be collected and, even, what constitutes a fact.
This debate between the two key approaches suggests a polar divide and lack of compromise that in
reality does not exist. In real world research, although researchers may have a leaning towards a
particular tradition, it is possible to generate a design drawing on concepts from a number of areas.
Much contemporary social research benefits from both traditions and is all the richer for using a mixed
methods design.

Induction Versus Deduction

While some research will simply aim to describe an observable situation, most research exists in some
sort of relationship to theory. The theory may be peculiar to the researcher or it may be a tried and
tested canon of social science. For example, a research project might seek to investigate the
relationship between a social science theory such as youth alienation and a criminological phenomenon
such as ‘joy riding’.
Testing a theory is a frequent starting point to a research project. Theories are also developed from
research; and new perspectives on how phenomena operate can also originate from research findings.
These two portrayals of the relationship between theory and research introduce the concepts of
induction and deduction.
In induction, research is conducted and theories are developed to explain the results of that research.
Research precedes theory and the ‘data’ that stems from the research assumes a critical place in the
derivation of theory.
In deduction the reverse is the case: theory precedes research. A set of probable relationships within
research data is inferred from theory. These relationships are expressed as hypotheses and tested
using research methods. Conclusions concerning the hypotheses can then be related back to the initial
theory. The deductive approach is the basis of the standard scientific method.

In the example of ‘joy riding’ given above, the process of testing the theory would be a deductive
process. (You might like to consider how you might approach ‘joy riding’ inductively.)
Induction attempts to build understanding on the basis of observation. It centres on a belief that prior
theorising will inevitably cloud the interpretation of research data and render any conclusion liable to a
certain amount of bias. This is a powerful argument and has informed a number of more and less
sophisticated approaches to research. Essentially, it is an argument that seeks to ‘let the facts speak
for themselves’. Thus adherents of grounded theory would collect information on a topic and interpret
that information at a later date and in the light of the opinions of those from whom the data was collected
and the indications of the data itself.
There are well established criticisms of both key approaches to research. With regard to deduction, it is
difficult to conceive of a research exercise which does not, in some way, have a prior base in theory and
is unbiased by the unwittingly subjective standpoint of the researcher. The notion of research
proceeding unclouded by preconceptions is simply not tenable. On the other hand, the inductive
approach, in generating theory from research is dependent upon the quality of that research. If the
research is poorly constructed and insubstantial it may soon be disproved by subsequent research.

2
It is perhaps inappropriate to see induction and deduction as completely separate and distinct
approaches to understanding the relationship between theory and research. They can be related.
Induction can be used to draw out key points from pre-existing research data or highlight patterns in
data in a descriptive sense. These actions can then be related to theory or used to derive a theory that
is then, deductively, tested.
Activity 2: Induction and deduction
1. Identify a topic which you might like to research.
2. Briefly outline the key premises on which you would proceed to investigate topic using:

2.1. an inductive approach; and

2.2. a deductive approach.
1. Which approach seems to be the most appropriate in your chosen area, and why?

Scientific Method

Deduction provides the foundation of the now standard scientific research method. The central figure in
this field is Karl Popper. He argues that scientists should express their theories in a form that can be
tested by reference to research data (Popper, 1972). The important notion is that of the test. Not only
should theories be tested using concrete evidence, they should also be capable of being disproved by
that test.
The method by which such testing takes place has been termed the hypothetico-deductive method. It
proceeds through the assumption that a researcher will have a theory that she or he wishes to test
against the ‘reality’ of observation. The theory will need to be framed as a series of operational law-like
statements or propositions. These are known as hypotheses and summarise key elements of the
theory in a testable form. For example, whether there are specific gender-based factors that could
result in women pleading guilty to crimes they have not committed (Jones, 2011). The hypothesis can
be tested by Interviews of women who have been tried, surveys of plea bargain cases and
questionnaires of prisoners.
Activity 3: Hypotheses
1. Using your topic from the previous activity, write a hypothesis in as concrete and testable a
form as you can.
2. Identify how you might collect data to test this hypothesis.

Clearly a good and valid test should be capable of going one way or another: it could prove or disprove
a hypothesis. Popper (1972) argues that good science should be rigorous and focus on falsification.
This means that every effort should be made to destroy a hypothesis. Conversely, a great degree of
certainty should be required for a hypothesis to be accepted and research said to have demonstrated
something. The principle of falsification is particularly evident in the way that inferential statistics work.
As you will see in Chapter 6 of this unit, hypotheses are crucial to statistical tests.
Of course, accepting a research hypothesis with one particular set of data, in one particular setting, at
one particular time, using one particular approach and perhaps one particular significance level, does
not mean that under other conditions a different verdict might not be reached. In the case of rapidly
changing and complex social phenomena, certainty can never be reached. Scepticism and doubt are
seen as central to the development of research knowledge. For a useful account of theory and its
relationship to research method using criminology-based examples, see Hagan (1993, pp. 1–25).

Paradigms and Revolutions

According to Popper (1972) research knowledge is never conclusive. Research can only ever be
accepted as conditional confirmation of a theory. Successive verifications of a theory may lead to a
fairly conclusive belief, but it can never be total. Subsequent research may modify and develop this
knowledge. This produces a linear progressive evolution of understanding.

3
 A rather different perspective has been outlined
by Thomas Kuhn. He suggests that Popper’s notion of knowledge as a process of progressive
unfolding is inaccurate. Instead, he suggests that, through time, particular forms of knowledge become
regarded as conclusive, institutionalised and accepted as ‘truth’. Research evidence that contradicts
this dominant paradigm is dismissed as improperly conducted or regarded as a temporary problem for
further study (Kuhn, 1970). The Popperian perspective is thus seen as inherently conservative: change
is slow and gradual, and inextricably linked to an existing paradigm.
Real progress occurs when a paradigm shift takes place. This may reflect a crisis in existing
knowledge: when falsified hypotheses can no longer be explained away. It will also require new critical
researchers to enter the scene, using methods and developing theories radically different from those
that went before. In time such a revolution will settle down and a new dominant paradigm will be
established.
In a practical sense Kuhn’s work provides a useful reminder that the ordered Popperian hypothetico-
deductive framework is perhaps a little too rigid for the analysis of something as complex as society.
Kuhn (1970) also offers a timely reminder that knowledge changes and shifts quite markedly as time
goes by. For example, the differing perspectives of Newtonian physics and Einsteinian relativity theory
constitute major paradigm shifts in physics. On a less spectacular scale, similar shifts have occurred in
criminology, as differing perspectives have competed with each other. For example, changes in
emphasis from the rights of the victim to the rights of the criminal, or protection of the public versus the
inadequacy of prison system, have some impact on the way in which the criminal justice system is
perceived. The dominant paradigm will influence media coverage and public attitude and this may be
reflected in the kinds of policies that are implemented by the government, resulting in, for example,
tougher sentences for rapists, lighter sentences for first time burglary and so on.
Activity 4: Criminology and research paradigms
Read the following research summary:
HORS 269 Distraction burglary amongst older adults and ethnic minority communities (Home Office,
2003)
http://webarchive.nationalarchives.gov.uk/20110218135832/rds.homeoffice.gov.uk/rds/pdfs2/
hors269.pdf
1. What was the purpose of this study?
2. What type of paradigm is dominant in this study?

3. What kinds of research evidence do you think are most respected by government?

4. What does this tell us about the dominant paradigm in government circles?

Since Popperian and Kuhnian ideas were first promulgated, they have generated substantial literature.
Works by both writers have brought their two perspectives much closer and a general theory of how
research operates to develop knowledge would now include elements of both approaches. Lakatos
(Lakatos & Musgrave, 1970), for example, argues that at any one time there is a core of indisputable
knowledge. Theory develops to explain and understand this core knowledge; but it is always subject to
falsification. Proof of a theory through research activity can contribute to the enrichment and
enhancement of core knowledge. Falsification, on the other hand, can lead to a questioning of core
assumptions and, eventually, to a paradigm shift (Lakatos & Musgrave, 1970).

The Researcher and the Researched

A final topic to consider under the heading of research and the development of knowledge is the
relationship between the researcher and the researched. Within the field of research methods, this
relationship is an area in which a major shift of emphasis is currently taking place – perhaps even a
paradigm shift.
The traditional scientific research method sees research as an independent value-free process of
inquiry. Research should be objective and analysis should be detached. Care should be taken to avoid
4
letting emotions, subjective interpretations and personal viewpoints intrude upon the interpretation of
research findings. These expectations concerning the role of the researcher are extended to the
relations between the researcher and the researched. The researcher is an expert who brings skills to
bear on the object, but has a separate and detached exercise from that object of study. Any
understanding of a particular situation will result from the researcher’s interpretations of research
material and from the researcher’s approach to collecting data about the object of study.
New paradigm approaches to research, which may be taken to include interpretative research, reject
this approach. They argue that it stems from a scientific model: the detached, white-coated (male)
scientist, measuring and collating information about observable but inarticulate phenomena in laboratory
settings. They suggest that in social research this model is inappropriate. Human communication
opens up a much greater possibility for understanding than can ever be open to a researcher working
on problems in the areas of, for example, chemistry or physics. They also reveal a much richer
complexity, a complexity that is perhaps not always best tackled through the traditional approach of the
scientific method.
The interpretative perspective on research rejects the stress on measurement and testing, preferring
to seek understanding. It is argued that the search for reason and regularity neglects the emotional and
serendipitous nature of many social events. For example, many thefts may be highly individualised
responses to the presentation of an opportunity to thieve. Examining data on 1000 convicted thieves
may contribute to testing a plausible theory, but may be limited by the questions asked and the data
available on these cases. Conversely, in-depth interviews with ten convicted thieves may provide more
understanding, provided that the interviewer effects and biases are kept to a minimum.
More usually and more importantly, however, interpretative approaches also seek a different
relationship between the researcher and the researched. It is argued that the artificial separation of
researcher from the object of the research detaches the reason associated with the objective tradition
from the mutual understanding, which flows from a closer relationship with the researched. Instead of
seeing people who are being researched as objects or sources of data they become partners in the
research, able to influence and frame the direction of the research, give their own interpretations and
own the results. The researcher, instead of being a rational, detached analyst, becomes an involved
participant with acknowledged interests and stakes in the topic of the research. The relationship
between the researcher and the researched undoubtedly raises some interesting questions for the
philosophy of the social sciences and the sociology of research.

Research Design
Research design refers to the way the collection of research data is organised. The way in which the
researcher approaches this topic is crucial to the success or failure of a research project. Data will need
to be collected in a way that reflects the demands of the overall study. The way in which the data is
collected will also determine the techniques that are used for analysis.
We are going to consider the three key ideal-typical research designs:
1. experimental;
2. case study; and

3. survey

In addition, we will discuss longitudinal research, which often involves surveys but can be conducted in
other research designs.
Remember, when you are reading this section, that mixed method designs (using methods that collect
different types of data) and hybrid designs (for example, survey followed by case studies) are actually
very common in social research. Such designs can be very useful ways of collecting data that will
provide a rich, rounded and comprehensive picture of the research subject.
A key argument for the use of more than one method and/or data from different sources is triangulation.
The justification for triangulation lies in the increased validity of such a technique. If more than one

5
method or source of data lead to the same conclusions, then it is safe to assume that the results are not
due to the particular method or source used.

The following extract argues the case for the multi-method approach to data collection, in
research investigating the interpretation and implementation of the Crime and Disorder Act
1998.

Crime and Disorder Act 1998

Methods of data collection

As will be explained in greater detail below, each of the three stages of research were undertaken in
different environments and therefore required different approaches to data identification and collection.
In order to achieve this, a ‘multiple method’ model of research was adopted, described by Robson
(1993, p. 290) as the ‘Complementary Purpose Model.’ He believed that:
you need not be a prisoner of a particular method or technique when carrying out an enquiry
because it could restrict the full potential of research. Particularly if the focus relied too much on
process of the research and not the outcome the research hoped to achieve.
The use of multiple methods for the research was considered to be advantageous for a number of
reasons. It allowed a cross-checking, or triangulation, of information to ensure validity (Bell, 1999, p.
102) by combining both quantitative and qualitative data. Furthermore the author considered that
through a combination of research methods he was better able to place ‘factual’ measurable information
within an organisational context and give real meaning to the information captured. Newman (2000, p.
124) supports the use of a combination of methods as he believes it can enable a researcher to:
look at something from several angles [rather] than look at it only one way.
However, it is recognised that not all social scientists are convinced of this argument, Payne (1997, p.
108) for example, accuses multi-method research of fudging the issues by forcing together non-
compatible data in order to support nebulous arguments.
Whilst the debate over the validity of single versus multi-method research will continue, the author
chose to use multiple method research, arguing that the adoption of a particular research method is
more often dictated by pragmatic circumstances. The author discovered that his research took him into
different working environments and that he had to adapt to the circumstances presented and use the
most appropriate method which gained access to the information sought. Accessibility to the research
subject is highly influential on the final selection of an appropriate method, which may not necessarily be
the preferred choice.
For example Taylor and Brogdan (1998, p. 88) suggest in-depth interviewing may be used instead of
participant observation, if timescales are against the researcher, assuming of course that access is not
problematic. Apart from the physical access to a subject other restrictions may also apply, for example
data may be stored in paper form or within a computer which will dictate the form of research, which, in
this case, must include some form of documentary analysis. Whilst that held within a person’s memory
is likely to be accessed through a formal or informal interview or by way of a prepared questionnaire.
These are just some of the issues that might dictate the method of research.
Having taken into consideration the above factors, the author adopted three methods of data collection
throughout his research. The research methods used throughout this project included that of
documentary analysis, participant observation and informal interviews of key participants. Each method
being used as appropriate during all stages of the research.

Extract 1. Triangulation (Fletcher, 2006)

Experiments
The experiment is a research design most strongly associated with the traditional scientific model of
research. Data may be collected using a variety of methods including questionnaires, interviews and
observations. The aim of the experiment is to identify relationships between variables, either causal or
correlative, in an attempt to identify patterns in the data that can lead to generalisations from the sample
to a whole population. We have discussed some of the criticisms of this tradition and it should be
stressed again that the major criticism is that this perspective is in danger of overlooking the wider
cultural and social circumstances within which it is embedded.

6
Experimental designs are used when a researcher has a degree of control over the issues that are
being researched. In particular, the researcher may control or manipulate the independent variable.
Although this situation is not common in social research, the experimental design (or more accurately
‘quasi-experimental design’) still occupies a prominent place, particularly in psychological studies and in
evaluative research where the researcher is seeking to gauge the impact of some innovation or
development.
The classic experimental design is known as the case-control design. The researcher studies two
groups: the case group who are exposed to the independent variable and the control group who are
not. As an example, consider a group of young offenders. The group is separated into a case group
and a control group, and both are asked about their attitudes to law-breaking (the pre-test). The case
group then attends a day seminar (the intervention) designed to change their attitudes; the control group
does not. Finally, the two groups are again questioned on their attitudes to law-breaking (the post-test)
and comparisons are made. Possibly the day seminar will be revealed to have made a significant
difference to attitudes.
A number of issues within such an experiment need further examination. Firstly it can only be certain
that it was the day seminar that brought about the change if the two groups are otherwise alike in every
aspect. If they are not, there is the possibility that some other extraneous factor brought about the
change. Controls are used in experimental designs to help counter the latter possibility. One such
control is: matching. This means the characteristics of a person in the case group must be exactly
matched by those of a person in the control group. If there is a nineteen-year-old White woman in one
group, there must also be in the other group. Clearly matching becomes more problematic the more
characteristics that are taken into account. Area, gender and ethnic group may be possible, but are
housing, employment and so on equally important? The alternative to the complexities of matching is
randomisation. With this approach individuals are randomly allocated between the two groups. When
this is done from a suitably large group, people with particular characteristics have an equal probability
of being allocated to either group. With either tactic, the similarity of the two groups can be verified by
comparing the pre-intervention assessments. Vast differences will mean that they are already dissimilar
before the intervention occurs. Randomised Controlled Trials (RCTs) are seen as the strongest design
within experimental designs. This sort of approach is usual in drug trials.
A second important issue arises from the ‘Hawthorne effect’ – the phenomenon whereby people in the
case group respond in a particular way simply because they are in the case group and know that
something is being done to them to which they are expected to respond. This may be counteracted by
‘blinding’: by concealing group membership and the purpose of the experiment from participants.
Equally, the researcher may subconsciously work particularly hard with the case group in order to prove
the effectiveness of a programme of intervention. This too can be avoided by excluding the researcher
from the process of administering the independent variable and limiting his/her role to the interpretation
of results: a process known as ‘double-blinding’.
Both ‘blinding’ and the process of intervention in experiments raise ethical issues. The first issue
concerns the lack of information when the ‘blinding’ process is applied. The participants may have to be
deceived as to the true nature of the research and this contravenes recommended ethical practice. The
second issue involves the withholding of a potential benefit from the control group. If there is a benefit,
what are the ethical implications in terms of the control group? Can you justify your actions? The
researcher will need to confront these dilemmas and develop an appropriate argument to defend their
response.
Activity 5: Ethics and control groups in experimental research
1. Consider a programme aimed at treating sex offenders, but with no prior information about its
effects. In order to evaluate the effectiveness of such a programme, a control group will be used to
compare changes in behaviour. The control group receives no form of treatment during this time.
2. What are the ethical issues raised by the withholding of treatment?

3. Reconsider this example after reading the material on research ethics in Chapter 3 of
this unit.

Decision-making in a Murder Case

This experiment aimed to identify the factors that might affect the decision-making process in a murder
case. A fictional murder was presented in a vignette giving unbiased details of the case. (See Social
Research Update, Issue 25, for more information on the use of vignettes in research
http://sru.soc.surrey.ac.uk/.) Participants were then either given the prosecution scenario or the defence

7
scenario. It was anticipated that the participants would make a shift of judgement depending upon the
type of scenario shown after the murder vignette.
The experiment
This experiment explored how participants were affected by one of two possible additional scenarios
presented subsequent to the basic case vignette. The suspect’s defence lawyer presented the
exculpating scenario, while the prosecution presented the incriminating scenario. Neither scenario
contained any new factual information.
Participants who, following the basic case, were presented with the incriminating scenario were
expected to make a shift in their judgement in the direction of the suspect’s guilt, while those presented
with the exculpating scenario were expected to alter their judgement in the direction of not guilty.
The elements that participants spontaneously selected as contributing to their judgement (see Method,
below) were categorised as incriminating if they exclusively fitted the incriminating scenario, exculpating
if they fitted only the exculpating scenario, and ambiguous if, depending on the interpretation given to
them, they fitted both scenarios. It was expected that elements that implied the suspect’s guilt would be
selected more often by participants who thought the suspect was guilty and, vice versa, elements that
implied the suspect’s innocence would be more often selected by participants who thought she was
innocent.
With respect to ambiguous elements, the main hypothesis to be tested is that when selected
participants who judge that the suspect was guilty would regard them as evidence of the suspect’s guilt.
They are expected to be crucial with regard to changes on judgement following the biased statement by
either prosecution or defence. Participants who initially select ambiguous elements as fitting one
scenario and then receive a statement that supports an opposing scenario are expected to alter their
judgement in the expected direction. They will subsequently interpret the ambiguous elements as
evidence supporting their new judgement.
Design
The experiment involved two experimental groups, of 101 and 99 participants respectively, to which
participants were randomly assigned. Participants in both groups were first presented with the basic
murder vignette. This was then followed by either the prosecutor statement (group 1) or the defence
statement (group 2).
Procedure
All instructions and experimental materials were presented to the participants in a printed booklet.
Following initial instructions, the murder case vignette was presented. On the next page judgements
related to the perceived guilt of the suspect were requested (initial guilt judgement) as well as answers
to questions exploring the reasoning underlying the judgement. The next page then presented either
the prosecution or defence statement. On the final page, guilt judgements (second guilt judgement) as
well as supporting reasons were again asked for.
The guilt judgement required indicating on a 9-point scale how guilty the participants rated the suspect
to be. Participants were instructed not to assume the stance of a judge who has to pronounce the
verdict, but merely express their judgement of whether they felt the suspect had actually committed the
crime. The question relating to reasons for the judgement was an open one and it attempted to explore
which elements of the case were most salient in contributing to the judgement. Participants were
allowed to list a maximum of three elements. Participants worked through the booklet individually, at
their own pace. They were asked not to turn back pages.
Basic vignette
A woman is suspect in a murder case.
She is apprehended in the apartment of the victim.
The victim was a woman.
The victim was an acquaintance of the suspect.
The victim was a psychiatrist.
The victim had recently received several threatening phone calls.
This was known to the police.
This caller was an anonymous man.

8
The caller always threatened to kill her with a knife.
The victim had told the police she thought this had to be a patient or a former patient.
The crime was committed during broad daylight.
The victim’s assistant returned from her lunch break to find the door locked from the inside.
The assistant became worried.
The assistant decided to call the police.
Two officers arrived in a few minutes.
They also found the door locked from inside.
After a few minutes of bouncing on the door and ringing the doorbell the suspect opened the door.
The suspect had cuts and bruises on hands and face.
The suspect also had bloodstains on her clothing.
She was in a state of shock.
The police found the body of the victim in the apartment.
The police found a knife next to the body.
The suspect identified the weapon as the murder weapon.
The only fingerprints on the weapon were the suspect’s fingerprints.
The suspect denies all charges.
She claims that an unknown intruder attacked the victim with a knife.
She claims that she tried to prevent the crime, and that she took the knife from the attacker.
She also claims she got hurt in doing this.
She claims that the intruder beat her to the ground and fled through the front door.
The suspect also claims that the attacker wore gloves.
A neighbour told police that he saw a man running in the street where the victim lived.
This neighbour said that this man did not wear gloves.
The police think that the suspect is not telling the truth.
The police think that the suspect killed the victim out of jealousy.
The police claim that the suspect thought that the victim was engaged in a sexual relationship with the
suspect’s partner.
Legal actor statements
Prosecution statement
The most likely scenario must be the following:
The suspect assumes the existence of a sexual relationship between her partner and the victim.
The assistant has confirmed this.
This makes the suspect very jealous.
The suspect therefore decides to confront the victim, and demands that she break up this relationship.
In order to strengthen her arguments, she brings a knife.
Suspect and victim know each other, so when the suspect rings the bell, the psychiatrist lets her in.
The meeting does not go as planned.
The argument is not resolved, and turns into a physical fight.
The suspect pulls a knife and stabs the victim several times.

9
The victim dies, the suspect herself is wounded.
To prevent detection the suspect locks the front door.
This buys her time to find another way out of the apartment, because the assistant now cannot come in.
A little later the assistant arrives to find the door locked.
She calls the police.
The police arrive within a short time.
After a few minutes of knocking on the door and ringing the bell, the woman composes herself, decides
what story to tell the police, and answers the door.
Defence statement
The most likely scenario must be the following:
The killer is a patient or an ex-patient of the victim.
This man wants to kill the psychiatrist, as evidenced by threatening phone calls.
On the day of the crime he rings the bell of the apartment.
The psychiatrist opens the door.
Then she sees the man holding the knife, she tries to close the door again, but fails.
She flees into the apartment, followed by the killer.
In the living room he gets hold of her.
He throws her to the floor, and stabs her several times.
At this moment, the suspect enters the apartment through the open front door.
She witnesses the fight, and gets involved in it.
She manages to grab the knife out of the killer’s hands, but it is too late to rescue the victim.
The suspect is wounded in this struggle as well.
The culprit pushes her down to the floor, and he escapes through the front door.
The suspect locks the front door from the inside, fearful of the return of the culprit.
After all, she is the only witness to the crime.
Understandably, she is very upset and in a state of shock.
A little later the assistant arrives to find the door locked.
She calls the police.
The police arrive within a short time.
After a few minutes of knocking on the door and ringing the bell, the woman composes herself and
answers the door.

Extract 2. Experiment (Pakes, 2001)

Case Study
Case study can mean a number of things: individuals, groups, communities, organisations and
institutions, particular events and so on (Robson, 1993). Robson (1993, p. 146) defines a case study
as:
a strategy for doing research which involves an empirical investigation of a particular contemporary
phenomenon within its real life context using multiple sources of evidence.

10
Case studies therefore offer a chance to achieve a more in-depth understanding of an issue or topic. As
with other types of design, case studies may involve the collection of a variety of data, utilising different
data collection methods. Some case study designs place less emphasis on this feature and opt,
instead, for a fuller more descriptive account of the research area. The case study offers an opportunity
to focus on one single case or a small group of similar cases and explore the characteristics of the
individual, thus generating idiosyncratic explanations for the observed phenomena.
Case study designs do not seek to be representative in the same way that experimental and survey
designs do. Instead, they are concerned with in-depth understanding. The selection of the case study
may be for a number of reasons: opportunity and familiarity often play a part, as do contacts and access
invitations. There must, however, also be a rational explanation for the choice of case study; there must
be a theoretical justification for the choice. This might involve the suggestion that the case study is in
some way typical, or might equally entail a reference to the case study being particularly unusual:
difference or similarity are the most frequent rationales for the selection of a case study. They provide a
more in-depth understanding of the particular case and can therefore be of use to the practitioner. They
have the advantage of close contact with ‘reality’ and can be used to present a naturalistic picture of the
research issues. However, they do rely heavily on the interpretation of the researcher and, as
suggested earlier, are often considered to be a poor basis for generalisation.
It is often advised that the case study should be used only for in-depth studies or pilot investigations.
Although many research texts still argue that you should not attempt generalisation from case study
research, limited generalisation is possible if knowledge accumulates in a particular area and similar
studies can be compared. A key strength of the case study is in generating detailed and specific
information about a particular phenomenon.
Many research studies using a case study design lack a formal hypothesis and simply focus on the
description of a particular situation. This need not be a shortcoming; much good exploratory work
operates at the simple descriptive level. Descriptive research aims to develop an overall picture or
understanding of a research topic through a systematic collection of information about the topic. A
typical approach would be to assemble a data set summarising key characteristics of a phenomenon.
For example, a researcher might collect data on the geographical distribution of robberies in a number
of neighbourhoods.
Such a single-group descriptive study, so-called because it is based on one area, may raise interesting
issues, but it is unlikely to have any great impact unless its results can be compared with somewhere
else. Two--group comparative descriptive studies might enable differences between urban and rural
divisions to be identified. They could also be used to assess the departure of one area from a national
norm. Multi-group comparative designs could also be developed.
Activity 6: Conducting research using a case study
You have been asked to conduct research on the assistance to repeat offenders and their families.
You decide to use a case study design.
1. Explain why you would use a case study approach.
2. What data would you collect and why?

11
Surveys
In its simplest form, the survey involves collecting the same standardized data from an undifferentiated
group of respondents over a short period of time. The respondents are almost always selected as a
representative sample from some larger population.

(Robson, 1993, p. 130)

The most common type of survey design is cross sectional, in which data is collected at one point in
time (De Vaus, 1991, p. 40). This is clearly useful for comparing different subsections of society and
individuals, but is less valuable in identifying trends over time. Surveys typically collect data in
standardised form from each individual, using a questionnaire or structured interview schedule.

Surveys, in general, are a useful method of generating a large amount of data about the attitudes,
behaviour and experiences of a sample of people, using questionnaires and interviews. Government
departments regularly carry out large-scale surveys. For example, the British Crime Survey (BCS) has
been carried out annually since 2000/01. This survey is based on interviews with a representative
sample of 40 000 people over the age of 16. It generates an account of crime rates, and attitudes
towards crime and the criminal justice system.

Another crime survey is the BRC Retail Survey, which examines crimes against businesses. It was
completed by 54 retailers, employing 1,304,931 staff and between them representing 48 per cent of total
retail turnover (For more information see http://www.brc.org.uk/brc_show_document.asp?
id=4189&moid=7233).

Surveys can be useful in detecting patterns of crime across areas and so on, but it should be noted that
the figures presented are likely to underestimate the actual volume of crime. This may be due to some
crimes not being reported, under-recording of some crimes at police discretion (complaint may be
subsequently withdrawn, incident judged trivial, several offences count as one, and so on), and the
exclusion of some offence categories by the surveys themselves (Blackburn, 1993). Crime that is not
revealed by these surveys is known as the ‘dark figure’.

Despite the acknowledged shortcomings of crime surveys, they can still be useful sources of crime-
related information. You may, for example, use published surveys as a starting point for an issue you
want to explore in more depth.

Longitudinal Research
Longitudinal research involves the collection of data over time. It is essential if the purpose of the
research is to measure some type of social change. Longitudinal research is a broad term and can
involve the collection of data over time in any of the above designs. Longitudinal research may be
either prospective or retrospective.
An obvious criticism of this design is that by its very nature it will be time consuming (often spanning
decades) and can be expensive with little short-term reward. A key problem is attrition of subjects –
people dropping out of a study over time. Longitudinal data sets do offer an effective tool for
investigating continuity and change over a period of time within subjects, and for identifying any
potential causal factors for these continuities or changes. Longitudinal surveys collect information about
individuals over a number of years, identifying patterns over time and providing a useful profile of the
lives of those surveyed. A good example of a longitudinal study is the Cambridge Study in Delinquent
Development, which is a prospective longitudinal survey of about 400 London males. They were first
studied at the age of eight years and have been interviewed nine times up to the age of 46 (see
Farrington, 2002).

