Methods and Steps of Research
UGC NET Paper 1: experimental, descriptive, historical, qualitative and quantitative methods, steps of research, hypotheses, sampling and validity.
- 17 min
On this page · 15 sections
- Prerequisites
- The Hook: One Question, Four Methods
- 1. Variables: the Vocabulary of Methods
- 2. The Experimental Method
- 3. The Descriptive Method
- 4. The Historical Method
- 5. Quantitative and Qualitative Methods
- 6. Steps of Research
- Memory Tricks
- Don’t Confuse
- Common Mistakes
- Practice
- Quick Revision Table
- Key Takeaways
- What’s Next
Prerequisites
- Research: Meaning, Types and Approaches — positivism lies behind quantitative methods, interpretivism behind qualitative ones.
- Research Aptitude: chapter overview — where this topic sits among the six official lines.
The Hook: One Question, Four Methods
1. Variables: the Vocabulary of Methods
| Variable | Role | Example (effect of a new teaching method on scores) |
|---|---|---|
| Independent (IV) | The presumed cause; manipulated or selected by the researcher | Teaching method |
| Dependent (DV) | The presumed effect; measured | Test scores |
| Extraneous | Any other variable that may affect the DV and must be controlled | Prior ability, teacher’s experience |
| Confounding | An extraneous variable that varies with the IV, so its effect cannot be separated | Better teacher assigned only to the new method |
| Intervening (mediating) | Links IV to DV; inferred, not directly observed | Student motivation |
| Moderator | Changes the strength or direction of the IV–DV relationship | Gender, type of college |
2. The Experimental Method
An experiment is the only method that can establish cause and effect with confidence. It has three defining features:
- Manipulation — the researcher deliberately changes the independent variable.
- Control — extraneous variables are held constant, and a control group that does not receive the treatment is compared with the experimental group.
- Randomisation — subjects are randomly assigned to groups, so that groups are equivalent at the start.
The logic of comparing a group that gets the treatment with one that does not goes back to J.S. Mill’s methods of agreement and difference (A System of Logic, 1843). Randomisation as the basis of experimental design was set out by the statistician R.A. Fisher in The Design of Experiments (1935).
Types of experimental design
Donald Campbell and Julian Stanley’s Experimental and Quasi-Experimental Designs for Research (1963) gave the classification still used in textbooks:
| Category | Random assignment? | Control group? | Examples |
|---|---|---|---|
| Pre-experimental | No | Absent or not equivalent | One-shot case study; one-group pretest–posttest; static-group comparison |
| True experimental | Yes | Yes | Pretest–posttest control group; posttest-only control group; Solomon four-group |
| Quasi-experimental | No (intact groups such as existing classes) | Usually | Non-equivalent control group; time-series design |
The Solomon four-group design uses four groups (two with pretest, two without; in each pair one gets the treatment) so the researcher can see whether taking the pretest itself affected the result.
Validity of an experiment
| Validity | Question | Strengthened by |
|---|---|---|
| Internal validity | Did the IV really cause the change in the DV? | Random assignment, control group, holding conditions constant |
| External validity | Can the result be generalised to other people, places and times? | Random sampling from the population, realistic (field) settings |
Campbell and Stanley’s threats to internal validity include history (outside events during the study), maturation (natural change in subjects), testing (effect of the pretest), instrumentation (changes in the measuring tool), statistical regression (extreme scorers drifting towards the mean), selection (non-equivalent groups) and experimental mortality (dropouts).
