Research Design: Definition

 A research design is a detailed blueprint or framework for conducting a

research project.
 It is the overall plan that specifies the methods and procedures for collecting
and analyzing the needed information.
 Its primary purpose is to ensure that the evidence obtained enables you to
effectively address the research problem.
 It acts as a guide, keeping the researcher on track and ensuring the study is
structured and systematic.
 A good design minimizes the time, cost, and effort required while maximizing
the reliability of the data.
 It defines the type of data required, the sources of data, and the data
collection methods.
 It outlines the strategies for measuring variables and for analyzing the
collected data.
 The design is decided after the research problem is formulated but before
actual data collection begins.
 The choice of design depends on the research objectives (e.g., to explore,
describe, or establish causality).
 It is the master plan that links all parts of the research study together logically.
2. Types of Research Designs
1. Exploratory Research Design:
o Used when the problem is not well understood or defined.
o Goal is to gain insights, ideas, and familiarity with the problem.
o Flexible and unstructured, often qualitative in nature.
o Common methods: Literature review, interviews, focus groups, case
studies.
2. Descriptive Research Design:
o Used to describe characteristics of a population or phenomenon.
o Answers the questions: Who, What, When, Where, and How?
o Requires a clear understanding of the problem beforehand.
o Pre-planned and structured, often using surveys and observations.
o Example: A study profiling the users of a new smartphone.
3. Causal Research Design (Experimental):
 o Used to determine cause-and-effect relationships between variables.
o Involves manipulation of one variable (independent) to see its effect on
another (dependent).
o Conducted under controlled conditions to minimize the effect of
extraneous variables.
o Answers the question: Why?
o Example: Testing if a new training program (cause) improves employee
productivity (effect).
3. Exploratory vs. Causal Research Design
Feature Exploratory Design Causal Design
Objective To gain insights and To determine cause-and-effect
understanding. relationships.
Problem Unclear and loosely Clearly defined and structured.
Definition defined.
Structure Flexible and unstructured. Rigid and highly structured.
Sample Size Small, non-representative. Large, representative.
Data Analysis Qualitative, subjective. Quantitative, statistical.
Outcome Hypotheses for future Conclusive answers.
research.
Methods Used Expert surveys, pilot Experiments, controlled testing.
studies.
Cost Relatively low. Relatively high.
Time Shorter duration. Longer duration.
Basis for Provides initial Provides a basis for final
Decisions understanding. decisions.

