12-Sapsford -3330-12.

qxd 11/15/2005 3:41 PM Page 272

12
Documents and Critical Research
Victor Jupp

This chapter is concerned with the use of documents in social research.

The range of documents upon which social scientists have drawn includes
diaries, letters, essays, personal notes, biographies and autobiographies,
institutional memoranda and reports, and governmental pronouncements
as in Green Papers, White Papers and Acts of Parliament. This list might
give the impression that such researchers are exclusively concerned with
non-quantitative (sometimes called qualitative) documentary sources.
However, this is not necessarily the case; they are also interested in the
way in which quantitative data are collected, assembled and analyzed in
order to reach conclusions.

Throughout this book a range of analytical strategies has been introduced. Some of
these are structured and formal (as, say, in the analysis of experimental data) and others
are less so; some strategies are founded on the principles of statistical analysis,
whereas others follow the traditions and practices of qualitative research. Analytical
strategies associated with structured and less structured data, and with quantitative
and qualitative analyzes, have found expression within the documentary tradition.
Although these strands will all be represented in this chapter, the main emphasis will
be upon critical analysis of documents. The chapter brings together a distinct
methodological approach – the critical – with particular forms of data – documents.
(It should be recognized, however, that critical research is not exclusive to, and
extends beyond, the use of documents.)
It is important to distinguish ‘criticism’, in its everyday usage, and ‘critical analy-
sis’ as used by social scientists. The former usually refers to an evaluation which is
negative, censorious or fault-finding. Critical analysis in social science involves an
examination of the assumptions that underpin any account (say, in a document) and
a consideration of what other possible aspects are concealed or ruled out. It can also
involve moving beyond the documents themselves to encompass a critical analysis
of the institutional and social structures within which such documents are produced.
For example, Anne Worrall’s Offending Women (1990) is concerned with the
12-Sapsford -3330-12.qxd 11/15/2005 3:41 PM Page 273

Documents and critical research 273

assumptions about femininity that are found in probation reports about women
offenders which are produced, and acted upon, in the criminal justice system in the
UK. As with criticism in its everyday usage, critical analyzes can involve being
censorious or fault-finding, perhaps in terms of rejecting in-built assumptions of
documents or seeking to overturn institutions or systems within which they are
produced. However, this is not a necessary part of critical analysis.
In contrast with some other sections of the book, this chapter makes much greater
reference to theoretical approaches. This is because the critical analysis of text
makes much more obvious and explicit use of theoretical concepts and ideas than
other approaches (for example, survey research, which can often collect data with-
out explicit reference to theory). Indeed, the distinction between theorizing and
empirical research is not one that is readily accepted by those who engage in critical
analysis.

Documents, Texts and Discourse

This chapter adopts the distinction between ‘document’, the medium on which the
message is stored, and ‘text’, the message that is conveyed through the symbols
which constitute writing. Documents can have a number of features. For example,
they may be made up exclusively of written words, or they may include statistics, as
in a survey research report. Documents may refer to particular individuals, as with
school records and reports about pupils, or may concern more ‘macro’ issues, as with
one of Her Majesty’s Inspectorate Reports on the physical state of schools. Further,
documents may refer to contemporary events and issues, as in the case of newspaper
reporting of a prison riot, for example, or they may relate to past events and issues,
as in a nineteenth-century report on conditions in British prisons. Finally, documents
may have been produced for purposes other than social research but none the less
be of interest to researchers, in which case they are sometimes termed unobtrusive
measures (Webb et al., 1966): ‘An unobtrusive measure of observation is any method
of observation that directly removes the observer from the set of interactions or
events being studied’ (Denzin, 1978: 256).
One rationale for the use of such measures is the belief that the effects of the
observer on the data are reduced, thereby improving internal validity. Unobtrusive
measures can derive from a number of sources, such as simple observations of
behaviour without the individuals concerned knowing, or physical traces of behav-
iour left behind by individuals, and can also include documents. An example of the
latter would be institutional memoranda, produced as a normal part of bureaucratic
functioning but to which the social scientist can gain access in order to study key
aspects of institutional processes. Punch (1979b, 1985), for example, has outlined
the use of police organizational records to study corruption among officers working
in a red-light district of Amsterdam. The problem with the use of unobtrusive mea-
sures to study a sensitive issue such as police corruption is that access is vigorously
denied by those who have an interest in doing so.
In other instances, documents may be solicited deliberately and explicitly by
social researchers and may even be produced by them, in which case they cannot be
12-Sapsford -3330-12.qxd 11/15/2005 3:41 PM Page 274

274 Data analysis

treated as unobtrusive measures. This is the case with many life histories and also
with detailed interviews which are recorded and transcribed by social scientists for
subsequent analysis.
There is one further term which deserves consideration in this section, namely
discourse. The dictionary definition of ‘discourse’ refers to talk, conversation and
dissertation. Within social science, it takes on a wider meaning as a result of its close
association with a particular theoretical and methodological position, namely dis-
course analysis (see also below). As with documents and texts, discourses are con-
cerned with communication. However, as Worrall points out, discourse goes much
further ‘to embrace all aspects of a communication – not only its content, but its
author (who says it?), its authority (on what grounds?), its audience (to whom?), its
objective (in order to achieve what?)’ (Worrall, 1990: 8).
‘Discourse’ encompasses ideas, statements or knowledge that are dominant at a
particular time among particular sets of people (for example, ‘expert professionals’)
and which are held in relation to other sets of individuals (for example, patients or
offenders). Such knowledge, ideas and statements provide explanations of what is
problematic about the patients or offenders, why it is problematic and what should
be done about it. In providing authority for some explanations, other forms of expla-
nation are excluded. Implicit in the use of such knowledge is the application of
power. In some instances, discourses may be viewed as imposed by professionals on
clients but this is not necessarily the case. Discourses really come into their own
when the client, for whatever reason and by whatever means, shares the profes-
sional’s analysis of the problem and the means of addressing it. As indicated earlier,
discourse involves all forms of communication, including talk and conversation. In
the latter, however, it is not restricted exclusively to verbalized propositions, but can
include ways of seeing, categorizing and reacting to the social world in everyday
practices, such as policing practices. Its relevance to this chapter is that discourse can
also be expressed in text through the medium of documents.