Other Strategies – Hybrid and Combined Designs

Robson (1993, p. 41) reminds us that the traditional research designs outlined above ‘do not provide a
logical partitioning covering all possible forms of enquiry’. Robson likens these designs to ideal types or
established ‘camps’ that signal a preference for a way of working. He warns, however, that researchers
can get used to conducting particular types of study and assume that every research problem requires
the same design. Robson suggests that hybrid and combined designs may be more appropriate for
some studies.

12
Hybrid designs may include aspects of the ideal type of designs, for example survey instruments used
to collect before and after data in an experiment. Here two ideal-typical designs are intertwined.
Combined designs may use more than one ideal-typical design, perhaps in sequence - for example a
survey followed by a case study or studies. The case studies may investigate further the findings from
the survey.
Activity 7: Home Office studies – research designs
Visit the Home Office website and choose three studies from the Police Research Series Papers.
http://rds.homeoffice.gov.uk/rds/policerspubs1.html
Construct a table with the headings below:
 Title (of the study)
 Design (for example, experiment, longitudinal survey, case study)

 Methods (for example, structured interview, participant observation) and

 Purpose (try to identify the main purpose of the research.)

Chapter 2 The Research Process

This chapter looks into the role and purpose of research. It draws on debates that have been taking
place in the social sciences over the last thirty years. It provides a background to understanding why the
research is conducted and the distinctiveness of different types of research.
The aim is to encourage you to think beyond the basic questions of how you will carry out your research
and collect your data, and to consider why you are doing it and what the impact of your findings might
be.
The chapter focuses on the development of knowledge, induction versus deduction reasoning, the
scientific method, the researcher and the researched, typical research design, such as, experimental,
case study, survey, and other strategies.
Despite the varying approaches to undertaking research, handbooks on research methods continually
stress what is often called the research process. The reality is that there are several research
processes, and the actual process of research will vary in some way for each and every research
project. Nevertheless most researchers find the research process outlined here to be useful. It provides
a ready framework for the development and management of a research study as well as a useful set of
guidelines for writing up the research. With a little judicious tweaking, it can also prove a reasonable
template for any study – even non-empirical literature-based work or theoretical research.
Essentially, the research process comprises six main building blocks:
1. identifying the main aim (or aims) of the study;
2. finding out what is already known (existing knowledge) and the gaps in knowledge;

3. identifying knowledge gaps;

4. choosing a method of study;

5. undertaking the study; and

6. drawing conclusions.

Building Blocks of the Research Process

Identifying the main aim (or aims) of the study

13
This stage may entail a problem statement, a research question or even a hypothesis. These will be the
end points of a lengthy and possibly frustrating process of formalising and sharpening-up an initial idea
for a research topic. The objective is to be able to define the focus for the research study in a clear and
unambiguous fashion. It involves a process within a process that can be represented as follows:

Research topic - In choosing your research topic, it is best to choose an area that you find personally
interesting, as you will be studying it for many months. Your initial ideas will stem from personal or
practical experience, or from reading.
Problem identification - From this topic area, identify the scale and nature of the problem or list the
problematic areas. These may, for example, concern debates about the effectiveness or the impacts of
change. Are there really problems? Are they important problems?
Refining the problem - Is the chosen problem researchable? Think about how it might be studied.
Can it be addressed within the available timescale? How will you measure the key concepts in the
field? At the end of this stage the extent and limits of the research should be clearly set.
The research statement - Write out precisely, but concisely, what the research intends to do. Use
phrases such as:
 ‘This research aims to...’
 ‘This research investigates the following question...’; or

 ‘This research tests the hypothesis that...’.

Sometimes researchers specify a broad aim for their research and then outline precisely how they will
go about achieving that aim by specifying objectives.
With empirical research, this initial stage of the research process is likely to involve a progressive
focusing down of the research. The scope of the topic should shrink from a broad to a fairly narrow
concern. This need not be the case with non-empirical research; a researcher may wish to maintain the
scale of the study or even broaden it. Even in such cases, however, the notion of formalising and
sharpening an initial idea remains appropriate.
Activity 1: Research aims and objectives
1. Think of an area that you would like to research.
2. Write out as simply as you can the main aim or aims of your proposed research. (The aim
should be a sentence long and begin with words such as ‘to critically examine’, ‘to critically
evaluate’, ‘to critically explore’.)

3. How will you go about achieving your aim(s)?

4. What are your specific objectives?

14
Finding Out What Is Already Known About the Subject
of Study
This equates to what is often called a ‘literature search’. Non-empirical research projects may place
great emphasis on this stage of the research process. Indeed in some situations, literature-based
inquiry may be the sole foundation for a research study, perhaps to aid theoretical development or to re-
conceptualise a problem.
Many researchers find it easiest to record this information on cards, with bibliographic material on one
side and information on content on the other. Some researchers now use computer database packages
(for example, PROCITE), while others prefer notebooks, which they later index.

Choosing a Research Design and Particular Method(s)

This phase of the research process addresses the issues discussed in the preceding chapter on
research design. It also draws on the literature search. The choice of method is likely to be governed
by three factors:
1. The goals of the research. Some methods are undeniably better for certain sorts of research
than others. The researcher will need to think long and hard, in the light of past work, about the
most appropriate method for a particular study.
2. The experience of the researcher. Most researchers will feel more comfortable with certain
methods in preference to others. One way to choose a research method is on the premise, which
may well be entirely unjustified, that if you are enjoying something you must be good at it. It is not,
however, as good a strategy as the previous goal-related approach.

3. The time available for the study. Most research projects that run into trouble do so because
the researcher attempts too much. Time is the practical limiting factor on the choice of method.
The researcher needs to choose an approach that will yield appropriate and relevant dividends
within the available time.

While this phase is most obviously of use in empirical studies, researchers working on non-empirical
projects will also find it worthwhile to spend some time reflecting on how they intend to conduct their
research. At this stage ethical issues need to be considered and taken into account when planning the
data collection. (See Chapter 3 for more on ethics.)

Undertaking the Study

Having chosen the approach to the chosen topic, that is deciding how the study will proceed; the
researcher must undertake the study using the chosen methods. In empirical research, the sensible
researcher will undertake a pilot study at this stage of the research process. This, in effect, is a
‘dummy run’ to see if what you plan is going to work in practice. For example, when designing a
questionnaire you should pilot it before using it to undertake the research project. This can be important
in identifying questions that do not make sense to other people or show whether you are using the
wrong terms in a particular work setting and so on. A pilot questionnaire may also identify new and
important questions that you have not thought of. Ideally, a pilot questionnaire, interview or observation
should be ‘piloted’ on a member of the group or population you hope to research (or situation you want
to observe). However it may be more practical for you to ask course or work colleagues to be your
participants in a pilot study.
Both the pilot and the ‘real’ activity can be subdivided into a number of separate sub-themes as shown
in Figure 1.

15
The need to proceed through an analysis or discussion stage in the research process is applicable to all
research involving any type of data collection and analysis. All research requires the researcher to
focus attention on the application and interpretation of data. In non-empirical research, the scope of this
element of the research process is, however, likely to focus more on the issue of discussion.
It is often tempting in empirical studies to try to address in sequence the sub-stages involved in the
actual ‘doing’ of research. Indeed, it is not uncommon to find research papers that actually use words
like ‘analysis’ and ‘discussion’ as headings. A little thought will reveal that this is not always a sensible
strategy. Reading and understanding research is often a great deal easier if analysis and interpretation
are undertaken at the same time, thus avoiding complex referencing back to earlier notes. Certainly,
when it comes to presenting research, the separation of analysis from discussion can be a hindrance to
the understanding of the reader.

In addition to these building blocks, the research process also includes the dissemination of research.

Drawing Conclusions
This is your chance to consider your research in light of what it has achieved and to generate a critical
evaluation of your work. This is often a brief discussion that draws together the key themes that arise
from your research, and considers the shortcomings of the work and future research possibilities.

Disseminating Research
It is of little value to conduct research without then disseminating this research. As we discussed in
Chapter 1 of this unit, research contributes to the production and development of knowledge; and
communicating the results of your research is integral to this contribution.

Writing Up Research
The research report, thesis or dissertation, the journal article, academic text and conference paper are
still the main means of communicating with other researchers and sometimes with a wider audience.
Although there are other ways of communicating findings, such as through an oral presentation, most
people do want to see the written word.

16
Writing a whole report, dissertation or thesis can seem like a daunting task looked at as a whole. Whilst
you do need to have an overall plan for the whole piece of work, you also need to break it down into
particular types of task as well as chapters. This should help to put you into a ‘can do’ mood, rather
than get overwhelmed with the size of the whole task. Furthermore, writing up should not be left to the
end of a research project. Writing up research will inevitably mean drafting and redrafting your work
many times. It should be a continuous process, in which you are learning from your research
supervisor, your peers and from your own mistakes.

Writing for Different Audiences

The academic dissertation or thesis is bound to be the first thing on the mind of Master’s
students. However, many researchers want to go beyond this and get to a wider audience
of, at least, academics and often beyond these confines in the field of criminology. Getting
published or getting coverage of your research (for example in the media) is largely a matter
of considering how to present your findings to a particular audience. Different audiences
are likely to have particular expectations from research, as illustrated in Table 1.

Audience Expectation

Academic colleagues Theoretical, factual or methodological

insights.

Policy-makers Practical information relevant to current

policy issues.

Practitioners A theoretical framework for understanding

clients better; factual information; practical
suggestions for better procedures or ways of
working; and reform of existing practice.

Lay audiences New facts; ideas for reform of services,

practices or policies that effect or interest
them; guidelines on how to get a better
service from practitioners or institutions; and
assurances that others share their own
experience of particular problems in life

Table 1. Audiences and their expectations (Source: Silverman, 2000, p. 272)

Academic Dissertations
The issue of an ‘audience’ is raised because the academic dissertation is a piece of writing for a
particular audience. You may need to remind yourself of this, particularly if your research is based on
your workplace and there is a more immediate and practical interest in what you are researching.
Therefore, in writing up your Master’s dissertation or thesis, you need to remember that you are writing
for an academic audience and thus adopt the expected conventions in style of writing and the specific
requirements as expected by the institution.
It is a good idea to look at actual examples of completed dissertations that relate to your broad area of
interest early on. The University of Portsmouth Library holds copies of past Master’s dissertations
(www.port.ac.uk/library); and you can borrow dissertations from other libraries at the cost of an inter-
library loan. (Do be aware that it is only recently that electronic versions of dissertations have been held
by libraries so you will need to search print versions as well.)

Academic Paper
Academic papers follow many of the conventions of a dissertation but they are written for a wider
audience and are much shorter. When you start trying to get published there are a number of ways of
improving your chances:
 Study the journal(s) you have in mind for possible publication.
 Make sure you follow the ‘notes for contributors’ or other guidelines for prospective authors.
17
 Consider whether your topic and the way you want to write about it is the kind of article the
chosen journal(s) would be likely to publish.

 Ring up the editor and say that you are considering an article on a particular topic and ask
whether they are likely to be interested. This strategy sometimes means that you will get some idea
of their interest (or not) in what you want to write or whether a similar article has just been accepted.

 Consider writing with somebody more experienced. You can learn many of the tactics of
‘getting published’ in this way.

 Expect to rewrite and resubmit. Everybody experiences rejection.

Activity 2: Student theses published online

The Internet Journal of Criminology (ICJ) is a free access online criminology journal. As well as 'peer-
reviewed articles', good quality undergraduate dissertations and postgraduate theses are published.
1. Go to the website for this journal http://www.internetjournalofcriminology.com/
2. Browse through the Master’s theses and dissertations. This will give you a ‘feel’ of what type of
theses are being accepted for publication by the ICJ.

Academic Conferences

Academic conferences usually ask for abstracts on one or more of the conference themes. In large
national and international conferences, abstracts are usually published in book form and may be
numbered in their hundreds. This can be both a way of getting yourself and your research known, as
well as finding out about relevant work in progress, which has not been published yet. Conference
papers can also be a good way of getting published, either through contacts made during the
conference or through conference proceedings. Conferences are a good opportunity to test and
develop your ideas. Conference papers can often be seen as the first draft of a piece of work you later
hope to get published.

Wider Modes of Research Dissemination

Research into crime and related issues has undoubtedly had an influence on present-day life and
thought. Tracking how ideas based on research are disseminated is, however, not easy to do.
Despite the obvious importance of making research findings known, a look at most research methods
texts will illustrate that relatively little guidance is given – most texts simply refer to the structure of a
research report or thesis as if the job is then done. Robson (1993) is an exception, devoting three
chapters to ‘making an impact’. Robson certainly makes clear the numerous difficulties in trying to make
an impact with research: even when people are interested in what you have found out they may not be
willing to ‘read the small print’ and will generalise findings from one situation to another, when this may
not be applicable (1993, p. 432).
A common criticism of research (and researchers) is that they fail to let people know their findings or
they produce reports that are too long and difficult to read. More importantly, practitioners and policy-
makers often find it hard to translate findings into something that they can apply to the real world.
Sometimes this is a case of simply wanting things to be handed to them ‘on a plate’, but nevertheless it
is a criticism worth considering further.
The first point to make is that not all research sets out to make an immediate impact – it may just aim to
change ideas, which may have an impact at a later date. Secondly, there is a time and practicality
issue. There are a lot of potential users of practical research findings: how are they to know that you
have done work of relevance to them? This is where there are numerous possibilities to disseminate
research information.
Executive summaries
Executive summaries consist of a few sides, summarising the research design and context, key
findings, recommendations and where the main report can be found. Many research organisations and
government departments provide free executive summaries of research findings through their websites.
Sometimes whole research reports can be downloaded from these sites.
Oral presentations

18
Oral presentations and feedback to organisations and research participants are excellent when carefully
thought through. Some researchers go as far as offering to send summaries (even reports) to research
participants. As well as considering the practicalities of doing this (for example, how many reports or
summaries) you also need to think about how you get the messages across meaningfully and how you
handle painful, uncomfortable or controversial findings. In addition, within an organisation you need to
consider how you manage issues of confidentiality. Some conferences and presentations are of course
explicitly organised with the needs of practitioners in mind. They are more likely (than an academic
audience) to want some practical ideas about what they can do about the issue you have researched.
Professional organisations and specialist bodies
Professional organisations and specialist bodies might act as a useful conduit for your work, by carrying
a feature on it or the executive summary. You can certainly lodge copies of a thesis with organisations.
This is especially useful if they have a specialist library where people may get to see and read you work.
The media
The media plays a part in how the general public might hear about a piece of research. Not surprisingly
some academic researchers can be nervous of getting involved with the media, not least because what
they say may be taken out of context. However researchers can work well with the media; and some
funded research projects expect a media launch of findings, in which case the researcher is expected to
do interviews with the media.

The key thing to note is that the media can tend to sensationalise issues as it has to simplify them in the
short time allowed to get ‘a story’ across. This is a problem if ‘the story’ does not reflect what you
intended to say and can, in turn, reflect badly on you as a researcher affecting your credibility within an
organisation, whether or not your research was based on that organisation. If your research is likely to
attract public interest, it is a good idea to take the advice of the media or public relations officer within
your organisation. The more local the media is, the more care you have to take in relation to basic
ethical principles of confidentiality, anonymity and no harm to participants. Local media, not
surprisingly, usually want a local angle: that can mean presenting an issue in an unhelpful way, which
can create adverse publicity about a locality or group of people. Of course, representatives of the media
are sometimes present and invited at certain conferences, in which case you may or may not get a
mention, whether or not you set out to do so.

Key issues to consider and remember in research

dissemination
The following are some of the key issues to remember in research dissemination.
 Type of audience:
 are they academic, practitioner, general public?

 how specific are the audience’s interests likely to be?

 why would they want to know about your research?

19
  how will you make the research accessible and meaningful to your particular
audience?

 Managing the message: the key messages from your research should remain consistent.
However, you will need to explore different ways of phrasing the same message for different
audiences.

 Be prepared to follow through by:

 sending copies of overheads, papers, executive summaries and any other materials
requested;

 making sure you are available to follow up calls if your work has been exposed in the
media.

 Keep up-to-date: research can date very quickly in the current climate, particularly if it
relates to a particular policy initiative that has since been dropped or changed in some way.
Making a link with a recent relevant announcement, event or issue when speaking to an
audience can engage their interest and persuade them of the current relevance of what you
have researched.

 Criticism: you are bound to get some criticism, comments or questions that feel critical as soon
as you present your research to an audience (whether your presentation is written or oral). In order
to make a considered response to any criticism, you need to:

 evaluate the validity and implications of the criticism for your research;

 compare each criticism with the other responses that your work has generated;

 modify the interpretation of your findings, if the criticism is valid; and

 consider changes in any future research you do.

Chapter 3 Research Ethics

ALL STUDENTS WHO INTEND TO APPROACH PARTICIPANTS VIA OBSERVATIONS, INTERVIEWS
OR SURVEYS MUST SUBMIT FORMS FOR ETHICAL APPROVAL PRIOR TO BEGINNING DATA
COLLECTION.
Ethics is typically associated with morality and thus matters of ‘right’ and ‘wrong’. Obviously there will
be some differences of opinion about this between individuals, and different groups and professional
bodies may each have their own particular code of ethics. (See the section: Codes of Ethics for further
information on specific codes of ethics.) However all research with human participants and with animals
is governed by ethical considerations and particular professions or groups often have some basis for
agreement. Consequently there are some important key principles in social research on which there is
some broad agreement across disciplines (which will be discussed in further detail):
 Voluntary participation and informed consent
 No harm to participants

 Anonymity and confidentiality

 Deceiving subjects

 Analysis and reporting

 Political considerations

20
 Role conflict

When conducting social research in the United Kingdom, you also need to take account of the Data
Protection Act 1998. Irrespective of the specific provisions of this Act, conducting research in an ethical
manner requires you to consider the following issues regarding data:
 obtaining consent before using data
 ensuring that data held is accurate

 holding research data securely (e.g. locked filing cabinet or drawer)

 observing confidentiality in reporting.

Key Principles in the Conduct of Ethical Research

Voluntary participation and informed consent
Social research often, though not always, represents an intrusion into people’s lives. Furthermore, such
research often requires people to reveal personal information about themselves. Thus voluntary
participation and informed consent from research subjects are ideals in social research. However it is
not always possible to follow through both of these ideals. For example, information held about large
numbers of people within an organisation where you intend to analyse this secondary data as part of
your research may make it impossible or impractical to contact everybody on whom data is held.
People may have given their permission for their data to be used at an earlier date but were not
informed about how their data might be used in the future. You may have data protection clearance to
analyse the data but you still need to think about the ethics of what you are doing. Principles such as
ensuring anonymity and that no harm will come to participants are particularly important in these
circumstances. Data gained through a research project should not be used for another purpose,
especially not a fraud or criminal investigation.
No harm to participants
Social research should never injure the people in a study, regardless of whether or not they volunteer to
take part. Perhaps the clearest instance of this norm, in practice, concerns revealing information that
would embarrass participants or endanger their home life, friendships, job and so on. Obviously the
subjects of research can be harmed in a variety of ways and for a participant to face some aspects of
their life that the researcher chooses to ask about may sometimes be unpredictably painful. In such
instances researchers have a responsibility to know where individuals may seek help or support if they
want it.
Anonymity and confidentiality
The most obvious concern in the protection of the subjects’ interests and wellbeing is the protection of
their identity, especially in survey research. If revealing their survey responses would injure them in any
way, adherence to this norm becomes all the more important.
Anonymity
Any respondent may be considered anonymous when the researcher cannot identify a given response
with a given respondent. For example, a postal questionnaire may be sent out with no identification
numbers put on the questionnaire before its return. Data may also be made anonymous after
collection. For example, a survey might be conducted using face-to-face interviews so that participants
are known to the researcher. However, during the process of analysis, the data may have all the
identifying features removed from it.
Confidentiality
In a confidential survey the researcher is able to identify a given person’s response, but essentially
promises not to do so publicly. For example, in an interview survey the researcher would be in a
position to make public the income reported by a given respondent, but the respondent is assured that
this will not be done.
Deceiving subjects
Handling your own identity when researching can be tricky, even when it is possible to conceal your
research identity. Deceiving people is unethical and within social research deception needs to be
justified on firm grounds. Sometimes researchers admit that they are doing research but fudge issues
about why they are doing it or for whom.
Analysis and reporting
21
Being honest about the shortcomings of your study and the findings you did not set out to look for is part
of your ethical obligation to fellow practitioners or researchers. In general, knowledge expands through
honesty and openness and is retarded by ego-defences and deception.
Political considerations
Although ethics and politics are closely intertwined, the ethics of social research deals more with the
methods employed, where as political issues are more concerned with the substance and use of
research. Furthermore, there are few generally accepted political codes of conduct comparable to those
discussed above. It is, for example, generally accepted that a researcher’s personal political orientation
should not interfere with or unduly influence their scientific research. It would be considered improper
for you to use shoddy techniques or lie about your research to further your political views. Although you
are permitted to have political views you are expected to hold them aside when conducting research.
The extent to which you can do this will be a measure of the objectivity of your research.
Both ethical issues and questions of detachment and rationality become particularly marked in
situations:
 where the researcher is being paid as a consultant to undertake the research;
 when the researcher is actually investigating his or her own practice or workplace; or

 when the researcher is undertaking action research – a form of research in which the
researcher plans and introduces a change into a particular work setting and then evaluates the
results.

In each of these situations, the researcher may be under pressure to deliver particular conclusions. In
the latter two scenarios, there may be difficulties over the researcher’s connection with the topic and
setting of the research, as well as role conflict for the practitioner researcher.
Role conflict
A particular issue that may arise if you conduct research within your workplace is the potential for role
conflict (between your work role and your role as a student researching for an academic qualification).
Related to this is the issue of power relations. Workplace status and power should not be used to
coerce people into being participants in your research. There can be a fine dividing line, in practice,
between using your contacts in order to gain access to do research and misusing a situation of
privileged access to potential research data or subjects.
If you want the cooperation of your work colleagues to do research, you need to persuade them of the
relevance of your chosen exercise. It is also crucial that you obtain official clearance from the relevant
line manager, particularly if you want access to client files and other confidential material. It is always a
good idea to offer to supply your colleagues with a copy of the final report or document when you do a
project that results in a full report.
Activity 1: Role conflict and ethical issues
Your line manager has asked you to undertake a study of staff use of computers for personal use during
work time. Your line manager has suggested that you would be given more time to do this if it was also
your Masters dissertation. You decide to go ahead with this study.
1. What are the potential conflicts of interest and ethical issues that could arise in such a study?

Codes of Ethics
The British Criminological Society (BSC) has its own code of ethics as does the Socio-Legal
Studies Association (SLSA) and the British Psychological Society (BPS).
The BSC ‘Code of Ethics for Researchers in the Field of Criminology’ can be found on
http://www.britsoccrim.org/codeofethics.htm
The SLSA Statement of Principles of Ethical Research Practice can be accessed from
http://www.kent.ac.uk/nslsa/images/slsadownloads/ethicalstatement/slsa%20ethics%20statement
%20_final_%5B1%5D.pdf
The BPS ‘Code of Ethics Conduct can be accessed from
http://www.bps.org.uk/sites/default/files/documents/code_of_ethics_and_conduct.pdf
You are advised to consult these codes because it is your responsibility as the researcher to ensure that
you have given full consideration to ethical issues.
22
However it can be a challenge to know how to incorporate these ethical considerations into your
research. The ‘rule of thumb’ is: when in doubt, consult. Your research supervisor is the first person that
you need to consult.
The next section of this unit: ‘Assessing risk in research’ provides you with further guidance on
complying with ethical requirements.

Assessing Risk in Research

Although risk assessments are routinely undertaken in relation to science research projects, it is only
recently that researchers in some areas of the social sciences have considered the risks to themselves,
their employers and their participants in undertaking research. Research in criminology may require
researchers to work on their own. For example, they may conduct an interview with somebody in a
private situation. The topic of the research may be sensitive or invoke strong reactions from
participants; or the researcher may be working with potentially higher risk groups, such as ex-offenders
or people with a history of psychological disturbance, or simply exploring issues where there is a
potential threat of violence (for example, studying homophobic violence).
If you are undertaking interviews or other forms of research that involve a private meeting then you may
be exposing yourself to the following risks:
 physical threat or abuse;
 psychological trauma or consequences, as a result of actual or threatened violence, or the
nature of what is disclosed during the interaction; and

 being in a compromising situation in which there might be accusations of improper behaviour.

The research participant may also be at risk of being exploited, frightened or psychologically harmed,
particularly if the researcher is not aware of how to deal with the situation. It is your responsibility, when
undertaking research, to understand the risks involved on both sides and consider what precautions you
can take to minimise them. This can be done by using a simple form of risk assessment:

Activity 2: Ethics and risk

Go through the section: Key principles in the conduct of ethical research again; and read the British
Society of Criminology Code of Ethics. Then use the following as a checklist for beginning to assess the
issues in your proposed area of research:
 anonymity and confidentiality;
 assessment of risk (physical, psychological, compromising situation);
23
 data protection issues;

 deceiving subjects;

 no harm to participants;

 potential conflict (particularly if researching within own workplace or organisation); and

 voluntary participation.

If your initial assessment of any of the above areas suggests worrying issues that need to be resolved,
you should discuss them with your Unit Co-ordinator and/or research supervisor.

Ethical Dilemmas in Research

An ethical dilemma is a situation in which there is no ‘right’ answer or decision, ‘only a decision that is
thoughtfully made and perhaps ‘more right’ than the alternatives’ (Hill et al., 1995, cited by Laine, 2000,
p. 3).
Codes and guidelines for the conduct of research are useful but they cannot prepare you for every
eventuality, some of which may be presented to you with great immediacy if your research involves
contact with real people in real situations. Laine notes how:
In actual fieldwork, researchers experience ethical dilemmas with an immediacy and personal
involvement that draws on intuition and empathy, feelings and emotion. (Laine, 2000, p. 3)
Ethical problems and dilemmas are an inevitable part of conducting research. Not all of these problems
and dilemmas can be anticipated. There are no easy answers to many of the situations you might find
yourself in when undertaking research. For practitioners with privileged access to certain forms of
information about people or situations, this is especially important to remember.

The Stanford Prison experiment (Zimbardo, 1973)

In a well known piece of psychological research conducted by Zimbardo (1973) in the 1970s in America,
a mock prison was set up in the psychology department of Stanford University. Subjects were recruited
through advertisements placed in a local newspaper. The advert asked for volunteers for a study of the
psychological effects of becoming a prisoner or a prison guard. Over 70 applicants were screened
through interviews and personality tests to eliminate people with psychological and other problems. The
final sample in the study consisted of 24 healthy, intelligent, middle-class males. Half the group were
randomly assigned to the role of guard, half to the role of prisoner.
The basement of Stanford University in California was converted into a simulated prison environment
and the experiment began when the ‘prisoners’ were arrested by the police on a Sunday morning, in
some cases in front of their neighbours. They were charged with a felony, told their constitutional rights,
searched, handcuffed and taken in the back of a police car to be ‘booked’. Following fingerprinting and
so on they were taken to the ‘prison’ blindfolded. They were then strip searched, deloused, issued with
uniform, bedding and so on.

24
The experiment was planned to run for two weeks but was stopped after six days because of the
pathological reactions of the ‘prisoners’ and of the ‘guards’. The ‘guards’ became sadistic and the
‘prisoners’ became depressed and showed signs of extreme stress.
To find out more about this experiment, go to www.prisonexp.org and www.zimbardo.com
The University of Portsmouth Library also holds copies of Zimbardo‘s film on the Stanford prison
experiment:
Zimbardo, P.G. (2004) Quiet rage: the Stanford prison experiment. Stanford: Stanford Center for
Professional Development, Stanford University.
[DVD versions shelved at 303.33/ZIM and 303.36/QUI]
Although this experiment is often praised for its demonstration of how easily people can assume the
roles of guard and prisoner, Zimbardo and colleagues were heavily criticised in relation to the ethics of
this experiment. There has been vigorous and ongoing debate about whether it is ethically acceptable to
place people into situations such as those created for the Stanford Prison experiment.
Haslam and Reicher: The BBC Prison Study – the controversy continues
Interestingly a similar experiment to the Stanford Prison Experiment was conducted and televised on
BBC in 2002. The study was run by two British academics, Professor Alex Haslam and Dr Steve
Reicher. The BBC programme, called ‘The Experiment’ followed 15 participants who were assigned the
role of prisoner or guard in a purpose-built prison-like environment. This time the experiment lasted 9
days, instead of the 10 days planned.
Comprehensive information on the BBC Prison Study can be found on www.bbcprisonstudy.org . This
site includes a series of movie clips taken from the original BBC programme, an explanation of the
rationale for the study (including ethical considerations) and a range of interesting related materials.
There is also a short video on ‘The Experiment’ available from the Open University in which issues
raised by the study are discussed:
http://deimos3.apple.com/WebObjects/Core.woa/Browse/itunes.open.ac.uk.2076249052.02076722690
(Note: This download is free but you need to have installed iTunes software on your computer in order
to access it.)
Alternatively you can download the Open University podcast:
http://podcast.open.ac.uk/oulearn/psychology/podcast-dd307-social-psychology#
In addition the University of Portsmouth Library holds a series of video recordings on ‘The Experiment’
[shelved at 303.36/EXP].
The Haslam and Reicher study has attracted debate over ethical considerations, just as Zimbardo’s
experiment did. Zimbardo himself was said to be sceptical about this replication of his experiment. To
read Zimbardo’s criticisms and the response from by Haslam and Reicher, go to
http://www.bbcprisonstudy.org/resources.php?p=90
25
Activity 3: Ethics and the Stanford Prison experiment
1. Visit the websites for the Stanford Prison Study www.prisonexp.org and the BBC Prison Study
www.bbcprisonstudy.org
2. Identify the ethical concerns that both of these studies raise.

