Manova Flashcards
what is manova
multivariate analysis of variance
when is this used
used when have one IV and many dependent variables
give an example of this type of anova
effect of pet ownership on mental health (IV) mental health cons= depression, anxiety ocd etc. (many DV’s)
what is the issue with having lots of DV’s in an experimental design
increase chance of type 1 errors (false positives)
what do a manova do
assess whether IV has effects on variables taken together
If there is an effect on variables together what do you do next
look at the second part of MANOVA, which aspects of mental health does it specifically effect- analyses individual effects
what are the assumptions of a manova
usual anova assumptions
iv- always between subjects
Levenes for homogeneity calculate each DV (non-signif wanted)
box test- tests equal in covariances matrices
homogeneity tests
explain the two homogeneity of variance test manovas produce
- box test - equality of covariances in matrices, (one test related to all. not signif = p value taken from multivariate box in output is reliable
- levene’s test for equality of error variances (homogeneity) (sep tests for each DV)
want it to be non signif p > .05 - not sig = each p value for each Dv taken from between subs contrast box = reliable
explain what to do with the output of an anova
gives multivariate test- asses whether IV has effect on DV’s (overall)
gives- F value, degrees of freedom and p value (like anovas)
what 4 other measures does the output give
- Pillai’s Trace
- Wilks Lambda
- Hotelling’s Trace
- Roy’s Largest Root
w/ 2 levels of IV doesn’t matter which is used f value = identical
what happens when you add another level of the IV
depends on data (variances etc) which post hoc test is used