1

1) enter image description hereIn a Mixed-ANOVA design, with one-between subjects (2 levels) and one within-subjects test (2 levels), do we normally check the Levene's output to ensure that we are not violating the assumption of homogeneity of variance in a Mixed-ANOVA? In this case, Mauchly's Test of Sphericity is not possible with 2 levels each.

2) Another way to verify that we are not violating the assumption of homogeneity of variances is to check if the ratio of the largest variance is not more than 4 times the smallest variance. Does this rule apply in a mixed-ANOVA design test, with a between-subjects and one within subjects test? For example, referring to the attached Table, in the case of a Mixed-ANOVA design, do I need to check the ratio of D/C and B/A (as they are two different groups)?

kjetil b halvorsen
  • 63,378
  • 26
  • 142
  • 467
Vyas
  • 309
  • 1
  • 4
  • 11
  • 2
    Not an answer, but if you consider using linear regression with robust standard errors, the assumptions are moot. See [this question](https://stats.stackexchange.com/questions/131401/how-to-get-anova-table-with-robust-standard-errors) for some discussion. – AdamO Jul 25 '17 at 17:36

0 Answers0