Effects that distort experiments
| Effect | What happens |
|---|---|
| Hawthorne effect | Subjects change behaviour because they know they are being studied (named after the Western Electric Hawthorne Works studies of the 1920s–1930s) |
| John Henry effect | The control group works harder to compete with the experimental group |
| Pygmalion effect | Higher expectations from the teacher lead to better performance (Rosenthal and Jacobson, Pygmalion in the Classroom, 1968) |
| Placebo effect | Improvement because subjects believe they are receiving a treatment |
3. The Descriptive Method
Descriptive research describes “what is”: the present status of a phenomenon. The researcher does not manipulate any variable.
| Type | What it does | Example |
|---|---|---|
| Survey | Collects data from a large sample by questionnaire, schedule or interview | Survey of hostel facilities in state universities |
| Case study | Intensive, in-depth study of one unit: a person, family, institution or community | Study of one autonomous college’s governance |
| Correlational | Measures the degree of relationship between two or more variables (coefficient from −1 to +1) | Relation between library use and grades |
| Developmental | Studies change with age or time: longitudinal (same group followed) or cross-sectional (different age groups at one time) | Language development from age 3 to 10 |
| Ex post facto (causal-comparative) | Studies possible causes after the effect has occurred, by comparing groups that already differ; the IV cannot be manipulated | Comparing achievement of children from single-parent and two-parent homes |
| Observational | Records behaviour as it occurs: participant or non-participant, structured or unstructured | Classroom interaction study |
Questionnaire, schedule and interview
| Tool | Who fills it | Strength | Weakness |
|---|---|---|---|
| Questionnaire | The respondent, often by post or online | Cheap, covers large and scattered samples | Needs literate respondents; low response rate |
| Schedule | The enumerator, who asks the questions and records answers | Works with non-literate respondents; higher response | Costly; enumerator bias |
| Interview | Face-to-face or by phone; structured, semi-structured or unstructured | Depth, clarification possible | Time-consuming; interviewer effect |
Other tools you should recognise: rating scales, the Likert scale (Rensis Likert, 1932; agreement on a five-point scale), checklists, sociometry (Jacob L. Moreno; mapping choices within a group as a sociogram) and the Delphi technique (rounds of anonymous expert opinion until consensus).
4. The Historical Method
Historical research studies past events to understand the present and anticipate the future. Its data are records, not experiments.
Sources
| Source | Meaning | Examples |
|---|---|---|
| Primary | Created at the time by participants or eyewitnesses | Original documents, letters, diaries, minutes, official records, inscriptions, artefacts |
| Secondary | Accounts written later by someone who was not a witness, using primary sources | Textbooks, histories, encyclopaedias, reviews |
Historical criticism
Every source is tested twice, and in this order:
- External criticism — Is the document genuine? Checks authenticity: authorship, date, place, material, handwriting, signs of forgery or later insertion.
- Internal criticism — Is the content credible and accurate? Checks the meaning of the statements, the author’s competence, bias and motive, and consistency with other sources.
Limitations: incomplete records, the researcher cannot observe events directly, and sources reflect the bias of those who wrote and preserved them.
5. Quantitative and Qualitative Methods
| Basis | Quantitative | Qualitative |
|---|---|---|
| Philosophy | Positivism / post-positivism | Interpretivism, constructivism, phenomenology |
| Aim | Measure, test hypotheses, generalise, predict | Understand meanings, experiences, context |
| Reasoning | Mainly deductive | Mainly inductive |
| Data | Numbers | Words, images, observations |
| Sample | Large, often random | Small, purposive; until saturation |
| Researcher | Detached; uses standardised instruments | The researcher is the main instrument |
| Design | Fixed in advance | Emergent, flexible |
| Tools | Tests, structured questionnaires, scales | In-depth interviews, participant observation, focus groups, documents |
| Analysis | Statistics | Coding, themes, interpretation |
| Quality criteria | Reliability, validity, generalisability | Trustworthiness: credibility, transferability, dependability, confirmability (Lincoln and Guba, Naturalistic Inquiry, 1985) |
Main qualitative traditions
| Tradition | Focus | Name to know |
|---|---|---|
| Ethnography | Culture of a group, through long participant observation | Bronisław Malinowski’s fieldwork in the Trobriand Islands (Argonauts of the Western Pacific, 1922) |
| Phenomenology | Essence of a lived experience | Edmund Husserl |
| Grounded theory | Building theory from the data, not testing a prior theory | Barney Glaser and Anselm Strauss, The Discovery of Grounded Theory (1967) |
| Case study | In-depth study of a bounded case | — |
| Narrative | Life stories and how people tell them | — |
Thick description, a term Clifford Geertz used in The Interpretation of Cultures (1973) after Gilbert Ryle, means describing behaviour together with its context and meaning, so an outsider can understand it.
