I'm really bad at stats and am running an experiment looking at growth rate or individuals over the period of 28 days. I collected data on Day 0, 7, 14, 21, and 28 so I have 5 data points. I would like to compare the two to the control group to see if they are statistically different and then compare the two different treatment groups to see if those are statistically different. To clarify, I would really like to look at the rate of growth. What is the best way to organize my data (do I need to subtract subsequent weeks?) and which type of stats tests should I run (one-way ANOVA or two-way i literally have no clue)? Any specifics would be great.

PS-I have access to and some familiarity with JMP and I have StatPlus for XCEL

  • 1
    $\begingroup$ The methods that are appropriate here are fairly advanced. Given your evident level of statistical sophistication, you should probably work with a statistical consultant. Such services are often available through your university. $\endgroup$ – gung - Reinstate Monica Sep 16 '15 at 20:03

When you collect data on the same subjects repeatedly you violate the assumption of independent errors that is part of both one-way and two-way ANOVA. There are various methods for dealing with this including

1) Repeated measures ANOVA (but this makes assumptions that are usually not tenable)

2) Multilevel models

3) Generalized estimating equations.

As @gung notes, these methods are fairly advanced and you probably will need help.

Also ... for the future: The time to figure out how you will analyze your data (or, at least, how you plan to analyze your data) is before you start collecting it.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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