|
|
Multiple comparisons and inflated typ... |
|
Message/Author |
|
|
A reviewer has asked about protections for type 1 error in my analysis. I have run 6 separate path models examining the cross-lagged associations of two variables at three time points, and include several demographic covariates (regressed on time 1). I have found only 2 papers looking at multiple comparisons in SEM (Cribbie, 2007, and Smith and Cribbie, 2010) who suggest an False Discovery Rate (FDR) approach. I am wondering if 1) is multiple comparisons often a concern with these types of models? and 2) any suggestions for how to implement an FDR approach - that is, would it include only the alpha's for the paths, or all of the parameters? Thank you |
|
|
You want to ask this general analysis question on SEMNET. |
|
Back to top |
|
|