|
|
hi everyone,
exploring my data, i came across two different ways of performing FRIEDMAN
Repeated Measures ANOVA (by SPSS for Windows, 11.5.1, on the one hand, and
by WinPEPI's module "PAIRSetc", part I-1).
It seems strange to me that for some of the subgroups of data, the two
programs' FRIEDMAN results differ from each other.
I know that in SPSS the significance level has to be chosen by the user
(i.e., either asymptotic, or Monte-Carlo, or exact test), and I have so far
been doing FRIEDMAN's test using Monte-Carlo simulation with default
configurations (C.I. 99% and 10,000 samples).
I'm now wondering what might be the significance level that WinPEPI uses
for doing the test, and if maybe the different significance levels might
cause different FRIEDMAN results?
sonja
|
Locked
