Research Methods in Rural

Development UNIT 2 DESCRIPTIVE AND

EXPERIMENTAL RESEARCH
Structure
2.0 Introduction
2.1 Objectives
2.2 Descriptive Research
2.2.1 Descriptive Research: Main Steps
2.2.2 Types of Descriptive Research
2.3 Experimental Research
2.3.1 Three Characteristics of Experimental Research
2.3.2 Steps Involved in Experimental Research
2.3.3 Designs of Experimental Study
2.3.3.1 Pre-experimental Design
2.3.3.2 True Experimental Design
2.3.3.3 Quasi Experimental Design
2.3.3.4 Factorial Design
2.3.3.5 Time Series Design
2.4 Let Us Sum Up
2.5 Check Your Progress: The Key

2.0 INTRODUCTION
As you studied in the previous Units, a ‘research method’ is a particular way of
studying a problem. The features of the research problem and also the field of
inquiry determine the ‘method’. In this Unit we shall discuss Descriptive Method
and Experimental Method of research in detail.

2.1 OBJECTIVES
After the completion of this Unit you should be able to:

describe the steps involved in descriptive research;

explain the characteristics and features of types of descriptive research such
as survey research, documentary analysis, correlational studies and causal-
comparative studies;

describe the steps involved in experimental research; and

identify and explain a few designs for experimental studies.

2.2 DESCRIPTIVE RESEARCH

Descriptive research studies are designed to obtain information concerning the
current status of a given phenomenon. They are concerned with the existing
conditions or relationships, prevailing practices, current beliefs, points of view or
26 attitudes, processes that are going on and their effects and the developing trends.
In short, it determines the nature of a situation as it exists at the time of study. The Descriptive and Experimental
aim of descriptive research is to describe “what exists” with respect to variables or Research
conditions in a situation.
The descriptive research method is appropriate in behavioural sciences. Many types
of behaviour that interest the researcher cannot be arranged in a realistic setting.
For example, it would be unthinkable to prescribe cigarette smoking for the purpose
of studying its possible relationship to throat or lung cancer, or deliberately arrange
accidents, in order to evaluate the effectiveness of seat-belts or helmets in preventing
serious injuries.
Although some experimental studies of human behaviour can be appropriately
carried out both, in laboratory and in the field, the prevailing method used in social
sciences in descriptive. Under the conditions that naturally occur at home, inside
the classroom, on the playground or within the community, human behaviour can
be systematically examined and analysed. This analysis may lead to the modification
of factors or influences that determine the nature of human interaction. It is through
this modification of factors that social institutions may become more effective
influences in promoting human welfare.

2.2.1 Descriptive Research: Main Steps

In descriptive studies, we do not present private convictions and data based on
casual or cursory observations. In descriptive study, we:
i) examine the problematic situation,
ii) define our problem and state our hypothesis,
iii) list the assumption upon which our hypothesis and procedures are based,
iv) select appropriate subjects and source material,
v) select or construct technique for collecting data,
vi) validate the data gathering techniques,
vii) make objective and discriminating observations,
viii) describe, analyse and interpret our data in clear, precise terms.
We, as researchers, collect evidence on the basis of some hypotheses, tabulate and
summarize the data carefully, and then analyse the results thoroughly in an endeavour
to draw meaningful generalization that will advance knowledge.
Collection of data
When presenting a descriptive research report, one must identify not only the kind
of data obtained but also the exact nature of its population. The units that constitute
a population may be people, items, events or objects. After identifying the
population, one must decide whether to collect data from (a) the total population
or (b) a representative sample of the population.
a) Total population: Obtaining information from every unit of a small population
is not difficult in most instances, but the findings are not applicable to any
population other than the group studied. After collecting information from
every student in one particular study center, you may, draw generalizations
about the average age or the kind of profession of students in the center; but,
you cannot claim that these generalizations will hold true for students in any
other study center. Similarly, after studying the attitude of teachers of the
School of Continuing Education in IGNOU towards the Distance Education
27
Research Methods in Rural Programme, a researcher cannot claim that the findings will hold true for
Development teachers in other schools of IGNOU.
b) Sample: Obtaining information from a large population, such as all teachers
in the state, is often impractical, impossible or exorbitantly costly. Contacting,
observing, measuring or interviewing every unit in the group may need so
much time that the data becomes obsolete before the study is complete. To
overcome these difficulties, investigators often collect information from a few
carefully selected units drawn from a population. While carrying out the
research on Study Habits of undergraduate students of IGNOU, the researcher
cannot collect data from all the 50,000 undergraduate learners. He/She will
have to select a representative sample of the population. However, the sample
will have to be selected from different categories of the population, viz., male/
female, employed/unemployed etc. If these sample units represent accurately
the characteristics of the population, generalisations based on the data obtained
from them may be applied to the entire group. But selecting a representative
sample is a difficult task. How to select a proper sample is the topic of
discussion in Unit 1, Block 3.
Descriptive data may be expressed qualitatively (in verbal symbols) as well as
quantitatively in mathematical symbols. A study may consist almost exclusively
of one form or may contain both forms. Qualitative data may predominate in
studies that examine the general nature of a phenomena. Qualitative studies
give social scientists useful information, but verbal symbols lack precision as
words do not hold the same meaning for all people, at all times, and in all
contexts. However, qualitative studies need not be looked down upon; for,
they help workers identify the significant factors to measure. Until these general
explorations are made, measurement cannot be utilised fruitfully.