3. Then answer the following question:

‘On balance, was what was learned from these experiments worth putting aside certain ethical
principles?’
Activity 4: Ethics and control groups in experimental research
1. Consider a programme aimed at treating violent prisoners, but with no prior information about
its effects. In order to evaluate the effectiveness of such a programme a control group will be used
to compare changes in behaviour. The control group will receive no form of treatment during this
time.
2. What are the ethical issues raised by the withholding of treatment?

3. Reconsider this with the knowledge that an experiment has already proved the programme’s
positive effects, but the Home Office wants a rerun of the experiment.

Chapter 4 Key Concepts in Research

In this chapter you will be introduced to key concepts and issues relating to the information (data)
collected for research. It will also help you to understand the data that is presented in research papers
and reports. Types of data, and terms such as validity; reliability; correlation; cause; sample; population;
probability and non-probability samples will also be discussed.

Understanding Data
Data collection and analysis: are they necessary?
Before looking at various kinds of data and how we measure data, it may be useful to discuss the
argument that the collection of information to support a research hypothesis is not really necessary. It
could be argued that knowledge can be obtained by simply reporting what is believed to be true or what
common sense indicates to be true. Even if data is available, the use of methods and techniques could
be said to complicate matters and may not help in exposing any patterns or trends within the research
area. The following two examples contradict these notions. The first example illustrates the importance
of data, whilst the second example outlines the importance of analytical techniques to make sense of
the data.
Example 1: ‘Common sense’ conclusion – the need for data
After seat belt legislation was introduced in the early 1980s, it was generally believed that road safety
had improved. Although the number of car drivers and front seat passengers killed during the first
quarter of 1984 was down 28% on the same period in 1982 (before seat belts had to be used), there
was an increase of 10% in fatal accidents to pedestrians and pedal cyclists hit by cars and light vans.
Research has suggested that drivers felt more secure when wearing a seat belt and were more likely to
speed. Furthermore, there was also a dramatic rise in the number of rear seat passenger casualties.
Here, data was needed to investigate the actual consequences of introducing seat belt legislation.
‘Before’ and ‘after’ data were needed to illustrate that the use of a ‘common sense conclusion’ does not
reflect the full picture – such a conclusion is neither valid nor reliable.
This example highlights the need for measurement and data collection in helping to understand how the
real world operates. In progressing to the analysis of such data, it might also be argued that such
analytical techniques are not really necessary. The next example illustrates that analytical methods are
the best means of making sense of numerical research data.
Example 2: Looking for patterns – the use of technique
Consider the information provided in Table 1 below.
Is there any pattern between the value of stolen goods in the two police areas?
26
 (Please note that the figures are hypothetical)

Value of Goods

Area 1980 1981 1982 1983

A 108.4 60.3 89.2 50.0

B 206.1 158.1 187.3 149.7

Table 1: Value of stolen household goods (£ million) – raw figures

Table 1 simply reports the raw figures for the two police areas. Uncovering any existing pattern from
these raw figures is difficult.

In Table 2 the technique of rounding to the nearest 10 is employed.

Value of Goods

Area 1980 1981 1982 1983

A 110 60 90 50

B 210 160 190 150

Table 2: Value of stolen household goods (£ million) – rounded figures

Here the technique of rounding up has shown that Area B = Area A + 100. Although this is a very simple
analytical ‘technique’, it nevertheless uncovers a major pattern in the data.

Types of Data
Distinguishing between different types of data is important for choosing the correct analytical method.
Certain descriptive and statistical methods are only appropriate for certain types of data. A full
understanding of the research data in question is therefore essential for carrying out an appropriate
analysis of that data.
Quantitative or Qualitative Data?
In terms of research data or research information, a number of distinctions can be made. One way of
distinguishing between types of information is whether it is quantitative or qualitative. Many researchers
choose to use both types of data, often within a mixed methods design. The analysis of qualitative data,
that is data in non--numerical form, will involve different techniques to quantitative data analysis.

27
Quantitative data simply refers to numerical information such as the actual number of crimes
committed in an area in a given time period, or the rate of crime committed per 1000 population in an
area in a given time period. Other examples of quantitative data include the average number of days
taken as ‘sick leave’ in a police division or the personnel age range in any one police station. Types of
quantitative data will be discussed in more detail later in this chapter.
In contrast, qualitative information refers to any type of data that is not numerical in nature and often
comprises descriptive statements, opinions or feelings on a particular issue. For example, taped
interviews with victims of house theft, regarding their immediate emotional reaction to such crime are
regarded as qualitative information. Similarly, a postal questionnaire containing written comments on
the perceived effectiveness of neighbourhood watch schemes would also be classed as qualitative data.
It is important to note here, however, that very often qualitative information may be converted to
quantitative information for ease of collection and subsequent analysis. For example, rather than ask
the general public to comment on their satisfaction of policing within their locality, it is much easier to
ask them to choose from a range of options that best summarises their opinion (for example, very
satisfied, not satisfied, or not at all satisfied). A number can then be assigned to each of the qualitative
categories and the information treated in a quantitative manner. Alternatively, the data can be analysed
using techniques specifically designed for such non-numerical data sets. Crabtree and Miller (1992)
define these techniques as quasi-statistical methods, template approaches, editing approaches and
immersion approaches. These will be briefly discussed in turn. Computer software packages are also
available to deal with qualitative data analyses (for example, NUD*ST), but you will need to spend quite
a lot of time learning how to use such packages.

Approaches to Qualitative Data

A number of approaches to qualitative data analysis can be identified, for example:
Quasi-statistical approaches: this involves the conversion of qualitative data into a quantitative form
by categorising the words or phrases and identifying frequency and relationships. Content analysis is a
typical method used.
Template approaches: using codes developed from theory, a research question or from an initial
reading of the data. The codes form a template for data analysis.
Editing approaches: this is a much more flexible approach, which relies on the researcher’s own
interpretation. Grounded theory is an example of this approach.
Immersion approaches: this method is the most unstructured of the four and aims to avoid
systemisation of the analysis. Again it relies, to an even greater degree, upon the interpretation of the
researcher.
(For further information and discussion on the analysis of qualitative data, see Chapter 7: Analysing
research data: Qualitative analysis).

Approaches to Quantitative Data

For quantitative data, there are numerous descriptive and statistical techniques available. (See Chapter
6: Analysing research data: Quantitative data.) Many of these can only be used if the available
information is in an appropriate format for that particular method or technique. In other words, certain
types of analysis can only be used with certain types of quantitative data. For this reason, it is
necessary to be able to distinguish between four levels of numerical information: nominal, ordinal,
interval and ratio data.
The division of quantitative information into these four levels is based on the amount of numerical detail
found within the data set in question. This categorisation is often referred to as the ‘scale of
measurement’. Nominal data is less detailed than ordinal data, which is less detailed than interval data.
Ratio data is the most detailed numerical information available.
Nominal Data

This type of data usually refers to counts in categories, where categories themselves have no
implicit order. Table 3 illustrates nominal data and shows the annual number of house
thefts in three adjacent streets. Note that the emphasis is on the use of categorisation
rather than the actual counts.
28
 Street No. of house thefts

A 8

B 12

C 5

Table 3: Nominal data – house thefts

Streets A, B and C, and the numbers allocated to them, represent nominal data. This is because the
information is presented as counts (of annual theft) in categories (Street A, B and C), which are
unordered (in other words, there is no ordered difference between the streets). A, B and C simply act as
labels and do not denote size, status of street, and so on.
Ordinal data

Ordinal data is a ‘step up’ from nominal data in terms of amount of numerical detail.
Ordinal data can be further subdivided into weak ordinal and strong ordinal, depending on
the amount of such detail. Weak ordinal data is very similar to nominal data. They comprise
counts in categories. This time, however, the categories are ordered and, because of this,
they portray more information than straightforward nominal data. The data in Table 3
(nominal data) has now been altered (see Table 4) so that the information is weak ordinal.

House value (£) No. of house thefts

50 000 – 100 000 8

101 000 – 150 000 12

151 000 – 200 000 5

Table 4: House value and house thefts

In Table 4, rather than simply report the number of house thefts by street labels, the information has
now been sorted so that it is given by category of house value. The data is now weak ordinal because
they represent counts in ordered categories. The information remains weak ordinal because the detail
of order within the categories is not available. Of the eight thefts found in the lowest category, for
example, it is impossible to say which house is valued at a higher or lower figure than another.
Activity 1: Strong ordinal data
1. Explain in your own words what is meant by ‘strong ordinal data’.
2. How could the information given in Table 4 be shown as strong ordinal data?

Interval and Ratio Data

These two categories are described together because their difference is minimal and is likely to be
unimportant for police-related data sources. Interval and ratio data are the most numerically detailed
information available, the only difference between them being that interval data has an arbitrary zero.
The obvious example of interval data is measurements of temperature: 0°C does not indicate a lack of
temperature and in this sense the zero is unreal. In ratio data, the concept of zero is real. Zero income,
for example, can occur. Both interval and ratio data provide a precise measurement of a variable on a
continuous scale. In Table 4, the exact value of each property involved in house thefts would be
regarded as ratio data. Other examples of ratio data would include the rate of juvenile crime by police
area, the actual number of crime victims seeking legal aid or the number of road traffic accidents
involving a fatality by county.
Activity 2: Ratio data
1. Explain in your own words what is meant by ‘ratio data’.
2. Think of examples of police-related ratio data.

29
Primary and Secondary Data
Another way of categorising the type of data collected or used in research is primary data (collected by
the researcher) and secondary data (already collected), see Figure1 below.
Primary data

Secondary data

(Source: adapted from Dunsmuir & Williams, 1990)

Figure 1 Types of data – how they can be used in research
Secondary Data
At this point the use of secondary data should be emphasised: that is, data that has already been
collected by another researcher or agency. Using another researcher’s data can be beneficial in terms
of cost, time and access to populations. The quality of the data gathered might also be higher than
collected by a researcher with less experience and/or knowledge of the area. Crime statistics, collected
by police forces, government departments and researchers are one of the most common sources of
secondary data. Researchers have also included analyses of newspapers, official documents, meetings
and other records, as well as personal documents such as letters and diaries, in their research. The
possibilities are numerous.
Further discussion and information on the use of secondary data in research is provided in Chapter 5:
Common Research Methods.

30
Key Concepts in Research Methods

Measurement: Validity and Reliability

Measurement is integral to research design. Measurement is not just confined to numerical matters:
qualitative information requires equally careful conceptualisation and recording. Two concepts are
particularly important with regard to research design and measurement: validity and reliability.
Validity
In research the concept of validity primarily involves two aspects:
 internal (or face) validity; and
 external validity

Internal (face) Validity

First, it is necessary to confirm that a research design is actually measuring the concepts that it is meant
to be addressing. This is called internal (or face) validity. It focuses on the researcher’s assumptions
during the creation of the research design. Are these assumptions justified; are they borne out by
reality? The researcher may believe that s/he has a design that will enable the study of some particular
phenomenon when, in fact, something rather different is being uncovered.
External Validity
The second aspect to validity concerns the extent to which a research design can generate results that
can be generalised to a wider population. This is termed external validity. It relates to the confidence
that the researcher is able to place on the results of a study and also to questions of transferability. It
may be a particular problem with small-scale studies or very artificial experiments.
Issues of validity are specifically important when considering survey research. Survey instruments need
specific attention with regard to both internal and external validity. They may also usefully be assessed
for two more aspects of validity:
1. criterion validity; and
2. construct validity

Criterion Validity
This would check that different versions of a question seeking the same information obtained the same
result.
Construct validity
This would check whether two questions resulted in similar relationships with a third variable.
Reliability
Reliability relates to the consistency of a research design in generating similar results under repeated
operation. If conceptual definitions and the research design remain constant, the results from a reliable
research study should be unchanging. If either changes, then there should be a change in the outcome
of the study.
Reliability will depend particularly on the quality of the data collection instruments. In the case of a
questionnaire survey, ambiguous questions may render a research study unreliable. Many methods
exist to test reliability and identify unreliability in questionnaires and, indeed, in research designs
generally. These revolve around repetitions of the study, and/or reframing the study in a different way,
but with the same objectives and then checking to see if similar results are obtained.
A specific case of both validity and reliability assessment is the approach known as triangulation. In this
approach, the researcher approaches a problem using several different research methods. Each
method is focused on the research topic, but proceeds in a slightly different way. For example, a
researcher interested in the impact of the credit fraud (such as defaulting on loans/mortgages) on the
economy may conduct a survey on a large and representative sample, undertake a small and focused

31
interview programme in selected areas across the UK, and peruse local and national policy documents
in order to gain differing perspectives on the research topic. These differing perspectives could then be
integrated in the final report of the research and used to present a balanced and hopefully more reliable
and more valid picture. For a discussion of validity, reliability and the use of triangulated studies within
criminal justice and criminology see Hagan (1993, pp. 248–268).

Explanation, Correlation and Cause

Many researchers, and certainly most sponsors of research, want more than description. They want to
know why differences exist between places or why a particular situation has arisen in a case study
area. This introduces the notions of explanation and cause.
In designs such as the case study, explanation flows from the process of research. By interviewing or
observing over a period of time within the case study area, the researcher gains a familiarity and
understanding of a situation that will enable him or her to appreciate the reasons for its occurrence.
Similarly explanation in a documentary case study builds up by perusing many different sources and
cross-referencing different views on the chosen topic.
In designs involving the collection of quantifiable data, explanation requires the collection of information
on at least two variables. The first variable is that which requires an explanation: this is known as the
dependent variable. The second, the independent variable, is that which is doing the explaining. For
example, the dependent variable could be the distribution of burglaries and an independent variable
might be the level of male unemployment. There can, of course, be more than one independent
variable: burglaries are unlikely to be explainable solely by the male unemployment rate.
Dependent variable: this is the measurement made by the researcher and depends upon the
independent variable.
Independent variable: this is the variable that potentially causes the change in the dependent variable.
Hover over the shape below for a recap
Clearly for male unemployment to ‘explain’ a local burglary rate there needs to be a causal mechanism
that links the two. This might be based on an existing understanding of the issue under investigation:
most burglaries are committed by men with available time when other residents are out. In addition to
theoretical plausibility, three other criteria need to be satisfied if an independent variable is to
satisfactorily ‘explain’ a dependent variable: consistency, strength of relationship and trend of
relationship.
Consistency: can the relationship be identified in other places and at other times?
Strength of relationship: how closely matched is the data on the independent and dependent
variable? In terms of one of the statistical methods covered in Chapter 6, how close is the correlation to
1.0? Of course, with more than one independent variable, the individual relationships may be quite low;
while cumulatively all the independent variables provide a satisfactory explanation.
Trend of relationship: is a change in the independent variable matched by a change in the dependent
variable? Is the direction of the change expected?
Neither theoretical plausibility nor the other three criteria (outlined above) on their own are enough to
suggest explanation. All need to be present.

Samples and Populations

One way of finding out about a group of people is to collect information from everyone in the group. The
alternative is to collect information from only some people in the group, in such a way that their
responses and characteristics reflect those of the group from which they are drawn. This procedure is
much cheaper, faster and easier than surveying all members of a group. This is the principle of
sampling. Sampling is an important consideration in all types of research and is also linked to concepts
of validity, reliability and generalisability.
Sampling is the process by which people or things are chosen from a population. A sample may be a
probability or a non-probability sample. In both cases you still need to justify how and why you chose
your sample and what you can infer from it.
In research based on people, a sample is the group of people you chose to approach in your research
out of a total population. This includes all the people you could ask. A population could literally be
everyone in Britain, all inmates or staff in a prison (or all inmates in all prisons in Britain), all staff at a
police station, all illegal drug users in a city and so on. A population may also be the total known

32
instances of something such as all road traffic accidents in a particular place. In other words, the limits
of the population upon which you base your study should be specified. Whether we can establish a
whole population will depend on the topic. Even the Census cannot provide data on people who do not
complete Census returns. However, the Census is as near as we will ever get to information about the
whole population of Britain and some sampling strategies will be based on what is known about the
British population from the Census.

Samples can reflect the populations from which they are drawn with varying degrees of accuracy.
Sampling is important in relation to the claims we make from the research we do. In its simplest form
sampling is linked to a search for typicality: the extent to which what we have found in a particular
situation, at a particular time, can be said to apply more generally (Robson, 1993, p. 135). A sample
that accurately reflects its population is called a representative sample (also called a probability
sample). If your sample can be said to be representative of the whole population you can say that the
results you obtained from your sample probably apply to the whole population. However, when we do
not know what the total population is we cannot claim to have drawn a representative sample; instead
we need to justify who and what we have researched as a non-probability sample.
You will probably quickly spot that it is impossible to establish the extent of the total population in some
research topics. For example, you could not expect to get an accurate list of all illegal drug users in a
city, although you could establish all registered drug users. In some research, sampling may be
restricted only to considerations like time or place.
It is important to remember that time or place may be crucial if you are trying to understand or estimate
the nature and scale of an issue. For example, estimating the scale of drunken behaviour in public, in
London, would be greatly affected by where and when you tried to obtain this estimate.

Probability and Non-Probability Samples

As already noted, there are two broad types of sample:
1. probability; and
2. non-probability

A probability sample is one in which each person in the population has an equal or at least known
chance (probability) of being selected, while in a non-probability sample some people have a greater,
but unknown, chance of selection than others. Probability samples are based on the mathematical
theory of probability. The surest way of providing equal probability of selection is to use the principle of
random selection. This involves listing all members of the population (this list is called a sampling
frame) and then, in effect, ‘pulling names out of a hat’, although you can use a random number table to
do this. There are still likely to be differences between the sample and the total population, but using a
probability sample means that this should be by chance alone. There are other methods of probability
sampling: see Table 5.
A non-probability sample may be more practical for some research projects. For example, there are no
lists of all the people who have committed a crime or particular offence, although there are lists of
convictions for individual offences. The issue then may be whether you are allowed access to such a
list for research purposes. Non-probability samples may be more open to the criticism of not
representing a population, but are chosen in situations where probability sampling techniques are either
impractical or unnecessary. Six frequently based non-probability samples are detailed in Table 6.
Table 5: Probability samples

Type of sampling Selection strategy

Simple random Select from a full list of the population

(sampling frame). Can use a random
number table to do this.

Systematic Start at random, at a point on the sampling

frame, and chose every tenth case (or some
other frequency) depending on sampling
frame size.

Stratified Sampling frame stratified (e.g. by class,

33
 race, sex) then random sampling within a
strata.

Cluster Population divided into units or clusters each

containing individuals in a range of
circumstances (e.g. different type of young
offender institutions could be sampled).

Multi-stage An extension of the cluster sample, in which

samples are drawn from within clusters (e.g.
from within young offender institutions
identified through cluster sampling).

Table 6: Non-probability samples

Type of sampling Selection strategy

Convenience Select cases based on their availability for

the study.

Most similar Select cases that are judged to represent

similar cases conditions or, alternatively,
very different conditions.

Typical case Cases selected that are known beforehand

to be ‘typical’ and not to be extreme.

Snowball Group members identify additional members

to be included in sample.

Quota Sample selected that yields the same

proportions as the known population on
easily identified variables.

Activity 3: Sampling
For each of these topics:
 victims of a particular crime nationally (for example, robbery);
 prison inmates in Britain;

 homeless people in a city; and

 intravenous drug dealers.

1. Identify the type of sampling that might be possible (i.e. probability or non-probability sampling
and the specific type of sampling).

2. Consider how accurate each type of sampling is likely to be.

3. Rephrase the population description so that a more accurate sampling frame could be
constructed.

Chapter 5 Common Research Methods

In this chapter we will be covering the most common methods for collecting primary data as well as
research using secondary data.

34
Each method will be considered and illustrated with examples of research undertaken by ICJS staff.
The main advantages and disadvantages of each method will be outlined. We have also included a
range of activities that should help you further your understanding of each method.

Questionnaires

Questionnaires are one of the most common ways (along with interviews) of conducting social
research. Questionnaires are used when we want to measure the characteristics or opinions of people.
Sometimes we set out to send our questionnaires to a representative sample (probability sample),
sometimes to a whole population. Questionnaires are therefore a method used when we want to be
able to say something about a whole population from the sample we have drawn. However,
questionnaires are also administered to non-probability samples, where we cannot identify the whole
population. Questionnaires are often seen as a positivist research device in that they usually
concentrate on quantifiable responses and their analysis may involve the use of statistical tests.
Questionnaires come in many forms and with a variety of question styles, from the ‘tick box’ or closed
question, through attitudinal scales to open-ended questions in which a written response is sought.
Questionnaires may be administered face-to-face as a structured interview, in a group interview,
through the post, over the telephone or via the internet. Questionnaires that are delivered via on-line
survey sites are becoming increasingly popular.
There are a number of key factors you need to consider in terms of designing a good questionnaire i.e.
a questionnaire that is both valid and reliable and which ensures a good response rate.
1. Authority and clearance: for example: who you are; your name and contact details; who gave
you clearance to undertake this work and where they can be contacted. This information is usually
provided in a covering letter or in an introductory paragraph.
Layout and appearance: for example, make sure there is enough space to write in answers; make clear
whether people must tick only one box in response to a particular question; group questions under
themes; and consider using coloured paper (a yellow or orange questionnaire) as it is easier for people
to recall when you follow-up with letters or phone calls.
2. Clear instructions and question wording: for example, ‘If your answer is ‘no’ to Q5, then go to
Q10’. Avoid presuming, leading or unclear questions.
3. If using a postal questionnaire include a stamped addressed envelope or another way of
returning the questionnaire for free, but confidentially (for example, internal post if within a
workplace or organisation).

4. Timing: avoid obvious holiday period or work ‘bottlenecks’ when sending out a
questionnaire.

Advantages Disadvantages

Large number of respondents If postal or internet questionnaire, possibility of

low response rate (few completed
possible questionnaires returned)
Representative sample possible Questions cannot be explained to respondent
Questions can be standardized and can be misinterpreted

Statistical analysis (descriptive and/or May be difficult to locate answers in real world
inferential possible context

Respondent has time to consider questions Likelihood of socially desirable responses to

certain questions
Inexpensive way to cover a large geographical
area
Questionnaires can be replicated and used in
later studies.

Table 1 Advantages and disadvantages of using questionnaires in research

Table1 shows the key advantages and disadvantages of using questionnaires in research. A key
advantage of questionnaires is that they allow data to be collected from a large number of people. This

35
in turn enables you to obtain a representative cross-section of a wider population In addition
standardised questions allow comparison between respondents’ answers to questions. The data
obtained can be displayed in tables and graphs, and depending on the nature of the data collected,
inferential statistical tests can be performed on the data to look at relationships between variables.
Questionnaires are not useful if you want very detailed information about the reasoning behind
respondents’ views, especially if you are asking questions about sensitive issues and people’s thinking
behind their views. In addition questionnaires often rely on people making a judgement or giving a view
at a particular point in time. People do not always remember things accurately and may change their
view if asked the same question at a different point in time. In these cases in-depth interviews would
usually be more effective. Getting people to respond to postal questionnaires can be very difficult and
time consuming. It can also be relatively expensive if you do repeat mailings, with SAEs (self-
addressed envelopes), to a large number of people. People may also answer some questions in a way
that presents them in a favourable light; we call this the socially desirable response bias.

Types of Question
There are two basic types of question:
1. open – a space or maybe a few lines are given to the respondent to fill in a written response;
and
2. closed – a number of alternative answers are given to the respondent and they are usually
expected to ‘tick a box’ or ‘circle a number’.

The advantage of open questions is that respondents can answer in the way that they want to; they are
not forced to choose between the alternatives you have given them. However, if the main purpose of
your questionnaire is to produce statistics and graphs, you should use closed questions because they
make it much easier to count up the different answers. This is made even easier if you give each
alternative answer a number before you give the questionnaires out. This is called pre-coding.
Activity 1: A badly designed questionnaire
The following questionnaire contains both open and closed questions.
Identify which questions are open questions and which are closed questions.
Next, go through the questionnaire below and identify some of the problems in way these questions
have been written. (All the questions have at least one problem regarding the way they are written.)
----------------------------------------------------------------------------------------------
Could you please answer the following questions?
1. What is your name and address?
2. How old are you? Under 15 15–30 30–50 50 or older

3. Sex?

4. What is your social class?

5. How long ago were you mugged?

6. How much money was stolen?

 £5 or less

 £3–15

 £15–25

 £25 or more

7. Do you ever buy stolen goods?

8. Who do you get these stolen goods from?

9. Would you agree that criminals are all bad and evil people who should be in prison?

10. Since being mugged is it true that you have been out of your house after dark less often?
36
Thank you
----------------------------------------------------------------------------------------------
These are some of the problems you might have identified:
1. leading questions: questions that encourage the person answering to say ‘yes’ or ‘no’ because
of the way in which they are worded (Question 7; Question 9)
2. presuming questions: questions that assume people have undertaken a particular action
(Question 5: it is not clear whether all the respondents have been mugged; Question 8; Question
10)

3. double questions: questions that have more than one point (Question1, Question 9)

4. unclear choices: questions where answer choices are unclear or vague (Question 4)

5. difficult wording: questions that use difficult words or assume knowledge or agreement about
the meaning of a term (Question 3: is this a question about the person’s gender or about their
sexual orientation?; Question 4: it is assumed that the respondent understands the term ‘social
class’)

6. overlapping categories: pre-coded questions in which the options given to the respondent
overlap (Question 2, Question 6).

Extract 1
The following extracts, from four different questionnaires, offer a useful insight into the
many techniques and question types that can be used to gather information from
respondents.

Investigative Interviewing
Background
PC Colin Clarke and Dr Becky Milne are conducting an evaluation of PEACE investigative interviewing,
with a Police Research Group award. As a first step in this project it is necessary for them to identify
how PEACE training is currently being delivered and which police forces in England and Wales practice
the management and supervision of investigative interviews: whether witness, victim or suspects. This
will provide background information for the report and help with the identification of sites at which to
undertake the main study. All the information will be treated as confidential by the researchers.
1. What is your role?
2. The PEACE course:
(a) What percentage of your force is PEACE trained?
(b) What is the length of your PEACE course?
(c) How long do you spend on witness/victim interviews?
(d) How long do you spend on suspect interviews?
(e) What are the selection criteria for officers attending a PEACE course?
(f) Is there anything else you would like to add regarding PEACE?
3. Supervision of interviews:
(a) What is your policy regarding interview supervision?
(Please attach document if necessary)
[If you currently have no supervision policy in place, please go to Question (f)]

37
(b) Who supervises investigative interviews (e.g. what ranks/grades etc.)?
(c) What form does the supervision take?
(d) What is the suggested sampling rate (e.g. 2 tapes per officer per year)?
(e) Is supervision conducted on?
(i) live cases or yes/no
(ii) those that have been completed yes/no
(iii) other (please specify)______________________
(f) Is supervision conducted on?
(i) suspect interviews yes/no
(ii) victim interviews yes/no
(iii) witness interviews yes/no
(g) If you currently have no supervision policy:
(i) Are there plans to introduce interview supervision in the future?
(ii) Does interview supervision take place anyway?
(iii) What form does this supervision take?
(h) Is there anything you would like to add regarding the supervision of investigative interviews?
4. Would you be interested for research into investigative interviewing to take place at a command unit
in your constabulary? Yes/No
If your answer is ‘yes’, please provide your name and contact telephone number.
Thank you for your cooperation.

Extract 1 Questionnaire (adapted from Clarke & Milne, 2001)

Extract 2

38
 Extract 2 Part of a questionnaire (Ellis & Tedstone, 2001)
his questionnaire also used a Likert scale to identify attitudes towards the statement in
question. The Likert scale presents a statement or a question and the respondent is offered a
selection of responses, such as ‘Strongly Agree, Agree, Neutral, Disagree, Strongly Disagree’.
The respondent can be offered five or seven choices thus allowing a ‘neutral’ or ‘Do not know’
option. Alternatively, an even number of choices can be offered, forcing the respondent to
choose a response in one direction or the other.
For example:
Where necessary, the Pre-Sentence Report informs me of treatment available for a young person with
mental health needs:
(Please circle only one option)
Strongly agree 1, 2, 3, 4, 5 Strongly disagree
This five-point rating scale allows a neutral response (3) for people who neither agree nor disagree with
the statement. The following extract (Extract 3) illustrates a forced choice design, in which the
respondent is forced to provide an opinion in one direction or the other.

39
Extract 3

Extract 3 Part of a questionnaire (Shawyer, 2001)

Consider now the following questionnaire, which aims to identify young people’s experience of criminal
and antisocial behaviour. This was targeted at a very different sample of people and has been
appropriately designed and worded.

Extract 4

40
41
 Extract 4 Part of a questionnaire (Hayden & Martin, 1998)
Activity 2: Questionnaire design
As the four extracts above illustrate, there are many different ways of asking questions on a
questionnaire and all bring with them particular considerations in terms of analysis.
Questionnaires that are used in well-established surveys can also provide you with examples of different
types of questions.

Introduction: Interviews

In many ways interviews are one of the most obvious ways of finding out about people because you are
talking to them. However, this apparent simplicity can be deceptive. In many ways a research interview
is a highly skilled activity. Robson (1993) characterises interviews as a kind of conversation: a
conversation with a purpose. Whilst the interview is a kind of conversation, it does demand different
conventions in terms of social interaction, in comparison with an ordinary conversation. In part, these
42
different conventions relate to the level of structure in the interview and thus the level of control the
interviewer exerts over the pace and direction of the interview.