Mixed methods and triangulation
Mixed-methods research combines quantitative and qualitative data in one study. Triangulation means checking a finding through more than one angle. Norman Denzin distinguished four kinds: data triangulation (different sources), investigator triangulation (different researchers), theory triangulation (different theoretical lenses) and methodological triangulation (different methods).
6. Steps of Research
The sequence below is the one Indian textbooks follow, slightly condensed from the research process in C.R. Kothari’s Research Methodology (which also lists “execution of the project” between data collection and analysis). In an MCQ, the order matters.
graph TD
A["1. Identify and formulate
the research problem"] --> B["2. Review of literature"]
B --> C["3. Formulate hypotheses"]
C --> D["4. Prepare research design"]
D --> E["5. Determine sample design"]
E --> F["6. Collect data"]
F --> G["7. Analyse data"]
G --> H["8. Test hypotheses"]
H --> I["9. Interpret and generalise"]
I --> J["10. Prepare the report"]Step by step
| Step | What happens | Exam point |
|---|---|---|
| 1. Research problem | Choose a researchable, specific, significant and feasible question | A good problem is clear, limited in scope and answerable with data |
| 2. Review of literature | Study earlier work | Avoids duplication, finds gaps, sharpens the problem, suggests hypotheses and methods; comes before the hypothesis |
| 3. Hypothesis | A tentative, testable statement of the expected relationship between variables | Must be testable, specific, consistent with known facts, stated simply |
| 4. Research design | The blueprint for collecting and analysing data | Exploratory, descriptive, diagnostic or experimental |
| 5. Sample design | Decide the population, sampling frame, method and size | See sampling below |
| 6. Data collection | Use tools such as tests, questionnaires, interviews, observation | Pilot study first to test the tools |
| 7–8. Analysis and testing | Code, tabulate, apply statistics; accept or reject hypotheses | Parametric vs non-parametric tests |
| 9. Interpretation | Explain findings, link to theory, state limits | Generalise only as far as the sample allows |
| 10. Report | Write the thesis or article | See topic 3 |
Hypotheses and errors
- Null hypothesis (H₀): no relationship or no difference. It is the hypothesis that is statistically tested.
- Alternative hypothesis (H₁): there is a relationship or difference.
- Directional hypothesis states the direction (“girls score higher”); non-directional does not (“there is a difference”).
| H₀ actually true | H₀ actually false | |
|---|---|---|
| Reject H₀ | Type I error (probability α, the level of significance) | Correct decision |
| Fail to reject H₀ | Correct decision | Type II error (probability β) |
A Type I error is a “false alarm”; a Type II error is a “missed effect”.
Sampling
| Probability (every unit has a known, non-zero chance) | Non-probability (chance of selection unknown) |
|---|---|
| Simple random — lottery or random numbers | Convenience — whoever is easily available |
| Systematic — every kth unit after a random start | Purposive (judgement) — chosen for a specific purpose or expertise |
| Stratified — divide into homogeneous strata, sample from each | Quota — fill fixed quotas by category, non-randomly |
| Cluster — randomly select whole groups (schools, villages) | Snowball — each respondent refers others; for hidden or hard-to-reach populations |
| Multistage — clusters within clusters in stages |
Levels of measurement
S.S. Stevens set out four levels in “On the Theory of Scales of Measurement” (Science, 1946):
| Level | Property | Example |
|---|---|---|
| Nominal | Categories only | Gender, religion, roll numbers |
| Ordinal | Order, unequal intervals | Ranks, Likert responses (strictly) |
| Interval | Equal intervals, no true zero | Temperature in °C, IQ |
| Ratio | Equal intervals and a true zero | Height, income, age |
Mnemonic: NOIR.