Types of Document
A wide range of documents has been used in social research, including the following.

Life histories The life history is similar to a biography or autobiography and is a

means by which an individual provides a written record of his or her own life in his
or her own terms. It can include a descriptive summary of life-events and experi-
ences and also an account of the social world from the subject’s point of view. There
is no concern with whether the account is ‘right’ or ‘wrong’; if a subject sees the
world in a particular way then that way is, for that person, ‘right’. The examination
of social perspectives and images is often a forerunner to the analysis of social
actions, the assumption being that actions are underpinned by the way in which the
social world is interpreted by actors.
Life histories may be written by the subject or by a second party, often a social
scientist. A major landmark in the development of life history as a method was
Thomas and Znaniecki’s The Polish Peasant in Europe and America, Volume 1 of
which comprises a 300-page life history of a Polish emigré to Chicago in the early
12-Sapsford -3330-12.qxd 11/15/2005 3:41 PM Page 275

Documents and critical research 275

part of the twentieth century (Thomas and Znaniecki, 1958, first published in
1918–20). It provides an account, not only of life in an American city at that time,
but also of the way in which it was experienced by the immigrant. Following
Thomas and Znaniecki’s work, the life history became an important element in what
was known as the Chicago School of Sociology of the 1930s, which focused
especially on the problems of urban life.

The diary Diaries have been used by both psychologists and sociologists but for
different types of analysis. For example, the psychologist Allport focused on diaries
as the prime means of uncovering the dynamics, structure and functioning of mental
life (Allport, 1942). Luria (1972) used diaries, among other accounts, to explore the
experience of brain damage, leading to short-term memory loss. The sociologist
Oscar Lewis (1959) used diaries to assemble data about the economy, life-style and
daily activities of individuals in poor Mexican families. More recently, the diary has
been used to gain insights into physical conditions and constraints of imprisonment,
and individuals’ subjective experiences, reactions and responses to them, as, for
example, in Boyle’s chronicle of his life in a number of Scottish prisons (Boyle,
1984). In a more formal sense, diaries can be used as part of survey methods, as in
the National Food Survey, in which families are asked to list their food intake over
a period of one week.

Newspapers and magazines The use of newspapers has been central in what is
usually referred to as media analysis. Media analysis has several interests, one of
which is an examination of the way in which stereotypes of categories of people or
types of action are created, reinforced and amplified with wide-ranging conse-
quences for those people and actions. For example, newspapers have been used to
examine the portrayal of ‘folk devils’ such as ‘mods’ and ‘rockers’ (Cohen, 1972),
and the creation and career of the label and stereotype of the ‘mugger’ in the British
press (Hall et al., 1978).

Letters Along with a life history, the analysis of letters played a central part in
Thomas and Znaniecki’s The Polish Peasant in Europe and America (1958). For
example, the authors were able to gain access to letters sent by emigrés to relatives
in Poland, which they used to gain insight into the experiences of assimilation into
American culture. The problem with the use of letters in social research is that they
have a tendency not to be very focused, though where they are they can be a valu-
able source of unsolicited data. This is the case, for example, with letters written on
a specific issue to newspapers, which can be used to identify differing and some-
times conflicting political viewpoints in relation to that issue.

Stories, essays and other writings Researchers can make use of essays or other
writings which are already in existence or can solicit such writings as part of their
research design. For example, analysis of children’s writings has been used to
explore their experience of home, family and social relations (Steedman, 1982),
whereas Cohen and Taylor’s (1972) examination of the subjective experiences of
imprisonment and strategies of psychological survival among long-term prisoners
12-Sapsford -3330-12.qxd 11/15/2005 3:41 PM Page 276

276 Data analysis

was in part founded on an analysis of essays and poems on topics suggested by

Cohen and Taylor themselves. This strategy was consistent with the qualitative,
naturalistic and discovery-based methodological approach, directed at uncovering
the subjective experiences of prisoners, and it was also appropriate on practical
grounds in so far as Cohen and Taylor gained access to the prison and to prisoners
in their role as part-time visiting lecturers. As such, they were in an ideal position to
solicit essays and other writings for research purposes.

Official documents and records A great many official documents and records on
a wide range of topics are available for analysis. An important part of government
activity relates to the production of official documents as, for instance, forerunners
to legislation (as in the case of Green Papers and White Papers), as part of the
regular review of activities of Departments of State or institutions under govern-
ment control (as with reports of Select Committees of the House of Commons), or
as part of official investigations into the running of affairs (as in the case of some
Royal Commissions). Official documents provide valuable data for the analysis of
official definitions of what is defined as problematic, what is viewed as the expla-
nation of the problem, and what is deemed to be the preferred solution. In this way,
analysis of such documents provides an important element in the critical analysis
of texts.
Apart from documents at a societal or macro level, there are other official docu-
ments at an institutional or micro level which can be just as important to the disposal
and destiny of individuals. These are organizational records which define what is, or
is not, problematic about individuals, which put forward explanations for behaviour
and actions and which record decisions relating to outcomes. Of course, such indi-
vidual records are not necessarily separate from official documents operating at a
societal level in so far as there is often a close connection between the formulation
of concepts, explanations and solutions at one level and such formulation and appli-
cation at another. Official records from a variety of settings have been examined. For
example, Kitsuse and Cicourel (1963), working within an interpretative approach,
analyzed pupils’ records to make assertions about the labels and stereotypes assigned
by teachers, and about the consequences of these. Anne Worrall (1990), working
within the critical approach, examined the discourses of solicitors, magistrates, psy-
chiatrists and probation officers, and the consequences these have for the disposal of
women offenders.

