Saral Shiksha Yojna
Courses/Behavioral Research: Statistical Methods

Behavioral Research: Statistical Methods

CG3.402
Vinoo Alluri•Monsoon 2025-26•4 credits

Answer Structure Templates

University exams reward formatting. Use these.

Define + formula + when to use

  1. Give 1-line definition.
  2. State the formula.
  3. State the assumptions / range / domain.
  4. Give an example use case.
  5. Note one common misinterpretation.

Pick-the-test

  1. Identify DV scale (nominal / ordinal / interval / ratio).
  2. Identify IV scale and #groups.
  3. Decide independent vs paired.
  4. Pick parametric default; check assumptions; drop to nonparametric if violated.
  5. State the test, df formula, and effect size measure.

Critical evaluation of a study

  1. IV/DV identification.
  2. Hypotheses.
  3. Test recommendation with justification.
  4. Assumptions check.
  5. Effect size + CI.
  6. Discuss limitations + recommendations.

Bayes update (Bayes' rule problem)

  1. List P(H), P(D|H), P(D|¬H).
  2. Compute P(D) = P(D|H)·P(H) + P(D|¬H)·P(¬H).
  3. Apply Bayes: P(H|D) = P(D|H)·P(H) / P(D).
  4. Compare to the prior — Bayes makes the update explicit.

Hypothesis-test report

  1. State H₀ and H₁.
  2. Choose α.
  3. State test + df.
  4. Report statistic, p-value, effect size, 95% CI.
  5. Interpret: statistical AND practical significance.

Multicollinearity diagnosis

  1. Inspect predictor correlations / VIFs.
  2. Note VIF > 5–10 = severe.
  3. Recommend: drop / combine via PCA / ridge / collect more data.
  4. Caveat: model may still predict well even when coefficients are unstable.

ANOVA analysis plan

  1. State H₀ and H₁ across the k groups.
  2. Check assumptions (Levene, Q-Q).
  3. Report F, df, p, η² / partial η².
  4. If significant → Tukey HSD post-hoc.
  5. If RM-ANOVA → Mauchly + Greenhouse-Geisser as needed.

Factor analysis pipeline

  1. Item screening + reliability (Cronbach's α).
  2. Sampling adequacy: KMO + Bartlett.
  3. Choose # factors: parallel analysis > scree > Kaiser.
  4. EFA → rotation → interpret loadings.
  5. CFA on held-out sample; report fit indices (CFI, RMSEA, SRMR).