Week 2 End Term MCQ's Flashcards
Q1. What is the best method for checking skewness in a large sample (e.g., N = 15,467)?
a) Calculate the Z-score and see if it’s beyond ±1.96
b) Use the Kolmogorov–Smirnov test
c) Check the p-value of the skewness statistic
d) Use visual inspection and absolute skew value
d
Q2. A Z-score for skewness is calculated as -3.56. What does this imply about the distribution?
a) The distribution is significantly positively skewed
b) The distribution is approximately normal
c) The distribution is significantly negatively skewed
d) Z-scores should not be used for this purpose
c
Q3. The Kolmogorov–Smirnov test produced a p-value of < .001 for a variable. What does this indicate?
a) The variable is normally distributed
b) The variable has equal variances
c) The variable significantly deviates from a normal distribution
d) The test is inconclusive for small samples
c
Q4. Levene’s Test for equality of variance gave p = .046 when comparing spelling scores by teacher. What does this mean?
a) Variances are equal and homogeneity is satisfied
b) Variances are not equal; use Welch’s t-test
c) The test is invalid due to small sample size
d) Levene’s test only applies to medians
b
Q5. Why is the Mann–Whitney U test sometimes used instead of an independent t-test?
a) It has higher power when data is normal
b) It is used for comparing categorical data
c) It does not assume equal variances or normality
d) It compares group variances directly
c
Q6. Which of the following is not an assumption of a parametric test?
a) Normal distribution of the DV
b) Homogeneity of variance
c) Independence of observations
d) Dependence of scores
d
Q7. What is the definition of a variance ratio?
a) One variance minus another
b) One variance multiplied by another
c) One variance divided by another
d) One mean divided by another
c
Q8. A variable is significantly skewed, but the sample size is small (N = 25). What is the most appropriate course of action?
a) Proceed with t-tests; skewness doesn’t matter
b) Transform the data or use a non-parametric test
c) Use Levene’s test instead
d) Ignore the issue if the histogram looks symmetrical
b
Q9. When is it not appropriate to use the Kolmogorov–Smirnov test?
a) When you have a small sample and want to test for normality
b) When testing for equality of variances
c) When comparing a distribution to a theoretical normal distribution
d) When checking assumption violations before a t-test
b
Q10. In SPSS, which menu path correctly produces skewness statistics?
a) Analyze > Compare Means > One-Sample T-Test
b) Analyze > Descriptive Statistics > Frequencies
c) Analyze > Descriptive Statistics > Descriptives (Options > Skewness)
d) Analyze > Descriptive Statistics > Explore
c
Q11. What does a Levene’s test p-value of .046 suggest?
a) There is homogeneity of variance
b) There is evidence of unequal group variances
c) There is a non-significant difference in group means
d) There is normality of the residuals
b
Q12. When is it appropriate to use Welch’s t-test instead of a standard t-test?
a) When the sample size is below 30
b) When variances between groups are significantly different
c) When data is nominal
d) When assumptions of linearity are violated
b
Q13. Which of the following is true of Mann–Whitney U test?
a) It compares medians directly
b) It assumes normality of the outcome variable
c) It tests whether two independent groups differ in distribution
d) It requires homogeneity of variance
c
Q14. What does a histogram with long left tail and Zskew = -3.56 indicate?
a) The data is symmetric
b) The data is positively skewed
c) The data is negatively skewed
d) There is kurtosis but not skew
c
Q15. Which transformation is least effective for correcting negative skew?
a) Square transformation
b) Log transformation
c) Reciprocal transformation
d) Cube transformation
b
Q16. What is the purpose of spread vs. level plots in SPSS?
a) To assess normality of residuals
b) To compare median scores across groups
c) To test homogeneity of variance visually
d) To determine whether variances correlate with means
d
Q17. Why is p = .000 never correct to report in APA or thesis work?
a) Because it’s more precise to report p = .0000
b) Because it means the test was inconclusive
c) Because p-values can’t actually be zero
d) Because it is only allowed in nonparametric tests
c
Q18. When is trimming or Windsorizing a valid correction method?
a) When dealing with missing values
b) When outliers severely distort the mean
c) When variances are unequal across groups
d) When group sample sizes differ
b
Q19. What happens if you forget to deselect ‘boxplots’ in the Explore function when testing K-S?
a) You get an error in SPSS
b) The output will contain unnecessary visual clutter
c) The K-S test is not performed
d) Levene’s test will run instead
b
Q20. What does a Mann–Whitney U result of U = 634.5, p = .206 imply?
a) The two groups differ significantly
b) The data violates normality
c) There is no significant difference between groups
d) The test is invalid due to sample size
c