2.2.2 Types of Descriptive Research

In this sub-section, we are going to discuss various types of descriptive research.
There is nothing sacrosanct about such a categorisation, but it helps us understand
the phenomenon (research) more clearly.
i) Survey studies: Quite often, descriptive research itself is termed survey
research, but it is better to think of it as a category under descriptive research.
‘Survey’ is probably the most widely used method for obtaining descriptive
and evaluative information in social research. When trying to solve problems,
non-governmental, governmental, industrial and political, organisations often
conduct surveys. Detailed descriptions of existing phenomena are collected
with the aim of employing the data to justify current conditions or practices or
to make more intelligent plans for improving them. The objective may not
only be to ascertain status, but also to determine the adequacy of status by
comparing it with selected or established standards. Researchers who wish to
improve existing status of health of children may survey how others have solved
similar problems. At times, we need to collect all three types of information:

Data concerning existing status, e.g., to find out the study habits of
postgraduate students of IGNOU,

Comparison of status and standards, e.g., to compare the achievement level
of distance education students and conventional education students, and

Means of improving status e.g., to find out measures to improve student
support services in distance education programmes.
On the other hand, we can also limit our studies to one or two of these types.
28
 Surveys may be broad or narrow in scope. They may be confined to a small Descriptive and Experimental
geographical area or whole state or even the country. The scope and the depth Research
of the survey depend primarily on the nature of the problem. The steps involved
in the case of surveys are presented as follows:
a) Selection of topic involves the determination of what topic is to be
investigated and what population is to be studied. At this stage, one also
decides the methods and procedures that will be used to collect the data.
b) Sampling involves decision making about which people from the population
are to be included in the survey. If one is to generalise from the sample to
the population, it is essential that the sample selected be representative.
c) Construction of a tool for data collection involves writing the questions
and planning the format of the instrument to be used. Among the data
collection techniques used in surveys are personal interviews, questionnaires,
rating scales, etc.
d) Collection of data includes pre-testing the instrument to determine
whether it will obtain the desired data, interviewing subjects or distributing
questionnaires to them, and verifying the accuracy of the data gathered.
e) Processing and analysis of the data includes tabulating the data, computer
processing and analysis of the data, interpreting the results, and reporting
the findings.
ii) Analysis of documents: From documents and records, we can unearth
pertinent data. Documentary analysis, which is sometimes referred to as
‘content’, ‘activity’or ‘informational’ analysis is very much like historical
research. Both methods of investigation require that researchers examine
existing records: historical research is primarily concerned with the more distant
past, but descriptive research is concerned chiefly with the present.
A wide variety of documentary surveys are made. Some scholars analyse
judicial decisions, state laws and rulings. Researchers may also collect data
describing existing NGO’s practices, processes and conditions from
administrative records, committee reports and minutes of meetings, budget and
financial records, etc. Documentary research produces valuable information,
but the method has certain limitations, and investigators may draw faulty
conclusions from the data.
Some documentary researches are of little value because the investigators fail
to analyse the representative sample of source materials. Many studies do not
provide information concerning the adequacy of the sample size or the
conformity of the universe of the sample to the universe. If an analysis is
made of newspaper editorials concerning the problems of agricultural labour,
for example, the researcher must judge for himself/herself whether the
newspapers selected represent the opinion of the different social, religious,
economic or political groups from all the parts of the country. Very often, the
write-ups are not representative in nature.
Documentary research in not without other pitfalls as well. Data do not become
true reflections of reality through the magic of publication. In using
documentary sources, one must bear in mind the fact that data appearing in
print are not necessarily trustworthy. One must subject each documentary source
material to the same rigorous external and internal criticism that a historian
does. Moreover, the categories used in the available statistical material do not
always coincide precisely with the variables which the researcher wants to
investigate. Sometimes the definitions of categories are ambiguous, and they 29
Research Methods in Rural may change from year to year. The boundaries of some units of analysis, e.g.,
Development
villages, districts, age cohort, etc. can also change; different agencies collecting
similar data do not always use exactly the same classificatory system. The
data collected always reflect the orientation, concerns, self-interests, and the
levels of accuracy preferred by the producers of the records, which may not
be an accurate reflection of reality or behaviour.

Check Your Progress 1

List the steps involved in survey studies. Describe briefly in about 30 words the
objectives of documentary analysis.
Note: a) Space is given below for your answer.
b) Compare your answer with the one given at the end of this unit.
................................................................................................................................
................................................................................................................................
................................................................................................................................
................................................................................................................................
................................................................................................................................
................................................................................................................................
................................................................................................................................
................................................................................................................................

iii) Correlational studies: Human behaviour at both, individual and the social
levels is characterised by great complexity. However, given the present state
of social research, we understand too little of this complexity. One approach to
a fuller understanding of human behaviour is to begin by testing out simple
relationships between those factors and elements which are supposed to have
some bearing on the phenomenon in question. The value of correlational research
is that it is able to achieve this end. We know that one of the primary purposes
of science, as conceived traditionally, is to discover relationships among
phenomena with a view ultimately to predicting and, in some situations,
controlling their occurrence.
Much of social sciences research is concerned at our present stage of
development with the first step in this sequence, i.e., establishing
interrelationships among variables. Correlational studies are concerned with
determining the extent of relationship existing between variables. They enable
us to measure the extent to which variations in one variable are associated
with variations in another. We may wish to know, for example, how delinquency
is related to social and class background, or whether a relationship exist between
the number of years spent in full-time education and subsequent annual income,
or whether there is a link between personality and achievement.
Correlational studies are generally intended to answer three questions. They
are:
a) Is there a relationship between two variables (or two sets of data)? If the
answer to this question is ‘yes’, then other questions follow:

30 b) What is the direction of the relationship? and

 c) What is the magnitude of the relationship? The magnitude of the relationship Descriptive and Experimental
is determined by the coefficient of correlation. Research

For instance, on the basis of his/her experience, a researcher may hypothesize that
there is a relationship between performance in an intelligence test and a test of
achievement in arithmetic. The correlational technique will help him test his/her
hypothesis about the relationship. Pearson’s product moment, one of the best known
measures of association, is a statistical value of the coefficient of correlation ranging
from –1.0 to +1.0, through zero and expresses relationship in quantitative form.
Where the two variables fluctuate in the same direction, i.e., as one increases so
does the other, a positive relationship is said to exist. A negative correlation or
relationship, on the other hand, is to be found when an increase in one variable is
accompanied by a decrease in the other variable. The values near zero indicate a
weak relationship between the variables, whereas values closer to either +1.0 or –
1.0 indicate a stronger relationship in either of directions. Thus, the coefficient of
correlation, tells us something about the relationship between two variables. However,
other measures exist which allow us to specify relationship when more than two
variables are involved. These are known as measures of multiple correlation and
partial correlation. (We will not go into details about these measures over here.)
One danger in interpreting correlations is to assume that because two variables are
related in a predictable fashion to one another with a high degree of probability,
they are also in a causal relationship. This is not necessarily the case. For one
thing there is never more than a probable relationship between variables in any
case . For another, it is quite possible for two variables to be related to one another
with a high degree of probability but with a third variable accounting for the nature
of relationship. Correlation must not be interpreted to mean that one variable is
causing the scores in other variable to be what they are. For example, it may be
found that there is a negative correlation between measures of anxiety and measures
of intelligence. It should not be interpreted that there is a causative relationship
between anxiety and intelligence, that is, that pupils are anxious because they are
unintelligent or that pupils appear unintelligent because they are anxious. It might
be that there are other underlying characteristics of individuals that tend to make
some appear unintelligent and anxious, and others, intelligent and not anxious.
Interpretation of such a correlation is difficult without experimental confirmation,
For example, the relationship between anxiety measures and intelligence measures
could be investigated experimentally by deliberately inducting anxiety in a testing
situation and determining the effect on intelligence test scores.
Characteristics of correlational studies
Correlational studies can be broadly classified either as relational studies or as
prediction studies. As a method, the former is particularly useful in exploratory
studies in fields where little or no previous research has been undertaken. It is
often a shot in the dark aimed at verifying hunches which a researcher has about a
presumed relationship between some characteristics or variables. Take a complex
notion like teacher effectiveness for example. This is dependent on a number of
complex factors operating singly or in combination. Factors such as intelligence,
motivation, person, perception, verbal skills, etc., come to mind as possibly having
an effect on teacher outcomes. A review of the literature of research will confirm
or reject these possibilities. Once an appropriate number of such possibilities has
been identified in this way, suitable measures may then be chosen or developed to
assess them. They are then given to a representative sample, and the scores obtained
are then correlated with the complex factor that is being investigated, namely,
teacher effectiveness. As it is an exploratory undertaking, the analysis will consist
of correlation coefficients only. The investigation and its outcomes may then be
used as a basis for further research or as sources of additional hypotheses.
31
Research Methods in Rural In contrast to exploratory research studies, prediction studies are usually
Development undertaken in an area having a firmer and securer knowledge base. Prediction
through the use of correlational technique is based on the assumption that at
least some of the factors that will lead to the behaviour to be predicted are
present and measurable at the time the prediction is made. For example, since
we know that IQ and General Achievement (GA) are positively correlated, we
can predict with some degree of accuracy that an individual with a high IQ
will probably have a high GA. To be valuable for prediction, the extent of
correlation between two variables must be substantial and, of course, the higher
the correlation, the more accurate the prediction.
iv) Causal-comparative studies: There is, at times the need to discover how
and why a particular phenomenon occurs, and not confine our investigation to
what a phenomenon is like. In this instance, the investigator tries to compare
the similarities and differences among phenomena to find out what factors or
circumstances seem to accompany or contribute to the occurrence of certain
events, conditions or practices.
Unlike a scientist working in a laboratory, a social researcher cannot always
select, control and manipulate factors that are necessary to study cause-effect
relations. An investigator cannot, for example, manipulate domestic background,
social class, intelligence, etc. in situations that do not allow researchers
manipulate the independent variable and establish the controls that are required
in “true experiments”, they may conduct a causal-comparative study.
In a causal-comparative investigation, a researcher studies a real life situation
in which subjects have experienced what he/she want to investigate. For
example, if an investigator wants to study emotional instability, he/she does
not place children in a situation where all factors are kept constant except one
variable which is manipulated to determine what causes a particular type of
emotional disturbance. Rather, he/she chooses children who according to a
selected criterion, are ‘disturbed’ and compares them with emotionally stable
children. After searching for factors or conditions which seem to be associated
with one group and not the other, he/she may present a possible explanation of
the underlying causes of the emotional problem.
Limitations of causal-comparative method
i) Lack of control is serious limitation and weakness of this method of research.
ii) It is usually difficult to identify the relevant factors causing a particular condition
or phenomenon. For instance, students’ liking for a teacher depends on a number
of factors and a researcher may not be able to identify all the factors. He/She
may only be able to identify good teaching and mastery of subject matter as
some of the factors effecting students’ liking for a teacher.
iii) The joint method of agreement and disagreement requires that a single factor
must be the cause for the occurrence or non-occurrence of the phenomenon.
But in the case of social phenomena, with which a researcher is usually
concerned, this condition does not come invariably. In fact, in these situations/
events usually have multiple rather than single causes. Furthermore, a
phenomenon may result not only from multiple causes but also from one cause
in one instance and from another cause in another instance.
iv) When a relationship between variables is established, it is difficult to distinguish
between the cause and the effect.
v) The classification of subjects into dichotomous groups for the purpose of
32 comparison also presents problems.
vi) In comparative studies of natural situations, the researcher does not have the Descriptive and Experimental
same control over the selection of subjects as he/she has in experimental studies. Research
It is difficult to identify existing groups of subjects who are alike in all respects
except for their exposure to one variable.
Though causal-comparative studies have many limitations, and they do not often
produce precise and reliable knowledge that can be gained through rigorous
experimental studies, they provide the means of tackling problems that cannot be
probed in laboratory situations. Furthermore, they yield valuable information and
clues concerning the nature of phenomena and are well suited to many types of
field studies seeking to establish causal relationships.