Research reports Finally, documents of particular interest to us in a book on

research methods are reports of social science research written by academics and
other researchers (perhaps government-sponsored researchers). The position taken in
this chapter is that such reports should not be treated as objective, accurate state-
ments of ‘fact’, but as documents which require examination and challenge in terms
of what they define as problematic, the way in which such problems are opera-
tionalized, the forms of explanation put forward and the policy implications which
flow from them. A critical analysis of particular research reports is important in
instances where such reports have a high profile or hold an influential position in the
public domain.
12-Sapsford -3330-12.qxd 11/15/2005 3:41 PM Page 277

Documents and critical research 277

Box 12.1 A classification of documents

Authorship
Personal Official
Private State
Closed 1 5 9
Access Restricted 2 6 10
Open-archival 3 7 11
Open-published 4 8 12
Source: Scott, 1990: 14

A Typology of Documents
Scott (1990) has produced a typology of documents based on two main criteria:
authorship and access (see Box 12.1). ‘Authorship’ refers to the origins of docu-
ments and under this heading he distinguishes ‘personal’ documents from ‘official’
ones (which have their source in bureaucracies). Official documents are further sub-
divided into ‘state’ and ‘private’ (non-state: for example, business annual reports and
accounts). The second criterion, ‘access’, refers to the availability of documents to
individuals other than the authors. ‘Closed’ documents are available only to a limited
number of insiders, usually those who produce them; ‘restricted’ documents are
available on an occasional basis provided permission has been granted; ‘open-
archival’ documents are those documents which are stored in archives and are avail-
able to those who know of them and know how to access them; ‘open-published’
documents are the most accessible of all and are in general circulation.
Such classifications can be useful in themselves. However, for Scott (1990) the
usefulness of a classification based on the criteria he suggests is that it poses four
key questions pertaining to the validity of particular documentary sources. Who has
and has not authored a document, and the degree to which a document is accessible
or withheld, influences its authenticity (whether it is original and genuine), its cred-
ibility (whether it is accurate), its representativeness (whether it is representative of
the totality of documents of its class) and its meaning (what it is intended to say).

Critical Social Research

Harvey (1990: 1) distinguishes critical social research as follows:

Critical social research is underpinned by a critical-dialectical perspective which attempts
to dig beneath the surface of historically specific oppressive social structures. This is con-
trasted with positivistic concerns to discover the factors that cause observed phenomena or
to build grand theoretical edifices and with phenomenological attempts to interpret the
meanings of social actors or attempt close analysis of symbolic processes.
12-Sapsford -3330-12.qxd 11/15/2005 3:41 PM Page 278

278 Data analysis

This quotation reveals some of the differences between critical research and, on the
one hand, positivism (which is often, but not exclusively, associated with quantita-
tive research such as surveys, experimentation and content analysis) and, on the
other hand, phenomenology, which is roughly equivalent to what we have termed the
interpretative tradition (and often, but not exclusively, associated with ethnographic
research). The differences that are highlighted are as follows: first, positivism
emphasizes explanations cast in causal terms, whereas critical research does not;
secondly, while both interpretative and critical perspectives are concerned with
social meanings, the former places emphasis on how these are generated in small-
scale interactions, whereas the latter seeks to analyze them critically in terms of
structural inequalities in society (for example, class, race or gender inequalities).
Within the social sciences, the critical tradition owes much to Marx or to rework-
ings of Marx by other writers. Critical research which is influenced by this source is
concerned with social structural inequalities founded on class inequalities. The work
of the American sociologist, C. Wright Mills, was influenced by the Marxist tradi-
tion but was less explicitly class-based in directing its attention at bureaucratization
in mass society and at the concentration of power in a power elite (see especially
Mills, 1956). During the 1970s, the critical tradition received impetus from the rise
of Black movements and from feminism. This led to the examination of structures
founded on race and gender inequalities.
There are variations within the critical tradition. Nevertheless, a number of central
assumptions are discernible. First, prevailing knowledge (for example, that provided
in official documents such as reports of Royal Commissions) is viewed as being struc-
tured by existing sets of social relations which constitute social structure. Secondly,
this structure is seen as oppressive in so far as there is an unequal relation between
groups within it and in so far as one or more groups exercise power over others.
Thirdly, the inequality, power and oppression are rooted in class, race or gender or
some combination of these. Fourthly, the aim of critical analysis is not to take pre-
vailing knowledge for granted or to treat it as some ‘truth’, but to trace back such
knowledge to structural inequalities at particular intersections in history. In doing so,
it is considered important to examine the role of ideology in the maintenance of
oppression and control and also the way in which social processes and social institu-
tions operate to legitimate that which is treated as knowledge. Ultimately, the aim of
critical research and analysis is to confront prevailing knowledge – and the structures
which underpin it – by providing an alternative reading and understanding of it.
A final point relates to emancipation. For Fay (1987), for example, it is not suffi-
cient that critical research enlightens oppressed groups by providing an analysis of
the root causes of such oppression. Such enlightenment should lead to emancipation:
By offering this complex set of analyses to the relevant group at the appropriate time in the
appropriate setting, a social theory can legitimately hope not only to explain a social order but
to do so in such a way that this order is overthrown. (Fay, 1987, in Hammersley, 1993: 36)

