0
$\begingroup$

I am hoping that I can get some advice on the best, or most appropriate, way to analyze a specific type of data.

I have collected data on brain volumes across two different species of mice, A and B. I want to compare subregions within the brain and know that one species is just bigger than the other so have larger brains in general. But, if I divide each subregion of the brain by total brain volume, I get proportional values. If I want to compare those values across the two groups, can I use a t-test? Obviously these aren't normally distributed, which is often claimed to not be a huge issue. The values will also be very similar in general within and between groups, and of course, the values across all different subregions within one subject will be correlated as they are a proportion of the total. I can't conceptualize if there are other issues, however. Any thoughts or advice would be greatly appreciated.

Here is an example of the data for one subregion, for example (though in my dataset I have about 100 samples in each group, I just didn't type out 100 rows):

A B
.01 .008
.009 .012
.012 .0075
Improve this question
$\endgroup$
0

1 Answer 1

Reset to default
0
$\begingroup$

Typically in neuroimaging studies of human brain size one conducts a regression and includes total brain volume or intracranial volume as a covariate. I think doing the same here would be appropriate.

Improve this answer
$\endgroup$
2
  • $\begingroup$ Completely agree with you! I suggested that, but was told to do it this way instead but will try to push for that. Thank you for the advice! $\endgroup$
    – kyro1021
    Commented Feb 2, 2023 at 16:24
  • $\begingroup$ Just to follow-up, it isn't an issue per se that values aren't normally distributed. The assumption of any regression (including a t-test) is that residuals are normally distributed. This is something you can check after you run analyses. There are other issues with using proportions that you should be aware of: stats.stackexchange.com/questions/103731/… $\endgroup$
    – David B
    Commented Feb 2, 2023 at 16:37

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

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