Check Your Progress 2

Explain briefly the purpose of correlational studies. List the weaknesses of causal-
comparative studies.
Note: a) Space is given below for your answer.
b) Compare your answer with the one given at the end of this Unit.
................................................................................................................................
................................................................................................................................
................................................................................................................................
................................................................................................................................
................................................................................................................................
................................................................................................................................
................................................................................................................................
................................................................................................................................

2.3 EXPERIMENTAL RESEARCH

Experimental research studies are designed for establishing causal relationships.
This method begins with a question concerning the relationship between two or
more variables. At the same time, the researcher advances one or more hypotheses
stating the nature of the expected relationship. The experiment is the event planned
and carried out by the researcher to gather evidence relevant to the hypotheses.
In its simplest form an experiment has three characteristics:
i) an independent variable is manipulated,
ii) all other variables except the independent variable are held constant, and
iii) the effect of the manipulation of the independent variable on the dependent
variable is observed.
The independent variable and the dependent variable(s) are important in an
experiment. The independent variable is manipulated or changed by the
experimenter. The variable upon which the effects of changes are observed is called
the dependent variable, which is observed but not manipulated by the experimenter.
The dependent variable is so named because its value is hypothesised to depend 33
Research Methods in Rural upon, and vary with, the value of the independant variable. For example, to examine
Development the effect of training for empowerment upon decision making, an investigator would
manipulate training, the independent variable, by using different training methods in
order to ascertain their effect upon decision making, the dependent variable.

2.3.1 Three Characteristics of Experimental Research

There are three essential ingredients in the conduct of an experiment: control,
manipulation and observation. We shall discuss each of them as follows:
i) Control: Control is the first essential ingredient of experimental method.
Without control, it is impossible to evaluate unambiguously the effects of an
independent variable. Basically, the experimental method rests upon two
assumptions regarding variables. These are:
a) If two situations are equal in every respect except for a variable that is
added to or deleted from one of the situations, any difference appearing
between the two situations can be attributed to that variable. This statement
is called the law of the single variable.
b) If two situations are not equal, and it can be demonstrated that none of
the variables is significant in producing the phenomenon under
investigation, or if significant variables are made equal, then any difference
occurring between the two situation after the introduction of a new variable
to one of them can be attributed to the new variable. This statement is
called the law of the only significant variable.
The main purpose of ‘control’ in an experiment is to arrange a situation in
which the effect of variables can be measured. The conditions to be fulfilled
under the first law can be obtained more easily in physical sciences. A high
degree of control is much easier to achieve in a laboratory setting than in
situation outside the laboratory. In the laboratory, there is only limited number
of variables which can be manipulated easily. However, as social research is
concerned with human beings, there are always many variables present in
situation. To attempt to reduce social problems to the operation of a single
variable, is not only unrealistic but perhaps impossible as well. Fortunately, we
do not require such rigorous control to be introduced in social settings, for
many factors involved in such a setting may be quite insignificant and irrelevant
for our study. To this extent, in social research, the law of the single significant
variable is more appropriate. For example, if we were to study the effect of
two methods of teaching alphabets to two groups of adult learners, we are
likely to select the two groups which are identical in every respect except in
the way they are taught alphabets. But it is impossible to have two groups that
are identical in every respect. So, the endeavour of the researcher should be
towards obtaining two groups that are as similar as possible, at least in those
factors that are thought to have an effect on learning alphabets. These could
be, general intelligence, motivation, reading ability, etc. Other variables that are
not likely to affect achievement in learning alphabets can be ignored. Thus, in
experimental studies in social research we need procedures that permit us to
use to compare groups in the basis of significant variables. ‘Control’ is to used
to indicate an experimenter’s ‘procedures’ for eliminating the differential effects
of all variables extraneous to the purpose of the study. (An extraneous variable
is variable that is not related to the purpose of the study but may affect the
dependent variable). The experimenter exercises controls, for instances, when
the groups are made comparable on extraneous variables that are related to
the dependent variable. If a variable is known to be unrelated to the dependent
variable, it cannot influence the dependent variable and we do not need to
34 control it for its effects.
ii) Manipulation: Manipulation of a variable is another distinguishing characteristic Descriptive and Experimental
of experimental research. It refers to a deliberate operation performed by the Research
researcher. In contrast to the descriptive research in which the researcher
simply observes conditions as they occur naturally, the researcher in the
experimental research actually sets the stage for the occurrence of the factors
whose performance is to be studied under conditions where all other factors
are controlled or eliminated. In social research and other behavioural sciences,
the manipulation of a variable takes a characteristic form in which the
experimenter imposes a predetermined set of varied conditions on the subjects.
This set of varied conditions is referred to as the independent variable; the
experimental variable, or the treatment variable. Then, different conditions are
designed to represent two or more values of the independent variable; these
may be differences in degree or differences in kind. That is, the independent
variable may assure two or more values and the difference in the values may
be of quantitative or qualitative nature. Methods of teaching, attitudes, socio-
economic status, personality characteristics, types of motivation, etc. are some
common examples of the independent variable in social research. For example,
if the researcher compares two methods of teaching, then method of teaching
is the independent variable and can be manipulated by the teacher. We may
manipulate a single variable or a number of variables simultaneously.
iii) Observation: In experimentation, we are interested in the effect of the
manipulation of the independent variable on a dependent variable. Observations
are made with respect to some characteristics of the behaviour of the subjects
employed in the research. These observations which are quantitative in nature,
may constitute the dependent variable. This needs some explanation.
The dependent variable in social research is often change of some type, such
as attitude towards learning. We are often interested in explaining or predicting
attitude. Since attitude cannot be measured directly, we can only estimate it
through measures like scores in a scale. Therefore, strictly speaking, the
dependent variable is scores or observations rather than change in attitude.