Structured and Unstructured Interviews

Highly structured interviews will have preset questions as well as pre-agreed probes or prompts if a
response is not forthcoming, and an expected sequence of questions. Highly structured interviews are
one way of administering a questionnaire and are a way of ensuring a better response rate than, say,
postal questionnaires. Semi-structured interviews typically have an interview schedule (a list of
questions with probes or prompts); but these are likely to be mostly open questions that do not have to
be followed in a particular order. Unstructured interviews may start with a single general question or
other stimulus (a picture or an account of a case or story, for example) and then allow the respondent to
talk about what they want, for example to tell the interviewer about an issue or experience.
There are a number of key things to remember about conducting good interviews, whether they are
structured or unstructured:
1. consider how you present yourself to the interviewee (for example, a formal interview with a
senior person is likely to call for a formal presentation of self, whereas an interview with a person in
their home might be much less formal);
2. listen more than you speak;

3. put questions in a straightforward, clear and non-threatening way;

4. avoid leading questions;

5. look like you are interested; and

6. take a full record of the interview (for example tape, after asking permission, or take notes or
shorthand if you know how).

‘In-Depth’ Interviews
In-depth interviews are not as ‘structured’ as questionnaire interviews. They are more like just talking to
somebody with some particular issues and questions you know you want to cover with that person.
Your interview schedule (list of questions and prompts) does not necessarily have to be followed in the
same order with each interviewee. By doing in-depth interviews you can allow the respondent to give
more detailed answers and express his/her own views. You may wish to tape the interview if you have
the equipment and the respondent is comfortable about having their interview taped.
The main reasons why you might prefer less structured (and particularly in-depth interviews) to
structured interviews in your research become clear when the two types of interview are compared:

43
 structured interviews emphasise reliability: i.e. how accurately different respondents’ answers
can be compared; whilst
 in-depth interviews emphasise validity: i.e. how close answers get to the respondents’ real
views (Langley, 1987, p. 24).

(The concepts of reliability and validity were discussed in Chapter 4: Key Concepts in Research
Methods.)

Advantages and Disadvantages of Interviews

An important advantage of research interviews is their adaptability. A skilled interviewer can follow up
the thoughts, feelings and ideas behind the responses given, in a way that questionnaire completion
cannot capture. Having obtained access to a person or group of people to interview, you are
guaranteed some sort of response. (This is in comparison to relying on people’s willingness to complete
and return a postal questionnaire, for example). Questionnaire responses have to be taken at face
value, whereas responses to interview questions can be clarified and expanded on the spot.
There are advantages specific to different forms of interview, as outlined in Table 2. In many ways the
structured interview shares many of the advantages of types of structured questionnaire (a potentially
large and representative sample; standardised questions and the possibility of statistical analysis;
relatively high on reliability). Validated structured questionnaires (that is, questionnaires that have
already been used and shown to measure what they say they do), which can also be administered as
part of an interview, are highly respected for their reliability. Less structured and in-depth interviews
allow for more exploration and understanding of responses (they are more likely to elicit valid
responses). With a skilled interviewer more open and honest answers can be forthcoming in
unstructured interviews, in comparison with structured interviews or postal questionnaires. Structured
interviews are generally easier to analyse than in-depth and unstructured interviews.
The most obvious disadvantage of conducting any type of research interview is that they are time
consuming, particularly if they are taped and fully transcribed.
Transcribing is the term used to describe the process of transferring the dialogue from a tape into a
written form. If you have recorded unstructured interviews then it will be necessary to transcribe them at
some point. Transcribing is an extremely lengthy process. It is unlikely, in fact, that you will be able to
transcribe entire interviews, unless you either have a lot of free time or help to do this. Instead, you will
need to listen to the tape carefully in order to identify the sections of the interview that are directly
relevant to the project. Once you have located them, then you need transcribe only these sections of
the interview. If you have made notes on the interviews then, again, you will need to look carefully
through your notes in order to identify the relevant parts.
As with the advantages of conducting research interviews, disadvantages vary with the type of
interview. Structured interviews inevitably limit responses and the data obtained may not be reliable if
there are faults in the way questions are asked or understood by the respondent. Structured interviews
can be very limiting in terms of allowing any real exploration or understanding of the responses given.
In highly structured interviews the key differences, in comparison with postal questionnaire completion
by the respondent, are that the researcher can explain questions that the respondent does not
understand or provide prompts to help the respondent answer the question. Standardised prompts are
agreed beforehand. The sample size for in-depth and unstructured interviews is generally small and
may not be representative of a particular population. Some would argue that it can be difficult to
compare the results of in-depth and unstructured interviews in that they may be very specific to a
particular interaction (the research interview itself).
Advantages Disadvantages

Structured interview Can reach a large sample Respondents are ‘forced’ to

choose between the alternative
answers the interviewer gives
them.

A representative sample is It can be difficult to obtain

possible and the results used reliable data on attitudes,
to makes statements about opinions and values (unless
the population validated questionnaires are

44
 used).

Questions are standardized The respondents may feel

so that respondents’ answers inhibited and not be entirely
can be more easily compared honest in a written response.
and in some cases analysed
using statistical methods.

The interviewer has to stick to

the agreed questionnaire, even
though interesting lines of
enquiry might emerge in the
interview.

Time-consuming – in terms of
data collection (compared with
postal questionnaires).

In-depth interview Respondents can answer Only a relatively small number

question in as much details as of interviews can take place
they want. because each one can last a
long time and generate a large
amount of data.

More valid information can be It is difficult to compare the

obtained about respondents’: results directly because each
attitudes; values and beliefs – interview is unique.
particularly about how people
explain and contextualize
these issues.

The informal atmosphere Because your sample size is

encourages the respondent to small, your results are unlikely
be open and honest. to be representative of a
particular population.

Flexibility – the interview can Time consuming – both in

adjust questions and change terms of data collection and
direction as the interview is data analysis.
taking place.

Table 2 Advantages and disadvantages of structured and in-depth interviews

Table 2 Advantages and disadvantages of structured and in-depth interviews
In practice many interviews are neither completely structured nor completely in-depth, but somewhere
between the two. Figure 1 illustrates the trade-off between reliability and validity in different types of
interview.

45
(Source: Langley, 1987, p. 24)
Figure 1 Considerations when choosing type of interview

Using Semi-Structured Interviews in Research

Semi-structured interviews allow more detailed responses to be collected than structured interviews.
The interviewer has a checklist of points or themes, sometimes referred to as an interview schedule,
which must be probed during the interview. However, the interview has a more flexible format in terms
of question order and type. The interviewer is also free to explore any issues in more depth if the
occasion arises. Extract 5 illustrates this type of interview.

The Presidential Team and the Past President Study

Questions for research
The research project, which forms the bulk of this study, set out with the central hypothesis that ACPO
(Association of Chief Police Officers) had become a more corporate, strategic and effective body in
recent years and, furthermore, that this process had been reflected in two areas of policy-making. On
the one hand ACPO had become a more influential body, in influencing and shaping force or service
level policing policy i.e. it had become more influential over the forty-three individual police forces which
make up the police of England and Wales. The traditions of British policing have it that policing is or
should be, essentially, a local and localised affair and that as a consequence policing policies can and
do differ from force to force. We hypothesised that within this framework ACPO had managed over time
to bring individual forces more and more into line with a common policy for most areas of policing. On
the other hand ACPO had become a more effective body, campaigning on the national scene to
influence policing and wider criminal justice policy, including legislation in these areas. In this case it
was hypothesised that ACPO had become more effective in the external politics of representation. With
this framework of general research
questions we set out to formulate the detail of more specific lines of investigation. In the final analysis
the project was differentiated into three sections (one of which is reproduced below).
The ‘Presidential Team’ and the past president study
In order to trace the development of ACPO over time and to capture the views of key role holders within
ACPO, we set out to interview those that had been, were or were shortly to become the President of
ACPO. First, this involved a number of senior ACPO members who constituted the ‘Presidential Team’.
The Team comprised the (then) current President, the immediate past President and the First and
Second Vice-Presidents. ACPO Presidents serve for a term of one year; the retiring President
continues to serve as part of the Presidential Team for one year after his or her term of office (to
maintain continuity). The First Vice-President is the member elected to serve as President in the
following year after that. All members of the Presidential Team agreed to be interviewed.
In relation to this we also interviewed members of the ACPO Secretariat, who provide support to the
Presidential Team and the administration of the organisation, in order to gain detailed background
information. We then set out to trace as many presidents of ACPO as possible. Of those we contacted
and approached all but one (who was in fact very ill at the time) agreed to be interviewed, generating a
further ten interviews, including one past president who was president in the late 1970s. Fourteen
current and past members were interviewed therefore as part of this area of research. Interviews were
conducted through the ‘semi--structured’ interview format. The interviews were framed around the

46
following major themes:
(a) career background, prior to and within ACPO;
(b) the extent of ACPO’s influence over the time of the respondent’s period within ACPO;
(c) ‘watersheds’ or key factors in ACPO’s development or change over time;
(d) views on how ACPO functions and/or how it did function, including sources of power within the
organisation; and
(e) the nature of ACPO’s relationships with other bodies associated with policing, including the Home
Office, Her Majesty’s Inspectorate of Constabulary, local police authorities and the Police Federation.
The ‘Presidential Team and Past President Study’ enabled us to gather experiences of and views on
ACPO over a considerable period of time. Even the (then) current Presidential Team could offer
observations on the organisation stretching back some considerable way into the past. Not only did
these members have considerable experience within ACPO as active members, they had in most cases
served as ‘staff officer’ to chief officers when they were more junior, and had gained experience of the
workings of ACPO as a result. Taken together, this part of the study provided information and comment
on the role and development of ACPO going back over twenty years. The interviews also provided rich
opportunities to question some of the most famous (if not notorious) former chief constables whose
actions and reputations had much to do with the heated debates over police accountability, including the
accountability of ACPO itself, during the 1980s.

Extract 5 Using semi-structured interviews (Savage, Charman, & Cope, 2000)

Extract 6 illustrates the conduct of a more structured type of interview.

In this study,
validated self-esteem and social skills questionnaires were administered at a different time
from the interview. Validated questionnaires are common in psychology: they are
standardised questionnaires that have already been used in other research and have shown
to be a good measure of the subject or issue of interest.

Evaluation of a Crime Diversion Scheme

The following research was conducted at a prison in the UK. It was an evaluation of a scheme run by
inmates at the prison. The scheme aimed to divert vulnerable or at risk young people away from crime,
and the benefit for the young people was already being evaluated. However, my research aimed to
investigate the advantages for the (adult) inmates running the programme, as I believed that improved
self-esteem and social skills would result from this work. I interviewed the inmates individually in a
secluded and private room in the prison, and all of the 15 participants agreed to be taped. Validated
self-esteem and social skills questionnaires were also administered to the inmates at a different time,
and it was hoped that through this triangulation of methods to gain a broader account of the benefits of
the scheme for the inmates.
Each participant received the following introduction, set of questions and closure. I aimed to be
consistent with each interviewee, and to ask the questions in a neutral manner.
Introduction: (This was read to each interviewee at the start of the interview)
Thank you for taking part in my study. [Introduce myself.]
I am carrying out an evaluation study of the Crime Diversion Scheme, and will be talking to all of the
inmates who are involved in the scheme. I would like to ask you twelve questions about your
experiences and feelings about being part of the programme. This should take less than half an hour.
You do not have to answer any questions that you do not feel comfortable answering or don’t want to
answer.
I would like to tape this interview, the tape will be destroyed once I have taken the information off it. Are
you happy for me to tape this interview?
Everything that you tell me will be in confidence and anonymous, I have allocated you a number, which
will be used in my report. If there is anything you are not sure about please let me know. There are no
right or wrong answers, I just want your opinion on the scheme. Is there anything that you want to ask
me before we start?
The interview questions:

47
 What is your role in the crime diversion scheme? Please give as many details of your
activities and responsibilities as possible.
 What skills did you have before starting on the scheme that have helped you in your work on
the scheme?

 What skills have you gained from doing your work on the scheme?

 Do you think that these skills will be useful in other areas of your life, both here and when you
leave the prison? In what way?

 What work would you like to do when you leave prison?

 Do you think that the skills you have gained here will be useful in that line of work? In what
way?

 Do you think that you have changed in any way since becoming a member of the crime
diversion scheme? In what way?

 What benefits, if any, are associated with the scheme for you?

 Do you think that being part of the scheme has changed the way you see yourself? In what
way?

 Do you think that being part of the scheme has changed the way you get on with other
people? In what way?

 Do you think that being part of the scheme has changed your attitude towards crime? In what
way?

 Is there anything else that you would like to tell me about the scheme: how you feel about the
way it is run, about your colleagues or what the scheme achieves for the inmates and the young
people?

Closure
That’s the end of our interview, thank you very much for taking part in my study. Your contribution will
go towards providing an overall picture of the scheme in terms of the benefits for the inmates.
I will be asking you to fill out some questionnaires in the next couple of weeks, and I will let you know
the findings from my study later in the year.
Are you feeling OK about everything we have discussed today? Is there anything else you would like to
ask me?
Thank you again and I will see you soon for the next stage of the study.

Extract 6 Structured interviews (Shawyer, 2001)

Activity 3: Considerations when conducting interviews
1. After reading the above extracts, consider how important the status and presentation of self by
the interviewer might be in each of these studies.
2. On the basis of the information presented in this chapter, consider what kind of interview would
you prefer to do and why (unstructured, semi--structured or structured).

3. Would you tape your interview?

4. Would you fully transcribe it?

5. Estimate the total time costs of conducting a one-hour taped and fully transcribed (word for
word) interview, with a professional located half an hour’s drive from your home or workplace.

Observations

48
The various types of observation used in social research have different and quite distinct histories.
Consider, for a moment, the difference between conducting an observation of a football match as a
member of the crowd and conducting an observation of a police or psychiatric interview through a two-
way screen. They are very different activities. The former type of observation has a long tradition in
social research and is frequently referred to as participant observation. This approach places the
researcher within the world they seek to investigate and understand. The approach originated in social
anthropology and was used in attempts to understand other cultures’ way of life. However, social
researchers, from the Chicago School in the 1920s and 1930s, began to use observation as a way of
researching crime and deviance, race relations and the development of distinct social spaces in urban
areas.
Observations are used in different research designs and for a variety of purposes. Sometimes
observations are used as part of the early or exploratory phase of research, when the researcher is
trying to find out more about their object of study and develop hypotheses and research ‘instruments’
such as questionnaires and interview schedules. In other types of research it may be one of a number
of methods used as part of a triangulation strategy; for example, to see whether what is observed is
what is recorded in agency documents or what is said at interview. Observation may be the main or
primary method used in some studies. Observations can also be undertaken in controlled conditions,
such as in a laboratory or in special situations where practitioners are undertaking an activity.
Whilst the ways observations are conducted vary a great deal, there are a number of key principles you
need to adhere to when conducting an observation:
 decide on your role in advance (participant or non-participant, overt or covert) and how you are
going to manage this role in practice;
 think through and have a clear strategy for dealing with ethical and safety issues that can arise
in the course of observations;

 consider who should (ethically and practically) be informed about your role;

 maintain an awareness of your impact (or potential impact) on the observation;

 be aware of how your own experiences and views will affect your observations; and

 decide how you are going to record (degree of structure) and analyse your observations.

Observation involves looking and listening very carefully. We all watch other people sometimes, but we
do not usually record our observations. Observations can be of two types:
1. non-participant observation – where you would observe people’s behaviour without joining in
any way; and
2. participant observation – where the researcher deliberately joins in with the activities of a
group while observing them, as in Extract 8.

You will need to work out how you are going to keep a systematic record of your observations. To do
this, observations can be further categorised into those that are structured and non-structured. As
with interviews there are different levels of structure between ‘structured’ and ‘non-structured’
observations. Non-structured observations generate more subjective ‘narrative accounts’, often
providing a more in-depth insight into the phenomena being observed. They can also be used as a pilot
to more structured coded observations. Structured observations involve the use of an observation
schedule, which is strictly followed by the researcher. These schedules can be evaluated for reliability
by using two observers and testing consistency of observations (inter-observer agreement) or by testing
the same observer at different times (observer consistency). There are standardised observation
schedules available.

Advantages and Disadvantages

A key advantage of conducting observations is that you can observe what people actually do or say,
rather than what they say they do. People are not always willing to write their true views on a
questionnaire or tell a stranger what they really think at interview. Observations can be made in real life
situations, allowing the researcher access to the context and meaning surrounding what people say and
do. There are numerous situations in the area of criminology where approaching people for an interview
or questionnaire completion is unlikely to yield a positive response, but where observations could yield
valuable insights (see Extract 8).

49
There are a number of problems associated with observational research. A very important one
relates to the role of the observer and what effect he or she has on the people and
situations observed. This is difficult to gauge. There is also the additional problem of being
able to write an account, as a researcher, when one is immersed in a situation or culture.
This latter situation can mean that the research is dismissed as too subjective. Observation
can be very time consuming. Some well known observational pieces of research took some
years of observation and immersion in a situation or culture. However, it is more common
in modern research to reduce the observation time substantially. Observation time may be
further reduced in experimental conditions (laboratory or simulation) or in real life
situations: in other words, controlled settings. An important potential disadvantage, in
conducting observational research, is the ethical dilemmas inherent in observing real life
situations for research purposes.

Advantages Disadvantages

Access to situations and people where Can be viewed as being too subjective
questionnaires and interviews are impossible
or inappropriate to use

Access to people in real-life situations Time consuming

Good for explaining meaning and context Role of researcher: if known, this may affect
the situation and thus the validity of the
finding; may also raise ethical concerns

Strong on validity, in-depth understanding High potential for role conflict where
researcher (observer) is also the practitioner

Often produces interesting and lively accounts

of the social world

Table 3 Advantages and disadvantages of conducting observational research

Activity 4: The researcher in the ethnographic account
Read this extract about participant observation
1. Consider your own personal characteristics and biography. How might these impact on a
research situation such as the one above?
2. Consider, in turn, another area you might wish to research. How might your own personal
characteristics and biography impact on the research?

3. Notice how the author integrates other academic arguments with quotes from people within the
research context and his own line of argument.

4. Identify the ways the author of this extract validates his observations.

Introduction: Secondary Data

The three research methods discussed this far in this chapter involved the collection of new or original
data ‘in the field’ (primary data). Another way of conducting research is to analyse or re-analyse data
that has already been collected, either by other researchers or for another purpose. This type of data is
called ‘secondary data’. Secondary analysis is defined as:
Any further analysis of an existing data set which presents interpretations, conclusions or knowledge
additional to, or different from, those presented in the first report.

50
(Hakim, 1982, quoted in Robson, 1993, p. 282)

Analysis of Crime Statistics Already Collected

Crime statistics can be re-analysed (through secondary analysis) to investigate and help answer
different questions from those originally intended. Indeed some criminology research is based totally on
secondary analysis of existing data sets.
On the one hand, crime statistics may simply be used to summarise what is already officially known
about an issue. In this sort of study an analysis or summary of existing statistics would be used as a
supplementary research method, rather than the main method. For example, if you were planning a
local survey on fear of crime you might want to refer to national data from the British Crime Survey as
the national context to what you were doing locally. On the other hand, existing statistics can also
provide the main data for a piece of research.
A key issue to consider when analysing or re-analysing statistical data is the purpose for which the data
was collected. What is missing or not collected (perhaps even not collectable) is as important to note as
what is collected. In other words, statistical data sets share some common problems identified in other
types of data sets or documents.

Advantages and Disadvantages

The advantages and disadvantages of analysing existing data sets of crime statistics are shown in
Table 4.

Advantages Disadvantages

Access to large samples Data quality may vary

Access to representative samples Limits to questions that are asked and thus
can be answered by further analysis

Relatively low cost Questions not always asked in the same way

Longitudinal analysis possible Data collection and recording can be influence

by political consideration and target setting

Focus can be on quality and depth of analysis Data may be needed to be adjusted for
comparability over time

Data collection already paid for

Can capitalize and build on the research

already done using the data

Table 4 Advantages and disadvantages of analysing existing data sets of crime statistics
A key advantage in the analysis of existing statistics is that the data is already collected for your use.
Data may be local, regional, national or even international. In other words, the data can cover a much
larger area than you could hope to do as a lone researcher. Some of the data will be drawn from
representative samples, enabling more sophisticated analysis and the use of particular statistical tests.
It is a relatively cheap form of research, in that it may not involve much (if any) travel. In many cases
you will be able to download existing data sets from data archives, the Internet or from within your
organisation. Some statistical data within your organisation may not be available outside the
organisation and may have been subjected to only very limited analysis. As data collection is very time
consuming, a big advantage of analysing data that is already collected is that you can concentrate on
the analysis. Furthermore, you can build upon the findings of other researchers using the same data
set.
A key problem with existing statistics is that they may not include answers to some of the questions that
interest you. The recording of some data may also be influenced by political and other priorities. It is
well known that official records of crime (and indeed other ‘official’ data) are not a record of all crime

51
committed, not least because not all crimes committed are reported. A key issue in the analysis of
existing statistics is the quality of the data collected. Is it accurately recorded and does it measure what
it claims to measure? For example, official statistics on illegal drug use can provide estimates based on
self--report data, convictions for possession and handling, and other matters of record. Such data
should take into consideration how and why the data was collected and what is not known or not
recorded.
One protection against the problem of reliability in crime statistics is awareness, knowing that the
problem exists. Being aware and critical of how particular crime statistics are collected will mean that
you should be able to make some assessment of the relative reliability of the data and its impact on the
conclusions you draw.
Activity 5: Crime statistics
1. Access the Home Office section; and find statistical information on a subject that interests you.
2. Consider how you might use available statistics if you were writing a report on organized crime.

Introduction: Documentary (Content) Analysis

A great deal of evidence about many research topics that police officers might wish to pursue may have
already been collected in the course of their work or the work of their colleagues. ‘Document’ is a term
used for ‘an impression left by a human being on a physical object’ (Travers, 1964, cited by Bell, 1989,
p. 53). ‘Documents’ are usually understood to be printed sources but can include film, video, tapes and
other non-written sources.
Documentary analysis is commonly referred to as content analysis, although there are some specific
aspects to content analysis that might separate it out from other forms of documentary research.
Documentary research in its various forms is usually an unobtrusive way of conducting research.
Instead of taking up people’s time by asking them to take part in an interview or complete a
questionnaire for the purposes of research, we use a document produced for some other reason as our
source. An exception is when researchers ask their subjects to keep a diary for research purposes. In
this latter case the document produced (that is, the diary) is still amenable to content analysis but the
method is no longer unobtrusive. By asking people to keep diaries, a change in their behaviour might
happen that affects what they write and what is available for the researcher to analyse (Robson, 1993).
Content analysis, as a research method, grew in importance in the social sciences during the early
twentieth century, via a series of quantitative analyses of the content of newspapers in the US. The
general tenor of the debate, arising from this research, was to do with the notion that gossip and cheap
journalism was reducing the coverage of ‘worthwhile’ news – a familiar contemporary debate. This type
of analysis was extended to radio, then television and other forms of media (Robson, 1993). The study
of violence in the media and the reporting of industrial disputes as well as representations of race and
gender, are all common examples of studies using this method. Content analysis is often seen as a
way of collecting quantitative data from texts or other parts of the media. It often involves establishing
categories and then counting how often items fall into these categories, in particular reports (Silverman,
2001). However, there is no reason why content analysis should merely be treated as a way of
collecting quantifiable data.
Like any other form of research, the nature of the data collected from documents and the way it is
collected and analysed depends on the main purpose of the research and the underlying epistemology.
Positivists are likely to prefer some form of content analysis. This may be particularly appropriate for
documents held in a workplace, in a standardised form. In many ways this type of documentary
research is very much like conducting a survey, using a questionnaire. The data is usually collected in a
standardised and quantifiable way such as a questionnaire or form of some sort, sometimes referred to
as a pro forma (if it is very brief). Interpretivists are more interested in people’s accounts of a situation,
how they present their view of the world and so on. In this case it is common to produce descriptive or
analytical accounts of cases, using direct quotes from people as they appear in the document. Of
course, it is also possible to theorise on the basis of these accounts. In practice some people will
choose to collect both quantitative and qualitative data in their documentary research.

Advantages and Disadvantages

Advantages Disadvantages

52
Access to inaccessible subjects Bias in content

Elimination of ‘the researcher effect’ Selective survival of documents

Longitudinal analysis possible Incompleteness

Larger sample size possible Access/availability of some documents

Can contain spontaneous data Potential sampling bias

Access to confessional style documents Limitation of written and verbal behaviour

Relatively low cost Lack of standardized format

Many documents are good quality, some are Coding difficulties

very detailed

Data may need to be adjusted for

comparability over time

Table 5 Advantages and disadvantages of analysing written documents for the purposes of
research
One of the key advantages in conducting documentary research is that you can get access to
information that would be difficult to get in any other way, such as people or cases that might not be
willing to talk in a formal research interview or might be difficult to track down. By using documents you
eliminate the effect that you, as an individual, have on a person or situation when you conduct research
(‘the researcher effect’). The effect you have on a situation or subject may be partly due to the
knowledge that you are there as a researcher. People will also be affected by how you conduct yourself
and how they perceive you. Issues of sex, age, race and other characteristics are likely to have an
impact on what people tell you or do when they know they are being researched.

Documents are often particularly useful for tracking change over time: that is, doing longitudinal
research. Longitudinal research may be done either prospectively (forward) or retrospectively (back
over time). Documents often make possible the collection of data over a longer period of time, as well
as larger samples than might be collected from questionnaires or interviews. Some documents may
contain spontaneous data, such as feelings, and refer to actions that are recorded in a specific context,
not with a view to answering a particular research question. Confessional documents may give us
insight into how people see things or how they want to present things. Either way, they provide a very
particular account of reality, whatever the confessor’s motivation behind their account.
Further advantages of using documents in research include the fact that such research is relatively low
cost, particularly when the documents are easily accessible and already located in your workplace.
Documents vary a great deal in quality, often related to the perceived importance of recording certain
information, but some types of document can be extremely detailed and yield much more information
than you could hope to gain from a single questionnaire or interview.
Nevertheless, documents are usually not designed with research in mind. The information recorded
may be idiosyncratic or incomplete (or illegible!). Documents get misfiled, left on people’s desks for long
periods or simply just do not get fully completed at all. Even standard ways of collecting data in a police
station may change, so there may be information that is available for one period of time and not
another. All of this will create gaps in data (missing data) as well as coding difficulties. The motivation
to collect certain types of data will vary over time, perhaps related to the decision by a particular team to
focus on a specific issue for a period of time or because of government targets.

Doing Your Own Documentary Research

A variety of documents may already exist in your workplace, such as minutes of meetings, reports on
cases and so on. With permission and due consideration to ethics, any of these types of document
might be subject to analysis for the purpose of research.

53
Firstly, you will need to decide what kind of topic and what kind of document interests you. Secondly, if
you chose to use documents in your workplace, you must obtain permission from your line manager to
use them for the purpose of your research. You will need to consider ethical issues and data protection
in reaching this agreement. Thirdly, you should look at the feasibility of access to documents at work:
when, where, how many? You need to ascertain whether documents are in continuous use. If so, when
are you going to be looking at the documents; others will need to know so that they can find them. If you
are going to use archived documents you will need to be clear about any changes in the way data may
have been collected and again sort out the logistics of access to these documents. Some types of
document may be very numerous, so you may need to either specify your sample to within a particular
time frame or look at all cases within a very limited time frame.
You may want to consider another form of document, such as newspapers. You could, for example,
consider analysing newspaper reporting of a particular incident (for example, a riot, stalking cases, a
murder hunt or other high profile event) or issue (for example, race relations) over a specified time
frame. In doing this, you would need to consider both tabloid and broadsheet reporting as well as the
well known political stances of different newspapers.
Activity 6: Content analysis – newspaper reporting and crime
1. Look at a newspaper reporting of an issue of interest to you within the criminal justice system,
on a single day in which there is reporting on this issue.
2. Read a tabloid and broadsheet newspaper account of the same event and then note the
following:

 Title – What was the title given to the article? What did this suggest? Did the titles
give the same message?

 Location and coverage – Where in the newspaper was this article located (for
example was it in the headline, front page or buried somewhere in the middle of the paper)?
How many column inches are devoted to this news item in each paper?

 Content – Are factual details the same in each paper? Are the same or different
quotes used? Is an illustration used? Is the illustration the same, different or not used in one
account? Would you come to the same or different conclusions about what happened or what
the issue was after reading these two accounts? Note the use of language in each account,
how it differs and how the message is conveyed.

3. What do you conclude from this exercise?

Extract 8a

Persistent young offenders and exclusion from school

In the following study researchers at the University of Portsmouth and
Hampshire Constabulary conducted some commissioned exploratory research into the association
between persistent offending and exclusion from school.
Background details to the study
The population: 495 juvenile offenders who made appearances at a youth court over a one year
period.
The sample: 44 individuals all of whom fulfilled the criteria of ‘persistent young offender’ (PYO). PYO
was defined as ten to seventeen year olds who had been sentenced by a UK court, on three or more
separate occasions, for one or more recordable offences (Home Office, 1998).
Research design: retrospective longitudinal survey, using a range of secondary data from different
sources, enabling triangulation of data on each case.
Data sources: documents – court records, education department admissions and exclusions; case files
and educational welfare case files; computerised records – police national computer and social services
client database.
This research was commissioned by the local authority. However, the researchers still had to develop a
clear protocol about who obtained access to certain types of data and how it was put together and held.

