Assertion (A): Random assignment strengthens the internal validity of an experiment. Reason (R): Random assignment makes groups equivalent, on average, on all extraneous variables before treatment, so later differences can be attributed to the independent variable. Choose the correct option.

Both A and R are true, and R is the correct explanation of A

Both A and R are true, but R is not the correct explanation of A

Which of the following are breaches of ethics in REPORTING research? Select all that apply.

Presenting a hypothesis formed after seeing the results as if it had been predicted

Reporting all the measures that were collected

Presenting a hypothesis formed after seeing the results as if it had been predicted

Debriefing participants after a deception study

Research Methodology I: Meaning, Purpose and Dimensions of Research; Problems, Variables, Operational Definitions, Hypotheses and Sampling; Ethics; Quantitative, Qualitative and Mixed Paradigms; and the Methods of Psychological Research

Unit 2 begins with the logic of research before any statistics: "Research: Meaning, Purpose, and Dimensions. Research problems, Variables and Operational Definitions, Hypothesis, Sampling. Ethics in conducting and reporting research. Paradigms of research: Quantitative, Qualitative, Mixed methods approach. Methods of research: Observation, Survey [Interview, Questionnaires], Experimental, Quasi-experimental, Field studies, Cross-Cultural Studies, Phenomenology, Grounded theory, Focus groups, Narratives, Case studies, Ethnography." This chapter treats these as a psychologist must — with the kinds of variables that confound psychological studies, the sampling designs that decide whom a finding applies to, the ethical codes and famous violations behind them, and the twelve methods side by side.

1. Meaning, purpose and dimensions of research

Research is systematic, controlled, empirical and critical investigation of propositions about presumed relations among phenomena (Kerlinger's definition). Its purposes are conventionally three — exploratory (to become familiar with a new area and generate questions), descriptive (to portray accurately the characteristics of a group or situation) and explanatory (to test why something happens, usually through causal hypotheses) — with evaluation research (did a programme work?) often added. W. L. Neuman organises research along four dimensions:

Four dimensions of research
Dimension	Types
Use	Basic (pure) research that advances knowledge, versus applied research that solves a practical problem (action research, evaluation, needs assessment)
Purpose	Exploratory, descriptive, explanatory
Time	Cross-sectional (one point in time) versus longitudinal — time series, panel (the same people repeatedly) and cohort (people sharing an experience, such as a birth year, though not necessarily the same individuals); plus case studies
Data collection	Quantitative (experiments, surveys, content analysis, existing statistics) versus qualitative (field research, historical-comparative)

2. Problems, variables, operational definitions and hypotheses

A good research problem is a question about the relationship between two or more variables, stated clearly and in a form that implies empirical testing (Kerlinger). A variable is a property that takes different values; a constant does not. An operational definition specifies the operations by which a variable is measured or manipulated — "anxiety" becomes "score on the State–Trait Anxiety Inventory", "frustration" becomes "being interrupted just before completing a puzzle" — and is what makes a study replicable.

Types of variable
Variable	Role
Independent (IV)	Manipulated or selected by the researcher; the presumed cause. An organismic or subject variable (age, sex, personality) is a selected, not manipulated, IV.
Dependent (DV)	Measured; the presumed effect.
Extraneous	Any other variable that could affect the DV; controlled by elimination, holding constant, matching, counterbalancing or randomisation.
Confounding	An extraneous variable that varies systematically with the IV, so its effect cannot be separated from the IV's.
Moderator	Changes the strength or direction of the IV–DV relation (when or for whom) — for example, stress predicts illness more strongly under low social support.
Mediator	The mechanism through which the IV affects the DV (how or why) — for example, exercise lowers depression through improved sleep (Baron and Kenny, 1986).

A hypothesis is a tentative, testable statement of the expected relation between variables. The null hypothesis (H₀) states no difference or no relation; the alternative (H₁) states that there is one, either directional (one-tailed: "girls score higher") or non-directional (two-tailed: "the groups differ"). A good hypothesis is testable, falsifiable, consistent with existing knowledge, stated in operational terms and parsimonious. S. S. Stevens's four levels of measurement decide which statistics the data will allow: nominal (categories), ordinal (ranks), interval (equal units, arbitrary zero — IQ, temperature in °C) and ratio (true zero — reaction time).

⚠️ Moderator or mediator?

Ask "when or for whom?" for a moderator and "how or why?" for a mediator. A moderator is usually a pre-existing characteristic that interacts with the IV; a mediator is caused by the IV and in turn causes the DV.

