Research design
A student finds an association between study hours and scores among volunteers. Define the question and target population; discuss selection and causality.
The questions cover the breadth of the course and ask you to justify decisions rather than merely recall functions.
Weekly student problem sets
Four 5-mark problems plus formative checks each week
Twelve revision questions
Variable types and missingness
A table includes ID, nominal group, ordinal 1–5 stress rating and quantitative score. Classify them and explain how to report three missing ratings.
Distributions
Groups A and B have the same mean but A has greater standard deviation. Why is a mean-only comparison misleading?
Probability
Explain the probability of two consecutive heads from a fair coin and why the third flip still has probability 1/2 of heads.
Confidence interval
A report says a frequentist 95% confidence interval assigns the fixed parameter a 95% probability of lying inside. Explain the issue.
Hypothesis testing
A study reports p=0.04. Why is this not P(H0 is true)? Why also report an effect magnitude?
t-tests
The same 20 people are measured before and after. Which t-test fits? Contrast this with independent groups.
Correlation and regression
A researcher observes r=0.65 and R²=0.42, then claims causation and perfect prediction. Identify two mistakes.
ANOVA
An omnibus F-test is significant for three groups. Why does this not prove that all three group means differ pairwise?
Categorical analysis
Distinguish within-row from full-sample percentages in a 2×3 table. What if some expected cell counts are small?
Alternative methods
For paired data with influential outliers, what sensitivity analysis might help? Why is it wrong to call it assumption-free?
Reporting and ethics
List the minimum items needed for a reproducible and privacy-conscious research report.
Capstone assignment
Using the 36-row synthetic dataset, formulate one descriptive or associational question. Submit a 250-word methods-and-results report, a descriptive table, at least two labelled plots, one justified analysis and a reproducible R script. Do not claim observational comparisons are causal. Describe two limitations.
Proposed 100-point rubric
| Criterion | Points |
|---|---|
| Research question and design | 15 |
| Data quality and method justification | 20 |
| Descriptive tables and two plots | 15 |
| Correct method and computations | 20 |
| Interpretation, uncertainty and limits | 20 |
| Reproducibility and ethics | 10 |
| Total | 100 |
This rubric is an original teaching aid, not official university grading policy. Instructor solutions are kept in the private Divan records.