54
For example, access to court records and the police national computer was negotiated by the
Hampshire Constabulary researchers and access to the educational and social services data was
negotiated by the University researchers who already had good working relationships with both
departments.
Essentially, the research involved putting together data on these 44 individuals from these sources. In
order to do this we developed a pro- forma, which is rather like a questionnaire. This pro forma ensured
that we were looking to find certain types of details from these records in a standardised way. We also
allowed for additional observations to be noted on a case.
Part of the pro forma we used to collect data from case files and elsewhere follows [see Extract 8b].

Extract 8a Using case file and other institutional data (Martin, Hayden, Turner & Ramsall, 1999)

Extract 8b

Extract 8b Part of a pro forma used to collect data

55
You will notice in the section of the pro forma in Extract 8b that ‘case number’ is written instead of
name and the form is marked ‘confidential’. These are both attempts to ensure the confidentiality of the
identity of an individual to the casual observer. If you were doing this sort of work you would give each
individual a case number and keep a list of names against case numbers separately in order to make
the person less readily identifiable should anybody come into possession of this document by mistake.
The date of birth (DoB) allows you another identifier or check when completing information on the pro
forma. You can make data collection instruments, such as this pro forma, even more structured if you
wish (for example, the question on ethnicity could follow Census categories or the question on special
educational needs could follow accepted categories and levels of need). Each pro forma allowed ample
space for additional observations and comments on the case.
Once you have collected data in the above way you then have to decide how you are going to analyse
and present it. Much of this data might best be analysed using an Excel spreadsheet. There is
guidance on the analysis and description of data later, with particular reference to online key skills
support.
The 44 cases in the study were summarised in a number of tables and descriptive accounts, but ten
cases were investigated in more depth. One of the ways of illustrating how a number of factors played
out across individual cases, was to construct and then interpret a table (see Extract 8c).

Extract 8c

Extract 8c Table to summarise some of the key factors and agency involvement in the case
studies
The researchers were also able to look in more detail at the exact timing of recorded offences and
produce a timeline of when offences took place in relation to whether young people had an educational
placement or were attending this placement and/or had been excluded from school or educational
placement. One way of illustrating this data was to produce a timeline of events; another was to
produce a summary of key factors in each case, as well as an outline description of each case. Extract
8d shows how one of the timelines looked.

Extract 8d

56
 Extract 8d. Timeline of events
Descriptive case histories were also produced, such as the one in Extract 8e.

Extract 8e

Young Person 3
Factual details
Year of birth: 1982
Ethnicity: white
Sex: female
Domestic circumstances: Lives with mother, stepfather and six siblings
Exclusion from school: Permanent exclusion in June 1996 (age 14)
School attendance: Has not attended school regularly since June 1996
(age 14)
Special educational needs: No evidence
First police records: (CYPs) starts at age 9
Additional issues/problems in evidence: started drinking in a problematic way since parents’ split in
spring 1997; reference to eating disorder on EWS file (April 1996)

57
Comment
There were records of concern (CYPs) about this young woman at a police station when she was only
nine years old. There is evidence of a variety of problems in this young woman’s background, including
the split in the relationship of her parents when she was 14 years old, following which she was said to
have started drinking in a problematic way. There is also evidence on file that she was referred for
specialist help on an eating disorder during the same period. At the same time as all these problems
were happening in her life, she was permanently excluded from school in June 1996. She never went
back to full time schooling. She received no formal education until November 1996, when she had a
few hours home tuition a week from an individual tutor, from the local education authority. This went on
until March 1996, when she was offered a place at a pupil referral unit (PRU).
During the period of home tuition there was no recorded offending but before this time, when no
education was received (June–November 1996), eight offences were recorded. Following the end of
her period of home tuition and after being registered at a PRU (pupil referral unit), 39 offences were
recorded. There is no record of her ever having attended the PRU. Two-thirds of her offences (29)
were for shoplifting. When the overall pattern of offending is looked at, in this case, there does appear
to be a link between records of increased offending and being out of education. However, there is a
long and complicated history of other problems in this case, which started prior to her being out of
school.

Extract 8e Descriptive case histories

As the descriptive case history in Extract 8e illustrates, a great deal of data already exists as part of the
work of those agencies that work with people. However, conducting a study such as the one illustrated
does require a clear, information-sharing protocol and adherence to the requirements of the Data
Protection Act. If you are employed by such agency you will need to consult the Data Protection Officer
for your workplace in order to conduct a survey such as this.
Activity 7: Access to documents in your workplace
1. Find out what kinds of documents you might use in your workplace, in an area or topic that
interests you.
2. Ask your line manager what restrictions might be placed on any of these documents, if they
were to be used for research purposes.

3. Take a look at a document of your choice (one which you have legitimate access for the
purpose of this exercise) and evaluate how you might use it for a research study.

 What kind of document is it? e.g. court report, case file, standardised form, report to
funders of a ‘special project’ or colleagues in another agency, minutes from meetings.

 What is the audience for the document?

 e.g. courts, line managers, external agencies.

 Who produced the document and why?

 Has the document set out to make a case for a line of action, continued funding or
way of working?

 Is compliance in producing this record or document an expected part of the job? If so,
why?

 When and in what circumstances was the document produced?

 Is it typical or atypical of its type?

 Is the document complete?

 Is there any evidence about how long after the event the document was produced?
What influence might this have had on what is in the document?

 Note how long it takes to read through the document.

 Note gaps in information and any other problems.

58
  Consider how you would analyse documents like this for research purposes.

4. Review the impact the above questions or issues have had on the document you have looked
at.

An important part of analysing documents critically is assessing what is factually correct (this may
include such things as ethnicity and nationality, or establishing that a particular event happened) from
what constitutes an opinion or judgement about an individual or event. When interpreting opinions and
judgements in documentary research one needs to consider the evidence presented. If you detect bias,
this does not mean that you dismiss the document for research purposes (Bell, 1987). You need to
analyse why this bias might have occurred. Is this happening systematically (for example, from a
particular source) or is it random? Can the apparently biased sources be compared with those that
appear to be unbiased?
In some ways it could be argued that the techniques of documentary analysis are just the application of
common sense (Bell, 1989). This is partly true, but experience of analysing documents systematically
and critically is more than common sense, in that it calls into play the kind of academic research skills
that should lead to critical scholarship and a greater level of understanding of the object of study.

Narrative Analysis
Systems of narration within crime literature and other texts that give accounts of crime and criminals
(court reports, newspaper reports, letters and so on) are all forms of text potentially available for
analysis for research purposes. Texts and talk can be viewed as creating their own realities. For
example, Propp (1968) argues that the fairy tale has established a narrative form that is central to all
story-telling. The apparent detail of a story can be reduced to an underlying structure in which key types
of character play a particular role or function within the overall structure of the story.
Narrative analysis is more concerned with the process or way in which talk and text depict reality, rather
than whether this depiction is true or factual (Silverman, 2001). A narrative thus becomes the way a
story or event is commonly told or explained. Certain types of event are often depicted in a particular
way; this is easily seen in the media reporting of events.
However, there are established narrative forms for all kinds of talk and text, including research reports.
The primary data collected through interviews may therefore be seen as one way of getting at individual
experience or perception. Narrative analysis of interviews and other types of talk and text are another
way of doing this. The similarity that can be found in certain accounts of reality lead some researchers
to question whether research is accessing true perceptions or meanings, or whether a narrative is being
actively constructed. This construction of a narrative is a two-way process, between the interviewer and
interviewee. Narrative analysis can be used with published accounts and stories or interviews from data
archives. Although it can be seen as a method that uses secondary data, this type of analysis can also
be used on primary data, from unstructured interview transcripts to the analysis of conversations and
other verbal exchanges. We will concentrate on narrative analysis as a form of secondary data
analysis.

There are similarities between content analysis and narrative analysis (Silverman, 2001). For
example, it is possible in both forms of analysis to quantify and tabulate particular themes or
terms used in an account. The key difference is in the attention to the structure of an
account, the social and other processes revealed, and the development of theory in
narrative analysis. One application of narrative analysis is examined in the following extract.

‘Art of the Ugly Soul’– an analysis of the writings of murderers

One is limiting art much too severely when one demands that only the composed soul, suspended in
moral balance, may express itself there. As in the plastic arts, there is in music and poetry an art of the
ugly soul, as well as an art of the beautiful soul.
Frederich Nietzche
Popular criminology surrounds us in its various forms. It includes: true crime literature; crime fiction;
newspapers; magazines, both specialist and general; television in the form of drama; drama
documentaries; straight documentaries; news; TV reality shows such as Police Action Live or Police
Camera Action and reconstructions such as programmes like Crimewatch. Then there is film and
theatre and, last but not least, the Internet. This genre is huge, amorphous and growing daily and yet it
is routinely neglected, at best, and ignored, at worst, by academic criminologists. This is what intrigued

59
me and led to my research. This concept that one form of criminology could be so widely followed and
yet pushed away by its more stately cousin. Popular criminology is full of possibilities for the academic
study of crime. It presents a rich seam of narratives and constructions that are ready, available,
accessible and crying out for attention, analysis and deconstruction. The chief aim of my research is to
promote this message, reverse this trend, address and perhaps, to some extent, explain this anomaly.
Most of the analyses undertaken in its various forms have used the accounts of crimes and criminals,
written by others. In this part of the study some research has been conducted that explores the self
perceptions of criminals with a view to discovering more about, as cultural criminologist Jack Katz
(1988) describes it, ‘the foreground of crime’ (p. 4). I am focusing on the most extreme and violent of
crimes, homicide, as the availability of material for analysis is much greater than for less publicised
offences and offenders. The aim is to find out something about how it feels to commit such a crime and
what motivates the perpetrator to move from the relatively ambivalent ‘before’ to the lethal ‘after’. As
Katz says:
whatever the relevance of antecedent events and contemporaneous social conditions, something
causally essential happens in the very moments in which a crime is committed. The assailant must
sense, then and there, a distinctive constraint or seductive appeal that he did not sense a little while
before in a substantially similar place. Although his economic status, peer group relations, oedipal
conflicts, genetic makeup, internalised machismo, history of child abuse, and the like remain the same,
he must suddenly become propelled to commit the crime. Thus, the central problem is to understand
the emergence of distinctive sensual dynamics.
(ibid)
Firstly, it is important to outline the methodology used. This is that of narrative analysis which explores
the ways in which a story is told. Essentially this form of analysis picks apart the ways in which stories
are told, in such a way as to shift attention from the content of stories to the structure and process of
their telling. This in turn tells us more about the teller, their motivations, tastes and discriminations
rather than allowing us to make assumptions about these subjects (Lusted, pp. 53-–55). It is, therefore,
a particularly useful analytical technique when one is exploring the reasons behind somebody’s actions.
The way they tell their own story and account for their own actions can tell us much about their
emotions, motivations and sense of self. When looked at in conjunction with other accounts
comparisons can be made and any patterns that exist will emerge and be easily identified.
Russian folklorist Vladimir Propp analysed a hundred fairy tales and the resulting findings led him to
establish an analogy between the structure of language and how narratives are organised. He identified
thirty-one functions or fundamental components that formed the basis of any tale. These were units of
‘narrative language’ and included elements such as, ‘a difficult task is proposed to the hero’ of which
there were twenty-five and ‘the villain is punished’ of which there were thirty. These various functions
were distributed amongst seven spheres of action such as villain, donor and helper. The structures that
appear in murderers’ stories are different but arguably they are narratives in the same way as Propp
(1968) proposed.
Narrative Structure
1. Childhood.
2. Relationship with parents.

3. First sexual encounters.

4. First relationships with women (or men).

5. Events that suggest later crime.

6. Crime.

7. Time up to arrest.

8. Interrogation.

9. Confession.
Extract 9 Narrative analysis (Bretherick, 2006)

Advantages and Disadvantages

60
Narrative analysis of secondary sources (narrative analysis can be conducted on primary data too) has
some key advantages common to other forms of secondary data analysis. Key advantages include the
fact that the data is already collected so that the focus can be on analysis rather than data collection.
This is a relatively low cost research method. Because the data is already collected the researcher
cannot influence the account given of an event or experience: that is the ‘researcher effect’ in terms of
data collection is eliminated. Narrative analysis can make possible access to inaccessible, or difficult to
access, research subjects (as in the previous extract). Perhaps because we are used to written
interpretations of events, narrative analysis can produce accessible and interesting insights to areas of
criminology.
Nevertheless, narrative analysis is open to the criticism that interpretation can be a highly subjective
activity. This latter disadvantage, coupled with the likelihood that accounts of lives and events
might have significant gaps, means that the possibilities for generalisation from such analysis
is open to question. Narrative analysis can also be time consuming when numerous,
detailed and lengthy accounts have to be read before analysis and interpretation can begin.

Advantages Disadvantages

Data already collected There may be important gaps in the data

collected

Focus can be on analysis Potentially high levels of subjectivity in

interpretation

Can produce lively and interesting insights into Time consuming

human experience

No impact by the researcher on the account

given

Table 6 Advantages and disadvantages of narrative analysis of secondary data

Activity 8: Narrative analysis
1. Read an account of the background and/or newspaper reporting on a well known murder
case. Newspaper websites will provide background details about cases about famous and
contemporary cases.
2. The Metropolitan Police provide details of well-known historical cases on their website
http://www.met.police.uk/history/index.htm

3. The crime library also provides detail of a wide variety of known criminals and historical cases
http://www.trutv.com/library/crime/index.html

4. Can you apply the above narrative structure to the case you have chosen?

5. Are there other themes in this case that you would like to add?

Systematic Reviews and Meta-Analyses

Good quality reviews of research that have already been completed can be extremely important
research in their own right. Here we will discuss two such types of review: the systematic review; and
meta-analyses.

Systematic Reviews
The systematic review is a newly popular approach to reviewing academic literature using empirical
research methods. The systematic review became popular in medical and health services research in
the 1990s, taking hold in the social sciences in the new millennium. Now preferred in some areas of
criminology, it is typically used for the purpose of evaluating the effectiveness of an intervention in order
to support evidence based policy making. In recent years there has been much discussion across the
public sector about the need for good quality research evidence that can inform the development of
policy and practice. Such reviews address the common failures of social science and other disciplines
to produce cumulative knowledge on given hypotheses, since studies carried out into a similar issue

61
have frequently provided differing results. Systematic reviews purport to bring together the ‘best’
research that is available on a particular topic, employing transparent and systematic criteria for
searching the available literature, evaluating the suitability of each source for inclusion in the review,
and synthesising the results. The approach followed in the systematic review is described as fully and
systematically as would be any empirical research.

What Does a Systematic Review Involve?

The systematic review offers a rigorous and replicable approach to reviewing the literature, for the
purpose of eliminating the subjectivities of reviewer bias. It is also seen to address the uncertainty as to
whether all the evidence has been identified and evaluated. The Campbell Collaboration is an
international research network that produces systematic reviews of social interventions in order to
inform policy-making, including criminal justice interventions, and which has been instrumental in raising
the status of the approach within criminology. On its website (www.campbellcollaboration.org) the
Collaboration outlines four components of a systematic review:
 clear inclusion/exclusion criteria;
 an explicit search strategy;

 systematic coding and analysis of included studies, and

 meta-analysis (where possible). (See the section at the end of Chapter 5 for a fuller
explanation of meta-analyses.)

The attractions of the systematic review for policy-makers are:

 the explicitness of the research question to be addressed;
 the transparency of the methods used for searching studies;

 that studies reviewed include unpublished work;

 the clear criteria set for including or excluding studies;

 the joint reviewing to avoid bias; and

 a clear answer to the question set. (MacDonald, 2000)

What Does a Systematic Review Look Like?

Box 1 provides an extract from the contents page of a systematic review by Farrington and
Welsh (2002) of research evaluating the effects of street lighting on crime, conducted for the

62
British Home Office. In this example, the parameters of the research are modest: rather than
interrogating a broad literature on street lighting and crime, the focus is establishing the
combined results of thirteen carefully selected American and British studies. All of the
studies adopt the same general strategy of evaluating the direct relationship between
interventions to improve street lighting in the study areas, and crime levels in those areas,
for the purpose of informing crime prevention policy making.

Box 1: What does a systematic review of the effects of improved street lighting on crime look like?

1.Background
Research on street lighting and crime
How might improved street lighting reduce crime?
Causal links between street lighting and crime
Determining what works to reduce crime
Characteristics of systematic reviews
Aims of this report

2.Methods
Criteria for inclusion of evaluation studies
Search strategies
Programmes not meeting inclusion criteria

3.Results
Key features of evaluations
Results of American studies
Results of British studies

4.Conclusions
Summary of main findings
Priorities for research
Policy implications

Extract from Farrington and Welsh (2002), available at http://www.homeoffice.gov.uk/rds/pdfs2/hors251.pdf

Strengths and Limitations of a Systematic Review

The strengths of the systematic review are also its limitations. A summary of these strengths
and limitations is presented in Table 4.

Strengths Limitations

Identifies the cumulative knowledge on a given The prescriptive approach precludes

hypothesis comprehensive coverage of the literature

63
Supports evidence-based policy making Can generate an excess of sources requiring
extensive filtering

Employs systematic and transparent literature Can be formulaic and dry in style
search criteria

Seeks to include only high quality research Is not impervious to political agendas despite
claimed methodological rigour

Provides a rigorous and replicable approach to

reviewing the literature

Seeks to eliminate reviewer bias

Addresses uncertainties as to whether all the

evidence has been identified and evaluated

Table 7 Strengths and limitations of systematic reviews

A systematic review can be a huge undertaking, generating tens of thousands of sources that need to
be filtered, depending on the scope of the topic. The field of study can of course be narrowed. Lösel
(2008) gives the example of a broad definition (sex offender treatment) versus a narrow one (hormonal
treatment of adult child molesters), the former of which will generate a large body of results and perhaps
one that is too large. Yet, he argues, the latter approach may produce only a limited number of studies
and present dilemmas of inclusion and exclusion of those limited studies, some of which may be limited
in their methodological rigour. An alternative approach based on similar principles is a rapid evidence
assessment, as employed by Joliffe and Farrington (2007) in relation to the impact of mentoring on re-
offending, in which the methods used in a systematic review are applied within a restricted timeframe.

Further Information on Systematic Reviews

Detailed information on systematic reviews can be found on the websites of the Campbell Collaboration
(www.campbellcollaboration.org) and the Cochrane Collaboration (www.cochrane.org), while Petrosino
et al. (2003) and Farrington and Welsh (2005) offer guidance on the systematic review as applied to
criminology. Examples of systematic reviews can be found on the websites of the UK Home Office
Research, Development and Statistics Directorate (www.homeoffice.gov.uk/rds), and the Swedish
National Council for Crime Prevention (Brottsförebyggande rådet - Brå) (www.bra.se/english), on such
topics as CCTV, street lighting, drug treatment and bullying.
Activity 9: Systematic reviews
1. Go to the websites for the Campbell Collaboration www.campbellcollaboration.org
2. Click on ‘The Campbell Library’ and in the ‘Search’ box, type in ‘Stress’.

3. Have a look at the reviews that are listed.

 [Tip: click on ‘Protocol’ to access the full review.]

4. Now go to the Cochrane collaboration www.cochrane.org

5. Click on ‘The Cochrane Library’ and in the ‘Search’ box, again type in ‘Stress’.

6. Notice how many reviews are generated.

7. Now try and refine your search by using the key words: ‘occupational stress police’.

8. Have a look at any of the reviews that are listed.

Meta-analysis
A ‘meta-analysis’ (a term coined by Glass in 1976) is a process used in summarising the results of a
number of different studies. It comprises a statistical synthesis of the findings. This is achieved by
identifying a common measure of effect size, such as a standardised mean difference (a way of
64
standardising a range of outcomes measured on different scales) or correlation coefficient (showing the
relationship between variables demonstrated by the combined results).
The bias, as with systematic reviews, is towards studies that are based on quantitative data, giving
particular attention to power calculations and statistical significance. Meta-analyses are usually based
on published studies, which have the disadvantage that the publication policies of some journals favour
positive results (Robson, 1993).
Criminological examples are provided by Andrews et al. (1990) on the effectiveness of correctional
treatment or Gore and Drugs Survey Investigators’ Consortium (1999) on young people’s illicit drug use.

Extract 10 is from the methods section of an article reviewing published and unpublished
empirical studies on the effects of the cognitive interview. It illustrates how the authors
went about their search.

The Cognitive Interview

Method
Sample of Studies
A literature search of Psychological Abstracts from 1984 onwards was conducted to obtain references
for studies on the effects of the cognitive interview. This was followed by a search of potentially relevant
references cited in any of the publications acquired during step one. In addition, several authors of
primary articles were contacted and asked to send any relevant work, published or unpublished and
names of other relevant researchers.
Criteria for including studies in the final sample were that (i) a full cognitive interview (either original or
the enhanced version) was conducted; (ii) its effects were compared with some form of standard
interview; (iii) the effects of the type of interview were measured in a recall task rather than a recognition
task, and (iv) the number of correctly and incorrectly recalled details were available. Thus, the studies
by Fisher, Gieselman and Amador (1989) and the field study by George (1991) were excluded from the
main analysis because, due to being field studies, correct and incorrect details could not be separated in
the recall scores (because no records were available of the witnessed events). In the Fisher et al.
(1987) study the enhanced version of the cognitive interview was compared with the original version
rather than with a standard interview and it was thus eliminated. Some studies (e.g. Fisher and Quigley,
1992; McCauley, 1993) report recognition data along with recall performance. In these cases only the
recall data were used for computing effect sizes.
The final sample consisted of 55 experimental comparisons from 42 empirical reports representing 2447
interviewees. Of these reports, 29 were published manuscripts (including those in press), 3 papers
presented at conference, 1 unpublished research report, 3 unpublished papers (including those
submitted for publication), and 6 unpublished theses or dissertations (see Table 1).

Extract 10 Meta-analysis (Kohnken, Milne, Memon, & Bull, 1999)

Characteristically, research findings about ‘what works’ in criminal justice (as in other parts of the public
sector) are mixed and often contradictory. This can lead to the conclusion that either ‘nothing works’ or
‘everything works, sometimes’ and in certain contexts (Tilley & Laycock, 2000, p. 216). The attractions
of the systematic review and meta-analysis in the latter context might seem to be obvious.
Systematic reviews and meta-analyses are time consuming if they are done thoroughly with agreed
protocols for including and excluding literature in the review. These ideas are presented here to
encourage you to think of how you can be more organised in the way that you review the existing
literature. Proposals and dissertations that are totally based on literature should pay particular attention
to how the search is structured, justified and reported.

Summary
You should now have a working knowledge of the key advantages and disadvantages of collecting data
in the most commonly used methods in social research. If you chose to use one or more of these
methods it is recommended that you do further focused reading on the method(s). There are, of course,
other methods and techniques of data collection that you may want to follow up. You will find
differences of emphasis and use of terminology across research methods texts, which can be a little

65
confusing. In addition research projects in criminology will often involve the use of more than one
method or data collection technique.
Three key things you need to remember are that:
1. whatever definition or perspective you take on research, you must reference it;
2. the research method or methods you use in any research project should be appropriate to the
kind of issue or problem you want to study; and

3. you should always be able to justify your choice of methods.

Chapter 6 Quantitative Data

The primary or secondary data that we have collected has to be analysed. In the process of deciding
upon our data collection method(s) and in collecting the data we should already have some idea of how
we will go about the analysis. We are making sense of reality in the process of data collection. Thus,
sometimes the raw data will be immediately open to analysis. However, in many cases this will not be
possible, due to the size of the data set and/or the format in which the data was collected. Some
processing will be needed. For example, the researcher may wish to convert his or her qualitative data
into quantitative form or detect trends and patterns in a large data set. This is when we can use
analytical techniques to help us identify and to work out what our data can tell us about our research
area. Our raw data might confirm our ‘gut instincts’, but we also want to know more.
Using simple descriptive statistical techniques on our data, such as tables, graphs and charts will
render some data into a more accessible format for interpretation and identification of patterns. You can
use a spreadsheet package or word processing programmes to produce tables, simple charts and
graphs. The most common spreadsheet programmes are Microsoft Excel, Microsoft Works, Lotus 123,
Quattro Pro and Apple Numbers. MS Excel tends to be the software that we recommend students use
for initial data analysis; and there are a number of versions of MS Excel in use.

Descriptive Statistics: Measures of Central Tendency

A large data set can appear daunting and it is almost impossible to make accurate interpretations by
simply looking at the raw data. It is necessary to carry out some simple descriptive statistics that will
provide a summary of the data, perhaps showing averages of sets of data. The average, or arithmetic
mean, is an example of a measure of level, and these measures can be regarded as a single value,
which best summarises the data set as a whole. The arithmetic mean is not the only measure of central
tendency or level. Calculations of the median and mode, as described in the paragraphs below, can also
provide a measure of central tendency. Which measure is used will depend upon the level of numerical
information (nominal, ordinal, interval or ratio).
The arithmetic mean (or ‘average’)
Also known as the average, the arithmetic mean is calculated by adding the set of numbers and then
dividing by the number of values in that set.

66
The mean is represented in many textbooks as:
x̄
The formula for its calculation is given by:

Here ∑, otherwise known as ‘sigma’, denotes the ‘sum of’ and indicates that the numerator consists of
summing all of the x values or all of the numbers in the set. The denominator consists of n, which
represents the number of observations or values in the set. This is a useful and valuable tool for use
with interval and ratio data, as all of the data values are known and can be calculated precisely.
Example 1: The mean
Imagine that we have following data set:

The arithmetic mean of this data set can be calculated as follows:

Therefore arithmetic mean of this data set is: 6.

Mean mode median
The median (or ‘middle value’)
When the data set is normally distributed, as in Example 1, the arithmetic mean is a very appropriate
measure of central tendency. However, when the distribution is skewed then the average will be
affected by the extreme values and will no longer represent the central values of the data set. (For an
explanation of the normal curve and normal distribution, see the section in this chapter on: Inferential
Statistics: Probability, hypothesis testing and the normal curve.) A more appropriate measure is the
median, or middle value, when the numbers are organised into ascending or descending numerical
order. If there is an even number of values, and therefore no middle number, then the two middle
values are ‘averaged’ – that is, totalled and divided by two. The median can also be used on interval or
ratio data.
Example 2: The median
To calculate the value of the median for the data set given in Example 1, you first need to re-arrange
the data so that the values appear in numerical order. The data set will then look like this:
1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11
You can now see that the value of 7 is the ‘middle’ number. Therefore the median for this data set is: 7.
The mode (or ‘the most common value’)
The mean and median cannot be calculated for nominal data. The mode, or most frequently occurring
category, can be used to identify central tendency or ‘level’. Modal values can also be found in ordinal,
interval and ratio. Put simply, it is the most common value.
Example 3: The mode
From the data set given in Example 2, it is clear that the most frequently occurring value is 7.
Therefore the mode for this data set is: 7.

Descriptive Statistics: Measures of Spread

In addition to summarising the central tendency of a data set, another descriptive statistical measure is
needed: measure of spread which describes how the rest of the data values are spread around the
centre. Are the values similar to the central measure (indicating a fairly compact distribution) or is there
a wide range in the values (indicating a fairly spread distribution)? Four values are commonly used to
summarise spread: percentage in the mode, range, midspread and standard deviation.
Percentage in the Mode

67
This is the only measure of spread that can be used for nominal data. Its calculation is simple and
consists of expressing the frequency found in the modal class as a percentage of the total number of
observations.
The Range
For ordinal, interval and ratio data a simple measure of spread is the range. This is calculated by
subtracting the smallest value in the distribution from the largest. The main disadvantage of the
range is that, by definition, it includes the extremes of the distribution and may not necessarily
describe the range of the bulk of the data. An alternative that combats this problem is
interquartile range, otherwise known as the midspread.
The Midspread
The midspread (dq) is the range of the middle half of the data, once the data has been placed in rank
order. This has the effect of discounting the bottom and top quarters of the data batch, thus ignoring
extreme values. To calculate the midspread, the difference between the values found at the upper
quartile (QU) and the lowest quartile (QL) is determined. The upper and lower quartiles are found in a
similar way to the median, but rather than dividing the data set into equal parts, quartiles divide into four
parts. The upper and lower quartiles are therefore the top and bottom numbers in the middle two
quarters. The equation to find the range of the mid-half of the data set is as follows:
dq = QU - QL
The Standard Deviation
The two previous measures of spread, the range and the midspread, use only two data values to
describe the spread. The standard deviation (σ but often written either as s.d. or d), in contrast, uses
all of the data values and is a measure that describes how all of the values differ from the arithmetic
mean, on average. In other words, it is a way of describing the average spread of all values around the
mean. The main disadvantage of the standard deviation is that it is affected by extreme values. It is,
however, an important statistical measure and is often used in many other statistical functions, so it is
worth understanding this measure. Its calculation is relatively simple by hand, but it is unlikely that a
researcher would need to calculate this measure manually as most pocket calculators and all computer
statistics packages include this function.
Activity 1: Key concepts in descriptive statistics
1. Write down a summary of why you might want to use a ‘measure of central tendency’ (mean,
median, mode) when you write up your research findings.
2. Why might you also want to calculate ‘measures of spread’ (percentage in the mode, the
range, the midspread, the standard deviation)?

