305 final

Question 1

Scheffé’s

Answer 1

post-hoc test (more conservative, less power than Tukey’s)

Question 2

Tukey’s HSD

Answer 2

post-hoc test (more power than Scheffé’s)

Question 3

Bonferroni

Answer 3

post hoc, adjusting Type I error rate and critical value (dividing by number of tests)

Question 4

Sidak

Answer 4

post hoc, less conservative than Bonferroni correction, also adjusts Type I error rate

Question 5

Dunnet

Answer 5

post-hoc test, comparing 1 group to the other k-1 groups

Question 6

Holm

Answer 6

post hoc, sequential mean comparisons using Bonferroni correction

Question 7

Fisher-Hayter

Answer 7

post hoc, starts with largest mean difference and keep going until H0 is retained
using Qcrit with df = k-1 (less conservative than Tukey)

Question 8

Newman-Keuls

Answer 8

post hoc, starts with largest mean difference and keep going down until H0 is retained
minimum absolute difference is re-calculated for every comparison

Question 9

Duncan

Answer 9

post hoc, starts with largest mean difference and keep going down until H0 is retained
minimum absolute difference is re-calculated for every comparison (same as Newman)
uses Sidak’s Fcrit

Question 10

post hoc tests

Answer 10

1. Scheffé
2. Tukey’s HSD
3. Bonferroni correction
4. Sidak’s correction
5. Dunnet
6. Holm
7. Fisher-Hayter
8. Newman-Keuls
9. Duncan

Question 11

tests for normality

Answer 11

• test for skewness
• Kolmogorov-Smirnov (quantiles)
• Shapiro-Wilk (quantiles)
• Q-Q plots
• histograms

Question 12

tests for homoscedasticity

Answer 12

• Hartley’s F-max
• Levene’s test (ANOVA on deviation scores from group means)
• Brown-Forsythe (ANOVA on deviations from the group medians)

Question 13

ways to correct a violation of homoscedasticity

Answer 13

• run anova on sqrt(outcomes) - weak
• run ANOVA on log(outcomes) - mild
• run ANOVA on 1/outcomes - strong
• Box-Cox transformation
• non-parametric tests (Kruskal-Wallis test uses medians)

Question 14

Effects of a mixed design

Answer 14

1. between-subjects (main effect of A - subject to assumption of homoscedasticity, normality, independence)
2. subject variation within levels of A (residual for between-subjects)
3. within-subjects factor (main effect of B - subject to assumption of sphericity and normality)
4. interaction effect between A and B
5. interaction between within-subject factor B and subjects nested within levels of between-subjects factor A (residual for main effect of B and interaction)

Question 15

effect of violating normality

Answer 15

decrease in Type I error rate than nominal (less power)

Question 16

effect of violating homoscedasticity

Answer 16

increase in nominal Type I error rate

Question 17

define residual variance in between-subjects ANOVA

Answer 17

variance within groups (random fluctuations in subject scores)

Question 18

orthogonal comparisons

Answer 18

independent portions of variance due to group membership (limited number of comparisons k-1, once all are computed = model variation SSm)

Question 19

non-orthogonal comparisons

Answer 19

could deal with overlapping pieces of the model variation, so could amount to more than SSm

Question 20

follow-up to a two-way ANOVA with no significant interaction

Answer 20

comparison of marginal means

Question 21

follow-up to two-way ANOVA with a significant interaction that is dominated by main effects

Answer 21

comparison of marginal means

Question 22

follow-up to two-way ANOVA with a significant interaction that dominates main effects

Answer 22

simple main effects
if those are significant, follow them up with simple comparisons if you have more than two levels (i.e. directionality not obvious based on cell means)

Question 23

relationship between Type I error and power

Answer 23

increasing nominal Type I error rate (less conservative alpha) = increasing type I error rates and increasing power
- decreasing nominal Type I error rate (more conservative alpha) = decreasing type I error rates and decreasing power

Question 24

assumptions of one-way repeated-measures

Answer 24

1. normality of DV in the population
2. homogeneity of variances
3. homogeneity of covariances
   assumptions 2&3 are compound symmetry - assumption of sphericity

Question 25

assumptions for between-subjects ANOVA

Answer 25

1. independence of observations
2. homogeneity of variances across levels of the IV
3. normality of DV in the population

Question 26

Keppel & Wickens recommendations

Answer 26

1. choose factor with the most levels (fewer effects to compute) - factor with most levels at each level of the other factor
2. choose the quantitative factor
3. choose the factor with the greatest SS for the main effect (accounts for more variability)
4. choose the experimentally-manipulated factor

Question 27

controlling alphas familywise when main effects are significant

Answer 27

when only the main effects of an omnibus ANOVA are significant, the main effect comparisons count as separate analyses (we set alpha familywise to .05, then use the Bonferroni correction for each individual comparison)

Question 28

controlling alpha familywise when interactions are significant

Answer 28

simple main effects: set alpha familywise to .10 (simple main effects include the variation due to main effects and the interaction)
simple comparisons: use the same alpha familywise as for simple main effects, then use the Bonferroni correction to calculate alpha for each individual comparison

Question 29

Kolmogorov-Smirnov test

Answer 29

for normality (comparing quantiles to a reference distribution)

Question 30

Shapiro-Wilk test

Answer 30

for normality (comparing quantiles to a reference distribution)

Question 31

Hartley’s test

Answer 31

for homoscedasticity (s2 (largest) / s2 (smallest))

Question 32

Levene’s test

Answer 32

for homoscedasticity (ANOVA on deviations from group means)

Question 33

Brown-Forsythe test

Answer 33

for homoscedasticity (ANOVA on deviations from group medians)