# Statistical test to analyse four variables over time across three groups?

#### Aith

##### New Member
Hi, I'm designing a study involving four measures of stress obtained six times across 12 weeks (baseline, day 2, week 3, 6, 9 and 12). It's based off of a similar study which used repeated measures ANOVA for statistical analysis. However, I'm additionally dividing the sample into three groups (differing environments) and therefore I believe I can't use the same test because across the groups the subjects are independent/unpaired?

I've tried to select a suitable test on my own but I know next to nothing about statistical tests (our statistics lessons don't begin until next term) and my efforts have mostly just confused me even more.

Thank you so much for taking the time to read this,
Aith.

#### Karabiner

##### TS Contributor
You could perform "mixed"MANOVA, i.e. a repeated-measures MANOVA (the 4 stress measured jointly represent the dependent variable,
which is measured 6 times), with an additional between-subjects factor "group".

With kind regards

Karabiner

#### Aith

##### New Member
You could perform "mixed"MANOVA, i.e. a repeated-measures MANOVA (the 4 stress measured jointly represent the dependent variable,
which is measured 6 times), with an additional between-subjects factor "group".

With kind regards

Karabiner
Thank you so much for the fast reply! Is it possible for four measures to jointly represent the dependent variable if they are measured on different scales (e.g. faecal cortisol in ng/g, weight in grams, and a 7-point behaviour-based stress score)? And do you know if it is possible to compute a repeated measures MANOVA in R?

Thanks again,
Aith.

#### fed2

##### Active Member
you should draw box plots, if it seems like they are different you can run a t-test

#### Karabiner

##### TS Contributor
Is it possible for four measures to jointly represent the dependent variable if they are measured on different scales (e.g. faecal cortisol in ng/g, weight in grams, and a 7-point behaviour-based stress score)?
My assumption was that the four measures should be analysed jointly
(they need not be measured on the same scale), but that was not intended?
If you want to analyse them separately, you can just perform several
mixed analyses of variance (grouping factor + repeated-measures factor).
And do you know if it is possible to compute a repeated measures MANOVA in R?
I bet.

With kind regards

Karabiner

#### Aith

##### New Member
My assumption was that the four measures should be analysed jointly
(they need not be measured on the same scale), but that was not intended?
If you want to analyse them separately, you can just perform several
mixed analyses of variance (grouping factor + repeated-measures factor).

I bet.

With kind regards

Karabiner
Thank you so much for your help, you've saved me from a real headache.

Best wishes,
Aith

#### fed2

##### Active Member
don't forget to draw a boxplot.