 |
 |
Dear all, I have a data set comprising of 36 cases....each case has been diagnosed as being language delayed (LD) or language normal (LN) at both age 2 and 3 years. At both ages, 6 of the cases have been diagnosed as LD and 30 have been diagnosed as LN...however, these are *not* the same 6 & 30 cases at both ages - there is some movement between the two diagnostic groups. For each case (at both ages), 7 'performance' measures have been recorded. Ultimately, I wish to determine if there is a difference in 'performance' between the 2 diagnostic groups (LN & LD); if age is a significant effect and if there is an interaction between the factors. At first glance it does look as though the mixed design ANOVA would be appropriate with the between subjects factor being 'diagnosis' and the within subjects factor being 'age'.....I could then look at *each* of the performance measures in turn....doing 7 separate mixed design ANOVAs.... BUT there is a problem as there are not the same children in each diagnostic group at ages 2 and 3. As I understand, for this method, the researcher should select 2 samples of subjects (LD & LN) and perform some repeated measures experiment on each group.... .between subjects factors subdivide the sample into discrete subgroups. Each subject has only one value for a between subjects factor (i.e. either LN OR LN). Also, in the texts I've been reading there always seems to be the same number of individuals in each of the 'between subjects' factor's levels. I'm not sure whether this has to be the case all of the time (?). My guess is that it would be best to look at the 7 recorded performance measures *together* but I'm not sure how. Any views? Thanks in advance for any suggestions, Kim. [log in to unmask] %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%