1

I am trying to analyse data of 2 different groups (control vs experimental, brain damage) and their performance on 3 different cognitive tests (the results of which are not correlated) to determine the effect of brain damage on different cognitive funcitons.

I have 2 questions

  1. Should I use a mixed ANOVA or a MANOVA or even some other analysis?
  2. Can uncorrelated measures (the cognitive tests) be assigned as levels to one DV as they would be if I used a mixed anova design which assumes a single DV?
ttnphns
  • 51,648
  • 40
  • 253
  • 462
Katie
  • 11
  • 2
  • Are the three cognitive functions comparable theoretically and/or by measurement units? I mean, can there be seen as "repeated measures" (factor "cognitive function" with 3 levels)? If yes, you could analyse it as multivariate setting or as split-plot univariate setting (using mixed linear model). See e.g. [1](http://stats.stackexchange.com/q/13197/3277), [2](http://stats.stackexchange.com/q/17378/3277). – ttnphns Dec 08 '15 at 19:50
  • See also [this](http://stats.stackexchange.com/q/185115/3277) recent Q about "long" and "wide" dataset formats. – ttnphns Dec 08 '15 at 19:52
  • In your question #2, did you mean to ask if the cognitive tests could be treated as levels of one *IV* (as opposed to DV)? If not, perhaps you can clarify what exactly you mean in this question. – amoeba Dec 08 '15 at 22:51
  • The three cognitive tests simply test participants scores on maths, reasoning and language skills but the I am primarily interested in comparing the control group vs the brain damaged group on their cognitive functions – Katie Dec 22 '15 at 17:05
  • Sorry yes if the IV could be treated as the test then with 3 levels of different test or if each test type could be treated as a separate IV – Katie Dec 22 '15 at 17:06

0 Answers0