QUESTIONS WITH SOLUTIONS GRADED A+
◉ The test statistic for testing a population proportion uses:
Answer: z distribution
◉ The normal approximation for proportion tests is appropriate
when: Answer: np ≥ 5 and n(1−p) ≥ 5
◉ The standard error for a one-sample proportion test uses:
Answer: Hypothesized p₀
◉ A pooled two-sample t-test requires: Answer: Equal population
variances
◉ If equal variances cannot be assumed, the correct test is: Answer:
Welch's t-test
◉ The degrees of freedom for Welch's test: Answer: Are calculated
by a Satterthwaite approximation
◉ If the 95% CI for μ₁ − μ₂ includes 0: Answer: Do not reject H₀
, ◉ A paired t-test analyzes: Answer: Differences between paired
observations
◉ A paired test is preferred over an independent samples test when:
Answer: Observations are naturally matched (before/after, twins)
◉ Paired tests generally have greater power than independent tests
because: Answer: They reduce between-subject variability
◉ The test statistic for difference in proportions uses: Answer:
Pooled standard error
◉ The pooled proportion for a two-sample test equals: Answer: (x₁ +
x₂) / (n₁ + n₂)
◉ A large difference in sample proportions will: Answer: Increase
power
◉ Testing a single population variance uses: Answer: χ² statistic
◉ Testing equality of two variances uses: Answer: F distribution
◉ Which assumption is required for F-tests for variances? Answer:
Normal populations