The twin concepts of ‘deconstruction’ and ‘reconstruction’ are central to much

critical research. Deconstruction is the process by which prevailing knowledge, or
any construct within it, is broken down into its essential elements. This can involve
the collection of empirical data and the examination of such data in relation to the
abstract constructs that constitute knowledge. Reconstruction involves the rebuilding
12-Sapsford -3330-12.qxd 11/15/2005 3:41 PM Page 279

Documents and critical research 279

of a construct in terms of the oppressive social structural arrangements which underpin

it and sustain it.
An example is required in order to illustrate what is otherwise an abstract set of
prescriptions for analysis. We can take the construct ‘housework’, which can be
deconstructed or broken down in terms of a set of activities and tasks which are
viewed within prevailing knowledge as constituting its essence (for example, wash-
ing dishes, ironing clothes). The process of reconstruction involves an examination
of this construct in terms of wider structural arrangements, especially gender
inequalities in society. It may also provide an analysis in terms of class (e.g. a study
of working-class housewives) and class and race (e.g. a study of Black working-class
housewives). Such reconstruction views ‘housework’ not as a set of activities, such
as washing dishes, making beds and so forth, but as an exploitative relationship
within a social structure with patterned inequalities and oppressions.

Critical Analysis of Documents

Critical analysis is explicitly theoretical. However, empirical work has been and is
carried out, including social surveys, detailed interviews, social history research,
participant observation and, of course, the analysis of documents. (For examples of
the use of each of these in critical research, see Harvey, 1990.) The contribution which
the analysis of documents can make within the critical research tradition is outlined
by Jupp and Norris (1993). Their exposition is based upon ‘discourse analysis’, an
important development within the critical paradigm, stemming from the work of the
French social theorist Michel Foucault. In general terms, discourse analysis has a
number of features. One key assumption is that discourse is social, which indicates
that words and their meanings depend on where they are used, by whom and to
whom. Consequently, their meaning can vary according to social and institutional set-
tings and there is, therefore, no such thing as a universal discourse. Secondly, there
can be different discourses which may be in conflict with one another. Thirdly, as well
as being in conflict, discourses may be viewed as being arranged in a hierarchy: the
notions of conflict and of hierarchy link closely with the exercise of power. The con-
cept of power is vital to discourse analysis via the theoretical connection between the
production of discourses and the exercise of power. The two are very closely inter-
woven and, in some theoretical formulations, are viewed as one and the same.

Box 12.2 A discourse-analytic research agenda

1 What public and/or institutional discourses are important in terms of

knowledge of what is ‘right’ and what is ‘wrong’?
2 In what kinds of documents and texts do such discourses appear?
3 Who writes or speaks these discourses and whom do they represent
or purport to represent?
4 What is the intended audience of such writing or speech?
5 What does a critical reading of these documents uncover in terms of:

(Continued)
12-Sapsford -3330-12.qxd 11/15/2005 3:41 PM Page 280

280 Data analysis

Box 12.2 (Continued)

(a) what is defined as ‘right’ and ‘wrong’ and therefore what is seen
as problematic;
(b) what explanation is offered for what is seen as problematic;
(c) what, therefore, is seen as the solution?
6 What does a critical reading of these documents tell us about
(a) what is not seen as problematic;
(b) which explanations are rejected or omitted;
(c) which solutions are not preferred.
7 What alternative discourses exist?
8 How do these relate to ‘internal differentiation’ within and between
semi-autonomous realms of control?
9 What does a critical reading of these alternative discourses tell us?
10 Is there evidence of negotiation with, or resistance to, dominant
discourses?
11 What is the relationship between the discourses and social conflict,
social struggle, hierarchies of credibility, order and control and, most
of all, the exercise of power?
12 Are discourses, knowledge and power pervasive or reducible to
class, class conflict and struggles refracted through one source, the
state?
Source: Jupp and Norris, 1993: 50

The Foucauldian approach to discourse analysis is distinctive on a number of

counts, including the position that discourse and power are one and the same:
‘Power produces knowledge, they imply one another: a site where power is exer-
cised is also a place at which knowledge is produced’ (Smart, 1989: 65). What is
more, Foucault’s position is that there is not one focus of knowledge and power (the
state) but several:
His viewpoint is that strategies of power and social regulation are pervasive and that the
state is only one of several points of control. This is an important divergence from
Marxist analysis. For Foucault there are many semi-autonomous realms in society, where
the state has little influence, but where power and control is exercised. In this way
Foucault’s notion of the pervasiveness of loci of regulation and control encourages
research about discourses in a range of institutional settings. (Jupp and Norris, 1993, in
Hammersley, 1993: 49)

The ways in which research may be carried out in such settings are laid out in the
‘research agenda’ suggested by Jupp and Norris (see Box 12.2). The ‘agenda’ brings
together questions which typically would be asked in a critical analysis of docu-
ments, especially with reference to discourse analysis. It is unlikely that any given
analysis will deal with all these questions; rather, it will tend to focus on some to the
exclusion of others.
12-Sapsford -3330-12.qxd 11/15/2005 3:41 PM Page 281

