0

I got trouble on a work problem, which I need to design an experiment that can compare different treatment means. I'm gonna use one way anova with 6 levels which are tel_check, applicant_check, and several other checks. I want to find all this checks single or combination work and find the best check combination that works well.

However the question is my dependent variable is count data that is not continuous which violates the assumption of anova. Also, I don't figure out alternatives if I can't use one way anova here. If one way anova can work, how can I estimate a rough value for the sample size that can my plan work?

I asked a lot of questions, sorry for that, but I didn't find suitable answers for my questions on the Internet.

Zhang Moyi
  • 1
  • since it's more stats than programming, you should probably ask the stats folks at http://stats.stackexchange.com – Rorschach Oct 16 '15 at 03:34
  • There are several methods to properly analyze count data. R supports many applications. You need to post example data and at least show what you want to estimate using tabular methods. – IRTFM Oct 16 '15 at 04:21

0 Answers0