Check Your Progress 3

Explain briefly the significance of control, manipulation, and observation in an
experimental study.
Note: a) Space is given below for your answer.
b) Compare your answer with the one given at the end of this Unit.
................................................................................................................................
................................................................................................................................
................................................................................................................................
................................................................................................................................
................................................................................................................................
................................................................................................................................
................................................................................................................................
................................................................................................................................
................................................................................................................................
35
Research Methods in Rural 2.3.2 Steps Involved in Experimental Research
Development
A number of steps are involved in experimental research. Here, we shall talk about
four steps to reach the stage of the ‘actual experiment’. Brief explanations are
needed for steps 3 and 4 only. The steps are :
i) Surveying the literature related to the problem,
ii) Identifying and defining the problem,
iii) Formulating a hypotheses is an important step in experimental research. They
suggest that an antecedent condition or phenomenon (independent variable) is
related to the occurrence of another condition, phenomenon, event, or effect
(dependent variable). To test a hypothesis, the experimenter attempts to control
all the conditions except the independent variable which he/she manipulates.
Then he/she observes the effect on the dependent variable presumably because
of the exposure to the independent variable.
iv) Constructing an experimental plan is the next step in experimental research.
This refers to the conceptual framework within which the experiment is
conducted. This would involve:

Selecting a research design,

Selecting a sample of subjects to represent a given population, assign
subjects to groups, and assign experimental treatments to the groups.
(Subject implies the respondent or living organism that is studied),

Selecting or constructing and validating instruments to measure the
outcomes of the experiment,

Stating the procedures for collecting the data and possibly conduct a
pilot or “trial run” test to perfect the instruments or design, and

Stating the statistical or null hypothesis.
The above steps bring the researcher to the stage when he/she actually conducts
the experiment, applies statistical measures to the data obtained, and then test the
significance of the results.
In the next sub-section, we shall take up the various designs involved in experimental
method.

2.3.3 Designs of Experimental Study

A research design is very important for the researcher. A well developed design
provides the structure and strategy to control the investigation and extract dependable
answers to the questions raised by the problem or hypothesis. It is the nature of the
problem that determines the appropriateness of the design.
Before we discuss the experimental designs, it will be relevant to look into the
terms and symbols which we shall make use of.
i) X represents the independent variable, which is manipulated by the researcher;
it is also referred to as the experimental variable or the treatment variable.
ii) Y represents the measure of the dependent variable. Y1 represents the dependent
variable before the manipulation of the independent variable X; it is usually a
pre-test of some type administered before the experimental treatment. Y2
represents the dependent variable after the manipulation of the independent
variable X; it is usually a post-test administered to subjects after the
36 experimental treatment.
iii) S represents the subject or respondent used in the experiment. Descriptive and Experimental
Research
iv) E group refers to the experimental group – t he group that is given the
independent variable treatment.
v) C group refers to the control group – the group that does not receive the
experimental treatment.
vi) R indicates random assignment of subjects to the experimental groups and
the random assignment of treatments to the groups.
There is a large number of experimental designs. Various authors have classified
experimental design into certain categories. Most common categorization comprises:

Pre-experimental Design

True Experimental Design

Quasi Experimental Design
Some authors like Donald Ary and others (1985) have added more categories namely

Factorial Design

Time Series
Various designs under the above mentioned categories are given in the table below:

Pre- True Quasi Factorial Time

Experimental Experimental Experimental Design Series

One Group
Randomized
Non-randomized
Simple
One

Pre-test ‘Subjects’ Post-test Control Group, Factorial Group
Post-test only Control Group Pre-test Post Design Time
Design Design test Design Series
Design

Static Group
Randomized Matched
Counter
Comparison Subjects, Post-test Balanced Design
Control
only Control Group Group
Design Time
Series

Randomized subjects Design
Pre-test Post-test
control Group Design

Solomon Four Group

Design

However, in this section, we will bring before you only a few most frequently used
designs, from each of the five categories.

