Normality, Post-Hoc, and Non-Parametric Tests

Question 1

Durbin-Watson Statistic

Answer 1

Used in linear and multiple regressions
Should range between 1.5-2.5
Checks independent observations

Question 2

VIF

Answer 2

Variance Inflation Factor
Ideal value is less than 10
Tests multicollinearity

Question 3

Levene’s Test

Answer 3

p = >0.05
Homogeneity of variances

Question 4

Shapiro-Wilk Test

Answer 4

ANCOVA’s:
p = > 0.05
Residuals normally distributed for each category of the IV
ANOVAS:
p = > 0.05
DV is normally distributed for each category of the IV
T-Tests:
p = > 0.05
DV is normally distributed

Question 5

Mauchly’s Test

Answer 5

p = > 0.05
Sphericity

Question 6

Wilcoxon rank-sum sign test

Answer 6

Non-para version of independent/unrelated (between-subjects) t-test
Rank non-normally distributed data regardless of group, then significance test (does one of the groups have more of the higher ranking scores than the other?)

Question 7

Mann-Whitney U Test

Answer 7

Non-para version of independent/unrelated (between-subjects) t-test

Question 8

Kruskall-Wallis Test

Answer 8

(for trends)
Non-para version of 1 way ind ANOVA

Question 9

Friedman Test/Friedmans ANOVA

Answer 9

Non-para version of one-way repeated (within-subjects) ANOVA

Question 10

Fisher’s Exact Test

Answer 10

Used when a Chi-Square test has a small sample size and expected frequencies are lower than 5

Question 11

Wilcoxon signed-rank test

Answer 11

Non-para version of repeated measures/related (within-subjects) t-test

Question 12

Jonckheere-Terpstra

Answer 12

Non-para version of 1 way ind ANOVA

Question 13

Loglinear analysis (cat outcome and participants as a factor)

Answer 13

Non-para version of multi-way between or within-subjects ANOVA

Question 14

Spearman’s Rho

Answer 14

Non-para version of correlation

Question 15

Kenall’s Tau

Answer 15

Non-para version of correlation

Question 16

Greenhouse-Geisser

Answer 16

Corrects for a lack of sphericity (which would have been tested with Mauchly’s Test)

Question 17

Huyn-Feldt

Answer 17

Corrects for a lack of sphericity (which would have been tested with Mauchly’s Test)

Question 18

Lower-Bound estimates

Answer 18

Corrects for a lack of sphericity (which would have been tested with Mauchly’s Test)

Question 19

Games-Howell

Answer 19

Post-Hoc Test
Used when Levene’s test of homogeneity of variance is significant

Question 20

Gabriel

Answer 20

Post-Hoc Test
Levene’s test of homogeneity is not significant, there are slightly different different numbers of participants in each condition/level

Question 21

Hochberg’s GT2

Answer 21

Post-Hoc Test
Levene’s test of homogeneity is not significant, there are very different numbers of participants in each condition/level

Question 22

Bonferroni

Answer 22

Post-Hoc Test
Levene’s test of homogeneity is not significant, there are NOT different numbers of participants in each condition/level, a few pairwise comparisons will be made

Question 23

Tukey

Answer 23

Post-Hoc Test
Levene’s test of homogeneity is not significant, there are NOT different numbers of participants in each condition/level, a many pairwise comparisons will be made