SIOP Research Methods & Data Analysis in I-O Psychology 5 — Questions and Answers
Question 1: Which of the following best describes the purpose of bootstrapping in mediation analysis?
- To increase sample size by duplicating observations
- To estimate the sampling distribution of the indirect effect without assuming normality (Correct answer)
- To replace missing data with plausible values
- To test moderation effects in the model
Correct answer: To estimate the sampling distribution of the indirect effect without assuming normality
Bootstrapping resamples the data repeatedly to construct an empirical sampling distribution for the indirect effect, providing more accurate confidence intervals.
Question 2: A moderator variable in a regression context is indicated by:
- A high zero-order correlation between predictor and outcome
- A significant interaction term between predictor and moderator (Correct answer)
- A suppressor effect on the main predictor
- A significant main effect for the mediator
Correct answer: A significant interaction term between predictor and moderator
Moderation is evidenced by a significant interaction term, indicating that the predictor-outcome relationship differs across levels of the moderator.
Question 3: An organization collects supervisor ratings and subordinate self-ratings of performance. The low correlation between them is best explained by:
- Common method variance
- Criterion contamination
- Rater perspective differences and unique information (Correct answer)
- Range restriction in ratings
Correct answer: Rater perspective differences and unique information
Supervisors and subordinates observe different behaviors and have access to different information, which typically produces low convergence.
Question 4: What is the key statistical assumption underlying the use of parametric tests like ANOVA and t-tests?
- Variables are measured on an ordinal scale
- Data are drawn from normally distributed populations (Correct answer)
- All predictors are uncorrelated
- Effect sizes must exceed Cohen's conventions
Correct answer: Data are drawn from normally distributed populations
Parametric tests assume that the data come from populations with normal distributions, though they are robust to moderate violations with large samples.
Question 5: Which approach to handling missing data is generally superior to listwise deletion when data are missing at random (MAR)?
- Mean substitution
- Pairwise deletion
- Multiple imputation (Correct answer)
- Single imputation
Correct answer: Multiple imputation
Multiple imputation creates several plausible datasets reflecting uncertainty about missing values, yielding unbiased estimates under the MAR assumption.
Question 6: A researcher conducting an organizational survey finds that employees in the same department have more similar responses than employees across departments. Which analysis is most appropriate?
- Simple linear regression ignoring grouping
- Independent samples t-test
- Multilevel modeling (HLM) (Correct answer)
- Factor analysis
Correct answer: Multilevel modeling (HLM)
Multilevel modeling accounts for the non-independence of observations nested within groups such as departments or teams.
Question 7: Which of the following is a key difference between exploratory factor analysis (EFA) and confirmatory factor analysis (CFA)?
- EFA requires larger samples than CFA
- EFA tests a pre-specified factor structure while CFA generates one from data
- CFA tests a pre-specified factor structure while EFA generates one from data (Correct answer)
- CFA cannot estimate factor loadings
Correct answer: CFA tests a pre-specified factor structure while EFA generates one from data
CFA tests whether data fit a theoretically specified factor structure, while EFA is data-driven and uncovers factor patterns without prior constraints.
SIOP Research Methods & Data Analysis in I-O Psychology 5
Which of the following best describes the purpose of bootstrapping in mediation analysis?