Hi!
I have data from an experiment in which there are 2 groups (n = 8 and n = 4), and for each animal in those groups, I have collected data at 15 time points (“pairings”). I am analyzing this data using a repeated-measures GLM, with the between-subjects factor = group (simple contrast) and the within-subjects factor = pairing (polynomial contrast). (I’ve attached the output from analyzing 1 of my dependent variables this way.) Note that the Multivariate Test results couldn’t be produced, and that Mauchly’s Test of Sphericity is missing the values for Approx. Chi-Square and Significance.
I’ve run the same type of analysis after dividing my data into 7 “bins” (blocks of 2 days) and discarded one anomalous pairing (the same pairing across all rats, nothing shady here). I have also attached an output from this analysis. The values which were missing from the other output (with 15 levels of the “within” variable) are now given.
First, why do I get the Multivariate test and Mauchly’s values only for the “binned” data? I understand that it has something to do with degrees of freedom, but I would like to be able to answer this question more specifically, and with references, if possible.
Second, when the Multivariate and Mauchly’s values are not produced, which values should I report for the within-subjects effects? I have been going by the Epsilon values; if epsilon is close to 1, I use the “sphericity assumed” values; otherwise I have been reporting Greenhouse-Geisser. Is this correct?
2012-06-10_CocaineGapes.spv2012-06-12_CocaineGapesBin2.spv
Locked