Documents and critical research 281

Critical analysis, and discourse analysis in particular, has a tendency towards the
theoretical. It is appropriate, especially in a book concerned with social research
methods, to consider how an abstract set of ideas and concepts can be converted into
a programme for research. This will be done in the following sections via a number
of case studies, each of which uses different types of documents and represents a
different selection of research questions, from the above agenda, with which to
address the documents.
The first case study is of a fictitious research proposal to carry out discourse
analysis on what – using Scott’s typology (Box 12.1) – can be called ‘state’ docu-
ments which are ‘open-published’; that is, they are in general circulation. This case
study is especially useful because it shows how a particular theoretical system can
be turned into a programme of research. The second case study shows the end prod-
uct of a critical analysis of an open state document. It illustrates the conclusion that
one social scientist, Mike Fitzgerald, reached after a critical ‘reading’ of a report on
prisons. The third case study is based on institutional records and transcripts of
detailed interviews with professionals in the criminal justice system to examine
decision-making regarding the disposal of women offenders. The final case study
involves a different form of document, a report of survey findings produced by
social researchers. This case study illustrates the difference between a critical
analysis of text and a ‘technical’ evaluation of research design and the findings
derived from it.

Case Study 1: A Proposal for Critical Analysis

Activity 12.1 (allow 30 minutes)

You should now read Box 12.3, an example of a research proposal based on
critical analysis of a text. Write notes on these questions:
1 Which, if any, of the research questions included in the agenda outlined
earlier are represented in the proposal?
2 What method of enquiry is advocated?

The research proposal puts the analysis of the report of the Woolf Inquiry at its
centre. In doing so, it enlists theoretical ideas from Foucault, particularly the view-
point that society comprises an array of discourses which express and produce moral
norms defining what are ‘right’ explanations and techniques of control. The report
of the Woolf Inquiry is one such official discourse relating to prisons. It provided
official definitions of what is wrong with prisons, why these problems exist and how
they should be solved. (The precise recommendations are not reproduced here: for a
useful summary and commentary consult Sparks, 1992.) The theoretical ideas
derived from Foucault generate research questions to be asked of the Report at two
levels.
12-Sapsford -3330-12.qxd 11/15/2005 3:41 PM Page 282

282 Data analysis

Box 12.3 A research proposal

Introduction
The transformation of a sequence of events into what becomes defined as
a deep-seated ‘social problem’ is often marked by the setting up of a pub-
lic inquiry. The theoretical analyzes and moral perspective of such
inquiries play a large part in determining public perception of ‘normality’
and ‘dangerousness’. In 1990 serious rioting took place in Strangeways
Prison, Manchester. The government of the day commissioned Lord
Justice Woolf to conduct inquiries into the rioting. Woolf decided to adopt
a broad interpretation of his terms of reference to address wider issues
which the disturbances raised (for example, physical conditions in prisons,
the use of local prisons to keep individuals on remand, the extent of over-
crowding). The Report of the Woolf Commission was published in 1991
(Woolf, 1991).
Research questions
Two sets of important questions can be asked about the Woolf Report, or,
indeed, any other official report. First, we can try to ascertain the nature
of the official discourse it represents:
• How does Woolf define the problems of prisons in the 1990s?
• What range of explanations does he consider?
• What does he propose as the control solution?
Secondly, and more generaIly, we could investigate the role of such public
‘voices’ as Woolf’s, perhaps by comparing the Report with other official or
quasi-official reports. For example, in relation to crime and criminal justice,
we could undertake ‘readings’ of the Scarman Report on the Brixton dis-
orders of 1981 (Scarman, 1981) or of the Taylor Report on the Hillsborough
disaster (Taylor, 1990). However, we need not restrict ourselves to this
area of concern. Instead, we can investigate a wide range of official
reports (on health, education and housing). The important questions to
ask are:
• What is the audience addressed by these official reports and for whom
do they speak?
• What influence do reports of this kind have on what happens in agen-
cies of social control?
Theoretical frameworks
Much of the interest today in official discourses stems from the influence
of Michel Foucault on social science. Foucault envisages society not as
something ‘out there’ which causes, and is in turn reacted upon by, cer-
tain kinds of knowledge or social policy. Rather, ‘society’ comprises an
array of discourses which exhibit and produce moral norms, theoretical
explanations and techniques of social control. These three aspects of
social regulation are, in Foucault’s view, quite inseparable. So, the first
three research questions listed aim to try to establish the various compo-
nents of official discourse about problems in prisons and the overall
moral climate such discourse creates.
12-Sapsford -3330-12.qxd 11/15/2005 3:41 PM Page 283

Documents and critical research 283

Box 12.3 (Continued)

The second set of questions gets us to think about who is represented in

public discourse of this kind. On whose behalf does Woolf speak: the liberal
professions, the ruling class, the Establishment? And whom is he address-
ing: the moral majority, the British public, the respectable white male citizen?
It is important to recognize here that, for Foucault, these ‘subjects’, on both
sides, are not concrete individuals or groups existing outside the field of the
discourse itself. Rather, they are ‘ideal’ positions which are produced in and
through such discourses, serving as powerful moral regulators. The last
questions further reflect Foucault’s view that official discourse is only one
type among others, and that the social priorities established in any given
discourse may well be undercut or qualified by those established in other
discourses (such as those of the media or the police).
Methods of enquiry
This project involved ‘reading’ and reflecting upon Woolf and similar offi-
cial reports, looking closely at the way in which language is used, and at
the values involved, so as to produce the typical ‘subject’ of the discourse.
Reports embody certain types of theory or knowledge – which may be
embodied in policy and institutions – about what or who is the problem,
about what is the explanation and about what is the ‘correct’ solution.
Considerations of power are deep-embedded in such theory and knowl-
edge. The purpose of ‘reading’ is to apprehend such theory, knowledge
and power. This type of approach does not accept a distinction between
the ‘theoretical’ and the ‘empirical’ modes of investigation.

