Lecture 5-6 Research Design Variable & Validity

KYAMBOGO UNIVERSITY
Faculty of Engineering
Department of Civil and Environmental Engineering
TCE 6101- Research Methods
Variables, Validity & Research Design,
Charles Onyutha
conyutha@kyu.ac.ug
Research design
Materials in this slides were all extracted from:
Kothari C.R. (2004). Research

methodology - methods and techniques.
New Age International (P) Ltd., Publishers.
ISBN (13) : 978-81-224-2488-1, 414pp.
2
C. Onyutha (2021). Research Methods: Lecture Slides, Department of Civil & Environmental Engineering, Kyambogo University, Uganda
VARIABLES
3
VARIABLES
What is a variable?
 A characteristic that takes on different values/conditions for
different individuals.
 its anything that varies/takes on more than one form.
Variables should have two distinct characteristics that is:

 should me mutually exclusive and should be exhaustive.
 Variables should have dimensions and dimensions must have
indicators e.g. sex, employment, reward management, marital
status etc.
4
VARIABLES
Types of variable
 Independent or experimental variable
 This is the variable that affects the dependent variables under study and
is included in the research design so that its effects can be determined,
it’s also known as predictor variable in certain types of research.
 They help to explain/account of variation in the dependent variables.
 Dependent or criterion variable

 This is the variable being affected or assumed to be affected by the
independent variable and does not include all the specific information.
 The dependent variable is the variable of interest; it’s the one that
causes the study to be inquisitive enough through the research.
 It’s the variable that is used to measure the effects of the independent
variables, it’s also known as an outcome research in certain types of
research.
5
VARIABLES
Types of variable
 Moderator variable
 This is a variable that may or may not be controlled but has an effect on the
research situation.
 When controlled- we call it control variable (effects are known),
 when uncontrolled- it is the intervening variable (effects are unknown)
 This variable helps to moderate the relationship between the independent and
dependent variables.
 Intervening variable
 A variable whose existence is inferred but which cannot be manipulated or directly
measured; it’s also known as nuisance variables, mediator variable or confounding
variables.
 This variable may interfere with the experiment if not taken into account or
consideration. E.g. attitude of workers can be an intervening variable.
 Intervening variables are hard if not impossible, to observe because they usually have
to do with an individual’s feelings like boredom, stress, fatigue, excitement etc.
6
VARIABLES
Types of variable
 Controlled Variable:
The effectiveness of an experimental variable is examined by
comparing with other variable, known as controlled variable.
 Confounding Variable:
Those aspects of study or sample, that might influence the dependent
variable (outcome measures) ,and whose effect may be confused with
the effects of the independent variable. They are of two types;
Intervening and extraneous variable.
7
VARIABLES
Types of variable
 Extraneous Variable:
 Independent variables that are not related to the purpose of the study, but
may affect the dependent variable are termed as extraneous variables.
 Suppose the researcher wants to test the hypothesis that there is a
relationship between children’s gain in social studies achievement and
their self-concept.
 In the above case, self-concept is independent variable and achievement in
social study is dependent variable. Intelligence may as well affect the social
studies achievement; but since it is not related to the purpose of the study
undertaken by the researcher, it will be termed as extraneous variable.
 Whatever effect is noticed on dependent variable as a result of extraneous
variable(s) is technically described as an ‘experimental error.’
8
VARIABLES
Types of variable
 Extraneous Variable:
 A study must always be so designed that the effect upon the dependent
variable is attributed entirely to the independent variables and not to some
extraneous variable(s).
 When the dependent variable is not free from the influence of extraneous
variable(s), the relationship between the dependent and independent
variable is said to be confounded by an extraneous variable(s).
 Extraneous variable can be controlled by removing the variable causing
distraction.
 Extraneous variable may be eliminated by selecting cases with uniform
characteristics and through randomization.
 In other words, compared with intervening variable, extraneous variable on
the other hand, are more readily observed or measured and thus are more
easily controlled.
9
VARIABLES
Types of variable
 Organismic Variable:
 There are some variables which cannot be manipulated.
 They are accepted by the researcher as they are.
 Examples are levels of intelligence, sex, class levels, and the like.
 The researcher can classify the subjects by sex but he cannot modify to
suit his research condition.
 If a researcher attempts to compare boys & girls on some learning task,
any differences might be attributed to gender but not necessarily so.
 The differences between boys and girls could be due to differences in
intelligence, training, motivation or a myriad of other conditions present in
all human beings and not necessarily to biological differences between sex.
 Those variables which cannot be manipulated and cannot themselves point
out causal relations are called organismic variables.
10
VALIDITY
11
VALIDITY
What is validity?
 Refers to when the study measures what it's supposed to
measure - how realistically or truly the variables have been
operationalized.
 The extent to which a test measures or predicts what it is
supposed to.
 IV actually causes changes in the DV, increases ability to draw
causal conclusions
12
VALIDITY
What is validity?
 Refers to when the study measures what it's supposed to
measure - how realistically or truly the variables have been
operationalized.
 The extent to which a test measures or predicts what it is
supposed to.
Types of validity
1) Internal Validity
2) External Validity
13
VALIDITY
Types of validity
1) Internal Validity
 Refers to whether the results of the study were really
due to the variables the researchers suggest were
tested by their methodology.
 the extent to which an experiment shows that changes
in behavior are a function of the independent variable
and not the result of extraneous variables.
 IV actually causes changes in the DV, increases ability
to draw causal conclusions
14
VALIDITY
Types of validity
2) External Validity
 Refers to whether the results can be generalized if
conducted in different environments or using different
participants.
 Extent to which we can generalize findings to real-
world settings.
 Generalizability, will it generalize to the target
population or other populations?
15
VALIDITY
Types of validity
3) Construct Validity
Construct validity is the extent to which the results
support the theory behind the research
4) Content Validity
Content validity is the notion that a test should sample
the range of the behavior that is represented by
the theoretical concept being measured.
5) Criterion-Related Validity
Criterion validity is the idea that a valid test should
relate closely to other measures of the same theoretical
construct. 16
VALIDITY
Types of validity
6) Transactional Validity
 An interactive process between the researcher, the

