2. What is Research Design
• Research Design is the plan and structure of investigation so conceived as to obtain answers to
research questions.
• The plan is the overall scheme or programme of the research. It includes an outline of what the
investigator will do from writing the hypothesis and their operational implications to the final analysis of
data.
• Structure is a framework or configuration of elements related in specified ways. The best way to specify
a structure is to write a mathematical equation that relates the parts of the structure to each other. Such
a mathematical equation, since its terms are defined and specially related by the equation or set of
equations is unambiguous. In short structure is a model of the relations among the variables of the
study.
• Your research design should express both the structure of the research problem and the plan of
investigation used to obtain empirical evidence on the relations of the problem.
Introduction
3. What is the Purpose of Research design
• Research design has two basic purposes:
• (1) To provide answers to research questions and
• (2) To maximise experimental variance
Two Basic Purposes
4. Symbolism and Definitions
• Before discussing designs, explanation of the symbolism to be used is given here:
• X means an experimentally manipulated variable or variables e.g. X1, X2, X3, and so on, though we will use X
alone, even when it can mean more than one independent variable.
• The symbol Ⓧ indicates that the independent variable is not manipulable – is not under the direct control of the
investigator
• The dependent variable is Y: Yb
is the dependent variable before the manipulation of X, and Ya
is the dependent
variable after the manipulation of X.
• With ̴X, means that experimental variable X is not manipulated
• Note that Ⓧ is a non-manipulable variable and ̴X is not manipulated though it is possible to manipulate it.
• 🅁 will be used for random assignment of subjects to experimental groups and random assignment of
experimental treatments to experimental groups.
The symbols
5. The Basic Designs
• Design 19.1: Experimental Group-Control Group: Randomised Subject
• Design 19.1, with two groups as above, and its variants with more than two groups, are probably the best designs for
many experimental purposes in behavioural research.
• 🅁 before the paradigm (or research design) indicates that subjects are randomly assigned to the experimental
group (top line) and the control group (bottom line). With randomisation, all possible independent variables are controlled
at least theoretically. Practically of course this may not be so. If enough subjects are included in the experiment to give
the randomisation a chance to operate, then we have strong control.
• Research design 19.1 can be extended to more than two groups.
Introduction
6. Notion of Control Groups
• The notion of control group can be explained as below.
• Assume that in an educational experiment we have four experimental groups as below.
• A1 is reinforcement of every response
• A2 is reinforcement at regular time intervals
• A3 is reinforcement at random intervals and
• A4 is no reinforcement
• Technically, there are three experimental groups and one control group in the traditional sense of the control group.
However A4 might be another experimental treatment: it might be some kind of minimal reinforcement. Then, in the
traditional sense, there would be no control group.
• The traditional sense of the term control group lacks generality. If the notion of control is generalised the difficulty
disappears. Whenever there is more than one experimental group and any two groups are given different treatments,
control is present in the sense of comparison previously mentioned.
7. Notion of Control Groups
• Thus the traditional notion that an experimental group should receive the treatment not
given to a control group is a special case of the more general rule that comparison
groups are necessary for internal validity of scientific research.
• If this reasoning is correct, we can set up designs such as the following:
• Figure 19.2a
• (Special Diet Experiment)
8. These designs will be more easily recognisable if they are set up in the manner as shown in the next slide.
Figure 19.2b
9. The design on the left is a simple one way analysis of variance design and the one on the right is a 2x2 two factorial design. In the
right hand design, X1a might be experimental and X1b control, while X2a and X2b be either a manipulated variable or a dichotomous
attribute variables. It is of course the same design as shown in Fig 19.2a.
Figure 19.3 a
Figure
19.3b
11. Test for “‘interaction,” or a possible effect due to the peculiar combinations of the
two nominal-scale variables.
12. Data for Factorial ANOVA from Blalock and saved in SPSS as Factorial ANOVA Data from Blalock
For conducting Factorial ANOVA in SPSS, use ANALYSE, General Linear Model, Univariate
13. Design 19.2
• The structure of design 19.2 is the same as that of design 19.1. The only difference is
that subjects are matched on one or more attributes. For the design to take its place
as an adequate design however, randomisation must enter the picture as noted by the
small r attached to the M (for ‘matched’)
Experimental Group-Control Group: Matched Subjects
14. Propensity score matching is a quasi
experimental method in which the
researcher uses statistical techniques to
construct an artificial control group by
matching each treated unit of similar
characteristics. Using these matches, the
researcher can estimate the impact of an
intervention.
Matching is a useful method in data analysis
to estimate the impact of a program for
which it is not ethically or logistically possible
to randomise
Propensity Score Matching
17. QUASI-EXPERIMENTAL RESEARCH
DESIGNS
• “Quasi-experimental methods are research designs that aim to identify the impact of a particular
intervention, program or event (a treatment) by comparing treated units (households, groups,
salaries, schools, firms etc) to control units.
• While quasi experimental methods use a control group, they differ from experimental methods in
that they do not use randomisation to select the control group. Quasi-experimental methods are
useful for estimating the impact of a program or event for which it is not ethically or logistically
feasible to randomise.
• Common examples of what is the experimental methods include difference in differences,
regression discontinuity design, instrumental variables, and propensity score matching.
• In general, quasi-experimental methods require larger sample sizes and more assumptions than
experimental methods in order to provide valid and unbiased estimates of program impacts.
Introduction
18. QUASI-EXPERIMENTAL RESEARCH
DESIGNS
• Like experimental methods, quasi-experimental methods aim to estimate program effects free of
confoundedness, reverse causality, or simultaneous causality. While quasi-experimental methods use a
counterfactual, they differ from experimental methods in that they do not randomised treatment assignment.
Instead they exploit existing conditions or circumstances in which treatment assignment has a sufficient
element of randomness, as in regression discontinuity design or event studies. Or simulate an experimental
counterfactual by constructing a control group as similar as possible to the treatment group as a propensity
score matching. Other examples of quasi-experimental methods include instrumental variables and difference
in differences.
• In general, quasi-experimental methods require larger samples than experimental methods.
• Further, for quasi-experimental methods to provide valid and unbiased estimates of program impacts,
researchers must make more assumptions about the control group than in experimental methods. For
example difference in differences relies on the equal trends assumption, while matching assumes identical
and observed characteristics between the treatment and control groups.
OVERVIEW
19. QUASI-EXPERIMENTAL RESEARCH
DESIGNS
• Video Links
• https://mru.org/courses/mastering-econometrics/introduction-instrumental-variables-p
art-one
• https://mru.org/courses/mastering-econometrics/introduction-differences-differences
• Web links
• https://dimewiki.worldbank.org/Regression_Discontinuity
• https://dimewiki.worldbank.org/Propensity_Score_Matching
• https://dimewiki.worldbank.org/Instrumental_Variables
Some useful links