Mixed ANOVA

Question 1

between-subjects

Answer 1

independent measures

Question 2

within-subjects

Answer 2

repeated measures

Question 3

what is mixed anova and what does it test?

Answer 3

1+ w-s factor and 1+ b-s factor

tests a main effect for each factor and an interactive effect of the factors

Question 4

do you check homogeneity of variances of sphericity in mixed anova?

Answer 4

both

Question 5

when is sphericity assumption violated?

Answer 5

when differenced aren’t equal (sig. value)

Question 6

Box’s test of equality of covariance

Answer 6

homogeneity of intercorrelations (for each level of w-s should be same)

box’s M statistic

violated if p<.001 (we want the stat to be non-sig. >.001)

unequal sample sizes

Question 7

SPSS outputs

Answer 7

• w-s factors (factor levels)
• b-s factors (factor levels)
• descriptive stats (means and SDs)
• Box’s test of equality of covariance matrices (want a non-sig stat)
• Mauchly’s test of sphericity (want a non-sig stat)
• tests of w-s effects (ANOVA stats for main effects/interaction)
• Levene’s test of equality of variances (want non-sign stat)
• tests of b-s effects (df, f ratio etc)
• estimated marginal means (main effects means)
• profile plots (non-parallel = sig. interaction)

Question 8

formally reporting ANOVA

Answer 8

• anova type, effect of factor on DV
• means and SDs in table or text
• assumptions (sphericity, homogeneity of variance, homogeneity of covariances)
• anova results (df, f ratio and p value)
• effect size and meaning for all main effects and interactions
• report comparisons

interpret

Question 9

follow up steps for mixed anova

Answer 9

(un)planned comparisons when 3+ factor levels

unplanned = no hypothesis - Bonferroni
planned = hypothesis - paired/independent samples t-test/planned comparisons