researched, and the collected data that is aimed at
achieving a relatively high level of accuracy and consensus
by means of revisiting facts, feeling, experiences, and
values or beliefs collected and interpreted.
 Member Checking- Used to enhance transactional validity.

Reviewing data/observations with the respondent to ensure
the information was collected accurately and to gather
further reactions from the respondent?
17
VALIDITY
Types of validity
7) Transformational Validity
 Evidence of transformational validity develops through the

research process and is not contingent on the particular
techniques used by the researcher.
 The real support for transformational validity of a study

comes from its influence on the actions of others (i.e. the
degree to which the study motivates some sort of action or
change in others).
18
VALIDITY
Types of validity
8) Concurrent Validity: whether results correlate with other valid
research; test for internal validity.
9) Face Validity: is the idea that a test should appear to any person
to be a test of what it is supposed to test.
10) Predictive Validity: whether results predict real life behavior.

.
11) Statistical Validity: refers to the degree to which the observed
results are not due to chance only.
19
VALIDITY
Threats to internal validity
 Unintentional sequence of events (carryover effects,
maturation, history),
 Nonequivalent groups (morality, not a true IV/not

randomized),
 Measurement errors (ceiling/floor effects, low reliability,

instrumentation, regression to the mean/ high or low
scores will score closer to the mean if tested again),
 Ambiguity of cause and effect (non-experimental studies)
20
VALIDITY
Threats to internal validity
 Events outside the lab: Time of day, day of the week, national
tragedies, etc.
 Maturation: People get older and behave differently
 Effects of testing: Learn how to take tests, become less anxious, etc.
 Regression effect: Regression towards the mean, those with extreme

scores tend to drift towards the center
 Selection: those chosen for one condition may be different in an

important way from those who were placed in the other condition
 Mortality: participants dropping out of the study may be different in

some way than those who complete it.
21
VALIDITY
Ways to determine if a test has construct validity
 Manipulation check, which aims to see that a variable is
working in the way you think it is i.e., using standardized tests
for measuring the variable as well as your own
 Use various measurements of the construct to determine if it,

and not another related variable is being altered (i.e., is the rat
hungry or not satiated?)
22
VALIDITY
Threats to external validity
 Other subjects: Findings may not apply to everyone as people
are different in many important respects and some groups are
over studied while others are understudied
 Other times: society changes over time, and thus people's

reactions to certain things will change
 Other settings: Lab work may not transpose to natural settings

and vice versa.
23
VALIDITY
Threats to construct validity
 Loose connection between theory and method: i.e., using poor
measures of theoretical constructs.
 Ambiguous effect of independent variable: participant

expectations can alter their responding, people have a good-
subject tendency, and often have evaluation apprehension.
The general strategy for obtaining construct validity in a piece of

research is to ask whether alternative theoretical explanations of
the data are less plausible than the theory believed to be
supported by the research.
24
Research design
What is a research design?
 A research design is simply the framework or plan for a study
that is used as a guide in collecting and analyzing the data.
 Research design is the plan, structure and strategy of
investigation conceived so as to obtain answers to research
questions and to control variance.
 A good design is often characterized by adjectives like flexible,
appropriate, efficient, economical and so on.
 Generally, the design which minimizes bias and maximizes the
reliability of the data collected and analyzed is a good design.

