Research Aptitude
Research Aptitude
Paper I checks whether you can classify a study, name its variables, judge its design and spot ethical breaches - the vocabulary every supervisor assumes.
Typology you must be able to sort
- By purpose: basic (theory-building), applied (solves a practical problem), action research (practitioner, cyclical: plan–act–observe–reflect, local solution).
- By design: descriptive, correlational, ex-post facto (effect studied after it occurred, no manipulation), quasi-experimental (intact groups, no random assignment), true experimental (randomisation + control + manipulation).
- Variables: independent, dependent, extraneous (held constant), confounding (varies with the IV), moderator, mediator.
- Measurement scales: nominal (labels), ordinal (rank, unequal gaps), interval (equal gaps, arbitrary zero), ratio (true zero).
- Sampling: probability - simple random, systematic, stratified, cluster, multistage; non-probability - convenience, purposive, quota, snowball.
- Validity: internal, external, content, construct, criterion. Reliability: test-retest, split-half, Cronbach's alpha, inter-rater.
Worked example
A researcher teaches two existing school sections by different methods and compares post-test scores. No random assignment was possible, so the design is quasi-experimental, not true experimental. The main internal-validity threat is selection (pre-existing differences between sections); a pre-test would partially control it. Calling it "ex-post facto" is wrong because the treatment was deliberately manipulated.
Traps that cost marks
- Treating reliability and validity as the same; a test may be reliable but invalid.
- Calling ordinal data interval because the gaps "look equal".
- Ignoring UGC's 2018 academic-integrity bands (0–10%, 10–40%, 40–60%, above 60% similarity) and the penalties attached to each.
- Missing that self-plagiarism and patch-writing are plagiarism too.