Most research methods textbooks that cover quantitative data collection analysis include a section that
introduces descriptive statistical methods. Excellent accounts are provided by Erickson and Nosanchuk
(1992); Marsh (1988); Hinton (1995); and Reid (1987). Wright (2002) gives a particularly straightforward
introduction to statistics. Similarly, there are many texts that cover graphical methods. Hibberd and
Bennett (1990) provide a useful summary of methods using police-related examples.

Descriptive Statistics: Graphs - 1

Bar Charts
One of the simplest ways of investigating and/or presenting nominal data (in other words, counts in
unordered categories) is with the use of a bar chart. Much of the contextual details collected through
survey methods are often nominal in nature (for example, sex, occupation, area of residence and so on)
and a bar chart is probably one of the most effective ways of describing the counts within these
categories. Figure 1 shows an example of a bar chart giving the score of most harmful drugs to self
and others. Figure 2a illustrates the scores of contributions from harm to self and harm to society (a
compound bar graph). Figure 2b is a stacked bar graph: a further variation in how you can display your
data with a bar chart. Computer packages, such as Microsoft Excel, can be used to produce bar charts
such as those in Figures 1, 2a and 2b.

68
 Figure 1. Bar chart showing the score of most harmful drugs to self and others

Figure 2a. Compound bar graph showing the score of most harmful drugs to self and others

69
 Figure 2b. Stacked bar graph showing the composite score of harmful drugs to self and others
The most effective diagrams are almost always the simplest ones and the main reason for using
graphics is to have an immediate impact on the reader. The examples shown in Figures 1, 2a and 2b
immediately indicate to the reader that alcohol is the most harmful drug to society and that the most
harmful drug to self is crack cocaine.
Histograms
Histograms can be used to investigate weak ordinal data. (Interval or ratio data can be converted into
ordinal data by constructing class intervals to which each value can be assigned. For example, if actual
age is collected within a survey, then class intervals can be constructed to summarise that information:
0-5 years, 6-12 years, etc.) As outlined earlier, weak ordinal data can be regarded as counts in
categories or classes that are ordered.
In essence, a histogram is the same as a bar chart, except that the arrangement of the classes is along
the x-axis (the horizontal axis) in order of class value (usually ascending order). Sometimes histograms
are referred to as frequency diagrams because they illustrate the frequency of occurrence within each of
the class values graphically.
Line Graph
A simple line graph can be used to plot individual ratio or interval values and is most often used to
display time-series information, such as the reported number of crimes each month (see Figure 3).

70
Figure 3. A simple line graph showing the frequency of crimes over time
Scatter plot
If we want to examine the relationship or association between two variables, we can use a scatter plot to
display the data. For example, Figure 4 illustrates a scatter plot, which shows that there is no
relationship between the x variable (number of cinemas) and the y variable (number of crimes
reported). The plot shows that the points are scattered fairly randomly. Figure 5, however, shows a
positive correlation between the x variable (population density) and the y variable (crime rate), as we
might expect if we were looking at population density and crime rate. That is, when the population
density of a given area increases, so too does the crime rate in that area. When x increases, then so
does y. Figure 6 indicates the opposite effect and displays a negative correlation: as x (number of
police on patrol) increases, y (crime rate) decreases. It must be remembered, however, that the
associations displayed in these scatter plots do not suggest a causal relationship between the two
variables; they only indicate the direction and strength of relationship between them. In order to predict
which variable is potentially affecting the other it would be necessary to manipulate one of the
variables. This may not always be possible and, indeed, in some cases it would be unethical to do so.

Figure 4. Scatter plot – no relationship

Figure 5. Scatter plot – a positive relationship

71
Descriptive Statistics: Graphs - 2

Figure 6. Scatter plot – a negative relationship

Presenting data: an example

Male Homosexuals and the Police

The author of this study devised this questionnaire in order to gain a further insight into the concerns,
opinions and experiences that male homosexuals living in two police force divisions, may have of the
police. The author set out to recruit one hundred participants in Force One Division (N) and Force Two
Division (B).
The Community Questionnaires were distributed as follows:
Force Two Division (B)
As previously discussed, this is a division with a large ‘gay community’. Permission was first obtained
from the licensing managers of numerous ‘gay venues’ located in the city before conducting this
enquiry. Here, males frequenting these venues were approached and informed:
‘I am a student currently attending the Institute of Criminal Justice Studies at the University of
Portsmouth. I am conducting research for my dissertation. This is investigating the perceptions of
homosexuals towards the police in this division. If you live in this division and you consider your sexual
orientation to be other than heterosexual, would you mind participating in my research by answering a
questionnaire?’
Only once respondents replied ‘yes’ to the given questions were they issued with a pen and a
questionnaire and asked to return it to the researcher once completed. One hundred and twelve out of
one hundred and twenty questionnaires were returned. Of these eight were incomplete and a further
four were eliminated from this research to provide a total of one hundred questionnaires.
Force One Division (N)
Once again permission was sought from the licensing manger of a ‘gay venue’ located here before
conducting this research. The researcher introduced himself to possible participants in the same
manner which was used in Force Two Division (B). One hundred and ten males who considered their
sexual orientation to be ‘other’ than heterosexual participated in the research. Of these, eight
questionnaires were found to be incomplete and a further two were eliminated from the study to provide
a total of one hundred completed questionnaires.
The following are bar charts presenting the results from two of the questions:

72
Question 3: Do you feel safe whilst walking around this city during the night?

Safety – walking at night

In reply to this question over half the respondents (55.0%) in Force Division N reported that they felt
‘safe’ walking around at night, whilst about a third (32.0%) were of the opinion that they felt ‘neither safe
nor unsafe’. One in eight (13.0%) considered themselves to feel unsafe whilst walking around at night.
In Force Division B a similar proportion to Division N (51.0%) considered they felt safe and a bigger
proportion (44%) reported feeling ‘neither felt safe or unsafe’. Only one in twenty (5.0%) in Division B
stated that walking around the area at night made them feel unsafe.
Question 4: Do you feel safe whilst walking around this city during the day?

Safety – walking in the day

73
A different pattern emerges if the same question is asked in relation to daytime. The great majority of
respondents in both divisions reported feeling ‘safe’ during the day: 91% in Division N and 89% in
Division B. A smaller proportion of respondents reported feeling ‘neither safe nor unsafe’ during the
daytime: 8% in Division N and 10% in Division B. Only 1% in each division reported feeling ‘unsafe.’
(Adapted from Blackbourn, 2002)

Note the expressions used in this example to describe what is shown in the graphs, such as ‘over half’,
‘the great majority’ and so on. You need to avoid repetitive descriptions of what graphs show and link
what is shown in different graphs. Following such depiction you would go on to offer possible
explanations of the differences illustrated, linked to other data collected in the project as well as
previous research and theory.
If the sample in each area did not add up to 100 in each case, as in the example above, it is usual to put
the number as well as the percentage in brackets, for example (30%, 45%).
Activity 2: Graphs
Which graphical method could be used to illustrate the following?
1. The salaries of forensic medical examiners grouped into salary classes?
2. The size of a local football crowd over the last twenty games?
3. The age/sex profile of the personnel in the local youth club?
Note: In certain instances, there may be more than one graphical method that is appropriate.

Inferential Statistics
Probability, Hypothesis Testing and the Normal Curve
Probability
As the name suggests, inferential statistics enable the researcher to make inferences about the
research data in relation to the general population, using probability theory. The data set is commonly
gathered from a representative sample of the population. Although the researcher can never be 100%
sure that the findings from the sample mirror those of the general population, inferential statistics allow
statements to be made that are concerned with the likelihood of the sample findings being similar or
dissimilar to the population. Inferential statistics can also be used, again relying on probability theory
(see for example Wright, 2002), to discuss the likelihood that the differences found in samples are
actually real and have not occurred simply by chance.
Hypothesis Testing
There is a specific convention that must be followed when using inferential statistics to test whether the
results are significant. Before carrying out the statistical test, both the null hypothesis and the
alternative hypothesis must be set out. The hypothesis, in general, is a prediction about the research
data. The null hypothesis (H0) always assumes ‘no difference’ and any observed peculiarities in
research findings are there because of chance. The alternative hypothesis (H1) assumes that the
observed findings are significant because the probability of them occurring due to chance is small. In
inferential tests, the null hypothesis is always investigated and, based on the results of the test, either
accepted or rejected. Therefore, a researcher can never say that the research has ‘proved’ that there is
a relationship between two events, only that they can reject the null hypothesis and that the findings are
significant. To understand the fundamental concepts behind all inferential tests, it is first necessary to
discuss the characteristics of something called the normal curve and its relation to probability theory.
The Normal Curve
In a perfectly normal curve the shape is symmetrical and there are an equal number of observations
either side of the average (see Figure 7). It is also the case in a normal curve that the mean, the
median and the mode are all equal. Therefore 50% of the observations fall either side of the average.
An alternative way of expressing this is that the chance of obtaining an observation higher than the
average is 50% and one lower than average is also 50%. This chance of 50% can be translated into a
probability of 0.5 (probability is always expressed as a decimal, unless it is 100% chance which is equal
to a probability of 1). In terms of the area under the normal curve, 50% of the area represents
observations above the mean and 50% represents observations below the mean. We call this type of
distribution a normal distribution.

74
Figure 7. The characteristics of the normal curve
The observations found under the normal curve can also be described in terms of standard deviation.
As noted earlier, the standard deviation indicates the average difference of all values around the mean.
In a perfectly normal curve (see Figure 7), 68.3% of the values are found within 1 standard deviation
either side of the mean; 95.4% of the values will be within 2 standard deviations and 99.7% of the
values will fall within standard deviations. The normal curve indicates that the likelihood of a value
falling in the tails of the curve (very high or very low observations – an ‘extreme event’) is quite small.
The following example is used to illustrate the ideas being used:
If it is assumed that the average height is 66 inches and that the standard deviation is 4 inches, then
68.3% of the population will be between 62 inches and 70 inches (that is, 1 standard deviation above
and below the mean). Likewise, approximately 95.4% of the people will be between 58 inches and 74
inches, and 99.7% between 54 inches and 78 inches. (These are not true height figures; they are only
used as an example).

A Note regarding p values; and Type I and Type II Errors

There obviously has to be a point when the researcher has to make a decision as to whether a
particular value is likely to have happened by chance or whether the value is significant. Much
published social research uses cut-off values of 95% or 99%. In other words, if a sample value could
have been obtained by chance 95% or 99% of the time, then it is not regarded as significant. Obviously
the use of a 99% critical level is stricter than a 95% level. In the literature, such levels may be indicated
by a p value: 95% equates with p=0.05; and 99% equates with p=0.01.
Erickson and Nosanchuk (1992) provide an excellent account of critical values and levels of confidence
in statistical testing. They also relate these concepts to the dangers of falsely accepting or rejecting a
null hypothesis and compare it to the courtroom situation of falsely finding the accused innocent or
guilty. Another quite ‘readable’ account of the relationship between probability levels and hypothesis
testing is provided by Hinton (1995, pp. 92–102). In both of these accounts, Type I and Type II errors
are introduced. A Type I error is when the null hypothesis has been rejected when really it should have
been accepted (that is, finding a ‘significant’ result when one does not really exist). Conversely, a Type
II error involves accepting the null hypothesis when really it should have been rejected (that is, failing to
recognise a ‘significant’ result).

Inferential Statistics: Choosing an Inferential Statistics

Test

75
A number of statistical tests exist to assess the extent to which any differences or associations in data
collected are ‘significant’ and likely to be real, as opposed to chance.
The data analysis and calculations required for these tests are usually conducted using software
packages such as SPSS/PASW. Therefore we have not provided these calculations here.
You will need to refer to a statistics textbook (such as Wright, 2002) if you want to know more about
these tests (and their calculation) than the brief outline below.
Irrespective of the type of inferential test that is used, the following general steps are employed in all
inferential tests:
 the null and alternative hypotheses are stated
 a test statistic is calculated and compared with a critical value derived from statistical tables

 if the test statistic is greater than the table value (at a specified value of p), then the null
hypothesis is rejected

 if the test statistic is smaller than the table value (at a specified value of p), then the null
hypothesis is accepted.

There are different statistical tests for the different levels of data (nominal, ordinal, interval or ratio data)
and different tests for different test scenarios (for example, testing the significance of one sample result,
testing the difference between two samples and testing the difference between many samples). Four
widely-used statistical tests are described below.

Inferential Statistics: T-test Analysis of Difference

If the problem is one of determining whether there is a difference between two samples, then a T-test
should be employed. In other words, in a two-sample T-test, the difference between the means of two
samples is investigated. If the difference is ‘significant’ then it is assumed that the samples have been
drawn from different populations. If not, then the samples are assumed to be drawn from the same
population. To carry out a two-sample T-test, a calculated T (based on the sample values) is compared
with a critical value found in T-tables. If the calculated value is greater than the table value, then the null
hypothesis is rejected – the null hypothesis being ‘no significant difference’. Alternatively, if the
calculated value is smaller than the table value, the null hypothesis must be accepted.
Most statistical computer packages contain the function to carry out T-tests and the calculation process
is therefore relatively unimportant. The formula calculates the difference between two sample means as
a proportion of the standard deviation of all the possible differences (termed the standard error). A high
value indicates that the observation is found in the tail end of the distribution and is unlikely to have
happened by chance. In order to determine whether this is large enough to reject the null hypothesis,
the table of critical T-values is observed (see Wright, 2002, pp. 134–135). Some additional statistical
terms need explanation. First, when reading the table of critical values, the degrees of freedom (df)
must be known. The degrees of freedom take account of the level of uncertainty in a statistical formula
by noting the size of the sample. In essence, as the sample size gets larger, the critical values get
smaller and the null hypothesis can be rejected more easily.
Note that the degrees of freedom are often denoted by ‘v’ in statistical tables.
Second, the test can either be one-or two-tailed. One- or two-tailed tests are relatively simple to
explain: if the alternative hypothesis states that one particular sample mean is thought to be greater or
less than another then the test is one-tailed. If the test is simply investigating ‘differences’ as stated in
the alternative hypothesis, where the difference could go either way, then a two-tailed situation exists.
The option to choose a critical value for a one- or two-tailed test is available in statistical tables. See
Wright (2002, pp. 34–135). Hover over the shapes below to test your understanding

Finally a t-test can either be ‘independent’ or ‘paired / repeated’. An independent t-test is used if the two
samples include different people, and are mutually exclusive – for example comparing males to females.
A paired, or repeated, t-test is used when the two samples are made up of the same people – for
example before and after training.
Two good examples of the t-test in action can be found in the following articles authored by ICJS staff:
Nee, C. and Ellis, T. (2005). Treating Offending Children: What Works? [Electronic version]. Legal and
Criminological Psychology, 10, 1-16.

76
Walsh, D., and Milne, R. (2008). Keeping the PEACE? - A study of investigative interviewing practices in
the public sector [Electronic version]. Legal and Criminological Psychology, 13, 39-57.

Inferential Statistics: ANOVA

Analysis of Variance - three or more samples
The analysis of variance (ANOVA) test, sometimes known as an ‘F-test’, is similar to a T-test in that it
investigates the difference between sample means for ratio or interval data. An F-test is used when
there are three or more samples, whilst a T-test is used for investigating the difference between only two
samples. Basically, an analysis of variance test compares the variance within each sample (the
variance is the standard deviation squared) with the variance between them. If the samples have been
drawn from the same population, the ‘within’ variance will be approximately the same as the ‘between’
variance because both will reflect the overall population variance. If the samples are significantly
different and have been drawn from different populations, the variance between the samples will be
significantly larger than the variance within.
The steps to ANOVA are similar to the statistical tests that have previously been described in this unit.
A statistic that provides the ratio of the between variance to within variance is calculated using the
sample values (an F-value). This calculated F is then compared with a critical value of F provided in
statistical tables. If the calculated value is greater than the table value, the null hypothesis is rejected. If
it is smaller, then the null hypothesis is accepted. F-tests, however, can only be carried out on
parametric (normally distributed) data, which is either interval or ratio in nature.
The F-test calculations look very complicated but are really quite easy to work through. Again, statistical
computer packages are able to provide ANOVA tests and it is unlikely that the value would have to be
calculated manually.
For example, we would use an ANOVA if we wanted to investigate whether gender and marital status
influenced ratings on a fear of crime scale. This may, for example, highlight that the impact of marital
status on fear of crime is different for males and females – this would suggest an ‘interaction’ of factors.
2.3 Degrees of freedom for the F-test
The degrees of freedom for the between sample variance are the number of samples minus one (k –
1). The degrees of freedom for the within sample variance are the total number of individuals in the
data minus the number of samples (N – k). The table value of F can be derived using Murdoch and
Barnes (1986, pp. 18–19). Within this version of the table, the degrees of freedom for the between
sample variance are found along the horizontal axis and the degrees of freedom for within sample
variance are presented vertically. Not all possible values for the degrees of freedom are given. The
nearest values should, obviously, be used. See Wright (2002, pp. 136–137) for the table.

Inferential Statistics: χ2 (chi square)

Analysis of difference between samples containing nominal or weak ordinal data
The T-test and F-test are used to investigate the difference between samples, based on a sample
means. The sample means are derived from ratio or interval type data. Much survey data, however, is
not in this form and much of the resultant data is comprised of counts in categories (ordered and
unordered). A sample mean cannot be derived for nominal data and can only be estimated for weak
ordinal data. To investigate the difference between samples containing categorical data, a 2 test
(pronounced ‘chi square’ - with a hard ‘ch’, as in ‘sky’) can be used.
In essence, a 2 test determines whether the observed values are significantly different from those
expected. The expected values are determined by assuming that the phenomena in question are
distributed equally.
For example, The British Crime Survey provides a wealth of data on victimisation, and we could use χ2
to determine whether gender (male / female) affects feelings of safety when walking in the dark (safe /
unsafe).
As previously stated, χ2 can only be used with nominal or weak ordinal data (that is, counts in
categories). In particular, the counts should be raw numbers and not rates or percentages. The data,
however, does not have to be normal – χ2 is a non-parametric test. If the number of categories is 2 then
the expected frequencies should be 5 or larger. If the number of categories is greater than 2, no more

77
than one-fifth of the expected frequencies should be less than 5. None should be less than 1. If these
conditions are not met, then the χ2 test is unreliable. See Wright (2002, pp. 138–139) for the table.

Inferential Statistics: Measures of Association

Pearson’s r and Spearman’s r
Attention now focuses on the use of inferential statistics to investigate the association between two
variables rather than difference. The section that covered: Descriptive statistics: Graphs illustrated how
two variables could either have a positive, a negative or no association with each other (see Figures 4, 5
and 6). These scatter plots provide a useful pictorial image of the data and allow the researcher to ‘see’
how the two variables are associated. This concept is extended here to measure the strength of that
association.
For example, suppose we want to identify if there is an association between the amount of graffiti and
the numbers of local crimes. We can compare different areas and see if there is a positive correlation –
ie higher levels of graffiti are associated with a higher number of local crimes; a negative correlation,
higher levels of graffiti associated with lower crime rates; or no association at all. It is important to
highlight here that association is not the same as ‘cause’. If there is an association, we cannot infer from
this exercise that the higher levels of graffiti have caused the higher or lower crime rates, or vice versa,
but we can say that there is an association.
We use of correlation coefficients to indicate the direction and strength of an association. As with all
other inferential tests, a null hypothesis and an alternative hypothesis are put forward and a correlation
coefficient (r) is calculated. A coefficient value of r =+1 indicates a perfect positive relationship, r =0
indicates no relationship and r =–1 indicates a perfect negative relationship. On the basis of this
relationship, the null hypothesis is either accepted or rejected.
The two most-commonly used tests for analysing correlations are:
1. Pearson’s product moment correlation (Pearson’s r) which is used for investigating the
association between interval or ratio data that are parametric in nature (that is, a histogram of the
data is approximately normal in shape). The plot of the relationship between the two variables must
approximate to a straight-line relationship.
2. Spearman’s rank correlation (Spearman’s r). Spearman’s r should be used when ratio or
interval data has a non-normal distribution (that is, non-parametric) or if the data is in strong ordinal
form (that is, ranked). In fact, Spearman’s rank correlation coefficient can only be used with rank
data; and so ratio or interval data must be ranked before using the test.

A good example of the ANOVA and correlation tests is provided in the following article:
Dando, C., Wilcock, R., and Milne, R. (2008). The cognitive interview: Inexperienced police officers'
perceptions of their witness/victim interviewing practices [Electronic version]. Legal and Criminological
Psychology, 13, 59-70.
Activity 3: Inferential statistics - a summary
Construct a table to summarise each of the techniques above to show:
1. name of test or technique;
2. purpose; and

3. with what type of data it can be used.

Chapter 7 Qualitative Data

As was discussed in Chapter 6, the data that we have collected has to be analysed; and
in deciding upon our data collection method(s) and in collecting the data we should
already have some idea of how we will go about the analysis. Indeed, most researchers
who collect qualitative data and who analyse some sorts of secondary data (documents,
for example) would say that data collection and analysis go hand in hand. Our raw data
might confirm our ‘gut instincts’, but we also want to know more.
78
The key issue with qualitative data analysis is meaning. Therefore any analysis
should involve the understanding and interpretation of these meanings. The raw data will
usually be in the form of words and phrases: what people have said in interviews, for
example. These words and phrases can be analysed by coding and reconstructing the
information to identify emerging patterns and themes.

Data Reduction
The main problem you are likely to face when making decisions about the handling and
presentation of qualitative data concerns the sheer volume of material you are likely to
have collected. Because of their open-ended nature, methods of data collection such as
interviews and participant observation are likely to produce data in very large quantities,
far more than you can possibly report.
You will, therefore, need to find ways of condensing and ‘editing’ data that is relevant, so
that it can be presented clearly and concisely. One fundamental criterion for deciding
whether data is relevant is the aims and objectives of the research. These should guide
decision-making at every stage while working on the presentation of data.
Qualitative data analysis can already start in the early stages of the data collection; and
this early analysis can inform and frame later, more in-depth, analyses (Miles &
Huberman, 1994). This ensures that as patterns and themes emerge from the data, they
are picked up and can be developed in the course of the data collection. It also makes
the data set more manageable and less daunting than if data analyses were left to the
end of the data collection stage.
Miles and Huberman (1994) describe eight methods of early qualitative data analysis
and they can be categorised into summarising (contact summaries, case analysis
meeting and interim case studies); coding (labelling); thinking about the data; and
producing reports. Some of these methods will be briefly described here.
Contact Summary Sheet
This form enables the researcher to reflect upon a particular event or data collection
contact, and to consider the implications of the contact in terms of what concepts and
issues have been raised. A sample sheet can be found in Miles and Huberman (1994, p.
53).
Coding
Data will be collected from a variety of sources and in a variety of forms, and the data-set
will be substantial. Therefore it is vital to organise the information and reduce it to a more
manageable size and form. Labels can be attached to chunks of data in the form of
words or phrases that may be connected in some way. These ‘units of meaning’ are
allocated to chunks of data by the researcher and will be influenced by the focus of the
research question.
Memos
Memos provide a means of tying together key themes and concepts that have been
identified by the researcher in the course of coding. As ideas and thoughts develop, it is
essential to note them down, with the date and link to the appropriate data. Memos,
suggest Miles and Huberman (1994, p. 74), can often provide ‘sharp, sunlit moments of
clarity or insight – little conceptual epiphanies’.
These early analyses or data reduction exercises can then be displayed in some format,
a matrix or graph for example, which enable the researcher to extract some meaning
from the data and proceed accordingly. The researcher can then draw conclusions about
the data, identifying patterns and emerging themes.

79
Presentation of Qualitative Data

Most qualitative data is likely to be presented in a written form. A helpful technique to

ensure relevance when presenting qualitative data is to use headings that refer
specifically to aspects of your project’s objectives. In this way you are less tempted to
move off the point. You will need to add new headings for themes that emerge in the
course of research.
Graphical methods of presentation can also be helpful. It may be possible to present
some of the data in a quantitative form through the use of simple tables and graphs.
The actual presentation of data drawn from interviews can be handled in one of two
ways:
1. Each interview might be presented in turn, possibly starting with a brief
background biography of each respondent. Responses can be organised using
headings relevant to the key themes that either guided the interviewer or ‘emerged’
at the interview, depending on how you went about conducting them.
2. More often, interviews are presented thematically: in effect, identifying themes
across interviews. This is a better way of preserving the anonymity of responses.

Whenever you are presenting data gathered from interviews, use actual quotations from
respondents wherever possible. This will give your presentation an authentic ‘feel’ and
avoid misinterpreting views by putting them into your own words. However, simply
quoting respondents is not enough; you will need to analyse their responses in the light
of your project’s aims and objectives.
One of the features of the interpretive approach to research is the concept that
knowledge is constructed by the individual, it is not something that can be objectively
measured. Thus, the techniques rely heavily on the researcher’s interpretation of the
data, which can be seen as very subjective. This may be a weakness or a strength
depending upon which school of thought you favour. Robson (1993) highlights the
problems of relying on the researcher as analyst and claims that data overload, influence
of first impression and confidence in judgements are just three of the areas that will affect
the analysis of the data. (For a more detailed discussion refer to Robson, 1993, pp. 274–
75).
Activity 1: Analysing and presenting qualitative data
1. Look back at the examples of qualitative research given in Chapter 5
2. Consider the different ways in which qualitative data has been presented in those
examples.

You will also find useful examples and explanations of different types of qualitative
research in the ‘Social Research Updates’ http://sru.soc.surrey.ac.uk
For further guidance on qualitative data analysis, you can consult Miles and Huberman
(1994) and Silverman (2000; 2001).

Chapter 8 Comparative Research

Comparative criminal justice research considers the differences between national
comparison, multi-national comparisons, and international comparative research. It also
explores the relativist and positivist approaches to comparative research. This chapter is
for students who are researching a topic on two or more countries and/or using
international data sources.

80
Aims of Comparative Research
The aims of this entry on comparative criminal justice are to:
1. Enable students to consider the objectives of comparative criminal justice research
criminal to be for the promotion understanding, reform, or both.
2. To consider relativist and positivist approaches to comparative research.
3. To consider differences between national comparison, multi-national comparisons, and
international comparative research.

Going abroad has an almost universal appeal. For centuries, this has been regarded as
instructive - broadening horizons and coming back with a better perspective on local
affairs. But what we need to consider is, what can be learned from abroad? Most
important in answering this question, is what we aim to achieve.

Part 1 - Aims of Comparative Research

Understanding or Reform?
Are you personally studying to be able to better understand? Or to change and reform
(what you think you understand)?

Some time ago Cole, Frankowski and Gertz (1987) identified the value of endeavour
towards comparative research to be in six areas, specifically in promoting:

1. international co-operation
2. understanding legal systems
3. reform of legal systems
4. unification and harmonisation of laws
5. scientific progress
6. classification of legal systems

The outcomes of these endeavours are not mutually exclusive, but a distinction can be
made between the aim of seeking to understand and of seeking to change. Change
without understanding is obviously hazardous, but there clearly is a difference between
the two. Understanding your own legal system, promoting scientific progress and the
classification of legal systems could be associated with ‘seeking to understand’.
Classification would hopefully lead to understanding, and both ‘understanding your own
legal system’ and ‘scientific progress’ are aimed at understanding. The aims in
international co-operation, reform of one’s own legal system and unification and

81
harmonisation of laws have a more applied flavour, relating to ‘change’ or ‘reform’. In this
perspective, comparative research is a means, rather than an end.

We consider systems of justice in another entry, but briefly consider classification of

justice systems.

Understanding or Reform?
Are you personally studying to be able to better understand? Or to change and reform
(what you think you understand)?
Some time ago Cole, Frankowski and Gertz (1987) identified the value of endeavour
towards comparative research to be in six areas, specifically in promoting:
1. international co-operation
2. understanding legal systems
3. reform of legal systems
4. unification and harmonisation of laws
5. scientific progress
6. classification of legal systems
The outcomes of these endeavours are not mutually exclusive, but a distinction can be
made between the aim of seeking to understand and of seeking to change. Change
without understanding is obviously hazardous, but there clearly is a difference between
the two. Understanding your own legal system, promoting scientific progress and the
classification of legal systems could be associated with ‘seeking to understand’.
Classification would hopefully lead to understanding, and both ‘understanding your own
legal system’ and ‘scientific progress’ are aimed at understanding. The aims in
international co-operation, reform of one’s own legal system and unification and
harmonisation of laws have a more applied flavour, relating to ‘change’ or ‘reform’. In this
perspective, comparative research is a means, rather than an end.
We consider systems of justice in another entry, but briefly consider classification of
justice systems.
Cole et al (1987) classified criminal justice systems into 3 types:
 common law, or adversarial systems (using the examples of England & Wales,
USA and Nigeria).
 civil law, or inquisitorial or Romano-Germanic systems. (originating in
continental Europe).

 socialist law systems (using examples from the former USSR, including
Poland).

As we shall see in more detail in a later entry, most systems are hybrid mixes that have
developed over time, perhaps haphazardly (an example being Japan).
Regardless of the systems of justice, different methodologies tend to underlie research
that is geared towards understanding on the one hand, and research that is aimed
towards achieving reform on the other:
 The relativist approach is better suited to research aimed at understanding.
 The positivist approach is more often associated with research involving an
agenda of change.

