Help on Repeated Measures - Multiple Comparisons

[#1]

Trying to assess differences within each time point though can't figure out how.  New JMP user, using non-Pro version.  Response variable is colonization rate over 30 and 60 days, the factors are site (3) and treatment (3).  I am using the Manova > Repeated Measures analysis.  What I would like to look at is for each site, does the treatment have an effect on the colonization rate for 30 days and then for 60 days (i.e. treatments within sites for each time point).  There is a good tutorial for Mixed Models and multiple comparisons but the non-Pro version of JMP does not have Mixed Models.     

[#2]

Fit Model can handle this no problem. You just need an ID column to uniquely identify the unit that is measured at 30 and 60 days and include that ID column as a random effect. The only limitation is you can’t specify a correlation structure for the covariance matrix

[#3]

Thanks for the response.  However, when constructing the model effects, I add the unique identifier with its attribute as a random effect but I get an error that says "Random effects are not supported by Manova". 

[#4]

Oh, right. Base JMP doesn’t have a lot of support for mixed models. You can basically do a mixed model in Fit Model with a Standard Least Squares personality. You mentioned only 1 response variable. MANOVA should be used with multiple dependent variables. If you really only have one response, you should be doing standard least squares anyway.

[#5]

Hi @janiakd,

It sounds like you have your repeated responses across columns (wide format), but in order to fit a mixed model with Fit Model you'll need your data stacked, or in "tall" or "long" format. Using Tables > Stack you should be able to get your data in the right format to fit your repeated measures model. I'm happy to give more detailed steps if you need them. Also, you might find value in the Repeated measures ANOVA Add-In, which can help you set up the model effects (and also includes additional details on data format).

@julian

[#6]

This morning I spent some time with the add-in and and figured out how to do the analysis.  I initially did have the response data in multiple columns for the multiple times points, hence the manova/repeated measures, but this was much more straighforward.  The add-in works great especially with all the fun Effect Details in the reports section.  Thanks!! 

[#7]

Fantastic, I'm so glad to hear it was helpful!

@julian