3. Sampling

The first four are probability designs; the last five are non-probability designs
Design	How units are chosen
Simple random	Every unit, and every combination of units, has an equal chance (lottery, random-number table).
Systematic	Every k-th unit from a list after a random start, where k = N ÷ n.
Stratified	Population divided into homogeneous strata (sex, class) and random samples drawn from each, proportionately or disproportionately.
Cluster and multistage	Naturally occurring groups (schools, villages) are sampled at random and all or some of their members studied; multistage repeats the step at several levels.
Convenience (accidental)	Whoever is available — the undergraduate subject pool.
Purposive (judgemental)	Chosen for a characteristic relevant to the study.
Quota	Non-random filling of set numbers per category — the non-probability analogue of stratified sampling.
Snowball	Participants recruit others — used for hidden or stigmatised populations.
Theoretical	In grounded theory, cases chosen to develop emerging categories until saturation.

🧠 Random sampling vs random assignment

Random sampling decides who is in the study and protects external validity (generalisation). Random assignment decides which condition each participant gets and protects internal validity (causal inference). A quasi-experiment lacks random assignment, not necessarily random sampling.

4. Ethics in conducting and reporting research

Modern research ethics grew out of abuses: the Nuremberg Code (1947) after Nazi experiments, the Declaration of Helsinki (1964), and in the United States the Belmont Report (1979) after the Tuskegee syphilis study, with its three principles of respect for persons, beneficence and justice. Psychology's own landmarks — Milgram's obedience studies (1963) and Zimbardo's Stanford prison study (1971) — shaped the rules on deception and on the right to withdraw. The APA Ethical Principles of Psychologists and Code of Conduct and, in India, the ICMR National Ethical Guidelines (2017) are the working codes, enforced through institutional ethics committees.

Conducting: voluntary informed consent (assent plus guardian consent for children); the right to withdraw without penalty; minimising risk and harm; confidentiality and anonymity; deception only when justified and unavoidable, followed by debriefing; special protection for vulnerable groups.
Reporting: no fabrication (inventing data), falsification (altering data or results) or plagiarism; honest authorship credit; reporting all conditions and measures; not HARKing (hypothesising after results are known) or p-hacking; sharing data for verification.

5. Paradigms and methods of research

Quantitative research tests hypotheses with numerical data and statistical analysis; qualitative research seeks meaning and experience in words, images and observation; mixed methods combine the two within one study. Creswell's core mixed designs are convergent parallel (both strands at once, then merged), explanatory sequential (quantitative first, qualitative to explain it), exploratory sequential (qualitative first, then a quantitative phase built on it) and embedded designs.

The methods named in the syllabus
Method	Defining feature
Observation	Naturalistic or controlled; participant or non-participant; structured (checklists, time sampling) or unstructured.
Survey	Interviews (structured, semi-structured, unstructured) and questionnaires (open or closed items), on samples, to describe attitudes and behaviour.
Experimental	Manipulation of an IV, random assignment and control — the only method that licenses causal conclusions.
Quasi-experimental	An IV is manipulated or compared but groups are not randomly assigned (non-equivalent groups, interrupted time series).
Field study	Research in real settings — a field experiment manipulates variables there; a field study observes relations without manipulation.
Cross-cultural	Comparison across cultures; the etic approach applies universal categories from outside, the emic approach studies a culture in its own terms (Pike; Berry).
Phenomenology	In-depth description of the essence of a lived experience from those who have had it.
Grounded theory	Glaser and Strauss (1967): theory generated from data through constant comparison, open, axial and selective coding, theoretical sampling and saturation.
Focus group	A moderated discussion among about six to ten participants; the interaction itself is data.
Narrative	Life stories and how people make meaning by storying experience.
Case study	Intensive study of one person, group or event — Phineas Gage, Freud's cases, H. M.
Ethnography	Prolonged immersion and participant observation in a cultural group, producing "thick description" (Geertz).

Key takeaways

Research is exploratory, descriptive or explanatory in purpose, and Neuman classifies it by use, purpose, time and data collection.
Operational definitions make variables measurable; confounds vary with the IV; moderators answer "when or for whom", mediators "how or why".
Probability samples (simple random, systematic, stratified, cluster) permit generalisation; non-probability samples (convenience, purposive, quota, snowball, theoretical) do not in the statistical sense.
Ethics rests on respect for persons, beneficence and justice: informed consent, withdrawal, confidentiality, justified deception with debriefing, and honest reporting without fabrication, falsification or plagiarism.
Only a true experiment, with manipulation and random assignment, supports causal claims; the qualitative methods — phenomenology, grounded theory, focus groups, narratives, case studies, ethnography — seek meaning rather than cause.

