Which stat test to use?

IrishKnight

Which stat test should I use for the following experiment to test for stat significance?

I treated cells with one of five different drugs and counted how many cells were alive at 24hr, 48hr, and 72hrs. The only variable which drug as I treated the cells with the same contraction etc. I left one sample untreated as my control.

Would a stranded t-test be okay to use? Or would I have to use a non-parametric test?

Also, I'd like to normalise my data to the control i.e. taking the control to be 1 and seeing the relative changes depending on the drug. Can I still carry out stat tests on this data or should I carry it out on the raw data?

Any help would be great!

Killbot2000

I think you should use an ANOVA. A t-test is used for comparing two means but here you have more than two treatments.

Your dependent variable i.e. the thing you're interested in looking at is the cell count which is continuous. Your independent variables i.e. the ones that are impacting cell count are the drug treatment and time. In this case they are categorical variables too, for instance, you've created blocks of time instead of having it continuous.

this link may help:
http://www.stat.ufl.edu/~ssaha/3024/Chapter13withExamples.pdf

IrishKnight

Thanks for that!