Exercise

82
Consider what system of justice you are most familiar with. Are you aware of others, in a
way that allows you to compare, effectively?
Being steeped within one society/jurisdiction may mean that we do not fully recognise or
consider others (similar or dissimilar). Consequently, it is even more important that we
fully consider and engage with ways of looking, analysing and comparing crime and
justice.
Now we move to some basic theoretical considerations in relativism and positivism in
comparative criminal justice research.

The Relativist Approach

Perhaps the oldest academic endeavours involving comparative research come from
anthropology - the science of humanity and human cultures. The anthropologist’s
concern with history, culture and social structure combines to illuminate dissimilarities
and to sharpen contrasts. Similarly, immersion in one’s own legal system tends,
inevitably, to solidify assumptions and blunt critical faculties. Laws appear ‘natural’,
modes of implementation ‘inevitable’ and relationships between criminal justice agencies
‘necessary’. The anthropological research approach guards against that.

Such research represents a relativistic philosophy underlying comparative research. The

way societies are organised does not correspond to certain templates or principles but is
a function of the environment. It depends on habits, policies, culture and history. Thus,
because a society works in a certain way that does not mean that other societies would
work in even remotely the same fashion. The clear implication here is that after having
learned how societies are organised in other countries, one could turn back to one’s own
society with a fresh perspective and realise that the way society is organised is perhaps
to an extent arbitrary and not necessarily the best possible way, and certainly not cast in
stone.
The equivalent in criminal justice has been called the interpretivist method of study.
Nelken (1994) defines this as ‘to show how crime and criminal justice are embedded
within changing, local and international, historical and cultural contexts’. Such studies
examine criminal justice or a certain aspect of it, and seek to understand how that aspect
has come into existence and why it seems to work in the way that it does.
The aim of interpretivist methods of research could be summarised as follows:
“to maximise understanding of alien cultures by honest-to-God field work, moral charity,
intellectual humility and a determination of the taken-for-granted assumptions of both his
own and others’ cultural milieu” (Dixon, 1977, p. 76).
The aim underlying comparative research from a relativist perspective is to gain an
understanding of criminal justice in a foreign system, so as to increase the understanding
of our own system.

The Positivist Approach

In contrast to the anthropological approach is the positivist approach. The philosophy
underlying this approach is that criminal justice systems have more in common with each
other than differences. There may be striking differences, but these are mostly
differences in form, not in core content. The underlying philosophy here is that justice is
a universal phenomenon and justice systems can be understood in general terms.
To quote Zedner (1995), positivist comparative criminal justice research is about:

83
“identifying similarity of form and substance in the systems under scrutiny. Many start
with the working assumption that criminal justice systems across countries are of the
same genus, differing only in varieties of form” (Zedner, 1995, p.16).
Following this approach, the main aims of comparative research are to reveal the core
set of principles underlying justice, from traits that are mere externalities.
What are the essential ingredients of a ‘proper’ criminal justice system?
Remember that some systems do not have a prosecution service, a jury or an appeal
system.
Positivistic studies may address issues such as:
 Justice systems around the world strive for expedience, and reduction in costs.
How do other countries go about achieving that?
 Drug abuse is a concern for every modern society. How do other countries use
the criminal justice system to fight drug abuse?Do prisons work? Are there examples
overseas where prisons are particularly effective

 Prostitution is a phenomenon that most people in most or all societies frown

upon. How is it regulated or repressed in other countries?

With a positivistic approach more straightforward questions tend to be addressed. It is

fair to say that the bulk of comparative research falls under this heading. These projects
tend to be more practical and inspired by current issues. Often there is a domestic
problem identified. This calls for scrutiny, and part of that scrutiny might involve looking
elsewhere for ideas of how the tackle the problem.

Considering Examples of Current Criminal

Justice Research
Consider the European Society of Criminology Annual Conferences. The Abstracts of 8th
Conference in 2008, for example, indicate the huge range of academic papers that were
presented as national papers to multi-national thematic panels, and several dozen that
dealt with comparative or international criminal justice issues. Some examples included:
European Society of Criminology Annual Conference 2008 topics:

 Detainee and Prisoner Data in the European Sourcebook

 Governing through Crime’ in an international context: The case of Bosnia and
Herzegovina

 Violence European Research group on National Reconviction rates

 Irregular influences against prosecutors and judges in Finland and Sweden

 Reflections on the evolving penal competence of the European Union

 Towards common EU crime definitions & an EU-level system of classification?

 Cross-National Collaboration to Combat Human Trafficking: Learning from the

Experience of Others

 Early Release of Prisoners - European Perspectives

 European Juvenile Justice Systems Current situation, reform developments and

good practices
84
 Ulster’s Policing Goes Global: The Police Reform Process in Northern Ireland
and the creation of a global brand

 Applying International Human Rights Standards as Measures of Juvenile / Youth

Justice

A quick glance shows that comparisons with foreign countries occur frequently. Just by
looking at the title, which of these presentations would you imagine to be relativist or
positivist in nature?
Finally, there are countless studies that basically have the theme: ‘let me tell you what
the state of affairs is about a certain aspect of criminal justice in my country’. Arguably,
these studies are not truly comparative. However, they are often aimed at an audience
that is not very familiar with those countries – such as the European Society of
Criminology conferences. They aim to facilitate understanding and comparison and
therefore can be said to be of a comparative nature anyway.

Further Reading
Aims of comparative research
The link below will direct you to key readings that are mounted on the library's online
reading list system. These will allow you to further consider these issues.
Online Reading List: aims of comparative criminal justice research
These readings entail:
Nelken, D. (2007). Comparing criminal justice [electronic version]. In M. Maguire, R.
Morgan & R. Reiner (editors). The Oxford Handbook of Criminology (4th Edition) (pp.139-
157). Oxford: Oxford University Press.
Pakes, F.J. (2010). Conducting comparative criminological research. In Comparative
Criminal Justice (2nd Edition) (pp.12–25). Cullompton: Willan Publishing.
Alternative Source:
Pakes, F.J. (2004). Aims and methods of comparative research [electronic version]. In
Comparative Criminal Justice (pp.13–25). Cullompton: Willan Publishing.

List of References Cited

Cole, F., Frankowski, S.J. & Gerz, M.G. (1987). Comparative criminal justice: an
introduction. In F. Cole, S.J. Frankowski & M.G. Gerz (Eds.) Major Criminal Justice
Systems: A Comparative Study (2nd Edition)(pp.15–26). Newbury Park, CA: Sage.
Dixon, K. (1977). Is cultural relativism self-refuting? British Journal of Sociology, 28, 75–
88.
Nelken, D. (Ed.) (1994). The Futures of Criminology. London: Sage.
Zedner, L. (1995). Comparative research in criminal justice. In L. Noaks, M. Levi, M.
Maguire (Eds.) Contemporary Issues in Criminology (pp.8–25). Cardiff: University of
Wales Press.

85
Part 2 - Comparative Research Methods

Introduction to Comparative Research Methods

The specific aims of this entry on comparative criminal justice are to:
1. To consider a range of research methods appropriate to comparative criminal
justice (including case studies, focused comparisons, truth tables and statistical
analysis).
2. Enable students to consider the inherent difficulties in comparative criminal
justice research, and some of the methodological challenges.

Comparative research can take place using a range of methods. In this section we will
be discussing some of these. It has been argued that methodology in comparative
criminal justice is underdeveloped. The following outline borrows therefore from
methodologies in comparative politics. This is an area in which methods have been
better described, and they are to a large extent valid for studies in comparative criminal
justice as well. In turn we will turn our attention briefly to case studies, focused
comparisons, truth tables and statistical analysis. First, we consider some of the key to
difficulties inherent in any comparative research.

Part 2 - Comparative Research Methods

Problems
Zedner (1995) notes the risks involved in comparative criminal justice research including
what she calls ‘criminological tourism’. This hazard is easy to understand and has
probably happened to many researchers when going abroad. Academic researchers feel
free of the worries and commitments of everyday life and are set to enjoy themselves. It
is then easy to find the locals very friendly, helpful and hospitable without realising that
your own happiness (not to mention your tourist money) might bring that about. Zedner
(1995) notes the possibility of misreading or simplifying local customs and to regard
exotic arrangements uncritically.

A second key caveat relates to linguistic difficulties. Zedner (1995) takes the position
that anything less than complete fluency leaves one vulnerable to misinterpretation.
This, however, has not stopped many prominent scholars from writing books about
foreign systems without having mastered the native language. In countries where most
people speak English as a second language the English-speaking researcher is probably
in less of a vulnerable position than in countries where the researcher’s language is not
widely spoken. Thus although the advantages of speaking the language are obvious, not
being able to do so should not automatically disqualify one from engaging in comparative
research.

A third difficulty relates to ‘touching base’. It relates to questions such as: ‘Who do you
talk to? What do you read? What do you observe?’ Many aspects of criminal justice
that are worth looking into are controversial. This is probably true wherever you look. It

86
is therefore important to try to assess both sides of any argument and not limit yourself to
talking to a limited set of people with a limited but shared set of opinions and knowledge.

These methodological hazards have pushed sceptics, such as the comparative criminal
justice scholar Vogler (1996), to argue that the entire enterprise of the comparative
analysis of foreign criminal justice systems is flawed.

Vogler emphasises that comparative research involves:

‘many attempts to translate the untranslatable and to introduce concepts completely alien
to English lawyers’ (1996, p.18).

It might well be impossible to reach a complete understanding of foreign systems, but

that should not stop us from seeking to understand and learning from their experiences.

The same is true for the study of foreign criminal justice systems where Vogler over the
years has subsequently made significant contributions.

Comparative Research Methods: Case

Studies
First, it is important to note that it is perfectly possible to carry out a comparative study
that treats only one jurisdiction in depth. But in order for such a case study to be
meaningful, the ‘case’ should be selected carefully and with justification. The case
selected should represent a wider category. For instance, when a student is interested in
examining suspects’ rights in inquisitorial criminal justice systems, picking ‘any old
inquisitorial system’ is not sufficient. In short, such a choice has to be justified.

Following Hague, Harrop and Breslin (1998), we can distinguish four types of cases to be
selected for a case study:
Representative cases are instances that are standard examples of a wider category.
They are the bread and butter of comparative research. A comparative study involving a
country with a low crime rate may choose to focus on Japan, as Japan has been well
known (at least until fairly recently) to have a low crime rate. Similarly, a study involving
high rates of imprisonment may consider the US, for much the same reason.

87
Prototypical cases are not representative cases, but might be expected to become so in
the future. In certain countries, decriminalisation of euthanasia or use of cannabis might
be at an advanced stage of development. These countries may lead the way for others
as to how to go about decriminalising certain behaviours and the associated pitfalls.
Study of such prototypical cases might yield valuable insight for countries that may tend
to go in the same direction but have not quite proceeded as far down the same path.
Deviant cases are studies to cast light on the atypical or unconventional. These can
involve causal relations, or lack of them. It is often thought that crime rates cause rates
of imprisonment. Thus when crime rates go up, one would assume that prison rates
would normally follow. The study of an atypical case (the Netherlands between 1950 and
1975) might show that this relationship is not one of causality, as prison rates went down
while crime rates consistently went up in this period.
Archetypical cases are cases that generate a category. The French inquisitorial system
of criminal procedure is a good example: all other European inquisitorial systems are
more or less derived from it. Thus when studying inquisitorial modes of justice, the
French one would be an appropriate choice as it can be said to be the quintessential
inquisitorial system.

Comparative Research Methods: Focussed

Comparisons
Focused comparisons are like case studies, but include more than one case. Most often
the number of countries compared is two or three. A key question is again how to select
jurisdictions for study. We will discuss two techniques, the most similar and the most
different design. As Hague, Harrop and Breslin (1998) define these designs:
“A most similar design takes similar countries for comparison on the assumption that
‘the more similar the units being compared, the more possible it should be to isolate the
factors responsible for differences between them’. By contrast, the most different
design seeks to show the robustness of a relationship by demonstrating its validity in a
range of contrasting settings” (Hague, et al, 1998, p.281).
Most similar designs tend to be easier to achieve. They often involve neighbouring
countries or countries that speak the same language. Similar designs sometimes involve
former colonies as well. Most different designs are often more difficult to achieve as they
tend to involve a selection of at least one jurisdiction that is very alien to the researcher,
with all the associated problems of familiarising oneself with such a system and all its
intricacies.

Comparative Research Methods: Truth Tables

Truth tables are a more quantitative form of analysis. It involves the comparison of a
relatively large number of jurisdictions on particular variables. It is an intriguing
method to help the researcher establish causal relationships by looking at a large number
of criminal justice systems. One could, for instance, look at countries that have
abandoned the death penalty in recent years. Other variables could be examined, such
as whether the country in question has gone through radical political transformations.
Such questions might shed light on when and why countries decide to abandon the death
penalty.
In most cases, this method can be too demanding from a practical perspective, but may
be particularly useful when the variables are dichotomous: when their value is either ‘yes’
or ‘no’, or ‘present’ and ‘absent’, making the inspection of such tables more
straightforward.
88
Organisations such as Amnesty International collect this kind of data
to inform their campaign for universal abolition of the death penalty. See for example
Amnesty International – Death Penalty: Countries Abolitionist for All Crime.

Comparative Research Methods: Statistical

Analysis
Statistics may occur in any study. The description of crime rates in any country will
probably involve at least descriptive statistics. In this context, however, statistical
analysis is meant to mean statistics as the main research philosophy.
The aim is to explore the relation between two or more variables that can be measured
quantitatively. In such a way, more can be learned about the relationship between such
variables, not in one country but in a range of jurisdictions.

But bear in mind that a statistical relation does not necessarily mean a causal one. Other
factors may be at work behind the scenes that actually cause the effect. A relation that
seems to imply a causal relationship but actually does not is called ‘spurious’.

Finally, a word of caution. The field of comparative criminal justice is not known for its
methodological rigour. Very often, choices of method to employ and selection of criminal
justice system to include in any study are made for reasons of convenience, or even
opportunism. Arguably there is no need to adopt too rigid a position with regard to
methodological imperfections. In the area of comparative criminal justice, perfectionism
is less important than a balanced assessment of the pros and cons of any
methodological design, so that subsequent findings can be evaluated according to their
merits.
These issues will be considered further in a later entry examining international official
data and international crime victim surveys.

Conclusion
People look at examples from elsewhere for different reasons. Generally broadening
horizons will be a learning experience. However, to avoid ‘tourism’, the purpose of the
comparison should always be made clear, otherwise comparative criminal justice
research might just read like someone’s holiday diary. The obvious risk for
anthropological research is that one gets too immersed in what is new, exciting and
exotic. The risk of positivistic research is that there will be too little actual immersion. As
with so many things, it is about striking a balance.
Zedner (1995, p. 19) describes the reality of comparative criminal justice research as
follows:
“Doing comparative research rarely entails selling one’s own home and tearing up one’s
passport, forever to live among the drug dealers of Delhi or the detectives of Düsseldorf.
Neither can one, with credibility at any rate, write about continental criminal procedure
without stepping outside the ivy-clad walls of an Oxford College. Rather the research
process entails developing a general theoretical (but distant) understanding at home-
89
base, punctuated by a series of forays (often of increasing duration) into the terrain of
study. This itinerary is matched by an intellectual journey which takes one from the
perspective of global structures to the minutiae of local detail and back and forth over the
course of the research in ‘a sort of intellectual perpetual motion’ (Geertz, 1983, p.235).
While periods of fieldwork provide for immersion in local culture (the court, the prison, the
police station), the journeys between make possible an intellectual distancing. Once
more library-bound, the researcher can engage in the detached reflections and distanced
evaluation which are the very stuff of comparison”.

Further Reading: Comparative Research

Methods
The link below will direct you to key readings that are mounted on the library's online
reading this system. These will allow you to further consider these issues.
Online Reading List: comparative research methods
These readings entail:
Heidensohn, F. (2008). International comparative research in criminology [electronic
version]. In R. D. King & E. Wincup (eds.), Doing Research on Crime and Justice (2nd
Edition). (pp. 199-228). Oxford: Oxford University Press. (note: also available from
Google books or browse via http://books.google.co.uk/books)
Mawby, R.I. (1999). Approaches to comparative Theanalysis: the impossibility of
becoming an expert on everywhere [electronic version]. In R.I. Mawby (Ed.) Policing
Across the World: Issues for the Twenty-First Century (pp. 13–22). London: Routledge.
Pakes, F.J. (2010). Conducting comparative criminological research. In Comparative
Criminal Justice (2nd Edition) (pp.12–25). Cullompton: Willan Publishing.
Alternative Source:
Pakes, F.J. (2004). Aims and methods of comparative research [electronic version]. In F.
Pakes, Comparative Criminal Justice (pp. 13–25). Cullompton: Willan Publishing.

Further Reading: Comparative Research

Methods
The link below will direct you to key readings that are mounted on the library's online
reading this system. These will allow you to further consider these issues.
Online Reading List: comparative research methods
These readings entail:
Heidensohn, F. (2008). International comparative research in criminology [electronic
version]. In R. D. King & E. Wincup (eds.), Doing Research on Crime and Justice (2nd
Edition). (pp. 199-228). Oxford: Oxford University Press. (note: also available from
Google books or browse via http://books.google.co.uk/books)
Mawby, R.I. (1999). Approaches to comparative Theanalysis: the impossibility of
becoming an expert on everywhere [electronic version]. In R.I. Mawby (Ed.) Policing
Across the World: Issues for the Twenty-First Century (pp. 13–22). London: Routledge.
Pakes, F.J. (2010). Conducting comparative criminological research. In Comparative
Criminal Justice (2nd Edition) (pp.12–25). Cullompton: Willan Publishing.

90
Alternative Source:
Pakes, F.J. (2004). Aims and methods of comparative research [electronic version]. In F.
Pakes, Comparative Criminal Justice (pp. 13–25). Cullompton: Willan Publishing.

List of References Cited

Hague, R., Harrop, M. and Breslin, S. (1998). Comparative Government and Politics: An
Introduction (4th Edition). Basingstoke: MacMillan.
Vogler, R. (1996). Criminal procedure in France. In J. Hatchard, B. Huber and R. Vogler
(Eds.) Comparative Criminal Procedure (pp. 14–95). London: B.I.I.C.L.
Zedner, L. (1995). Comparative research in criminal justice. In L. Noaks, M. Levi, M.
Maguire (Eds.) Contemporary Issues in Criminology (pp. 8–25). Cardiff: University of
Wales Press.

Part 3 - International Data Sources

International Criminal Justice Data Sources

The aims of this entry are to:
1. Alert students to a range of international criminological and criminal justice data
sources.
2. To consider issues of reliability, bias and selectivity in the creation of international
criminal justice statistics.
3. To highlight the existence and value of the International Crime Victimisation Survey
(and similar variants), European Union data sources and the range of studies undertaken
by the United Nations.

Why Consider International Data Sources?

As criminology students, you may be interested in the range of sources of information
when conducting criminal justice research:
 Criminological data: gathered by researchers, or data gathered by criminal
justice system, survey data from victimisation surveys.
 Criminal justice data: the operation of the criminal justice system (conviction
data, or possibly the numbers of ‘stops and searches’ for example) or criminal justice
agency composition (number of officers, etc).

This concise entry seeks to highlight the range of forms of these, focusing upon a
number of international organisations that have roles to gathering, collating and often
analysing this ‘multi-nation data’. As such, much of the data would be comparative – i.e.
provided by national authorities to international organisations (voluntarily, or as part of an
international treaty obligation).
These sources of information, often gathered over long periods of time, can be invaluable
and are often overlooked and under-researched. Overall this can and does lead to
biases and misconceptions about the ‘true’ nature of crime and criminal justice in the

91
world. International data sources can provide evidence for similarities between countries,
but also differences. They can also highlight areas where best practice or success is
being achieved, or on the contrary where poor progress and deterioration is occurring. If
done well, understanding and drawing upon these data sources, and placing your other
work into this kind of context, can be innovative, informative, and enlightening.
Consider the people on your course. Where do they come from? What part of the UK,
part of Europe, part of Africa, etc. Are they from an urban context or a rural environment?
Whatever the answers, they will have a different perspective, a differing experience of
crime and criminal justice, and a different conception of what is important to them,
compared to you. As a criminal justice scholar: this is an opportunity for you to reflect
upon your own experiences - an opportunity to perhaps address inherent biases and
preconceptions of crime and criminal justice in the world today. In conducting research,
looking to international data sources can help you to positively address, confront and
undermine inherent biases and preconceptions of crime and criminal justice research.

Range of International Data Sources

The following sections outline a range of potential data sources to draw upon in
international and comparative studies. As many sources are new and developing then
this will not be a comprehensive guide but one that will seek to demonstrate that sources
to support your research may be available if you look for them. Those highlighted will be
quantitative as well as qualitative data sources or studies, drawing on the ‘outcomes’ of
the criminal justice system, or as the product of surveys (victimisation surveys, public
opinion surveys).
The organisations producing or collating these range from the Council of Europe,
European Union, bodies within the United Nations system. First, however will consider
briefly issues of reliability and bias, so that you can en-gage in draw upon these studies
in an informed fashion. After reviewing some European Union sources, where the
countries are relatively well re-sourced, we consider the issues raised in less developed
countries where resources are scarce, and how the international community is seeking to
overcome the resultant shortcomings in access to data.

Reliable Criminal Statistics

Often we think of decriminalisation in the sphere of drugs, but relatively recently many
forms of homosexual behaviour were illegal, as they still are in many jurisdictions in the
Middle East. Criminalisation and decriminalisation occur frequently, and are sometimes
telling events with regard to a certain period’s Zeitgeist. This may mean that at a
fundamental level, efforts at comparison would be stymied if we unduly relied upon
official statistics.

92
(for more information visit United Nations on Drugs and Crime)
Official statistics may be unreliable in a comparative context, due to the way in which
they are compiled differently between countries. Police organisations differ
fundamentally between countries – and the fragmentation of the police in many countries
makes it difficult to obtain reliable country-wide statistics. In addition, policing in certain
countries is characterised by a great deal of secrecy and apprehension. Two implications
that this might have is that, firstly, performance figures, such as detection rates, would
not be made public, and often crime-reporting levels would equally be kept below the
surface. Secondly, in countries where the police are estranged from communities, and
working as a force against them rather than to serve them, the public is loath to report
crimes, so these figures would hardly be meaningful anyway.
Consider these issues by following the link below:
UN Office on Drugs & Crime: Compiling and comparing criminal justice statistics

International Crime Victimisation Survey

The limitations of official recorded crime statistics has led many criminologists and
international organisations such as the United Nations to focus on other measures such
as crime victim surveys. Since 1989 there have been a number of measurement waves
of the International Crime Victimisation Survey (ICVS)
The International Crime and Victimisation Survey (ICVS) is normally performed via a
telephone survey of 1000-2000 people per participating country. In 2004-5, 11
industrialised countries took part (17 in 2000).
Consider the range of methodology issues raised in the surveys
Consider what has been done to improve the validity of the sample, and what
compromises are made that may well lead to certain biases in the sample population.
On the positive side, the ICVS yields rich data, allowing for longitudinal analysis, by
comparing results from surveys compiled at different times. It obviously provides good
input data regarding comparative research, but can also allows for the testing of
hypotheses regarding differences between nations.
For more on these issues see the following text:
Pakes, F. (2010). Comparing Crime. In F. Pakes, Comparative Criminal Justice (2nd
Edition) (pp.26-39. Cullumption: Willan Publishing.

93
The European Union as a Criminal Justice
Data Source
The European Union provides a more focussed effort to the development of data
sources. For example, Eurostat embraces European Union, Candidate Countries and
those within the European Economic Area in its reviews of EU Statistics on crime and
criminal justice. In 2004 the EU’s Hague Programme committed the EU to develop more
comparable data sources, reflected in the subsequent 2006 Commission Communication
Developing a comprehensive and coherent EU strategy to measure crime and criminal
justice: an EU Action Plan 2006-2010.
Exercise–investigating Eurostat
1. Go to the Eurostat homepage, and type in ‘crime’ into the search menu at the top right
hand of the page. You’ll see that the results in the three columns encompass:
Publications
Datasets
Meta data
Considering the data sets that are available, then you can see potentially useful criminal
justice data resources available to you going back to 1950!
Crime, violence or vandalism in the area (Source: SILC) Publish Date: 15-NOV-2010
11:00 AM
Crimes recorded by the police Publish Date: 09-SEP-2010 11:00 AM
Crimes recorded by the police: historical data (total crime) 1950-2000 Publish Date: 09-
SEP-2010 11:00 AM
Crimes recorded by the police: homicide in cities Publish Date: 30-MAY-2009 11:00 AM
2. Now search for data using your own search term.
Public opinion surveys can also be useful in criminal justice research. European Union
through its EuroBarometer conducts regular comparative/multi-nation public opinion
surveys, many of which are criminal justice related. The different EuroBarometer surveys
use differing methodologies. A number of selected examples are set out below:
EB73.2, Domestic Violence against Women. Survey 02-03/10. Report 09/10
EB65.4, The role of the European Union in Justice, Freedom and Security policy
areas. Survey 06-07/06. Report 02/07.
EB65.4, The role of the European Union in fighting against organised crime.
Survey 06-07/06. Report 11/06.
Exercise–investigating EuroBarometer
1. Go to the EuroBarometer Homepage. Locate the ‘Standard EB’ menu on the left
hand page. See the latest ‘wave’ of the survey. How high does crime come on the
survey of citizens main concerns? Do you wonder how has this changed over time?
(Compare with previous waves…)
2. Go to the EuroBarometer Homepage. Locate the ‘Special EB’ menu on the left
hand page. Here you can see topics directly related to criminal justice issues, but
also others that could help inform a research project (mental health, e-
communications household survey, EU citizen’s attitude towards alcohol, or
corruption, etc.)

94
3.The Impact of Resources: Developing
Countries
4. The International Crime Victimisation Survey focuses on industrialised countries
but in the same context, interesting data from developing countries and countries
in transition have also been compiled.
5. The participating countries have included Argentina, Brazil, China, Costa Rica,
Egypt, India, Indonesia, Papua New Guinea, the Philippines, South Africa,
Tanzania, Tunisia and Uganda. Considering the example of burglary, the
patterns of victimisation revealed by these surveys varies widely:
6. “As regards the developing countries participating in the ICVS the highest rates of
burglary and attempted burglary were observed in Tanzania. One-year burglary
rates above 10% were also observed in some other countries in the African
region (Uganda, Zimbabwe and Botswana). The countries with the lowest risk of
burglary were India, Brazil, the Philippines, China and Egypt. The average
deviation from the regional mean was lowest in Asia and highest in Africa. In
fact, in the latter region victimisation rates ranged from a minimum of less than
3% in Egypt to a maximum of 19% in Tanzania (Alvazzi del Frate, 1998, p. 30).”
7. Setting out to explain such disparities is no small task. A number of macro-
economic characteristics might have a bearing on offending in general and
possibly on property crimes in particular. Alvazzi del Frate also presents data on
corruption and consumer fraud, together with some data on the distribution of
wealth in society. The conclusion of Alvazzi del Frate’s report is clear and
concise:
8. “The levels and effects of victimisation are more pronounced in the developing
countries than in the rest of the world… The ICVS shows that the overall
victimisation risks of citizens in developing countries are higher for all property-
related types of crime, while the risk for assault with force is equal in the
industrialised countries, countries in transition and the developing world “(Alvazzi
del Frate, 1998, p. 133).
9. It must, however, be noted that whilst such a description is valuable, explaining it
is not quite as easy. Van Dijk notes that the availability of goods is an important
predictor for levels of property crime. He particularly makes this point with regard
to bicycle theft: a bicycle in a bicycle-rich society such as the Netherlands is more
likely to be stolen than any bicycle in a country in which they are less common.
That places the findings on property crime in developing countries possibly at
odds. Explaining crime rates, even when international crime survey data are
used as a basis, remains difficult. Nevertheless, there are signs that economic
progress does relate positively to property crime:
10. “The relationship between crime and development has been the object of many
studies over the years, but no clear conclusion as regards the consequences of
socio-economic growth on crime has ever been reached. While a traditional
belief suggested that technological progress and a more equal distribution of
economic wealth would reduce social conflict, other theories envisaged that
socio-economic growth and modernisation would necessarily involve an increase
in overall crime rates, and in particular in crimes against property. This theory
was supported by the observation that developed countries generally showed
higher theft rates and lower homicide rates than developing countries” (Alvazzi
del Frate, 1998, p. 135).
11. Overall, students should be keenly aware of the biases in reporting and accuracy
within international criminal justice data sources that may be amplified by access
95
 to resources and expertise. Within the UN efforts to recognise and overcome
these biases include the ‘Data for Africa initiative’ considered below.
12. Serious consideration of these issues was spurred by the comprehensive report
from the UN below, which showed how levels of crime and victimisation were
adversely affecting development:
13. UNODC (2005). Crime and Development in Africa report

14. The Council of Europe

15. Since its inception, the Council of Europe has been keen to promote criminal and
judicial Corporation and research in the European region. To appreciate some of
this work, then it is useful to review the international comparative data sources,
has produced such as:
16. European Sourcebook of Crime and Criminal Justice Statistics
17. These sources have been studied by criminologist in many years, and represent
one of the best-developed international and comparative data sources available
for those studying and researching criminology and criminal Justice.
18. For those particularly interested, then the following journal articles provide useful
reviews of these data sources:
19. Alvazzi del Frate, A. (2003). The Voices of Victims of Crime: estimating the true
level of conventional crime. Forum on Crime and Society, 3(1&2), 127-140.
Retrieved from http://www.unodc.org/pdf/crime/forum/forum3_note4.pdf
20. Killias, M. & Rau, W. (2000). The European Sourcebook of Crime and Criminal
Justice Statistics: A New Tool in Assessing Crime and Policy Issues in
Comparative and Empirical Perspective [Electronic version]. European Journal on
Criminal Policy and Research, 8(1), 3-12.