25
Research design
 A design is used to structure the research, to show how all of the

major parts of the research project –the samples or groups,
measures, treatments or programs and methods of assignments and
how they work together to try to address the central research
question
 The design which gives the smallest experimental error is supposed
to be the best design in many investigations.
 Similarly, a design which yields maximal information and provides
an opportunity for considering many different aspects of a problem
is considered most appropriate and efficient design in respect of
many research problems.
26
Research design
 Thus, the question of good design is related to the purpose or
objective of the research problem and also with the nature of the
problem to be studied.
 A design may be quite suitable in one case, but may be found
wanting in one respect or the other in the context of some other
research problem.
 One single design cannot serve the purpose of all types of research
problems.
27
Research design
Considerations for a research design to be appropriate
 the means of obtaining information;
 the availability and skills of the researcher and his staff, if any;
 the objective of the problem to be studied;
 the nature of the problem to be studied; and
 the availability of time and money for the research work.
28
Research design
Purpose of a Research Design
(i) To minimize the expenditure:

Research design carries an important influence on the reliability of the
results attained .It therefore provides a solid base for the whole research.
This makes the research as effective as possible by providing maximum
information with minimum spending of effort, money and time by
preparing the advance plan of all about the research.
(ii) To facilitate the smooth scaling:

Research design is needed because it facilitates the smooth scaling of the
various research operations, thereby making research as efficient as
possible yielding maximal information with minimal expenditure of effort,
time and money.
29
Research design
(iii) To collect the relevant data and technique:
Research design stands for advance planning of the methods to be
adopted for collecting the relevant data and the techniques to be used in
their analysis, keeping in view the objective of the research and the
availability of staff time and money. Poor preparation of research design
upset the entire project.
(iv) To provide blue print for plans:

Research design is needed due to the fact that it allows for the smooth
working of many research operations. It is like blue print which we need
in advance to plan the methods to be adopted for collecting the relevant
data and techniques to be used in its analysis for preparation of research
project. Just as for better economical and attractive construction of a
house need a blue print and a map of that, similarly we needs a blue
print or a design for the smooth flow of operation of research.
30
Research design
(v) To provide an overview to other experts:
A research design provides an overview of all the research process and
with the help of the design we can take the help and views of experts of
that field .The design helps the investigator to organize his ideas , which
helps to recognize and fix his faults.
(vi) To provide a direction:

A research design provides a proper or particular direction to the other
executives and others who are helping us into the process. The researcher
studies available, literature and learns about new (alternative approaches.
31
Research design
Characteristics of Good Research Design
 Generally a good research design minimizes bias and maximizes the
reliability of the data collected and analyzed.
 The design which gives the smallest experimental error is reported
to be the best design in scientific investigation.
 Similarly, a design which yields maximum information and provides
an opportunity for considering different aspects of a problem is
considered to be the most appropriate and efficient design.
32
Research design
A good research design possesses the following characteristics
(i) Objectivity:
It refers to the findings related to the method of data collection and
scoring of the responses. The research design should permit the
measuring instruments which are fairly objective in which every
observer or judge scoring the performance must precisely give the
same report. In other words, the objectivity of the procedure may
be judged by the degree of agreement between the final scores
assigned to different individuals by more than one independent
observer. This ensures the objectivity of the collected data which
shall be capable of analysis and interpretation.
33
Research design
(ii) Reliability:
It refers to consistency throughout a series of measurements. For
example, if a respondent gives out a response to a particular item,
he is expected to give the same response to that item even if he is
asked repeatedly. If he is changing his response to the same item,
the consistency will be lost. So the researcher should frame the
items in a questionnaire in such a way that it provides consistency
or reliability.
34
Research design
(iii) Validity:
Any measuring device or instrument is said to be valid when it
measures what it is expected to measure. For example, an
intelligence test conducted for measuring the IQ should measure
only the intelligence and nothing else and the questionnaire shall
be framed accordingly.
35
Research design
(iv) Generalizability:
It means how best the data collected from the samples can be
utilized for drawing certain generalizations applicable to a large
group from which sample is drawn. Thus a research design helps
an investigator to generalize his findings provided he has taken due
care in defining the population, selecting the sample, deriving
appropriate statistical analysis etc. while preparing the research
design. Thus a good research design is one which is
methodologically prepared and should ensure that generalization is
possible.
36
Research design
(iv) Generalizability:
For ensuring the generalization we should confirm that our
research problem has the following characteristics;
a) The problem is clearly formulated.
b) The population is clearly defined.
c) Most appropriate techniques of sample selection are used to
form an appropriate sample.
d) Appropriate statistical analysis has been carried out.
e) The findings of the study are capable of generalizations.
37
Research design
(v) Adequate Information:

The most important requirement of good research design is that it
should provide adequate information so that the research problem can
be analyzed on a wide perspective.
An ideal design should take into account important factors like;

(i) Identifying the exact research problem to be studied
(ii) The objective of the research
(iii) The process of obtaining information
(iv) The availability of adequate and skilled manpower and
(v) The availability of adequate financial resources for carrying research.
38
Research design
(vi) Other Features:

Some other important features of a good research design are
 flexibility,
 adaptability,
 efficiency,
 being economic
A good research design should minimize bias and maximize reliability

and generalization.
39
Research design
APPROACHES TO RESEARCH
Research can either be quantitative or qualitative
Quantitative Research
This type of research that describes the phenomena in numbers and measures
instead of using words, quantitative research relies on the principles of
verifiability i.e. Knowledge emerges from what can be proven by direct
observation.
Quantitative approaches are applicable when,

 data is expressed in numbers,
 data is analyzed using statistical methods,
 frequencies are sought to explain meanings of presentations made,
 generalization is needed in order to chive representativeness.
There is need to control as many variables as possible.

40
Research design
Steps in quantitative research process
1) Preparation
2) Selection of the research problem
3) Data collection
4) Data processing
5) Reporting
41
Research design
Qualitative Research
 This the type of research where the phenomena are described in words
instead of numbers, it seeks to describe and analyze behavior from the point
of view of those being studied.
 Its purpose is majorly to promote greater understanding of the phenomena
and the methods of data collections are: Extensive observations, interviews
and focus group discussions.
Application of qualitative research

1) Its used where the subject matter is unfamiliar
2) Dealing with hard to research respondents
3) The researchers seeking for meanings rather than frequencies
4) Where there is need for an in-depth analysis
5) Where flexibility is needed to allow for discovery of unexpected information
42
Research design
TRIANGULATION APPROACH TO RESEARCH
This is another research approach that tries to combine the weakness of

both the qualitative and quantitative approaches.
 This approach attempts to integrate both the qualitative and
quantitative methods
 Inadequacies of each methods are minimized
 more valid and reliable outcomes are obtained
 it conforms that data generated is not due to chance or
circumstances etc.
43
Research design
CATEGORIZATION OF RESEARCH DESIGNS
1) Experimental
2) Quasi Experimental
3) Non Experimental
1) EXPERIMENTAL RESEARCH DESIGN

Experimental designs provides a systematic and logical methods of answering
questions if it’s done under controlled conditions and what will happen and has
the following characteristics:
 It involves carefully manipulating certain variables, treatment or
environmental conditions and observe how conditions or behaviors of
the subject are affected/ changed.
 The design examines the effect of the independent variables on one or more
dependent variables.
 The focus is on the testing of the hypothesis.
 There is total control of extraneous variables.
44
Research design

 The designs enable a researcher to examine issues more critically –to go
beyond description and prediction, beyond the identification of relationships
common in all qualitative studies.
 The groups to be manipulated are equivalent and homogeneous.
 It generates empirical findings that can be generalized with internal
and external validity.
Some of the characteristics of Experimental resign designs

1) Random assignments to the groups
2) Similarity of the subjects.
45
Research design
CATEGORIZATION OF EXPERIMENTAL DESIGNS
 There are various types of experimental designs that vary in complexity and
quality, they are mainly differentiated by the number of groups involved.
 The validity of the study depends on the match between the comparison
groups and the hypothesis, the control groups in the design greatly
strengthen the researcher’s ability to isolate the casual connection between
the independent and the dependent variables.
 The common designs under this designs are:
 The post-test only control group
 Two groups pre-test post-test control group
 Two groups pre-test only control group
 Two groups post –test only control group
 Solomon’s four group designs and factorial designs 46
Research design
2) QUASI EXPERIMENTAL DESIGN
These are designs where a researcher does not randomly assign subjects into the
control and experiment group, quasi experiment studies are sometimes called
correlational designs. In this study the researcher is part of the study
situation, interacts and interviews people, has no capacity to control the
extraneous variables, this study is normally conducted outside the laboratory.
For example you can conduct a field study to establish the effects of particular
training program using an experimental and control group. All field studies or
experiments fall under this category.
Types of quasi experimental designs includes:
1) One group post-pretest designs
2) One group post -test only
3) Two groups static designs

