Statistics #5

We use Welch's F or Brown-Forsythe F if: a. Levene's test is significant b. Variances are not equal c. Observations are not normally distributed

A and B

*Which assumption(s) must be met to conduct an Analysis of Variance (ANOVA)? a. Normality b. Homogeneity of Variance c. Independence

All of these

Which of the following parameters needs to be included when reporting the results of ANOVA test a. F-statistics b. Degrees of freedom (df) c. P-value d. Effect size

All of these To see some useful examples on how to report the results of ANOVA test properly, see section 12.11 page 419

If we want to compare means of three separate groups using analysis of variance (ANOVA), a significant F means: a. All three sample means are significantly different b. The mean of groups 1 and 2 are similar to each other but different from that of group 3. c. The mean of groups 2 and 3 are similar to each other but different from that of group 1. d. The mean of groups 1 and 3 are similar to each other but different from that of group 2.

Any of these

True or False: Post-hoc tests consist of pairwise comparisons that are designed to compare all different combinations of groups means.

True

True or False: the F statistics (F-ratio) is calculated by dividing the model mean squares by the residual mean squares.

True

*All the following tests are post hoc tests EXCEPT a. Tukey b. Hochberg's GT2 c. Games-Howell d. REGWQ e. Gabriel's f. Welch's g. Benferroni

Welch's

When do we use REGWQ or Tukey post hoc test in ANOVA? a. If we want a guaranteed control over type 1 error rate b. When sample sizes of groups are different c. When we have equal sample sizes and similar group variances

c. When we have equal sample sizes and similar group variances