RESEARCH METHODS – WEEK 4 Bias– e.g.

test worms after deworming without pretest can

EXPERIMENTAL METHOD have loss of basis-> already healthy students is misinterpreted
• is used to determine the effectiveness of a treatment or an as cured
intervention or the cause and effect” relationship of certain PRETEST-POSTTEST DESIGN OR BEFORE-AFTER SURVEY
phenomena under controlled condition. The subjects of the • The study wants to know if there is a change in
study are randomly assigned to the experimental group and characteristics of the study population in a given area. In this
to the control groups and both groups are exposed to similar design, the data is collected before an intervention is
conditions except for the intervention/treatment. introduced, and after the period of the study. The pretest and
• Among the different research methods, the experimental the post test is then compared.
method yields the most conclusive research findings. Bias-e.g. a respondent can know the post test if it is the same
Not all experiments are experimental (experimental only of pretest
you have intervention) TRUE EXPERIMENTAL DESIGNS
SURVEY METHOD • Subjects are randomly assigned to the experimental group
• this study obtains the data to determine specific and the control group to achieve pre-intervention/treatment
characteristics of a group. Its purpose is to get a general equality of the two groups. With this, validity threats are
picture of the characteristics of the study population at a avoided. – all validity threats are avoided
particular time. This is appropriate for most descriptive and STATIC GROUP COMPARISON
correlation studies. • Two groups are involved, experimental group and control
HISTORICAL METHOD group. The experimental group is exposed or receives to the
• used to determine the growth and development of a group, intervention or treatment. The measurement is then followed
organization or institution. The description is based on to which the result is then compared to the result of the
information about some past aspects of the group, control group.
organization or institution. Data from this study are usually • The problem with this design is the validity threat of
collected from secondary sources like records, documents, selection and mortality. It is possible that the two groups
accounts, written materials, etc. greatly differ from each other in the basis of main variables of
Retrospective Type of Study; Positive or Negative results- both the study.
published but latter is difficult (selection threat)-e.g. sect A is exp, sect B, is control; post test
CONTENT ANALYSIS is given but there could be a bias that sect A have more
• Usually used when the intention of the researcher to smarter students already, or (mortality) due to loss
ascertain the quality of message or information found in a -no post test, no fair distribution/selection
document or in mass media. This is also used to test the level PRETEST-POSTTEST CONTROL GROUP DESIGN
of readability of some books before they are printed for -same to static, but have post test and pretest(baseline)
distribution. This is also used in determining the authenticity • The experimental group is exposed to or covered by an
of documents used in literary research. intervention or treatment, like training or new strategy, while
NON/PRE-EXPERIMENTAL DESIGNS the control group is left alone or given another kind of
• This is appropriate for collecting descriptive information treatment. Before the intervention/treatment is introduced to
about a population or subjects of the study. They are the experimental group, a survey/observation/testing is
appropriate for descriptive studies, like profile studies, conducted for both experimental group and control group
exploratory studies, and for doing small case studies. They are using the same device or instrument.
ideal for diagnostic studies or situation analysis and • The baseline (pretest) and end line (posttest) data are
recommended for evaluation studies intended to determine compared. If the change in the “impact or effect indicator/s”
the impact of a certain intervention or treatment. or dependent variable/s is significantly better in the
If less than 10 – case studies, more than 10 – case series experimental area/group than the change in the control
POSTTEST ONLY / AFTER-ONLY SURVEY area/group, then the intervention is considered effective. If
• This is also called as the one shot survey because the data not, then the intervention is said to have had no effect.
are only collected once. This design is used when the study POSTTEST ONLY CONTROL GROUP DESIGN
objective is to describe a situation/condition of a study -static group but has more fair distribution
population as it exist, or to determine/describe the • This is used to determine the effects of an intervention or
characteristics of a population/respondents. There is no treatment introduced to a group of subjects (people or
baseline data. objects).
• The design is cheap and easy to conduct, but results cannot • The experimental group or area is exposed to or covered by
be conclusive in terms of causality or effect of an intervention. an intervention/treatment, while the control group is left
It is not however, recommended for evaluation studies that alone. No pretest/pre-intervention study is conducted. The
intend to measure the effect of a program intervention, like experimental and the control groups are assumed to have
training. similar characteristics at the start of the study. After the
introduction of an intervention in the experimental group or
area, an evaluation survey/observation/testing is conducted SELECTING A STUDY DESIGN
on both experimental and the control groups or areas, using • Ethical Issues - avoid violation of people’s rights and dignity.
the same “fair” instrument. Secure a consent form. –(but need clearance from committee)