47
Research design
3) NON EXPERIMENTAL RESEACH DESIGN
These describe the attitude, behavior, characteristics of a group or individual,

they are characterized by non-random assignment. Most studies especially
qualitative studies take up this and has no controlled environment.
Types of research design under this are
1) Longitudinal
2) Cross-Sectional Surveys
3) Case study
4) Descriptive
5) Retrospective
6) Prospective
7) Retro-prospective (retrospective-prospective)
48
EXPERIMENTAL DESIGN
What is experimental design?
 It may be defined as the study of the relationships among variables-

those manipulated and those measured.
 It simply enables the researcher to improve the conditions under

which the researcher observes and thus to arrive at a more precise
results.
 It is a scientific method; give more precise, accurate and reliable

results.
 It is just like an observation under controlled conditions. It acts on

the law of single variable and causing factors.
 It studies cause and effect relationship. It is a systematic and logical

method for answering the questions.
49
EXPERIMENTAL DESIGN
Characteristics of experimental design
1) Based on the law of single variable.
2) This method of research is maximum in use in educational / social
researches where the factors can be controlled.
3) Experimental method is a method of testing of clear specific
hypothesis of different intensions.
4) It is a bias free estimation of the true effect.
5) It emphasizes control of conditions and the experimentation of
certain variables in controlled conditions.
6) It sets out more or less of the causal type relationship between the
phenomena.
7) It uses standardized tool for experimentation and makes the
evidences very much objective.
8) The sample is selected with great precaution and very care is taken
to safe guard extraneous factors.
9) This method helps in developing laws, postulate and theories.
10)It allows for precision and definiteness.
50
EXPERIMENTAL DESIGN
Elements of experimental design

1) Manipulation
2) Observation
3) Replication
4) Control
(1) Manipulation
In the experimental method manipulation is done to set the stage for the
occurrence of the factor whose performance is to be studied under
conditions in which all other factors are controlled. Variables which can
be manipulated may be personality characteristics, attitudes, teaching
methods, type of motivation etc.
(2) Observation
Another element of the experimental method is that the effect of the
manipulation of independent variable on the dependent variable is
studied or observed. The technique of observation is applied if
measurement is not possible.
51
EXPERIMENTAL DESIGN

(3) Replication:
In spite of attempts of controlling extraneous variables, some extraneous
variable and some discrepancies remain and influence the results. Thus
replication is a matter of conducting a number of sub-experiments
within the frame work of an overall experiment design.
Features
a) It is based on law of the single variable. It means if one element is
added or subtracted in one of the two situations and other situation
is kept as it is, the change in the two situations is the result of that
single variable.
b) It is applicable when significant factor or conditions can be
controlled.
c) It is not a perfectly precise method in educational research as in
scientific method.
d) Control group and experimental group are never identical as they
should be for an exact experiment.
52
EXPERIMENTAL DESIGN

(4) Control- The extent to which different factors are accounted for
a) Purposes of Control
 Intervening variables are isolated so that they may effect the
dependent variable.
 Not only intervening variables are to be isolated from independent
variables but it is also to be ascertained how much it contributed.
 The magnitude is not only ascertained terms of larger or small but
also in quantitative terms.
b) Degree of Control
It has to be remembered that in social researches high degree of control is
not possible as in laboratory situation.
c) Methods of Control
The researcher directs efforts towards controlling the variables which are
related to the independent variable otherwise it will be difficult to
ascertain which is responsible for effect on dependent variable.
53
EXPERIMENTAL DESIGN
Methods used for controlling
(i) Random assignment of subjects to groups

It means that subjects are assigned in such a way that every member
has an equal opportunity of being chosen.
(ii) Matching subjects with random assignment:

The subjects are assigned to groups to match individual’s subjects on
as many extraneous variable as the researcher can identify. They may
be; (a) Subject to subject matching, (b) matching for mean and standard
deviation,(c)Ranking of subjects on matching variable.
(iii) Random assignment on the basis of homogeneous selection:

