RESEARCH METHODS

Dr. Manaswini Dash

 RESEARCH
 Research - search for knowledge using methods of scientific inquiry.
 Three important functions of scientific research:
Assessment, Prediction, and Control
 Scientific inquiry differs from common sense in:
 use of conceptual schemes
 empirical verification
 notion of control
 systematic investigation
 objective investigation
 explanations of observed phenomena.
 Scientific research is systematic, controlled, empirical, and critical
investigation of hypothetical propositions about the presumed
relations among natural phenomena
Sources of Knowledge

• EXPERIENCE (Subjective, cannot access all forms of

knowledge)
• AUTHORITY (can be wrong, authorities differ)

• DEDUCTION (must begin with universal knowledge,

not sufficient for discovering new truth)
• INDUCTION (imperfect because of incomplete
observation)
• SCIENTIFIC APPROACH (mixture of induction and
deduction, testing of hypotheses)
Facts about Research?
– Research aims to seek answers to questions.
– Research is based on empirical or observable evidence. It
rejects anything that is subjective, and only accepts
evidences that can be objectively observed.
– Research is systematic, objective and logical. A researcher
uses systematic methods to frame designs, collect and
analyze data, and reach conclusions. He eliminates any
form of bias to attain objectivity, and logically examines all
his procedures.
– Research findings are replicable. The researcher himself or
another researcher can verify the findings.
– Research involves the use of reliable and valid instruments
for data collection, and accurate analysis of the collected
information.
 Purposes of Research
• FUNDAMENTAL / BASIC RESEARCH (knowledge for the
sake of knowledge, universalistic temper, high
generalizability, testing theory)
• APPLIED RESEARCH (conducted in field of common
practice, Application of research-based knowledge,
mission-oriented)
• EVALUATION RESEARCH (determines the effectiveness
of a given practice, on a particular practice at a given
site, limited generalizability)
• ACTION RESEARCH (Oriented to problem solution or
management, no universal validity, more of a social
welfare activity, may not be termed a s research)
Qualities of a Good Research
• Good research is systematic.
• Good research is empirical.
• Good research is verifiable.
• Good research is logical.
• Good research develops theories and principles.
• Good research is replicable.
 Typical Stages in Research
• Identify and Define the Problem
• Formulate Hypotheses
• Select a Research Design / Strategy
(i) Select the sample
(ii) Develop tools of measurement
(iii) Collect data
(iv) Decide techniques for data analysis and
interpretation
• Tabulate and Analyze Data
• Interpret the Results and Draw Conclusions
• Prepare a Research Report
 Step I. Identify and Define the Problem:
• Select a problem that interests you
• Select a problem that can be investigated within the time you
have at your disposal.
• Define the problem in operational terms i.e., in a way that can be
investigated. The problem is usually stated in in the form of
relationship among two or variables.
• The variables must be clearly defined in measurable terms. Avoid
any philosophical descriptions.
• You should have some knowledge of the prior work on the
problem. Prior work related to the research problem is referred to
as review of literature.
• You should review the existing literature before deciding on a
problem. The review would help you define the variables,
conceptualize your problem neatly and accurately, and relate your
findings to the existing literature on the topic.
How to Find a Research Problem?