2.3.3.1 Pre-experimental Design

The two designs classified as pre-experimental designs offer minimal control of
extraneous variables. Still they are used quite often in social research. These designs
help to illustrate the advantages of more rigorously controlled designs that are
presented later.

Design 1: One Group Pre-test Post-test Design

When this design is employed, the dependent variable is measured before the
independent variable or treatment is applied or withdrawn, and then measured yet
again. The one group design usually involves three steps: 37
Research Methods in Rural
a) administering a pre-test measuring the dependent variable,
Development
b) applying the experimental treatment X to the subjects, and
c) administering a post-test again measuring the dependent variable.
Differences attributed to application of experimental treatment are then determined
by comparing the pre-test and post-test scores.

Pre-test Independent variable Post-test

Y1 X Y2

Design 1: One Group Pre-test Post-test Design

To illustrate the use of this design, let us assume that we want to evaluate the
effectiveness of a particular self-instructional material in Rural Development for
post graduate students. How may we go about this task?

At the beginning of the academic year, the students are given a standardized test
that measures the objectives of the course quite satisfactorily, following which the
distance teacher then introduces the self-instructional material. At the end of the
year, the students are administered the standardized test a second time. Comparing
the scores of the two tests would reveal what difference the exposure to the SIM
has made.

However, using only one group, as in Design 1, gives us superficial control. The
major limitation of the one-group design is that, since no control group is used, the
experimenter cannot assume that the change between the pre-test and the post-
test scores is brought about by the experimental treatment alone. It is quite possible
that some extraneous variables account for all or part of the change. For example,
students experience changes with the passage of time; they grow mentally as well
as physically, or they may acquire additional learning experiences that would affect
the dependent variable. This extraneous variable can be thought of as maturation
i.e., with the passage of time students get maturity and this in turn may effect
achievement level. Another type of extraneous variable that can operate between
the pre-test and the post-test scores and which cannot be controlled is history.
History as a source of extraneous variances refers to the specific events that can
occur between the pre-test and post-test other than the experimental treatment. In
the example cited above, not receiving material regularly or illness just before the
test, could decrease achievement scores. Similarly, a crucial research finding in
history could increase widespread interest and hence affect the test scores. In fact,
history and maturation become increasingly influential sources of extraneous variance
when the time interval between Y1 and Y2 is long.

Another short coming of Design 1 is that it offers no way of assessing the effect of
the pre-test Y1 itself. We know that “practice effect” exists when subjects take a
test a second time or take an alternate form of the test. In other words, subjects do
better the second time even without any instruction or relevant discussion during
the interval. This is true not only for achievement and intelligence tests but also for
personality tests. In the case of personality tests, a tendency towards better adjustment
is generally observed.

To sum up, Design 1 has little to recommend it; without a control group to make a
comparison possible, the results obtained in a one group design are basically
uninterpretable. The results of the experiment would have been dependable if there
38 could be a comparable group i.e. control group to which SIM had not been given.
Design 2: Static Group Comparison Descriptive and Experimental
Research
Design 2 utilizes two or more groups, only one of which is exposed to experimental
treatment. The groups are assumed to be equivalent in all relevant aspects, they
differ only in their exposure to X.
This design is often used in social research, For example, achievement of adult
learners taught by a new method is compared with that of similar class taught by
a traditional method.
Design 2 has a control group or groups, which permit (s) the comparison that is
required for scientific respectability. If the experimental group is superior on the Y2
measure, the researcher then has more confidence in his/her conclusion that the
difference is due to experimental treatment.
However, there is a basic flaw in this design. Since neither randomization nor
even matching is used to assign subjects to the experimental and control groups,
we cannot be sure that the groups are equivalent prior to the experimental and
control groups, we cannot be sure that the groups are equivalent prior to the
experimental treatment. They may differ on certain relevant variables, and it may
be these differences rather than X that are responsible for the observed change.
Because we cannot be sure that the groups are equal with regard to all the factors
that may influence the dependent variable, this design is considered to be lacking in
the necessary control and must be classified as pre-experimental.

Group Independent Variable Post-test

E X Y2

C — Y2

Design 2: Static Group Comparison

2.3.3.2 True Experimental Designs

The following two designs, belong to the ‘true experimental’ design, because of the
control that they provide. i.e.
i) Random assignment of subjects to the groups.
ii) Random assignment of treatment to the groups.
iii) Post-testing all the groups.
Design 3: Randomized Subjects, Post-test only Control Group Design
This particular design requires two groups to which subjects are randomly assigned
and each group is assigned to a different condition. No pre-test is used; randomization
controls all the possible extraneous variables. This does not mean that randomization
procedures (like drawing names out of a hat, or flipping a coin) remove the
extraneous variables, such as the IQ or age, which may affect the dependent variable,
or control their presence. These extraneous variables still affect the inquiry; but,
now, it is the laws of chance rather than the personal feature of E that operate. In
fact, the larger the number of subjects used, the more equivalent or similar the
groups will tend to be. Suppose a researcher wants to study the effect of instructional
material on achievement in a course during a contact programme. He/she may
randomly assign the students to the groups and provide treatment to one of the
groups. The assigning of the treatment will be random. At the end of the contact
programme he/she may test both the groups. 39
Research Methods in Rural After the subjects are assigned to the groups, only the experimental group is exposed
Development to the experimental treatment. Otherwise, in all other respects, the two groups
remain similar. Members of both groups are then measured on dependent variable
Y2. Scores are then compared to determine the effect of X.