It is done to make groups comparable on an extraneous variable so as
to select groups those are as homogenous as possible on the variable.
The variable may be like socio-economic –status and sex etc.
54
EXPERIMENTAL DESIGN
Methods used for controlling
(iv)Techniques of analysis of Co-variance:

It is attempted to control the variation within the groups .A co-variate is
a variable which the researcher has not been able to control.
(v) Methods of using subjects as their own controls:
It is done to assign the same subjects to two experimental treatments

and then to obtain measurements of the subjects under one treatment
and then on the other. But there is practical difficulty in this method in
some cases.
55
EXPERIMENTAL DESIGN
EXPERIMENTAL VALIDITY
Two types
 Internal validity
 External validity
(i) Internal Validity
It is the minimum without which an experiment is useless. The crucial

point is whether the independent variable created the effect on the
dependent variable. It means that some uncontrolled extraneous factors
do not affect the dependent variable.
56
EXPERIMENTAL DESIGN
(i) Internal Validity
The following eight extraneous factors are to be controlled for obtaining

internal validity;
a) Contemporary confounding factor.
b) Maturation of the subject with the passage of time,
c) Experience in pre-test reflecting itself in the post test,
d) Change in the calibration of the measuring instruments,
e) Tendency for extreme scores to regress towards the mean,
f) Difference in the selection of subject in the pre-test and post-test,
g) Differential loss of subjects from the experimental and control groups tend to
bias the results of the experiment,
h) Interaction of selection and maturation and selection and history.
Maturation is about time and the effect of the time on people. even on short
term, like just few hours, one can change his mood from being happy to sad.
This could be due to hunger, boredom, etc.
57
EXPERIMENTAL DESIGN
(i) External Validity

The crucial point is what is the applicability of the findings beyond the limits of
particular experiment. Are the results valid in general for students?
The following factors are to be controlled for obtaining external validity;

a) Selections of sample with some special characteristics make the
application restrictive.
b) Pre-testing may sensitize the subject to the experimental factor.
c) Effects of experimental procedures restrict the generalize ability of the
findings.
d) Effects of a treatment on a subject previously exposed to other
treatments cannot be generalized to other subjects who have not
undergone the same sequence of treatments.
58
EXPERIMENTAL DESIGN
Types of experimental design
It is to the researcher what a blue –print is to an architect. The following

are the factors upon which the selection of designs depends;
(i) What is the nature of the experiment?
(ii) What is the main purpose of the experiment?
(iii) What types of the variables are to be manipulated?
(iv) What is the nature of data?
(v) What is the competence level of the experimenter?
59
EXPERIMENTAL DESIGN
Classifications of experimental design
(a) Pre-experimental designs:
In this type of design there is no control over extraneous or situation

variables.
They are divided into the following categories;
• One group, pre-test-post-test design
• Two group, static design
60
EXPERIMENTAL DESIGN
61
EXPERIMENTAL DESIGN
 Two group, static design, the two groups are assigned but
neither on the basis of randomization nor matching.
 There is no pre-test so the researcher cannot measure the
difference on account of treatment.
 In this design, comparison is made on the basis of post-
test, in experimental group and control group.
62
EXPERIMENTAL DESIGN
(b) True experimental designs:
 In these designs, the researcher attempts to control the

effects of history, maturation, testing, measuring
instruments etc.
63
EXPERIMENTAL DESIGN
64
EXPERIMENTAL DESIGN
65
EXPERIMENTAL DESIGN
66
EXPERIMENTAL DESIGN
67
EXPERIMENTAL DESIGN
68
EXPERIMENTAL DESIGN
(c) Factorial design:

This design enables the experimenter to evaluate or manipulate two or more
experimenter to evaluate or manipulate two or more variables simultaneously
in order to study the effects of number of independent factors singly as well as
the effects due to interactions with one another.
Design No.8: Single Factorial Design

There are two independent variables and each of the independent variables
has two values. The first independent variable which is manipulated and
has two values is called the experimental variable. The second independent
variable which is divided into levels may be called control variable.
In this design the impact of more than one variable can be studied
simultaneously. In this not only the significance of difference of different
levels is studied but interaction effect can also be studied.
However, if too many variables and two many levels are studied, the
experiment and statistical analysis becomes too difficult to manage.
69
EXPERIMENTAL DESIGN
Steps of experimental design

