multivariate designs lec 5 Flashcards
what is a univariate design
one DV
what is a multivariate design
more than one DV
what are the limitations of a univariate design
might be other factors, might be multifaceted
what is an example of two DVS
bmi and bis
what is the problem with doing 2 ANOVAS for this type of design
the two dvs are likely to be correlated and two of them leads to an increased chance for type 1 error
what is then used for an analysis
MANOVA
what does a one-factor independent MANOVA do
combines the two new dvs to do one initial univariate test
what is a covariate
the two dvs added together
why do the two DVS need to be closely related to the same concept
to say that the overall ‘compotite’ has an effect on the experimental study (eg treatment)
what do we use to report this from SPSS
wilkas lamba
what happens to a non significant MANOVA, and how is it writen up
no effect on any if the DVS so written up like
design type and factors
composite dv, how this
what did MANOVA reveal
state the wilkas lamba
what does the independant MANOA suggest if the result is significant result
suggests the composite has an effect on at least one of the DVS
what occurs after a significant MANOVA
split the data and do univariate MANOVA for the two DVS but use bonferroni correction
what does MANOVA have to meet assumption wise to be trusted
homogeneity of covariance
what is homogeneity of covariance
this is the correlation between the DVS being the same in each experimental group, if not it is violated and we can’t trust it