Reliability and validity of tools
- Reliability = consistency. Types: test–retest, parallel forms, split-half, inter-rater; internal consistency is often reported as Cronbach’s alpha.
- Validity = measures what it claims to measure. Types: face, content, criterion-related (concurrent and predictive) and construct.
- A tool can be reliable without being valid (a clock that is always five minutes fast), but it cannot be valid without being reliable.
Parametric and non-parametric tests
| Parametric (assume normal distribution, interval or ratio data) | Non-parametric (distribution-free; nominal or ordinal data) |
|---|---|
| t-test, z-test, ANOVA (F-test), Pearson’s r | Chi-square, Mann–Whitney U, Wilcoxon, Kruskal–Wallis, Spearman’s rank correlation |
Memory Tricks
Don’t Confuse
| This | vs | That |
|---|---|---|
| Random sampling: how subjects are drawn from the population → external validity | vs | Random assignment: how subjects are put into groups → internal validity |
| Experimental: researcher manipulates the IV | vs | Ex post facto: IV already happened; no manipulation |
| True experiment: random assignment | vs | Quasi-experiment: intact groups, no random assignment |
| External criticism: genuineness of the document | vs | Internal criticism: credibility of its content |
| Primary source: eyewitness, contemporary | vs | Secondary source: later account |
| Questionnaire: respondent fills | vs | Schedule: enumerator fills |
| Longitudinal: same group over time | vs | Cross-sectional: different groups at one time |
| Type I: reject a true H₀ (α) | vs | Type II: retain a false H₀ (β) |
| Hawthorne: subjects know they are observed | vs | John Henry: control group competes harder |
Common Mistakes
Practice
All questions below are original practice questions written in the UGC NET format. They are not previous-year questions.
Level 1: Foundation
- In a study of the effect of yoga on stress, which is the dependent variable?
- How many groups does the Solomon design use?
- Checking whether a manuscript is a forgery is part of ______ criticism.
Answers
- Stress (the measured outcome); yoga is the independent variable.
- Four (Solomon four-group design).
- External criticism.
Level 2: Exam-style MCQs
Q1. (Practice) A researcher compares the lung capacity of people who have smoked for twenty years with that of non-smokers, to study the effect of smoking. The method is:
- (a) True experimental
- (b) Ex post facto
- (c) Action research
- (d) Historical
Answer and explanation
(b) Ex post facto. The “treatment” (smoking) has already happened and cannot ethically be assigned; the researcher compares existing groups to infer causes after the fact.
Q2. (Practice) Which experimental design allows the researcher to check whether the pretest itself influenced the results?
- (a) One-shot case study
- (b) Static-group comparison
- (c) Solomon four-group design
- (d) Time-series design
Answer and explanation
(c). By including groups that receive no pretest, the Solomon design separates the effect of testing from the effect of the treatment.
Q3. (Practice) A study of the problems of migrant construction workers who have no fixed address finds respondents through earlier respondents. The sampling method is:
- (a) Stratified random sampling
- (b) Cluster sampling
- (c) Systematic sampling
- (d) Snowball sampling
Answer and explanation
(d) Snowball sampling, used when no sampling frame exists and members of a hidden population can lead the researcher to others.
Q4. (Practice) Consider the following statements about qualitative research:
- It is mainly inductive.
- The researcher is regarded as the main instrument of data collection.
- It requires large random samples to ensure generalisation.
Which of the statements given above is/are correct?
- (a) 1 only
- (b) 1 and 2 only
- (c) 2 and 3 only
- (d) 1, 2 and 3
Answer and explanation
(b) 1 and 2 only. Qualitative research uses small, purposive samples and aims at depth and transferability, not statistical generalisation.
Q5. (Practice) Rejecting a null hypothesis that is in fact true is called:
- (a) Type I error
- (b) Type II error
- (c) Sampling error
- (d) Standard error
Answer and explanation
(a) Type I error, whose probability is the level of significance (α).
