 |
 |
Suppose that I am reviewing a paper that has necessary design multicollinearity (e.g., Cue always precedes response), but wants to make explicit claims about one portion of the trial (Cue). Most methods sections do not report measures such as VIF. But, multicollinearity at the time-series level cannot be assessed at the group-level SPM (inflates first level standard errors, but does not bias point estimates). Obviously, multicollinearity measures like VIF can be requested--when is it appropriate to do this? (when a paper has plausible multicollinearity? Why not all the time?) What resolution of VIF should be requested? At the run level (points are estimated at this level)? What if subjects have different time series (should you ask for a VIF for each run, nested in each subject)? How does one aggregate VIFs across different levels of a hierarchical analysis, and how do you assess the effect of heterogeneous VIF measures within a level? Andrew Mattarella-Micke University of Chicago