(i) Identifying, defining and delimiting the problem:
The problem that can be verified or refuted by the experimental data

should be selected first. The variables to be investigated are defined in
operational terms.
(ii) Reviewing the literature:
After selecting the problem, the related literature and experiments should
be reviewed and the method of approach or experimental design to be
pursued should be planned or outlined.
(iii) Formulating the hypothesis and deducing their consequences:
The problem has to be converted into a hypothesis that can be verified or

refuted by experimental method.
70
EXPERIMENTAL DESIGN

(iv) Drawing up the experimental design:
This section should place primary emphasis on the question of control,

randomization, replication, place of the experiment, duration of the
experiment, selecting or constructing and validating instruments to
measure the outcomes of the experiment and conducting pilot for trial
run tests to perfect instrument.
(v) Defining the population:
It is necessary to define the population precisely so that there can be no

question about population to which the conclusions are to apply. It may
consist of selecting a sample of subjects to represent a given population
and pairing of subjects to secure homogeneity.
71
EXPERIMENTAL DESIGN

(vi) Administration of test / Carrying out the study:
It is necessary to insist on close adherence to plans and experimental

design. This will involve controlling variable or non-experimental factors,
applying the experimental factors and keeping the careful record of steps in
the procedure.
(vii) Measuring the outcomes / Collection of Data:
Careful consideration must be given to the selection of the criterion on the

basis of which the results are measured.
(viii) Analyzing and interpreting the outcomes;
The result pertaining to the factors under study should be clearly noted.
The analysis of data demands expert use of statistical procedures. Only
then the analysis becomes the basis for valid interpretation.
72
EXPERIMENTAL DESIGN

(ix) Drawing the conclusions:
The conclusions of the study must be restricted to the population

actually covered; care must be taken not to over generalize the results.
The result also pertains to the conditions under which they were derived.
Care must be taken to restrict the conclusions to the conditions actually
present in the experiment.
(x) Reporting the results:
The study must be reported in sufficient details.
73
NON-EXPERIMENTAL DESIGN
Longitudinal studies
 This is a study that involves a series of sample measurements taken
over a period of time, the study focuses on a specific respondents or
sample over a specific period of time.
 The study examines the behavior of a respondent/groups noting the

relevant changes for example it could be a case study of human
development through childhood over a specific period of time. Or
monitoring the effects of ARV on the beneficiaries or how a certain
particular programme impacts on the lives of people over a given
period of time.
 Longitudinal studies gives a deeper insight into cases or issues at

hand, it may however give a negative impact or setback if the
respondent or participant dies or transfers before the end of the study.
74
Longitudinal studies
 Longitudinal study is further sub divided into Panel and trend studies.
 Panel Study: This occurs when the same group of people or

individual is used in the study, one sample or cohort is studied
at different point in time during the study period.
 Trend study or time series: occurs when different homogeneous

groups or samples are involved in phases over the study period,
and later on the results got from the study compared and
analyzed accordingly. Example of this include classes that
change every academic year, change in level, distribution,
relationship and degree.
75
Cross-sectional studies
 A type of survey design that involve a study of sample (s) at the same
point in time.
 The fundamental difference between cross sectional and longitudinal

studies is that cross section studies take place at a single point in time
and that a longitudinal study involve the series of measurements taken
over a longer period.
 Cross sectional studies are studies are used in most branches of

science, in the social sciences and other fields as well. Cross sectional
research takes a slice of its target group and bases its overall findings
on the views or behaviors of this targeted assuming them to be a typical
of a whole group.
 It may employ a single survey or multi surveys. A cross sectional of

respondents is involved in this study and the respondents are scattered
over a wider geographical area i.e. varying in age groups, tribe,
nationality, religion etc. It’s conducted over a specified and limited time
period.
76
Descriptive studies
This is concerned with the describing the characteristics of an
event, community or a region providing data about the population
or region studied, the description shows what, how, when, where,
of a situation at a time.
Case studies
A case study survey studies a social phenomenon through
analysis of individual cases, it could be a person, group, process,
community or any other aspect of social life. All relevant data to
the case are collected and organized in relation to the study.
77
Case studies
 The main advantage of this is that it provides an opportunity
for the intensive analysis of many specific details often
disregarded by other methods.
 The choice of this design is based on the assumption that the

problem under study is typical of cases of a certain so that after
intensive analysis, its generalization will be applicable to other
similar cases.
78
DESIGN BASED ON REFERENCE PERIODS
1) Retrospective
This studies and investigates a phenomenon, situation, problem or
issues that occurred in the past, it can be done either based on the
data from the period of time or consulting those who have got
knowledge on the same subject matter.
2) Prospective research designs

