0
$\begingroup$

I am struggling on which type of ANOVA to use for my work, I will provide as many details below about the study design, at the moment I am thinking mixed ANOVA but im encountering some problems which suggest that I have chosen the wrong type.

I have two groups, one with a particular defect and one without, both hold 40 participants. Each group had to take part in three tests, each relating to a particular intellect, i.e. logical ability. Each participant completed the same tests and therefore each participant had three different scores to their name. The aim is to determine whether the defects led to cognitive functioning problems, as determined by the three test scores. To do this I need to perform an ANOVA, but am uncertain which type.

Since there was an independent sample (two separate groups), but they all took part in the same three tests, would this be a mixed design? Or should I perhaps be using a two way ANOVA?

$\endgroup$
1
$\begingroup$

Classically, this is a multivariate analysis of variance design and would be fit via MANOVA. However, it is very often easier and better to fit this design as a mixed model. Your data will be coded as (for example):

Subject   Group   Test   Score
1001      1       1      34
1001      1       2      55
1001      1       3      36
1002      2       1      23
...

You would fit the model: Score = Test Group Test*Group and would specify a repeated factor with the subject and allow a unstructure variance model.

$\endgroup$
  • $\begingroup$ @JakeP If the provided answer is useful for you consider upvoting it and if it addresses you question adequately consider accepting. That way other users can see how StatNoodle gets love all around. :) (+1 to this post by the way, I think using a mixed model is a very good suggestion.) $\endgroup$ – usεr11852 Jan 5 '16 at 0:36

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.