The United Nations and Affiliated Research

Bodies
Since 1980 United Nations has been collecting data on the operation of criminal justice
systems within its global membership. Whilst national reporting is not comprehensive,
and subject the flaws discussed above, the data potentially provides a rich seam of time
series data can be drawn upon in many research projects. See in particular:
United Nations Surveys on Crime Trends and the Operations of Criminal Justice
Systems (CTS)
To overcome some of the biases in criminological and criminal justice data studies
across the world, the United Nations has invested in a specific project called ‘Data for
Africa’. This is a relatively new resource, but a developing one with a broad-based range
of studies on particular crime and victimisation issues and themes affecting Africa.
Data for Africa Publications
The United Nations Office of Drugs and Crime (UNODC) is a key body within the United
Nations with a task of developing the capacity of less developed States, sharing best
practice, and acting as a research focus. The UNODC’s Work Programme is focused on
three areas:
 Field-based technical cooperation projects to enhance the capacity of Member
States to counteract illicit drugs, crime and terrorism
96
 Research and analytical work to increase knowledge and understanding of
drugs and crime issues and expand the evidence base for policy and
operational decisions

 Normative work to assist States in the ratification and implementation of the

relevant international treaties, the development of domestic legislation on drugs,
crime and terrorism, and the provision of secretariat and substantive services to the
treaty-based and governing bodies

Its research and analytical work will be of particular interest to you here. For access to
the full range of publications, then follow the link below:
The United Nations Office of Drugs and Crime publications
The following is a collection criminological research bodies that are affiliated to the United
Nations.
Africa
UN African Institute for the Prevention of Crime and the Treatment of Offenders
(UNAFRI), Kampala, Uganda. www.unafri.or.ug
The Institute for Security Studies (ISS), Pretoria South Africa. www.iss.co.za/
Australia
Australian Institute of Criminology (AIC), Canberra, Australia, www.aic.gov.au
Asia
UN Asia and Far East Institute for the Prevention of Crime and the Treatment of
Offenders (UNAFEI), Tokyo, Japan, www.unafei.or.jp/english/index.htm
Europe
UN Interregional Crime and Justice Research Institute (UNICRI), Turin, Italy www.unicri.it
International Institute of Higher Studies in Criminal Sciences (ISISC), Siracusa, Italy,
www.isisc.org
European Institute for Crime Prevention and Control (HEUNI), Helsinki, Finland.
www.heuni.fi/
International Scientific and Professional Advisory Council of the United Nations Crime
Prevention and Criminal justice Programme (ISPAC), Milan, Italy. http://ispac.cnpds.org/
Raoul Wallenberg Institute of Human Rights and Humanitarian Law, Lund University,
Sweden, www.rwi.lu.se
North America
International Centre for Criminal Law Reform and Criminal Justice Policy (ICCLR and
CJP), Vancouver, Canada, www.icclr.law.ubc.ca
National Institute of Justice (NIJ), Washington, DC, USA. www.ojp.usdoj.gov/nij
International Centre for the Prevention of Crime (ICPC), Montreal, Canada, www.crime-
prevention-intl.org
South America
UN Latin American Institute for the Prevention of Crime and the Treatment of Offenders
(INALUD), San Jose, Costa Rica, www.ilanud.or.cr

97
For a comprehensive text on the United Nations, its criminal justice and human rights
bodies, and research work, then see:
Weiss, T.G. & Daws, S. (Eds.). (2007). The Oxford Handbook on the United Nations.
Oxford: Oxford University Press.

Further Reading
Issues in Comparative Criminal Justice Research
The link below will direct you to key readings that are mounted on the library's Online
Reading List system. These will allow you to further consider these issues.
Online Reading List: issues in comparative criminal justice research
These sources entail:
Nelken, D. (2009). Comparative Criminal Justice: beyond ethnocentrism and relativism
[Electronic version]. European Journal of Criminology, 6(4), 291-311.
Westfelt, L. & Estrada, F. (2005). International crime trends: sources of comparative
crime data and post-war trends in Western Europe [Electronic version]. In J. Sheptycki &
A. Wardak (Eds.), Transnational and Comparative Criminology (pp.19-48). London:
Glasshouse Press.
Goldstein, R. J. (1986). The Limitations of Using Quantitative Data in Studying Human
Rights Abuses [Electronic version]. Human Rights Quarterly 8(4), 607-627.
Young, P. (2005). The use of national crime statistics in comparative research: Ireland
and Scotland compared [Electronic version]. In J. Sheptycki & A. Wardak (Eds.),
Transnational and Comparative Criminology (pp.49-68). London: Glasshouse Press.

List of References Cited

Alvazzi del Frate, A. (1998). Victims of Crime in the Developing World. UNICRI
Publication No. 57. Rome: United Nations Interregional Crime and Justice Research
Institute (UNICRI).
Pakes, F. (2010). Comparative Criminal Justice (2nd Edition). Cullumption: Willan
Publishing.

Chapter 9 Abstract
Introduction
This section will explain expectations from your abstract and will provide you some
examples of abstracts from studies that have already been completed. Please note that
the abstract you will submit should discuss the study you INTEND to do. Thus, there will
not be any reference to results or conclusions.
FOR INFORMATION ON SUBMISSION DATES PLEASE REFER TO YOUR
PERSONAL STUDY SCHEDULE

Abstract Guidelines
98
The abstract forms the first "content" of your dissertation and is therefore the first thing to
be read by the supervisor and forms the first impression of your work. It will have NO
impact on your final mark but it is a useful tool for you to form your ideas and for your
allocated supervisor to understand what you aim to do for your dissertation.
The aim of the abstract is to be a concise and clear summary of:
 What you set out to do and why
 How will you do it

 What do you hope to find

 Indicate the thoroughness and integrity of your research

The abstract should be 250 words maximum. Length is very important. 250 words will be
adequate provided you write concisely and are summarising your ideas.

Write a title for the proposed dissertation, in bold, 14 pt.

We recommend you follow the next 10 steps:
1. Clearly identify the aims of your differentiation, in a sentence format beginning
with ‘to critically assess’; ‘to critically examine’, etc..
2. Define key terminology (only where relevant). Do not assume the reader is
familiar with the topic

3. Identify the methodology you intend to use

4. Identify potential major results or areas of debate

5. Assemble the above information into a single paragraph

6. Omit detailed background information, literature review, and detailed description

of methods

7. Remove extra words and phrases

8. Revise the paragraph so that the abstract conveys only the essential information

9. Use Times New Roman font, 12pt, double space throughout

10. Give the abstract to someone who is not familiar with your work and ask him/her
whether the abstract makes sense

Abstract Examples
Below are examples taken from MSc student e-dissertations
Akpala, James (2009) Trafficking in persons and human smuggling in Nigeria: a
contemporary manifestation of slavery and organised crime.
Trafficking in persons and human smuggling are crimes assumed to be difficult to
measure due to the scope and magnitude, which appears to be a growing phenomenon.
The trade in human beings has become so lucrative such that it generates more profit
than trafficking in drugs. This perhaps could be as a result of victims been used, traded
and sold numerous times and even if caught penalties are lower in contrast to trafficking
in drugs. Trafficking in persons and human smuggling are intractable problems
occasioned by greed, poverty, poor legislation, corruption, and huge demand for

99
services, weak values systems, and deceptive representation of the “heavenly” life in
cities and abroad. This research focuses on the processes and the extent of trafficking
and human smuggling in Nigeria. The study was equally intended to add to limited
knowledge and research on trafficking and human smuggling. The methodology includes
a focused literature review and telephone interview survey. Thus mixed methods were
felt to be suitable and appropriate for this research considering the primary objective
which was intended to be exploratory. The telephone interview surveys were used to
elicit information from law enforcement agents regarding the trends in trafficking and
smuggling of human beings as well as barriers and challenges faced in combating both
crimes. The results were analysed using both quantitative and qualitative methods in
order to add depth and meaning to the results obtained. The successful completion of
these objectives has proved that the extensive border between Nigeria, Chad, Niger,
Benin and Cameroon, has created easy passage and easy transportation of young
Nigerians through transit countries in North Africa, en-route Europe and other parts of the
world. This and other results were critically discussed, linking references to the ideas and
arguments presented in the focused literature review.
Bates, Catherine (2009) Watching the watchdog: examining public confidence in
the Independent Police Complaints Commission.
The Independent Police Complaints Commission (IPCC) commenced work on 1st April
2004. Created following calls for an independent body to investigate complaints against
the police, the IPCC has sought to raise awareness and increase public confidence in the
police complaints system in England and Wales. This study examines the progress made
by the IPCC during its first five years and determines whether the anticipated increase in
public confidence has materialised. Through the examination and triangulation of existing
literature and research, this dissertation has found that the IPCC has succeeded in
raising awareness and public confidence in its independence and impartiality during its
first five years. Despite this accomplishment, the increase in public awareness and
confidence is lower than anticipated and a significant gap still exists regarding the
willingness of some people to lodge complaints against the police. This is most keenly
felt amongst young people and ethnic minority groups who are the most likely to be
dissatisfied with the contact they have had with the police. Furthermore, those whose
complaint has been independently investigated by the IPCC state that consequently they
have a reduced level of confidence in the police complaints system as a whole. This
dissertation presents a timely review of the developments made by the IPCC and
provides a sound knowledge base upon which future research may be built.
Taylor, Katy (2010) One drink too many: evaluating 'what works' in reducing
alcohol related crime and disorder in Manchester city centre.
Alcohol related crime and disorder within the night-time economy on a Friday and
Saturday night is a major problem in towns and cities all over the country. It is estimated
that it costs £7.3 billion each year with at least 90 per cent of assaults being alcohol
related with the victim and/or offender having consumed alcohol in the 4 hours leading up
to the incident. This study aims to evaluate ‘what works’ in reducing alcohol related crime
and disorder within the night-time economy by reviewing the relationship between alcohol
and crime and applying this to licensed premises in town and city centres. This is further
compounded by a case study of various areas of Manchester city centre including
interviews with licensees, the police and consumers themselves. The results highlight
that government strategies in the past have been contradictory and confusing based not
on evidence but on what will win them the most votes. Whilst this may change with the
election of the Coalition Government in May 2010 it is clear from the research that crime
reduction strategies are more successful when implemented on a local level. With
regards to the future, targeting off-licenses which sell cheap alcohol in bulk must be a
priority so as to tackle the issue of 'pre-loading' at home and the associated problems.
Using a multi-agency approach to target problem premises on an individual basis through
enforcement and prevention, particularly crime prevention through environmental design
100
is also recommended as is working with universities to reduce the prevalence of student
binge drinking and its implications.
Glass, Dion Blair (2012) The success of geographic profiling is all too often
perceived by taking a narrow view of what constitutes its success. Does
geographic profiling have more to offer to investigators than merely an offenders’
residence?
The overall aim of this research is to investigate the relevance and accuracy of the
criminological theories that underpin geographic profiling. In doing so, the research aims
to ascertain whether or not (by inference) geographic profiling would have assisted and
supported investigators in solving cases. The overall aim is not only to establish whether
or not geographic profiling would have pinpointed the offenders’ residence, but whether
or not geographic profiling as a broader discipline would have enhanced the investigative
process.
The study examines how the criminological theories could have explained the offenders’
behaviour, and in turn, how the offenders’ behaviour fitted into the explanations offered
by the theories. By the same token, the study examines where the theories were out of
sync with the offenders’ behaviour, and where the offenders’ behaviour did not fit into the
theories. The study goes on to examine the implications of this for geographic profiling as
an investigative tool.
The approach used is that of case studies. Rossmo (2005, p. 3) maintains that there are
three effective ways to determine the efficacy of geographic profiling software, of which
one way is to examine performance against solved cases. Two solved South African
serial rape cases were used in this research.
The results showed that it was not be probable that geographic profiling would have
given an accurate estimation of either of the offenders’ residence. The results did
however show that if one considers the utility of geographic profiling (as opposed to only
applicability and performance), then geographic profiling would have been a useful tool in
the investigation processes in both cases.
Below are links to a few academic journals where you can view published abstracts
 British Journal of Criminology
 Journal of Qualitative Criminal Justice & Criminology

 International Journal of Investigative Psychology and Offender Profiling

 Computer Fraud & Security

Chapter 10 Literature Review

This assessment is designed to help you develop the focus of your research proposal
and should form the basis of the literature review you will need to do for your final
dissertation. Structured literature reviews are becoming increasingly common in
academic and policy research environments because of the shear volume of information
and research evidence available. In a structured review you set the limits on what
literature you are going to review, from where and over what timescale. This helps
create parameters for what you do and makes the review more systematic rather than
haphazard or idiosyncratic.

Literature Review
What Area(s) of Literature are you going to Review?
101
What is your aim? Identify the aim and the argument you are making. Then move from
the general topic area to more specific ideas directly affecting your study. Specify the
broad literature/area you are going to review, for example youth justice, then make clear
that you have a more specific focus, such as offending behaviour programmes in young
offenders institutions.
What Type of Literature?
Remember you are doing a piece of academic work. Therefore you need to focus
strongly on academic debate and research within your chosen area. Focus particularly
on academic journals – remember that the University of Portsmouth library has a wide
range of online journals and bibliographic databases. The use of journals will save you
time and, because they have been peer-reviewed, they are of a higher quality and
contain more precise information.
Reviewing policy documents can be seen as a distinct part of a review (where such
documents are relevant). However, policy documents from government departments are
not the ‘literature’ for an academic study. You need to be able to make clear links to
academic literature. Indeed you might see policy documents as a source for
documentary analysis. Also, when possible, use official statistics to support your point.
Seminal Works/Key Authors
Make sure you have identified academic authors who are well known in your area. You
can do this in a number of ways:
1. by looking at who writes and is cited in The Oxford handbook of criminology;
2. by looking through journals and citations indexes in libraries;

3. Google scholar can be useful as an initial search tool (www.scholar.google.com);

and

4. look at the bibliographies of recent works to help you identify other relevant
studies.

Search Terms
Decide on initial search terms and refine them by looking at the key words specified after
the abstract in some academic journal articles. You can use Boolean operators to
include and/or exclude parts of literatures. Use truncation where you are coming up with
very little literature. For example, ‘crim*’ will bring up all words starting with these letters.
Search different Subject areas and Databases
Criminology is inherently inter-disciplinary. Often you will need to search in other
disciplinary areas such as social policy, sociology, psychology, law, social care and so
on.
Structure and Limit Electronic Searches
Decide on the structure and limits of an electronic search and create a summary of your
literature search strategy for this assessment. You will be able to add to this as you do
your dissertation.
Use the following headings as a guide to structuring your review and look at the example
given within the unit text:
 timescale;
 key authors;
102
 key words;

 search engines;

 bibliographic databases;

 key journals;

 websites and organisations; and

 any inclusion/exclusion criteria, for example you may only want to include
empirical work (you don’t have to specify inclusion/exclusion criteria at this stage if
you find it unhelpful).

Writing and Presenting the Review

Your review should show that you know broadly what kind of research has been
conducted in your field both in terms of topic or theme as well as methodology. You
should know when your area became a focus for academic research and how the
research has developed in your field. You should show awareness of the terminology
used and any debates about this. Your analysis should include some evaluation and
criticism of the work reviewed. Link the literature and its gaps with the aims of your
study. Argue why your research is relevant and important.
You are advised to present you review thematically, with clear sub-headings for the
themes addressed – for example ‘debates about terminology’ ‘theoretical debates’ and so
on.
Your study should have a clear search strategy. This can be presented as a one-page
annex to the written part of the review and will be included within the word limits (see the
example on page 19 of this unit).
Annotated bibliographies or descriptive accounts summarising individual studies are not
acceptable.

103
Make the distinction between references and a bibliography. Your references will be
those works that are directly referred to in your account. Your bibliography may be wider
and include studies that you do not refer to directly in your account but which have
informed your thinking.
Remember to use the Harvard APA referencing system throughout.

Unit Bibliography

Bibliogprahy
Adamson, A. (1986). A student's guide for projects: field studies and research. Oxford:
Tharneson.
Allan, G. and Skinner, C. (Eds) (1991). Handbook for research students in the social
sciences. London: Falmer Press.
Andrews, D.A., I. Zinger, R.D. Hoge, J. Bonta, P. Gendreau and F.T. Cullen (1990) 'Does
correctional treatment work: a clinical relevant and psychologically informed meta-
analysis', Criminology, 28(3): 369-404.
Bell, J. (1989). Doing your research project: a guide for first time researchers in
education and social science (2nd ed.). Milton Keynes: Open University Press.
Berry, R. (1986). How to write a research paper (2nd ed.). Oxford: Pergamon.
Blackbourn, D. (2002). To serve with favour or affection, malice or ill-will. An investigation
into police officers' attitudes towards homosexuality. A comparison of two police forces in
England and Wales. MSc. dissertation, Criminal Justice Studies, University of
Portsmouth.
Blackburn, R. (1993). The psychology of criminal conduct: theory, research and practice.
Chichester: Wiley.
Bottomley, K., & Pease, K. (1986). Crime and punishment: interpreting the data. Milton
Keynes: Open University.
Bradshaw, Y., Twigg, L., Moon, G., & Savage, S. (1992). Metropolitan Police Forensic
Medical Examiner Service (Research report). Portsmouth: Health Information Research
Service/Institute of Police and Criminological Studies, University of Portsmouth.
Bretherick, D. (2006). 'Art of the ugly soul'. An analysis of writings of murderers. Extract
from PhD.
Canter, D. (1995). Psychology of offender profiling. In R. Bull, & D. Carson (Eds),
Handbook of psychology in legal contexts. London: Wiley.
Chalmers, A. (1982). What is this thing called science? Milton Keynes: Open University
Press.
Clarke, C., & Milne, R. (2001). National evaluation of the PEACE investigative
interviewing course. Police Research Award Scheme. London: Home Office.
Cohen, L., & Holliday, M. (1982). Statistics for social scientist: an introductory text with
computer programs in BASIC. London: Harper and Row.

104
Crabtree, B.F., & Miller, W.L. (1992). Doing qualitative research. London: Sage.
Dunsmuir, A., & Williams, L. (1990). How to do social research. London: Unwin.
Ellis, T., & Tedstone, C. (2001). Establishing a base line of sentencer satisfaction in the
youth court: Hampshire sentencer questionnaire 2001. Portsmouth: ICJS, University of
Portsmouth.
Erickson, B., & Nosanchuk, T. (1992). Understanding data. Milton Keynes: Open
University Press.
Evans, K.M. (1984). Planning a small-scale research project: a practical guide for
researchers and students (3rd ed.). Windsor: NFER-Nelson.
Farrington, D.P. (1995). The psychology of crime: influences and constraints on
offending. In R. Bull, & D. Carson (Eds), Handbook of psychology in legal contexts.
London: Wiley.
Farrington, D.P. (1996). Understanding and preventing youth crime. Social Policy
Research, 93. York: York Publishing Services.
Farrington, D. (2002). Risk factors for youth violence. In E. Debarbieux, & C. Blaya (Eds),
Violence in schools and public policies (pp. 13-32). Oxford: Elsevier Science.
Farrington, D.P. and Welsh, B.C. (2002) Effects of Improved Street Lighting on Crime: A
Systematic Review, Home Office Research Study 251, London: Home Office,
http://www.homeoffice.gov.uk/rds/pdfs2/hors251.pdf (accessed 2/4/09).
Farrington, D.P. and Welsh, B.C. (2005) What works in Preventing Crime: Systematic
reviews of Experimental and Quasi-Experimental Research, London: Sage.
Glass G. (1976) 'Primary, secondary, and meta-analysis of research', Educational
Researcher, 5(10):3-8.
Fletcher, R. (2006). Implementing government policy: a case study of the Metropolitan
Police Services' ability to implement the Crime and Disorder Act 1998. Portsmouth: PhD
Research, ICJS, University of Portsmouth.
Ford, L. (2002, 19 April). Tried and tested. Retrieved on 22 April 2004, from within the
Education Guardian website:
www.guardian.co.uk/education/2002/apr/19/academicexperts.highereducation
Frosdick, S. (1998). Public safety and risk management in stadia and sporting venues: a
holistic approach. Fieldwork notes for MSc. dissertation. Cranfield: Cranfield University
School of Management.
Gilbert, N. (1993). Researching social life. London: Sage.
Glass, G. (1976). 'Primary, secondary and meta-analysis of research', Educational
Researcher, 5 (10):3-8
Gore, S. and Drugs Survey Investigators' Consortium (1999) 'Effective monitoring of
young people's use of illegal drugs', British Journal of Criminology, 39(4): 575-603.
Hagan, F. (1993). Research methods in criminal justice and criminology (3rd ed.). New
York: Macmillan.
Handel, J.D. (1978). Introductory statistics for sociology. Upper Saddle River, NJ:
Prentice Hall.
Hart, C. (2000). Doing a literature review. London: Sage.
Hayden, C. and Martin, T. (1998). Safer cities and exclusion from school. Journal of
Youth Studies, 1(3), 315-332.

105
Hibberd, M. (1990a). Writing and research: a manual for police officers. London: Police
Foundation.
Hibberd, M. (1990b). Basic statistics: a manual for police officers. London: Police
Foundation.
Hibberd, M., & Bennet. M. (1990). Questionnaire and interview surveys: a manual for
police officers. London: Police Foundation.
Hill, M., Glaser, K., & Harden, J. (1995). A feminist model for ethical decision-making. In
E.J. Rave, & C.C. Larsen (Eds), Ethical decision-making in therapy: feminist perspectives
(pp.18-37). New York: Guilford Press.
Hinton, P.R. (1995). Statistics explained: a guide for social science students. London:
Routledge.
Home Office (2010). Information sharing aimed at reducing violent crime: A survey of
Community Safety Partnership. (HORS 45). London: Home Office.
Joliffe, D. (2008) 'Researching bullying in the classroom' in R. King and E. Wincup (eds)
Doing Research on Crime and Justice, Oxford: Oxford University Press.
Jolliffe, D. and Farrington, D. P. (2007) A Rapid Evidence Assessment of Mentoring,
London: Home Office.
Jones, S. (2011) Under pressure; women who plead guilty to crimes they have not
committed. Criminology and Criminal Justice, 11(1), 77-90.
Jupp, V., Davies, P., & Francis, P. (Eds) (2000). Doing criminological research. London:
Sage.
Katz, J. (1998). Seductions of crime: moral and sensual attractions in doing evil. New
York: Basic Books.
Kent, J., Williamson, E., Goodenough, T., & Ashcroft, R. (2002). Social science gets the
ethics treatment: research governance and ethical review. Sociological Research Online,
7(4). Retrieved 12 July 2006, from: www.socresonline.org.uk/7/4/williamson.html
Kohnken, G., Milne, R., Memon, A., & Bull, R. (1999). The cognitive interview: a meta-
analysis. Psychology, Crime and Law, 5, 3-27.
Kuhn, T.S. (1970). The structure of scientific revolutions. Chicago: University of Chicago
Press.
Laine de, M. de (2000). Fieldwork, participation and practice: ethics and dilemmas in
qualitative research. London: Sage.
Lakatos, I., & Musgrave, A. (1970). Criticism and the growth of knowledge. Cambridge:
Cambridge University Press.
Langley, P. (1987). Doing social research. Ormskirk: Causeway Press.
Locke, I.F., Spirduso, W.W., & Silverman, S.J. (1987). Proposals that work: a guide for
planning dissertations and grant proposals. London: Sage.
Lösel, F. (2008) 'Doing evaluation research in criminology' in R. King and E. Wincup
(eds) Doing Research on Crime and Justice, Oxford: Oxford University Press.
Lusted, D. (1991). The media studies book. London: Routledge.
MacDonald, G. (2000). Social care: rhetoric and reality. In H.T.O. Davies, S.M. Nutley, &
P.C. Smith (Eds), What works? Evidence-based policy and practice in public services.
Bristol: Policy Press.
McKenzie, I.K., & Bull, R. (2002). Criminal justice research: inspiration, influence and
ideation. Aldershot: Ashgate.
106
Madsen, D. (1983). Successful dissertations and theses: a guide to graduate student
research from proposal to completion. London: Jossey -Bass.
Magee, B. (1973). Popper. London: Fontana.
Marsh, C. (1988). Exploring data: an introduction to data analysis for social scientists.
Cambridge: Polity Press.
Martin, T., Hayden, C., Turner, D., & Ramsall, K. (1999). Out of school and into trouble?
Exclusion from school and persistent young offenders (SSRIU Report No.39).
Portsmouth: Hampshire Constabulary/University of Portsmouth.
May, T. (1993). Social research: issues, methods and process. Milton Keynes: Open
University Press.
Miles, M.B., & Huberman, A.M. (1994). Qualitative data analysis. London: Sage.
Miller, D.C. (1991). Handbook of research design and social measurement. Thousand
Oaks, CA: Sage.
Moser, C.A., & Kalton, C. (1971). Survey methods in social investigations. London:
Heinemann.
Murdoch, J., & Barnes, J.A. (1986). Statistical tables for science, engineering,
management and business studies. London: Macmillan.
Nachmias, C., & Nachmias, D. (1992). Research methods in the social sciences. London:
Edward Arnold.
Newman, O. (1972). Defensible space: crime prevention through urban design. New
York: MacMillan.
Pakes, F. (2001). Spiders in the web: public prosecutors at work. PhD. Leiden University,
the Netherlands.
Parsons, C.J. (1973). Theses and project work: a guide to research and writing.
London: George Allen and Unwin.
Petrosino, A., Boruch, R., Farrington, D., Sherman, L. and Weisburd, D. (2003) 'Towards
evidence-based criminology and criminal justice: systematic reviews and the Campbell
Collaboration Crime and Justice Group', The International Journal of Comparative
Criminology, 3(1), 18-41.
Phillips, E.M., & Pugh, D.S. (1994). How to get a PhD: a handbook for students and their
supervisors. Milton Keynes: Open University Press.
Polsky, N. (1967). Hustlers, beats and others. Chicago: Airline.
Popper, K. (1972). The logic of scientific discovery (3rd ed.). London: Hutchinson.
Potter, S. (2002). Doing postgraduate research. London: Sage.
Propp, V. (1968). The morphology of the folk tale. Austin, TX: University of Texas.
Prowse, R. with Pike, S. (2003). The Dove Project. The Basingstoke Domestic Violence
Family Group conference project. Portsmouth: University of Portsmouth.
Punch, K.F. (2000). Developing effective research proposals. London: Sage.
Reid, S. (1987). Working with statistics: an introduction to quantitative methods for social
scientists. Cambridge: Polity Press.
Robson, C. (1993). Real world research: a resource for social scientists and practitioner-
researchers. Oxford: Blackwell.
Ruderstarn, K.E., & Newton, R.R. (1992). Surviving your dissertation: a comprehensive
guide to content and process. London: Sage.
107
Ruspini, E. (2000). Longitudinal research in social sciences. Social Research Update.
Retrieved on 30 April 2004, from: www.soc.surrey.ac.uk
Savage, S., Charman, S., & Cope, S. (2000). Policing and the power of persuasion.
London: Blackstone Press.
Sayer, A. (1992). Method in social science. A realist approach. London: Hutchinson.
Schofield, M. (1969). Social research. London: Heinemann Educational.
Shawyer, A. (2001). Evaluation of the psychological benefits for inmates running a
prison-based crime diversion scheme for young people. MSc. dissertation. Surrey:
University of Surrey.
Silverman, D. (2000). Doing qualitative research. A practical handbook. London: Sage.
Silverman, D. (2001). Interpreting qualitative data. London: Sage.
Silverstone, D. (2003.) An ethnographic study of drugs and clubs. Popular and academic
criminology. PhD thesis. London: London School of Economics.
Stacey, M. (1969). Methods of social research. Oxford: Pergamon.
Sullivan, D. (2003). Drugs and clubs. PhD thesis. London: London School of Economics.
Tilley, N., & Laycock, G. (2000). Joining up research, policy and practice about crime.
Policy Studies, 21(3), 213-227.
Turner, B.S. (1996). The Blackwell companion to social theory. Oxford: Blackwell.
Vaus, de D. (2001). Research design in social research. London: Sage.
Wakefield, A. (2010) 'Undertaking a criminological literature review' in P. Davies, P.
Francis and V. Jupp (eds.) Doing Criminological Research (2nd ed.), London: Sage.
Walker, R. (1985). Doing research: a handbook for teachers. London: Methuen.
Webb, E., Campbell, D.T., Schwartz, R.D., & Sechrest, L. (1966). Unobtrusive measures:
non-reactive research in the social sciences. Chicago: Rand McNally.
Whyte, W.F. (1955). Street corner society. Chicago: University of Chicago Press.
Wright, D.B. (2002). First steps in statistics. London: Sage.
Yeomans, K.A. (1976). Statistics for the social scientist: introducing statistics.
Harmondsworth: Penguin.
Youngman. M.B. (1986). Analysing questionnaires. Nottingham: University of
Nottingham.
Zimbardo, P.G. (1973). On the ethics of intervention in human psychological research
with special reference to the Stanford Prison Experiment. Cognition, 2(2), 243-55.

108