• Personal Observation and Experience

(Good for beginning Researchers. Would attract interest and be
Practical)
• Deductions form Theory
(Deduce tentative hypothesis and test empirically)
• Review of Literature
(Helps putting research in perspective; General references, Secondary
sources, Primary Sources)
• Conventions and Professional Meetings
(Sharing of ideas and familiarity with a new vision)
• Personal Communication
(Feedback from knowledgeable peers)
• Non-Professional Sources
(Movements outside the profession lead to new research)
Checklist for Research Problems
• Is the problem researchable? (Can be investigated empirically and
objectively.)
• Is it significant? (Makes a significant contribution to your chosen field of
knowledge.)
• Is it clear and unambiguous? (Key terms, variables, and purpose of the
investigation clearly described)
• Is it feasible? (Must have the time, energy, financial resources, and all
the resource materials to carry out research on the problem.)
• Is it stated in the form of a research question? (Stated in the form of a
research question.)
• Is it appropriate to my interest? (Sincerely and passionately interested
to carry out research. No research is possible without interest and
curiosity.)
• Is it adequate? (Should not be too broad or too narrow.)
• Is it possible to review the literature on this problem? (Prior research
work and documents are necessary for formulating the problem and
writing the report.)
• Is it ethical? (Should not invite unethical practices such as deceiving or
punishing the subjects, or creating harmful consequences.)
 Step II. Formulate Hypothesis
• Hypothesis is a scientific guess about the relationship between variables. It is
like a suggested answer to the problem. It is stated in the form of a declarative
sentence. Hypothesis will provide direction for your research. Use the
following checklist to examine the adequacy of your hypothesis.
• Is it a statement regarding relationship between two or more variables?
(Hypothesis is a tentative statement showing a relationship between variables
under study.)
• Is it directly related to the research question? (Hypothesis should emanate
from the research question.)
• Is the hypothesis empirically testable? (Hypothesis should be stated in such
a way that the researcher can collect observable evidence to say that it is
either probably false or probably true.)
• Is it stated in the form of a declarative sentence? (Hypothesis should make a
declaration about the relationship between the variables.)
• Is it derived from known facts and theories? (Hypothesis should be deduced
from evidences present in the existing literature. It is not a personal opinion.)
• Is it reasonable and expressed in simplest possible words? (Hypothesis
should be free of technical jargons and should be stated in way that a
common educated person can comprehend it.)
 Step III. Select a Research Strategy / Design
Research design is the plan, structure, and strategy of investigation
so conceived as to obtain answers to research questions. It is the
blueprint of research.
• Plan – Overall scheme of the program of research. Includes an
outline of what the researcher will do from writing hypotheses to
the final analysis of data.
• Structure – More specific. The framework or scheme of the
operation of variables. We build structural schemes for
accomplishing operational research purposes.
• Strategy – More specific than plan. Methods used to collect and
analyze data. Specifies the sources and types of information
required to answer the research question.
It is a sequence of steps taken ahead of time to ensure that the
relevant data would be collected in a way that permits objective
analysis of hypothesis formulated with respect to the research
 Types of Research
• EXPERIMENTAL RESEARCH
Manipulates the IVs in a controlled setting to determine the effect on the DVs; Focuses on
Internal and External Validity; Most restrictive and artificial.
• HISTORICAL RESEARCH
Collects, interprets and synthesizes events of the past to reach conclusions and
generalizations; Depends on primary and secondary sources; Anticipates future; Biases in
interpretation and conclusions.
• CAUSAL-COMPARATIVE /EX POST-FACTO RESEARCH
The IVs cannot be directly manipulated; Observes an exciting phenomenon and searches
back through the data to discover possible cause-and-effect relationships; Main weakness
is lack of control; Hence explanations are confounded by several causes.
• DESCRIPTIVE RESEARCH
Describes what is; Aims to generalize; Sub-categories are: (i) Surveys, (ii) Correlational
studies, (iii) Developmental studies, (iv) Documentary analysis, and (v) Follow-up studies.
• QUALITATIVE RESEARCH
Tries to explain the process by observing persons and events in their natural settings;
Logical analysis of the phenomenon; Hypothesis develops as the study progresses;
 Reliability
• Reliability of a test refers to the consistency of scores
obtained by an individual on different occasions or
with different sets of equivalent items (Anastasi,
1957).
 Methods of Estimating Reliability
• External Consistency Procedures: (Compare findings from two
independent process of data collection)
1. Test-retest reliability
2. Parallel Forms Reliability
• Internal Consistency Procedures: (Items measuring same
phenomena should produce similar result)
1. Split-Half Reliability: Longer tests have higher reliability. Effect
of shortness is corrected by spearman-brown’s formula
2. Kuder-Rechardson Estimate of Reliability: K-R formula 20:
Applicable to where each test item is scored 0 or 1.
3. Cronbach’s Alpha (Extends the application of K-R formula to
items that are not scored 0 or 1. It is the mean of all possible
split-half coefficients.
 Validity
Validity refers to the degree to which a test measures what it
claims to measure.
Types of Validity:
1. Content Validity : The test should adequately sample
range of behaviour that is represented by the theoretical
concept being measured.
2. Criterion-related Validity
i.) Concurrent validity
ii) Predictive validity
3. Construct Validity
i) Convergent validity
ii) Discriminant validity
4. Face Validity
Validity of Research Designs
• Internal Validity (Did the experimental treatment produce a difference?)
– History – impact of extraneous events
– Maturation – changes within participants over time
– Effect of Pre-testing – effect of prior measurement on later
measurement
– Instrumentation decay – changes in measuring instrument, human
raters /interviewers
– Statistical Regression – groups selected on thee basis of extreme pre-
test scores
– Differential Selection – resulting from differential selection of subjects
– Experimental Mortality – differential loss of subjects from experimental
group
– Selection-maturation Interaction – interaction of selection with history
and maturation
Validity of Research Designs (Contd…)
• External Validity (Are the findings generalizable?)
– Selection-treatment interaction – characteristics of
subjects interacts with treatment conditions
– Reactive Effects of Pre-testing – Pre-testing may increase
or decrease subjects’ sensitivity to experimental
treatment
– Reactive Effect of Treatment – presence of observers may
limit generalizability
– Multiple-treatment interference – effects due to multiple
treatments