Group Independent Variable Post-test

(R) E X Y2
(R) C — Y2
Design 3: Randomized Subjects, Post-test only Control Group Design

The main advantage of Design 3 is randomization, which assures statistical

equivalence of the groups prior to the introduction of independent variable. Design
3 provides controls for the main effects of history, maturation and pre-testing;
because no pre-test is used, there can be no interaction effect of pre-test and X
(treatment).
Design 4: Randomized Matched Subjects, Post-test only Control Group
Design

This design is similar to Design 3 except that it uses a matching technique, rather
than random assignment, to obtain equivalent groups. Subjects are matched on one
or more variables that can be measured conveniently, such as IQ or reading scores.
The matching variables used are generally those that have a significant correlation
with the dependent variable. On the basis of these variables subjects are paired so
that opposite member’s/scores’ are as close as possible; and then, one member of
each pair is randomly assigned to one treatment and the other to the second
treatment.

Group Independent Variable Post-test

(Mr) E X Y2
(Mr) C — Y2

Design 4: Randomized Matched Subjects, Post-test only Control Group Design

Matching is most useful in studies where small samples are to be used and where
Design 3 is not appropriate. Also, the matched subjects’ design serves to reduce
the extent to which experimental differences can be accounted for by initial
differences between groups. However, for matching to really become a means of
control, the matching of all the potential subjects must be complete, and assignment
of the members of each pair to the groups must be determined randomly. If one or
more subjects should be excluded because an appropriate match could not be found,
this would bias the sample. When using Design 4, it is essential to match every
subject, even if only approximately, before random assignment is effected.
2.3.3.3 Quasi Experimental Design
One of the Quasi Experimental Designs is Non-randomized Control Group, Pre-
test Post-test Design. You would notice that randomized control group pre-test post-
test design is a true experimental design which we have presented before. The
only difference on the quasi experimental design is that the groups are not
randomized. Hence they are unlikely to be comparable. In fact, it is on this ground
that the design becomes quasi experimental and not true experimental. Since the
40
rest of the design related characteristics remain common with the randomized control Descriptive and Experimental
group pre-test post-test design of the true experimental design category, we do not Research
need to provide any further details on this design.

Group Pretest Independent Variable Post-test

E Y1 X Y2
C Y1 _ Y2

2.3.3.4 Factorial Designs

A factorial design is one where two or more variables are manipulated simultaneously
in order to study the independent effect of each variable on the dependent variable
as well as the effects due to interaction among the several variables. Factorial
designs are of two types. In the first type, one of the independent variables may be
experimentally manipulated. The researcher is primarily interested in the effect of
a single independent variable but he/she must take other variables into consideration
which may influence the dependent variables. In the second type of design, all the
independent variables may be experimentally manipulated. Factorial designs have
been developed at varying levels of complexity, the simplest factorial design is the
2 by 2 (2 × 2) Design. The two independent variables have two values.

Level 1 subjects receive Treatment A and others Treatment B. Some level 2 subjects
receive Treatment A and others Treatment B.

Attribute Experimental Variable X1 Variable X2

Variable X2

Treatment A Treatment B

Level 1 Cell 1 Cell 3

Level 2 Cell 2 Cell 4

The strength of the factorial design is that it can achieve in one experiment what
might otherwise require two or more separate studies.
2.3.3.5 Time Series Design
We have already discussed pre-test post-test designs. They generate one time data
on the dependent variable before and after the experimental treatment. There are
instances where it becomes necessary to compare changes in the trend of a particular
phenomenon or process or product. For example, let us assume that learners
behaviour to attitudes, achievements etc. changes over a period of time. If a specific
treatment is introduced in an institution to study the change in attitude or achievement
it is useful to study the trend through measurement at certain intervals before the
introduction of the treatment. Instead of one time pre-test, the test is repeated
three or four times before the treatment is administered. This generates data on the
trend of behaviour. Similarly after the treatment is administered instead of one time
post-test, the post-test is administered several times at intervals. This provides data
to derive the trend in the change in behaviour. Since both, pre-tests and post-tests
are used over a time series design the effect of the treatment on the dependent
variable is tested by comparing the trends. This can be represented in the following
form:

Y1 Y2 Y3 Y4 Y Y5 Y6 Y7 Y8
41
Research Methods in Rural What we have described above is one group time series design. If you add a
Development control group and repeat the same time series measurement without the treatment
of the control groups it becomes control group time series design. Similarly control
group time series design is represented as:

Group

E Y1 Y2 Y3 Y4 X Y5 Y6 Y7 Y8
C Y1 Y2 Y3 Y4 X Y5 Y6 Y7 Y8

Check Your Progress 4

Draw and compare the figures representing pre-test post-test experimental design
and one group time series design.
Note: a) Space is given below for your answer.
b) Compare your answer with the one given at the end of this unit.
................................................................................................................................
................................................................................................................................
................................................................................................................................
................................................................................................................................
................................................................................................................................
................................................................................................................................
................................................................................................................................
................................................................................................................................