Level 3: Assertion–reason, match and sequence
For assertion–reason questions, use these options:
- (a) Both A and R are true, and R is the correct explanation of A
- (b) Both A and R are true, but R is not the correct explanation of A
- (c) A is true, but R is false
- (d) A is false, but R is true
Q1. (Practice) Assertion (A): In historical research, external criticism is applied before internal criticism. Reason (R): The accuracy of a document’s content need not be judged until the document has been shown to be genuine.
Answer and explanation
(a). Establishing authenticity first is exactly why external criticism precedes internal criticism.
Q2. (Practice) Assertion (A): Random assignment of subjects to groups strengthens the internal validity of an experiment. Reason (R): Random assignment ensures that the sample represents the population.
Answer and explanation
(c). A is true: random assignment makes groups equivalent, so differences can be attributed to the treatment. R is false: representing the population is the job of random sampling, which concerns external validity.
Q3. (Practice) Match List I with List II.
| List I (Sampling method) | List II (Description) |
|---|---|
| A. Stratified | I. Whole groups such as schools are selected at random |
| B. Cluster | II. Respondents identify further respondents |
| C. Quota | III. Population divided into homogeneous strata, random selection from each |
| D. Snowball | IV. Fixed numbers from each category chosen non-randomly |
- (1) A-III, B-I, C-IV, D-II
- (2) A-I, B-III, C-IV, D-II
- (3) A-III, B-I, C-II, D-IV
- (4) A-IV, B-I, C-III, D-II
Answer and explanation
(1) A-III, B-I, C-IV, D-II. The stratified/quota pair is the trap: both use categories, but only stratified selects randomly within them.
Q4. (Practice) Arrange the following steps of research in the correct sequence:
A. Review of literature
B. Formulation of hypotheses
C. Identification of the research problem
D. Collection of data
E. Analysis and interpretation
(1) A, C, B, D, E
(2) C, A, B, D, E
(3) C, B, A, D, E
(4) C, A, D, B, E
Answer and explanation
(2) C, A, B, D, E. Problem → literature → hypotheses → data → analysis. The review of literature precedes the hypothesis.
Quick Revision Table
| Fact | Detail |
|---|---|
| Experiment | Manipulation, control, randomisation; only method for firm causal claims |
| Mill / Fisher | Methods of agreement and difference (1843) / The Design of Experiments (1935) |
| Campbell and Stanley (1963) | Pre-experimental, true experimental, quasi-experimental |
| Solomon four-group | Controls the effect of the pretest |
| Internal / external validity | Random assignment / random sampling |
| Hawthorne / John Henry / Pygmalion | Being observed / control group competes / teacher expectation |
| Descriptive | Survey, case study, correlational, developmental, ex post facto, observation |
| Ex post facto | Causes studied after the effect; no manipulation |
| Historical | Primary vs secondary sources; external (genuine?) then internal (accurate?) criticism |
| Qualitative | Inductive, purposive sample, researcher as instrument, trustworthiness (Lincoln and Guba, 1985) |
| Grounded theory | Glaser and Strauss (1967) |
| Thick description | Geertz (1973) |
| Triangulation | Denzin: data, investigator, theory, methodological |
| Steps | Problem → literature → hypothesis → design → sample → data → analysis → testing → interpretation → report |
| Errors | Type I: reject true H₀ (α); Type II: retain false H₀ (β) |
| Probability sampling | Simple random, systematic, stratified, cluster, multistage |
| Non-probability | Convenience, purposive, quota, snowball |
| Measurement | Stevens (1946): nominal, ordinal, interval, ratio (NOIR) |
Key Takeaways
What’s Next
- Thesis and Article Writing, Referencing and ICT — the last step, writing it up, and the tools used at every step
- Research Ethics and Plagiarism — consent, confidentiality and honest reporting at each stage
- Research: Meaning, Types and Approaches — revisit positivism and interpretivism behind these methods
- Quick Revision