First of all, one set of questions is asked of the document itself. These questions are
concerned with what is defined as problematic (and, by implication, what is not
defined as problematic); the explanations or theories that are provided (and, by
implication, the explanations that are omitted or rejected); the solutions that are
offered (and, by implication, the solutions that are rejected). These are typical of
questions 5 and 6 of the research agenda given in Box 12.2.
A second set of questions relates not to the document itself but to the ‘subjects’ on
either side, asking on whose behalf the report speaks and to whom it speaks. These
are close to questions 3 and 4 of the research agenda. Note that, in contrast to an
interpretative approach, the focus in the approach advocated in this proposal is not
upon the actual person who wrote the report, nor is it upon the actual people who
read it. Rather, it is upon ‘ideal’ positions that are produced in and through such dis-
courses, serving as powerful regulators.
With regard to methods of enquiry, the position adopted is in complete contrast to
that of positivist content analysis of documents (which is similar to the analysis of
survey data). There is no reference to formal protocol of categorization, coding and
counting of words, themes, headlines or column inches. Rather, the project involves
‘reading’ and ‘reflecting’ and is founded upon an approach that does not accept that
there are two separate yet interrelated activities of theorizing and empirical research
carried out by two different kinds of people: theorists and research technicians.
12-Sapsford -3330-12.qxd 11/15/2005 3:41 PM Page 284

284 Data analysis

Case Study 2: Critical Analysis of a Public Document

The second case study is based on an article entitled ‘The telephone rings: long-term
imprisonment’, by Mike Fitzgerald (1987), which represents a critical analysis of a
public document about prisons: the Report of the Control Review Committee.
Fitzgerald focuses on a number of official reviews, inquiries and policy papers that
have followed disorders in prisons, with a view to uncovering the principles of penal
policy that underpin their recommendations. In terms of the typology outlined earlier,
the documents can be classified as official state open published.
In his consideration of the Report of the Control Review Committee, in particular,
Fitzgerald focuses on three main areas that make up what he calls the ‘general orien-
tation’ of the Report. This general orientation, and its three sub-areas, constitute the
object of analysis. First, he is concerned with the concepts that the Committee
employs to define that which it sees as problematic, and he then goes on to decon-
struct some of these concepts to see how they themselves are defined. For example,
he notes that ‘control’ is central to the Committee’s conceptualization and also that
control is defined in terms of the control of the prisoners themselves, instead of in terms
of problems in the Prison Service. In short, this form of analysis asks why certain kinds
of concepts, defined in certain ways, are placed on the public agenda.
Secondly, he is concerned with the kinds of solution that emerge from the
Committee’s thinking. As he points out, such solutions are largely in terms of new
prison designs to increase security and control and not in other terms, such as
improving prisoner-staff relations. The preferred solution does not stand in isolation
but flows directly from the way in which the Committee conceptualizes what is seen
as being problematic in the prison system.
Thirdly, Fitzgerald analyzes the implicit theory of management that underpins
both the conceptualization of the problem and the preferred solution and its imple-
mentation. As Fitzgerald argues, he is not against management per se but questions
the rigid hierarchical theory of management that dominates the thinking of the
Committee.
At this point it is appropriate to summarize some of the key features of the criti-
cal approach to documents in the light of Fitzgerald’s analysis of the Report of the
Control Review Committee.
1 The sources of data are often, but not always, official texts which are important
at a macro level in so far as they put forward conceptualizations regarding, in this
case, the prison system, although they could refer to any other element of the
social system.
2 The method does not exhibit the formal protocols of quantitative content analy-
sis (categorizing, coding, counting), but is a critical reading of texts aimed at
uncovering how problems are defined, what explanations are put forward and
what is seen as the preferred solution. It also seeks to bring to the surface that
which is rejected in the text and that which does not even appear: what is not seen
as problematic, what explanations are not considered, and what are not the
preferred solutions. In other words, the analysis is concerned with how official
documents frame the public agenda.
3 Fitzgerald is not solely concerned with analyzing the definitions, explanations
and solutions put forward in official documents, but seeks to challenge them and
12-Sapsford -3330-12.qxd 11/15/2005 3:41 PM Page 285

Documents and critical research 285

suggests alternative proposals and viewpoints. In this sense, the methodological

approach is not exclusively a critical reading of the text, but is also a challenge
to the text.
4 The paper is concerned solely with the ‘communication’; that is, with the text. It
could have gone beyond this by examining the ‘senders’ and the ‘recipients’. Had
it done so it would not have been concerned with the identity of individual
authors or with the meanings they bring to the text as someone from the inter-
pretative tradition would be, but with the section of society for whom the docu-
ment speaks, and with the consequences for the prison system and its inhabitants.