2.4 LET US SUM UP

In this Unit, we studied two important research methods, viz., Descriptive Method
and Experimental Method. Descriptive research describes what is the condition
and involves the description, recording, analysis and interpretation of conditions that
exist. We also studied various types of descriptive research, like survey, documentary
analysis, correlational and causal comparative studies. Experimental research
describes what will be when certain variables are carefully controlled or manipulated.

2.5 CHECK YOUR PROGRESS: THE KEY

1) The steps involved in survey studies are: selecting the topic, sampling, selection/
construction of tools of data collection, collection of data, processing of the
data and analysis and interpretation of data.
Documentary analysis can show us the existing conditions and practices of an
institution and the importance of problems in an organisation.
42
2) Correlational studies are useful to: Descriptive and Experimental
Research

determine the relationship between variables and

measure the extent to which variations in one variable are associated with
the variations in another variable. The weaknesses of causal comparative
studies are: lack of control, difficulty in identifying the relevant causal
factors, determining their number in given phenomenon, classifying subjects
into dichotomous groups for the purpose of comparison, lack of control
over the selection of subjects.
3) Control is crucial to (i) evaluate unambiguously the effects of an independent
variable and (ii) arrange a situation in which the effect of variables can be
measured.
Manipulation controls or eliminates the irrelevant factors and arranges a
situation in which only relevant factors can be studied.
Observations are made to study specific characteristics in the behaviour of the
subjects employed in experimental research.
4) Pre-test Post-test Experimental Design

Group Pre -test Treatment Post-test

E Y1 X Y2

One Group Time Series Design

Y1 Y2 Y3 Y4 Y Y5 Y6 Y7 Y8

Compared to one test each before and after the treatment in pre-test post-test
experimental group design, tests are repeated at specified intervals in one group
time series design. Whereas time series designs compares the trends of change in
the dependent variable, the pre-test post-test experimental design tests one time
gain or change in the dependent variable.

References and Further Readings

Ary, D., Lucy, C. Jacobs and Razavich, A. (1972), Introduction to Research in
Education, 3rd Edition, Rinehart and Wdition Inc., Holt, New York.
Ackoff, R.L. (1953), The Design of Social Research,University of Chicago Press,
Chicago.
Bailey, Kenneth D. (1978), Methods of Social Research, The Free Press, London.
Baker, L. Therese. (1988), Doing Social Research, McGraw Hill, New York.
Black, James A. and Champion, Dean J. (1976), Methods and Issues in Social
Research, John Wiley, New York.
Cohen, L. and Manion, L. (1989), Research Methods in Education (Third Edition),
Routledge, London.
Campbell, D.T. and Stanley. J.C. (1963), Experimental and Quasi Experimental
Designs for Research,Houghton Mifflin, Boston.
Chapin, F.S. (1974), Experimental Design in Sociological Research, Harper,
New York.
43
Research Methods in Rural Epstein, I & Tripodi, T. (1974), Research Techniques for Program Planning,
Development Monitoring and Evaluation, Columbia University Program, New York.
Festinger, L. and Katz. D. (eds.) (1953), Research Methods in the Behavioral
Sciences,The Dryden Press, New York.
Goode, W.J. and Hat. P.K. (1952), Methods in Social Research, McGraw Hill,
New York.
Kerlinger, Fred R. (1964), Foundations of Behavioral Research,Surjeet
Publications, Delhi.
Kidder, Louise H. ( 1981), Research Methods in Social Relations,Holt, New
York.
Kothari, L.R. (1985), Research Methodology, Vishwa Prakashan, New Delhi.
Lal Das, D.K. (2000), Practice of Social Research: A Social Work Perspective,
Rawat Publications, Jaipur.
Mill, J.S. (1930), A System of Logic, Longmans 8th ed., New York.
Miller, D.C. (1964), Handbook of Research Design and Social Measurement,
David Mckey Co., New York.
Monette, Duane R. (et. al.) (1986), Applied Social Research: Tool For the
Human Services, Holt, Chicago.
Moser, C.A. and Kalton, (1975), Survey Methods in Social Investigation,
Heinemann Educational Books, London.
Nachmias D. and Nachmias C. (1981), Research Methods in the Social Sciences,
St. Martins Press, New York.
Rubin, Allen & Babbie E. (1989), Research Methodology for Social Work,
Belmohf, California, Wadsworth.
Sellitz, G. (et. al.) (1973), Research Methods in Social Relations: Holt, Rinehart
and Winston (3rd edition), New York.
Shah. V.F. (1977), Research Design, Rachna Prakashan, Ahmedabad.
Stouffer, S.A. (1962), Social Research to Test Ideas, Free Press of Glencoe,
New York.
Suchman, E.A. (1953), The Principles of Research Design. In John T Doby et.
al., An Introduction to Social Research, The Stackpole, New York.
Wilkinson, T.S. and Bhandarkar, P.L. (1977), Methodology and Techniques of
Social Research, Himalayan, Bombay.
Yin. Robert K. (1982), Case Study Research Design and Methods, Applied Social
Research Methods Series Vol. 5 Sage, New Delhi.
Young, P.V. (1953), Scientific Social Surveys and Research, (4 th edition),
Englewood Cliff, Prentice Hall, N.J.

44