Practice questions (10)

Attempt each one before opening the answer. Every explanation names the tempting wrong option as well as the right one, because that is where marks are lost.

A researcher finds that stress predicts illness strongly among people with little social support but weakly among those with much support. Social support here is a:
1. Mediator variable
2. Dependent variable
3. Confounding variable
4. Moderator variable
Show answer
Answer: D — Moderator variable
Social support changes the strength of the stress–illness relationship — it tells us for whom the effect holds — which defines a moderator. A mediator would be the mechanism through which stress causes illness.
A college has 600 men and 400 women. A researcher draws a proportionate stratified random sample of 50 students by sex. How many women should be in the sample? Type a number.
Numerical answer — type the value.
Show answer
Answer: 20
Women are 400 of 1,000, or 40 per cent of the population. Proportionate allocation gives 0.40 × 50 = 20 women (and 30 men).
From a list of 1,200 employees, a systematic sample of 60 is to be drawn. What is the sampling interval k? Type a number.
Numerical answer — type the value.
Show answer
Answer: 20
k = N ÷ n = 1,200 ÷ 60 = 20. After a random start between 1 and 20, every 20th name is taken.
Which of the following is a NON-probability sampling design?
1. Quota sampling
2. Stratified random sampling
3. Systematic sampling
4. Cluster sampling
Show answer
Answer: A — Quota sampling
Quota sampling fills categories non-randomly, so selection probabilities are unknown. Cluster, systematic and stratified random sampling all give each unit a known chance of selection.
Which of the following are principles of the Belmont Report (1979)? Select all that apply.
1. Respect for persons
2. Statistical significance
3. Justice
4. Beneficence
Show answer
Answer: A — Respect for persons; C — Justice; D — Beneficence
The Belmont Report rests on respect for persons (autonomy and informed consent), beneficence (maximise benefit, minimise harm) and justice (fair distribution of burdens and benefits). Statistical significance is not an ethical principle.
In grounded theory, the stage of coding that relates categories to their sub-categories around a central axis is:
1. Content coding
2. Axial coding
3. Open coding
4. Selective coding
Show answer
Answer: B — Axial coding
In Strauss and Corbin's scheme, open coding breaks data into concepts, axial coding links categories and sub-categories, and selective coding integrates everything around a core category.
A study compares the reading scores of two existing classrooms, one of which receives a new teaching method, without random assignment of pupils. This design is:
1. Quasi-experimental
2. Ethnographic
3. Phenomenological
4. A true experiment
Show answer
Answer: A — Quasi-experimental
The treatment is compared across intact, non-equivalent groups, which is the defining feature of a quasi-experiment. A true experiment would require random assignment of pupils to conditions.
Assertion (A): Random assignment strengthens the internal validity of an experiment. Reason (R): Random assignment makes groups equivalent, on average, on all extraneous variables before treatment, so later differences can be attributed to the independent variable. Choose the correct option.
1. Both A and R are true, and R is the correct explanation of A
2. Both A and R are true, but R is not the correct explanation of A
3. A is true, but R is false
4. A is false, but R is true
Show answer
Answer: A — Both A and R are true, and R is the correct explanation of A
Both are true and R explains A: equivalence of groups before treatment is exactly what rules out selection and other confounds, which is what internal validity means.
Studying a culture in its own terms, with categories drawn from within it, is the:
1. Etic approach
2. Nomothetic approach
3. Emic approach
4. Operational approach
Show answer
Answer: C — Emic approach
Pike's emic perspective is the insider's, culture-specific view; the etic perspective applies outside, supposedly universal categories. Nomothetic research seeks general laws across individuals.
Which of the following are breaches of ethics in REPORTING research? Select all that apply.
1. Reporting all the measures that were collected
2. Falsification of results
3. Presenting a hypothesis formed after seeing the results as if it had been predicted
4. Debriefing participants after a deception study
Show answer
Answer: B — Falsification of results; C — Presenting a hypothesis formed after seeing the results as if it had been predicted
Falsification and HARKing misrepresent what the research found. Debriefing after deception and reporting all measures are required good practice, not breaches.