Dear Friends,
I have just performed a repeated measures anova (T0, T1, T2) and asked for a post hoc analysis.
Something went wrong in the post hoc, all "SE" were reported with the same value.
Just because it looked strange to me I performed the same analysis with Jasp and R. The results were different .
Please find attached a screenshot of the results and the file with the data.
Very best regards
Giuseppe
[SOLVED] anova repeated measures - post hoc problems
[SOLVED] anova repeated measures - post hoc problems
- Attachments
-
- AttPhys.csv
- (789 Bytes) Downloaded 528 times
Re: anova repeated measures - post hoc problems
Hi Giuseppe,
How did you calculate the post-hoc tests in R? We use lsmeans to calculate the post-hoc tests which uses the pooled variance from all groups that are being compared. Therefore, all standard errors have the same value. Using the pooled variance has some benefits such as a higher power, and it's a bit more robust when there are only a few counts in certain cells. I think JASP is just doing multiple t-tests and thus is using the variance of the difference between the two groups that are being compared.
Cheers,
Ravi
How did you calculate the post-hoc tests in R? We use lsmeans to calculate the post-hoc tests which uses the pooled variance from all groups that are being compared. Therefore, all standard errors have the same value. Using the pooled variance has some benefits such as a higher power, and it's a bit more robust when there are only a few counts in certain cells. I think JASP is just doing multiple t-tests and thus is using the variance of the difference between the two groups that are being compared.
Cheers,
Ravi
Re: anova repeated measures - post hoc problems
Hi Ravi,
many thanks for your clear reply, I apologize for thinking it was a mistake. Generally I use the multcomp package in R, but i must confess that for the example above I used R Commander with some of the Rcmdr plugins. The post hoc test are indeed based on multiple t-test with the bonferroni correction.
Thanks again for your time.
Cheers,
Giuseppe
many thanks for your clear reply, I apologize for thinking it was a mistake. Generally I use the multcomp package in R, but i must confess that for the example above I used R Commander with some of the Rcmdr plugins. The post hoc test are indeed based on multiple t-test with the bonferroni correction.
Thanks again for your time.
Cheers,
Giuseppe
Re: anova repeated measures - post hoc problems
No problem, happy to help