Case Study 3: Critical Analysis of Decision-making

The next example, also within the critical tradition, is slightly different from the pre-
ceding two case studies, in so far as it is not concerned with macro state-originated
documents. Rather, it involves transcripts of detailed interviews with ‘experts’ within
the criminal justice system and institutional records. What is more, it is not solely
concerned with one discourse, namely that represented in official state reports on
prisons, but with a multiplicity of interacting discourses which have consequences
for the decisions made regarding women who offend.
Anne Worrall’s Offending Women (1990) is concerned with the ways in which
female offenders experience different treatment from that experienced by male
offenders. She examines how this situation occurs and is perpetuated, and also charts
the implications for women offenders. This is done by analyzing the discourse
surrounding women’s deviancy. This discourse is accessed by interviews with pro-
fessionals in criminal justice and welfare agencies – probation officers, magistrates,
solicitors, psychiatrists – as well as women offenders themselves. The transcripts of
such interviews represent documents appropriate for critical analysis. In addition,
Worrall had access to case records of offenders made available by probation officers.
The position taken by Worrall eschews social science which, on the one hand, is
concerned with the search for universal properties and causes and, on the other hand,
is solely concerned with social meanings. She is not interested in questions of what
is the ‘truth’, but rather with ‘the relationship between those who claim to know the
“truth” and those about whom they claim to know it’ (Worrall, 1990: 6). In turn, this
relates to the question, ‘What is it that endows certain individuals to have such
knowledge and to apply it?’ These are the hallmarks of critical analyzes in general
and of discourse analysis in particular.
The relationship between power and knowledge is vital to such analyzes.
Worrall’s viewpoint is that knowledge does not of itself give power. Rather, those
who have power have the authority to know. In this context, such people are magis-
trates, probation officers, solicitors and psychiatrists. Power is not reducible to one
source, class (as Marxist analyzes would have it), but exists in all social relations. In
this case, it exists in the relations between women offenders and those who make
decisions about them, and also in the relations between such decision-makers them-
selves. Within this analysis, the exercise of power is not the naked oppression of one
group by another but the production and subtle application of coherent ‘knowledge’
about other individuals which has consequences for what happens to these individuals
(for example, Social Inquiry Reports, written by probation officers about offenders,
12-Sapsford -3330-12.qxd 11/15/2005 3:41 PM Page 286

286 Data analysis

which can influence decisions taken about such offenders). This is discourse as discussed
in the section on ‘Documents, Texts and Discourse’.
Discourses have implications for practice in terms of programmes, technologies
and strategies: that is, coherent sets of explanation and solutions, ways of imple-
menting these solutions and strategies of intervention. The discussion of discourse
analysis (see ‘Critical Analysis of Documents’ above) indicated that there can be
differing and competing discourses. In this respect, Worrall suggests that the power
of the offender lies in the ability to resist, and even refuse, the coherent and homo-
geneous discourse of ‘experts’: ‘By demonstrating the existence of heterogeneity
and contradiction, the speaking subject is helping to keep open the space within
which knowledge is produced’ (Worrall, 1990: 10). In the main, however, women
offenders remain markedly non-resistant and ‘muted’.
The methodological approach is one of a case study of detailed interviews with
magistrates, probation officers, psychiatrists and solicitors and of institutional
records and reports. It has no claims to randomness or representativeness (as, for
instance, a social survey would) and it seeks to generalize via theorizing rather than
by reference to probability theory (again, as a survey would): ‘The adoption of this
particular mode of theorizing women’s experiences calls for a method of research
which rejects notions of generalizability through probability in favour of generaliza-
tion through theoretical production’ (1990: 12). As with the first case study, there is
rejection of the viewpoint that there are two distinct activities, theorizing and empir-
ical enquiry.
The main conclusion of Worrall’s work is that women are ‘muted’ within the
criminal justice system by being subject to the multiple discourses of the ‘experts’ who
are authorized to present coherent knowledge concerning problems, explanations
and solutions and who deny legitimacy to the discourses of the women themselves.
Worrall’s analysis involves deconstructing the discourses of the ‘experts’. Despite
the power and authority of such discourses, offenders develop means of resisting
them by exploiting construction within them:
Yet, while much of the women’s resistance is individualistic, inconsistent and, in some, self-
destructive, it has the important effect of undermining the authority of official discourses and
keeping open the possibility of the creation of new knowledge about them – both as women
and as law-breakers. (Worrall, 1990: 163)

The contribution which this case study makes to the discussion of critical analysis
is that, in comparison with the other examples, it shows that there can be a multipli-
city of discourses, that these can operate in subtle ways, that there can be resistance
to prevailing discourses, and that outcomes have a good deal to do with the positions
of particular discourses in the hierarchy of legitimacy and authority.

Case Study 4: Critical Analysis of a Research Report

The final case study relates to the critical analysis of a research report. This is done
to emphasize that such reports are themselves documents and as such are not
immune from critical analysis. It is also done to emphasize that social science
research reports are social constructions with their own in-built assumptions about
what is problematic and why.
12-Sapsford -3330-12.qxd 11/15/2005 3:41 PM Page 287

Documents and critical research 287

Almost any output from social science research would be appropriate by way of
illustration. Here the selection is from the Cambridge Study of Delinquency
Development, a prospective longitudinal study of 411 males in the London Borough
of Camberwell, started in 1961. At the time when they were selected for inclusion in
the sample the target males were eight years old.
The primary aim of the programme of research was to describe the development
of delinquency in inner-city males and to investigate how far delinquent and crimi-
nal behaviour can be predicted in advance. In addition, the researchers wanted to
explain why certain individuals continue offending into adulthood. The sample
members were interviewed at various points of time from the age of eight up until
their thirties. About one fifth of the sample had been convicted of criminal offences
as juveniles. These members differed significantly from the remainder of the sample
on many counts: for example, on a scale of anti-social behaviour. Over one third of
the sample had been convicted of a criminal offence by the time they were 32 years
old. Those who were convicted as juveniles tended to be the persistent offenders at
a later age. Six variables about which data were collected over the period of the
survey are suggested by the researchers as predictors of delinquency: poverty, poor
parenting, family deviance, social problems, hyperactivity (plus impulsivity and
attention deficiency), and anti-social behaviour. There have been many outputs from
the Cambridge Study. One summary of findings is provided in Farrington (1989).

