Saral Shiksha Yojna
Courses/Behavioral Research: Statistical Methods

Behavioral Research: Statistical Methods

CG3.402
Vinoo Alluri•Monsoon 2025-26•4 credits
Sample Papers/Mock Paper 5 — All Units Mix

Mock Paper 5 — All Units Mix

Duration: 120 min • Max marks: 100

Section A — MCQ (20 × 1 = 20)

20 marks
  1. 1.Scale that permits 'Person A's response is exactly twice Person B's': (a) Nominal (b) Ordinal (c) Interval (d) Ratio1 m
  2. 2.Large n yields p = 0.001 and r = 0.05. Conclude: (a) Strong evidence of a large effect (b) Real but tiny effect, practical importance limited (c) Probably p-hacking (d) Computation error1 m
  3. 3.A demand characteristic is: (a) Statistical assumption (b) Participants guessing the hypothesis and behaving accordingly (c) Sampling bias (d) ANOVA assumption1 m
  4. 4.χ² independence compares: (a) Observed vs theoretical means (b) Observed vs expected cell frequencies (c) Sample variances (d) Ranked values1 m
  5. 5.Binomial(n, p) mean is: (a) p (b) np (c) n + p (d) p/n1 m
  6. 6.'Uninformative prior' in Bayesian inference means: (a) Strong prior beliefs (b) Prior with little impact on posterior (c) Posterior with no data (d) A point mass1 m
  7. 7.qnorm(0.975) returns approximately: (a) 0.975 (b) 1.96 (c) 2.58 (d) 0.0251 m
  8. 8.High VIF for X₁ in regression: (a) X₁ unimportant (b) X₁ strongly collinear with other predictors (c) X₁ has high effect size (d) Missing data1 m
  9. 9.'Interaction effect' in two-way ANOVA: (a) Both main effects are significant (b) Effect of one factor depends on the level of another (c) Sample sizes unequal (d) Sphericity violated1 m
  10. 10.A CI that excludes the null value: (a) Cannot exist (b) Implies the result is significant at the corresponding α (c) Has 100% coverage (d) Means H₀ is proven false1 m
  11. 11.'Fishhook' anomaly in a residual-vs-predicted plot: (a) Constant variance (b) Linearity (c) Misspecified non-linearity (d) Independence1 m
  12. 12.Hawthorne effect refers to: (a) Instrument calibration drift (b) Behaviour change due to awareness of being observed (c) Outlier removal (d) Order effect1 m
  13. 13.Operationalisation is: (a) Computing means (b) Defining a measurable instance of an abstract construct (c) Choosing a test (d) Sampling1 m
  14. 14.Cronbach's α = 0.91 for a 10-item scale indicates: (a) Strong internal consistency (b) Strong construct validity (c) Large effect (d) High Type I error1 m
  15. 15.Sample with deliberately oversampled rare subgroups: (a) Simple random (b) Stratified (c) Cluster (d) Convenience1 m
  16. 16.Factor analysis suggests 3 factors via parallel analysis. Next step: (a) CFA on independent data (b) Re-run with more factors (c) PCA instead (d) Increase α1 m
  17. 17.Power = 0.5 study finds p = 0.07 with small effect. Likely: (a) No effect (b) Small effect possibly missed due to low power (c) Strong evidence (d) Sampling error1 m
  18. 18.BF = 1 indicates: (a) Strong for H₁ (b) Strong for H₀ (c) No evidence either way (d) Posterior = Prior1 m
  19. 19.Example of external validity: (a) Random assignment (b) Reliable measurement (c) Generalising to a different population (d) Pre-registration1 m
  20. 20.In ANOVA, df_between for k groups: (a) k − 1 (b) N − k (c) N − 1 (d) k + 11 m

Section B — MSQ (10 × 2 = 20)

20 marks
  1. 1.Advantages of pre-registration: (a) Prevents HARKing (b) Distinguishes confirmatory from exploratory (c) Eliminates all bias (d) Increases transparency (e) Improves replicability2 m
  2. 2.Sources of variability partitioned in one-way ANOVA: (a) Between-group (b) Within-group (c) Total (d) Interaction (e) Subject variability2 m
  3. 3.Increase statistical power: (a) Larger n (b) Larger effect size (c) Smaller variance (d) Smaller α (e) Within-subjects (vs between)2 m
  4. 4.Qualify as descriptive statistics: (a) Mean (b) Median (c) SD (d) p-value (e) IQR2 m
  5. 5.Threats to internal validity: (a) History (b) Maturation (c) Selection (d) Sampling from a single university (e) Mortality / attrition2 m
  6. 6.Methods to deal with multicollinearity: (a) Drop one of the correlated predictors (b) Ridge regression (c) Combine into a composite via PCA (d) Increase α (e) Center/standardise variables (for interactions)2 m
  7. 7.Apply to binomial: (a) Discrete (b) Two outcomes per trial (c) Fixed n (d) Constant p (e) Trials independent2 m
  8. 8.Appropriate to summarise a heavily skewed dataset: (a) Mean (b) Median (c) IQR (d) Boxplot (e) SD2 m
  9. 9.Limitations of frequentist NHST: (a) Binary reject/fail-to-reject thinking (b) Cannot quantify evidence for H₀ (c) Sensitive to optional stopping (d) Provides explicit prior beliefs (e) Often relies on threshold α = .052 m
  10. 10.Correct uses of regression: (a) Predicting an outcome from predictors (b) Estimating strength of associations (c) Establishing causation from observational data alone (d) Testing whether a coefficient differs from 0 (e) Forecasting2 m

Section C — Short descriptive (6 × 5 = 30)

30 marks
  1. 1.Differentiate sensitivity and specificity. Why does PPV depend on prevalence?5 m
  2. 2.Random sampling vs random assignment. Why each matters.5 m
  3. 3.A priori (planned) comparisons vs post-hoc tests — when use each?5 m
  4. 4.PCA vs FA — assumptions, interpretation, use cases.5 m
  5. 5.Why is the CLT so central to inferential statistics?5 m
  6. 6.r = 0.45 — compute R² and interpret. What does it not tell us about causality?5 m

Section D — Long descriptive (3 × 10 = 30)

30 marks
  1. 1.Walk through the conceptual logic of an independent-samples t-test. (i) Hypotheses. (ii) Formula + components. (iii) Assumptions. (iv) Statistic → p-value. (v) Interpretation. (vi) Assumption violations.10 m
  2. 2.Critically evaluate α = 0.05 as a significance threshold. (i) Origin. (ii) Strengths. (iii) Limitations. (iv) Alternatives.10 m
  3. 3.Design a within-subjects study on sleep deprivation × cognitive performance: 5 tasks (0–100), normal sleep vs sleep deprivation. (i) Hypotheses. (ii) Controls + counterbalancing. (iii) Sample-size justification. (iv) Analyses. (v) Reporting + ethics.10 m

Track your attempt locally — score and time are recorded in your browser. (Coming soon: timed-attempt mode.)