4. Descriptive & Experimental Research Design

 Descriptive Research Design:
o Purpose: To describe the current state of affairs.
o Approach: Observes and records without interfering.
o Variables: Measures variables as they exist; no manipulation.
o Time Frame: Can be cross-sectional (one point in time) or longitudinal
(over a period).
 o Example: A market survey to determine the market share of different
brands of soap.
 Experimental Research Design:
o Purpose: To establish a causal relationship.
o Approach: Actively manipulates one or more independent variables.
o Variables: Controls for extraneous variables that could influence the
result.
o Key Components: Treatment (manipulation), control group, and
random assignment.
o Example: Testing two different advertisements (Treatment) to see
which one leads to higher recall (Effect).
5. Types of Experimental Design
1. Pre-Experimental Designs:
o Simple, weak designs with little control over extraneous variables.
o One-Shot Case Study: A single group is exposed to a treatment and
then observed. (X O)
o One-Group Pretest-Posttest: A single group is measured before and
after a treatment. (O₁ X O₂)
2. True Experimental Designs:
o Strong designs with maximum control, using random assignment.
o Pretest-Posttest Control Group: Two randomly assigned groups (one
control, one experimental). Both are pre-tested, the experimental group
gets the treatment, then both are post-tested. (R O₁ X O₂; R O₃ O₄)
o Posttest-Only Control Group: Two randomly assigned groups. The
experimental group gets the treatment, then both are post-tested. No
pre-test is used. (R X O₁; R O₂)
3. Quasi-Experimental Designs:
o Used when random assignment is not possible.
o Lack full control but are better than pre-experimental designs.
o Example: Using existing groups (e.g., two different company
branches) for testing.
6. Validity of Findings: Internal & External Validity
 Validity refers to whether the research actually measures what it intended to
measure and whether the conclusions are correct.
  Internal Validity:
o The extent to which we can be confident that the independent
variable caused the change in the dependent variable.
o High internal validity means the study is well-controlled and
alternative explanations for the result can be ruled out.
o Threats: History, maturation of subjects, testing effect, instrument
decay, selection bias.
 External Validity:
o The extent to which the results of a study can be generalized to other
situations, people, settings, and times.
o High external validity means the findings are applicable to the real
world.
o Threats: Interaction of selection and treatment, reactive effects of
testing, artificial setting.
 Trade-off: Often, there is a trade-off between internal and external validity.
Tightly controlled lab experiments have high internal validity but may lack real-
world applicability (low external validity). Field studies have higher external
validity but lower control (lower internal validity).
7. Variables in Research
 Variable: A measurable characteristic that can take on different values.
 Independent Variable (IV): The variable that is manipulated or changed by
the researcher to observe its effect. (The presumed cause).
 Dependent Variable (DV): The variable that is measured; it is expected to
change as a result of manipulating the independent variable. (The
presumed effect).
 Extraneous Variable (EV): External factors that are not the focus of the study
but could affect the results (the DV). They can confound the results if not
controlled.
 Control Variable: An extraneous variable that is held constant or managed to
prevent it from influencing the outcome.
 Intervening/Mediating Variable: A theoretical variable that explains the
relationship between the IV and DV. (e.g., IV: Training -
> Intervening: Learning -> DV: Performance).
8. Measurement and Scaling
 Measurement: The process of assigning numbers or labels to objects,
people, or events according to specific rules. It is about what to measure.
  Scaling: The process of placing the measured objects on a continuum (e.g., a
scale from 1 to 5). It is about how to measure the intensity of the
characteristic.
 The goal is to create a map between the empirical world (the object) and the
formal world (numbers/symbols) for analysis.
9. Different Scales (Primary Scales of Measurement)
1. Nominal Scale:
o Numbers are used only as labels or for identification (e.g., 1=Male,
2=Female).
o No quantitative meaning. Basic classification only.
2. Ordinal Scale:
o Numbers indicate the relative position or rank order of objects, but not
the magnitude of difference (e.g., 1st, 2nd, 3rd place; customer
satisfaction rankings).
3. Interval Scale:
o Numbers indicate both order and the exact magnitude of difference
between them.
o There is no true zero point. (e.g., Temperature in °C or °F, IQ scores).
o We can add and subtract, but cannot multiply or divide (e.g., 20°C is
not twice as hot as 10°C).
4. Ratio Scale:
o Possesses all the properties of an interval scale with a true absolute
zero point.
o Allows for all arithmetic operations (e.g., Weight, Height, Age, Income,
Sales in units).
o Statements like "10 kg is twice as heavy as 5 kg" are meaningful.
10. Construction of an Instrument & its Validity/Reliability
 Instrument Construction (e.g., Questionnaire):
1. Define Objectives: Clearly state what information is needed.
2. Item Generation: Write clear, simple, and unbiased questions.
3. Determine Scale: Choose the appropriate measurement scale (e.g.,
Likert Scale, Semantic Differential).
4. Layout & Flow: Organize questions logically from general to specific.
 5. Pilot Testing: Test the questionnaire on a small group to identify
problems.
6. Finalize: Revise and finalize the instrument based on feedback.
 Validity of Instrument: Does it measure what it's supposed to measure?
o Content Validity: Ensures the instrument adequately covers all
relevant parts of the topic.
o Criterion Validity: How well the measurement predicts or correlates
with a current or future outcome.
o Construct Validity: How well the test measures the theoretical
construct it was designed to measure.
 Reliability of Instrument: Does it provide consistent results?
o Test-Retest Reliability: Administering the same test to the same
respondents at two different times.
o Internal Consistency Reliability: Assessing if all questions measuring
the same construct yield similar results (e.g., using Cronbach's Alpha).
o Equivalent Forms Reliability: Using two different but equivalent
versions of the instrument on the same people.
 A good instrument must be both Valid and Reliable. A reliable instrument
is not necessarily valid (it can consistently measure the wrong thing), but a
valid instrument is always reliable.