1. organize data in HLM format

2. regress Cov (one long column) on DV (one long column) and save unstandardized residuals

3. add constant (estimated in regression) to unstandardized residuals

4. can now re-organize data into repeated-measures format and estimate marginal descriptive stats and compute pairwise tests w/ one caveat -- you must subtract 1 df from the error term and re-compute the p-values for analyses performed on the 'corrected' data