This focuses into the future events of the subject matter, the likely
prevalence of a phenomenon, situation, problem, attitude or outcome
in the future.
3) Retrospective-prospective research designs

This is one that focuses on the past trend of events of a phenomenon
and studies into the future. There is a linkage between the past and
the future to draw conclusions for example development plans, we
need an analysis of the past to make a better projection.
79
Ethnographical studies
 Ethnographical is the in-depth study of naturally occurring behavior
within a culture or social group, it seeks to understand the relationship
between culture and behavior; with culture referring to the beliefs,
values, and attitudes of specific group of people.
 The ethnographic research method was developed by anthropologists

as a way of studying and describing human cultures.
Some of the characteristics of a good ethnography as being
1) Extended participant observation
2) Long time at site
3) Collection of large volume of materials such as notes, audios, video

tapes which means having no specific hypotheses or even highly
specific categories of observations at the start of the study.
80
Phenomenological studies
 Phenomenological literally means the study of phenomena.
 It’s a way of describing something that exists as part of the world in

which we live.
 Phenomena may be events, situations, experiences or concepts.
 This branch studies things as they are perceived not as they are, it
begins with acknowledgement that there is gap in our understanding
and that clarification will be of benefit.
Grounded theory
Based on facts, reasons, reliable evidence which is in touch with reality
and at times it’s based on personal feelings.
81
RESEARCH DESIGN SUMMARY
QUANTITATIVE RESEARCHES
82
QUALITATIVE RESEARCHES
83
QUALITATIVE RESEARCHES
84

Lecture 5-6 Research Design Variable &amp; Validity

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Lecture 5-6 Research Design Variable &amp; Validity

Uploaded by

Copyright:

Available Formats

KYAMBOGO UNIVERSITY

TCE 6101- Research Methods

Variables, Validity & Research Design,

Kothari C.R. (2004). Research

Variables should have two distinct characteristics that is:

 Dependent or criterion variable

 Refers to when the study measures what it's supposed to

measure - how realistically or truly the variables have been

 The extent to which a test measures or predicts what it is

 IV actually causes changes in the DV, increases ability to draw

 Refers to when the study measures what it's supposed to

measure - how realistically or truly the variables have been

 The extent to which a test measures or predicts what it is

 An interactive process between the researcher, the

 Member Checking- Used to enhance transactional validity.

 Evidence of transformational validity develops through the

 The real support for transformational validity of a study

10) Predictive Validity: whether results predict real life behavior.

 Nonequivalent groups (morality, not a true IV/not

 Measurement errors (ceiling/floor effects, low reliability,

 Ambiguity of cause and effect (non-experimental studies)

 Maturation: People get older and behave differently

 Regression effect: Regression towards the mean, those with extreme

 Selection: those chosen for one condition may be different in an

 Mortality: participants dropping out of the study may be different in

 Use various measurements of the construct to determine if it,

 Other times: society changes over time, and thus people's

 Other settings: Lab work may not transpose to natural settings

 Ambiguous effect of independent variable: participant

The general strategy for obtaining construct validity in a piece of

What is a research design?

 A research design is simply the framework or plan for a study

that is used as a guide in collecting and analyzing the data.

 Research design is the plan, structure and strategy of

investigation conceived so as to obtain answers to research

questions and to control variance.

 A good design is often characterized by adjectives like flexible,

appropriate, efficient, economical and so on.

 Generally, the design which minimizes bias and maximizes the

reliability of the data collected and analyzed is a good design.

What is a research design?

 A design is used to structure the research, to show how all of the

What is a research design?

 Thus, the question of good design is related to the purpose or

 A design may be quite suitable in one case, but may be found

wanting in one respect or the other in the context of some other

Considerations for a research design to be appropriate

 the means of obtaining information;

 the objective of the problem to be studied;

 the nature of the problem to be studied; and

 the availability of time and money for the research work.

(i) To minimize the expenditure:

(ii) To facilitate the smooth scaling:

(iv) To provide blue print for plans:

(vi) To provide a direction:

 Generally a good research design minimizes bias and maximizes the

reliability of the data collected and analyzed.

 The design which gives the smallest experimental error is reported

to be the best design in scientific investigation.

 Similarly, a design which yields maximum information and provides

an opportunity for considering different aspects of a problem is

considered to be the most appropriate and efficient design.

(v) Adequate Information:

An ideal design should take into account important factors like;

Lecture 5-6 Research Design Variable & Validity

Lecture 5-6 Research Design Variable & Validity