Behavioral Research: Statistical Methods
CG3.402Vinoo Alluri•Monsoon 2025-26•4 credits
PYQ-style paper · Paper B
Duration: 120 min • Max marks: 50
Section 1 — Objective Questions (15 marks)
0 marks- 1.Vinay measures the temperature in his lab at 0°C, 10°C, 20°C, 30°C across the day. Which statement is FALSE about this scale? (a) Differences are meaningful (b) Order is meaningful (c) "20°C is twice as hot as 10°C" is a valid statement (d) The zero is arbitrary, not absolute
- 2."Maturation effect" in a longitudinal study refers to: (a) Participants becoming familiar with the test (b) Natural biological/psychological changes over time that confound the IV (c) Differential dropout (d) Researcher influence on results
- 3.A researcher tests if a coin is fair by flipping it 100 times. The most appropriate test is: (a) Chi-square goodness-of-fit (b) Pearson correlation (c) t-test (d) ANOVA
- 4.Which statement about p-values is TRUE? (a) p is P(H₀ is true) (b) p is the probability of observing data this extreme or more, assuming H₀ is true (c) p < 0.05 means the alternative hypothesis is proven (d) Smaller p means a larger effect
- 5.Which is NOT a measure of central tendency? (a) Mean (b) Median (c) Mode (d) Standard deviation
- 6.Levene's test in the context of ANOVA checks for: (a) Normality of residuals (b) Homogeneity of variances across groups (c) Sphericity (d) Independence
- 7.A Bayes Factor BF₁₀ = 0.15 implies: (a) Strong evidence for H₁ (b) Strong evidence for H₀ (since 1/0.15 ≈ 6.7) (c) Inconclusive (d) Significant at α = .05
- 8.Cohen's d = 0.78 for a treatment vs. control comparison indicates: (a) Small effect (b) Medium effect (c) Large effect (d) No effect
- 9.In a repeated-measures ANOVA, Mauchly's test is non-significant. This means: (a) Sphericity is violated; apply Greenhouse-Geisser correction (b) Sphericity is satisfied; standard RM-ANOVA results are valid (c) The omnibus F is significant (d) Use Friedman test instead
- 10.A scree plot in factor analysis is used to: (a) Compute factor loadings (b) Decide the number of factors to retain by finding the elbow (c) Test for normality (d) Compute Cronbach's α
- 11.Manav at IIIT-H is interested in whether a new tutoring program (Yes/No) is associated with passing a coding contest (Pass/Fail). 240 students sampled. Best test: (a) Independent t-test (b) Chi-square test of independence (c) ANOVA (d) Pearson correlation
- 12.A regression yields R² = 0.04, p < 0.001 with n = 8000. Best interpretation: (a) Important and large effect (b) Real but very small effect; practical significance is minimal (c) Probably p-hacking (d) Compute error
- 13.A psychologist measures heart rate (bpm) before and after watching a horror film for 50 participants. Best test: (a) Independent t-test (b) Paired t-test (c) Chi-square (d) Pearson correlation
- 14.A correlation of r = 0 between height and IQ in a sample tells us: (a) Height and IQ are independent (b) No linear relationship in this sample; a nonlinear relationship may still exist (c) Height does not affect IQ (d) The sample size is too small
- 15.Pratham fits a multiple regression and finds adjusted R² = 0.62 while R² = 0.71. This gap suggests: (a) The model is excellent (b) Some predictors don't contribute much; adjusted R² penalizes them (c) Multicollinearity is severe (d) The model is over-fitted heavily — drop variables
- 16.The Welch ANOVA differs from one-way ANOVA in that it: (a) Does not require homogeneity of variances (b) Uses ranks instead of raw scores (c) Is for paired data (d) Requires sphericity
- 17.Kavita peeks at her data after every 5 new participants and stops when p < .05. This is: (a) Standard practice (b) Optional stopping → inflates Type I error rate (c) Required for replication (d) A way to save resources
- 18.A psychologist runs PCA on a 30-item personality questionnaire and retains 5 components. PCA differs from Factor Analysis in that: (a) PCA assumes latent factors cause items; FA does not (b) PCA maximizes total variance explained; FA models common variance with an error term (c) PCA is only for confirmatory analysis (d) FA is computationally simpler
- 19.A study reports "F(3, 196) = 12.4, p < .001, η²_p = .16." The η²_p indicates: (a) The model overfits (b) A large effect — about 16% of variance explained by the factor (c) Small effect (d) Non-significant
- 20.A researcher uses 1.5 × IQR rule on a boxplot and flags 3 data points as outliers. Best next step: (a) Always delete them (b) Investigate whether they are data-entry errors or natural extremes; report analyses with and without them (c) Replace with the mean (d) Increase α
Section 2 — Short Descriptive (15 marks)
0 marksSection 3 — Long Descriptive (20 marks)
0 marksTrack your attempt locally — score and time are recorded in your browser. (Coming soon: timed-attempt mode.)