Activity 12.2 (15 minutes)

Chapters 1 and 2 of this book cover various aspects of survey sampling. (Indeed,
the Cambridge Study is outlined in Chapter 1 as an example of a longitudinal
study.) In using the principles addressed in these chapters, either to plan survey
research or to assess research reports, one would be led to what can be called a
technical evaluation. This involves asking the fundamental question: ‘What
would fellow researchers and policy-makers ask of the research design of any
study in order to gain an assessment of the validity of the findings derived from
it?’ Look back over these chapters and consider what features of the research
design of the Cambridge Study you would address in order to assess its validity.
(If possible, refer to the summary presented by Farrington, 1989.)

Typically, a technical evaluation would be concerned to examine the representative-

ness of the sample and the validity of making generalizations from the sample to
wider populations. The survey involved a sample of 411 working-class, predomi-
nantly white boys. This makes generalizability to girls, to the middle class and to
ethnic minority groups questionable. A second question to be asked in a technical
evaluation relates to sample attrition or drop-out, which is especially relevant to a
longitudinal study requiring long-term commitment on the part of sample members
and detailed procedures to ensure that they can be traced and contacted. Two aspects
are particularly important: the size of the drop-out can seriously deplete the sample,
12-Sapsford -3330-12.qxd 11/15/2005 3:41 PM Page 288

288 Data analysis

and the drop-out may affect the representativeness of the sample in so far as the
drop-outs and those who are difficult to trace may have distinctive features and expe-
riences which correlate with delinquency. Other aspects to consider in a technical
evaluation include the possibility of fading relevance (the issues addressed at the
beginning of a longitudinal study may have little relevance at the end), expectancy
effects (the sample members may become so accustomed to the themes of the study
that they answer questions in the ways which they believe are expected of them), and
causal inference (the researchers may make fallacious inferences regarding causal-
ity on the basis of correlational evidence).

Activity 12.3 (15 minutes)

The reason that attention has been paid to matters to be addressed in a techni-
cal evaluation is that it facilitates comparison with matters to be addressed in a
critical analysis. Reflect back over this chapter and consider what questions
should be asked in carrying out a critical analysis of what you know about the
Cambridge Study.

The sophistication of such a critical analysis will obviously depend on the level of
knowledge of the Cambridge Study and of criminology in general. However, at an
elementary level, a number of key points can be made. For example, the agenda pre-
sented in Box 12.2 provides a starting point, and within that agenda questions 5 and
6 are especially crucial. Question 5 concerns what is seen as problematic, and what is
seen as the solution. In the Cambridge Study, the central problem is conceptualized
in terms of delinquent behaviour and subsequent criminal behaviour in adulthood in
modern British society. Farrington (1989) describes those engaging in such behaviour
as working-class males who are typically characterized by tattoos, heavy drinking,
heavy smoking, drug-taking and fighting. The explanation is in terms of the six pre-
dictors mentioned above about which the researchers collected data throughout the
study. The solutions are based upon using the six predictors to identify those who are
likely to offend, followed by intervention in their lives in order to influence what are
seen as the causal agents, thereby diverting individuals from potential criminal
careers. Some of the ‘solutions’ that have emanated from this kind of analysis,
although not necessarily directly from the researchers themselves, have included edu-
cation and guidance for parenthood, pre-school education for children of disadvan-
taged parents and family support programmes for so-called ‘problem families’.
Asking question 6 in the research agenda – what is not asked in the research –
gives a critical edge to the analysis. This is not to be critical of the authors as indi-
viduals or as researchers; rather, it is to question what is missing in order to uncover
and clarify the assumptions built into the discourse of the research report. With
regard to the Cambridge Study, what is not identified as problematic is the range of
other types of criminal activity typically associated with white-collar and corporate
crime; the people who are not seen as problematic are middle-class men who wear
12-Sapsford -3330-12.qxd 11/15/2005 3:41 PM Page 289

Documents and critical research 289

suits, drink wine and work in key financial institutions; and the kinds of explanation
that are not assembled are those in terms of crime as an outcome of power, opportu-
nity and structural inequalities in society. Instead, the explanation emphasizes
individual and familial variables, which is consistent with the dominant discourse
about crime and other forms of behaviour in the 1990s.

Conclusion
Critical research is not confined to the analysis of documents (see, for
example, Harvey, 1990). However, the focus in this chapter has been on
critical analysis of documents, especially in terms of documents as media
for discourses. The four case studies have illustrated a range of key points
regarding such analysis. For example, it can involve official public docu-
ments addressing macro issues, or it can involve institutional or personal
documents influencing the disposal of individuals at a micro level. Critical
analysis is characterized by not taking for granted what is being said in a
document and what is often assumed to be ‘knowledge’.
There are no formal protocols to the strategy of analysis, as there are,
say, in the design of an experiment or a survey. Instead, critical analysis
involves uncovering what is being treated as knowledge – often by
addressing what is not being treated as knowledge – and examining the
consequences of such knowledge. Critical analysis is different from techni-
cal evaluation, and as such asks different questions from those asked by
other styles of research. In all of these senses, critical researchers go about
their business in different ways from those employed by other researchers.

Key Terms
Critical analysis examining the assumptions that underpin an
account, and probably the institutions and social structures within which
it was produced and/or is being used.
Deconstruction breaking down a prevailing knowledge into its essen-
tial elements.
Discourse (here) ideas, statements or knowledge dominant at a
particular time among a particular set of people.
Reconstruction rebuilding a construct to emphasize or to overcome
underlying oppressive elements.

Further Reading
Fairclough, N. (1992) Discourse and Social Change, Cambridge, Polity.
Garland, D. and Young, P. (eds) The Power to Punish, Aldershot, Gower.
Harvey, L. (1990) Critical Social Research, London, Unwin Hyman.