• The data gathered is then compared as to how significant • Practical and Administrative Issues – Limited resources
the results are. results in adoption of a less ideal design.
QUASI-EXPERIMENTAL DESIGNS TECHNICAL ISSUES:
• In field studies, it is very difficult to meet the random • Experimental and control groups should be randomly
assignment criterion of a true experimental design. It is nearly assigned.
designs are nearly the same as the true experimental designs • When random assignment is not possible, try to find a
except that the former do not have restrictions of random comparison group that is nearly equivalent to the
assignment. experimental group.
• The two most commonly used quasi-experimental designs • When neither randomly assigned control group nor a similar
are the non-equivalent control group design and the time comparison or group is available, try using time series design
series design. that can provide information on trends before and after a
QUASI-EXPERIMENTAL program intervention.
Non-equivalent Control Group Design • If time series cannot be used, try to obtain baseline
• In field research, it is possible to compare an experimental information that can be compared against post program
group with a similar, but not necessarily equivalent group. information (pretest-posttest).
The two groups need only to have “collective similarity,” • If baseline (pretest) information is unavailable, bear in mind
which means that they should have more or less the same that the type of analysis that you can use is limited.
characteristics in terms of aspects which are relevant to the • Always keep in mind the issue of validity.
study. What is a Variable? -Week 5
Time Series Design  characteristic or feature that varies, or changes within a
• This is similar to the non-experimental pretest-posttest study.
design except that, it has repeated  Is a concept that stands for a variation within a class of
observation/measurements before and after the intervention. objects or persons.
Before the introduction of the intervention/treatment, a  Is a characteristic or property that can take different values
measurement/observation with respect to the impact/effect or attributes.
indicators will be conducted several times at a regular  Are basic elements which are measured in a study which
interval. are observable and measurable.
Examples of variables: Age Sex Marital status Income
OTHER TYPE/DESIGNS OF RESEARCH Types of variables:
• Action Research – focuses on pragmatic and solution driven Dependent Variable
• Case Study and Case Series - detailed contextual analysis of  The outcome variable or the responding variable.
a limited number of events. (not naturally occurring)  Is the assumed effect of the variable
• Causal Design - measure what impact a specific change will  It is the change that occurs in the study population when
have on existing norms and assumptions one or more factors are changed or when an intervention is
• Cohort - conducted over a period of time involving introduced.
members of a population with similarity. –(ginpangtingob)  it responds to a change in the independent variable, so you
• Cross Sectional - measure differences between or from can think of it as depending on the independent variable.
among a variety of people, subjects, or phenomena rather Independent Variable
than change. – e.g. many categories but 1 study  Assumed cause of the problem
• Meta-Analysis - epidemiological study design used to  Reason for any change or variation in a dependent variable.
systematically assess the results of previous research to  Also called antecedent variable or controlled variable.
derive conclusions about that body of research. Mostly uses  The independent variable is the condition that you change
clinical trials. – e.g. done by PHd in an experiment. It is the variable you control.
• Longitudinal - follows the same sample over time and Intervening Variable
makes repeated observations.  Is a factor that works between the independent and
• Philosophical - uses the tools of argumentation derived dependent variable
from philosophical traditions, concepts, models, and theories  It can weaken or strengthen the effect of the independent
to critically explore and challenge on the dependent variables.
• Sequential - carried out in a deliberate, staged approach  Also known as facilitating, moderator or a control variable.
[i.e. serially] where one stage will be completed, followed by  In experimental research, an investigator manipulates one
another, then another, and so on. (e.g. TB programs)
variable and measures the effect of that manipulation on
another variable.

 Example: Deworming – Ind., handwashing – intervening,

results - dependent
 Example:  You want to compare brands of paper towels, to
see which holds the most liquid. The independent variable
in your experiment would be the brand of paper towel. The
dependent variable would be the amount of liquid absorbed
by the paper towel.
 Example:  If you want to know whether caffeine affects
your appetite. The presence/absence of a given amount of
caffeine would be the independent variable. How hungry
you are would be the dependent variable.
 Examples:  In an experiment to determine how far people
can see into the infrared part of the spectrum. The
wavelength of light is the independent variable and whether
the light is observed (the response) is the dependent
variable.
 Example:  You want to determine whether a chemical is
essential for rat nutrition, so you design an experiment. The
presence/absence of the chemical is the independent
variable. The health of the rat (whether it lives and can
reproduce) is the dependent variable. If you determine the
substance is necessary for proper nutrition, a follow-up
experiment might determine how much of the chemical is
needed. Here, the amount of chemical would be the
independent variable and the rat health would be the
dependent variable