Hi List,

I'm having some methodological challenges that I hope you can help me out with. I'm researching the difference in how children with academic parents adjust to the transition to high school compared to children with non-academic parents. I have had students fill out surveys before and after this transition. My intention is to look at the change in motivation relative to pre-transition levels for the two groups. Additionally, I want to carry out a mediation analysis to assess whether choice of track, self-efficacy and parental support mediates the effect of having academic parents.


After having read up on the subejct it seems I have three choices; difference-scores, ancova (incorporate pre-transition level of motivation as covariate in regression of motivation after transition), and residualized gain scores. Residualized gain scores are calculated by running a linear regression with the follow up score as the outcome and the baseline score as the predictor and saving the residual values. These values are then used as dependent variables in the analysis. It seems that the two first approaches are widely critized as they can yield differen results and are prone to regression towards the mean.

This leads me to consider residualized gain scores. However, I am very uncertain as to what unit this variable is measured in and ultimately how to interpret it? And can I trust postive/negative signs infront of the coeficients to mean growth and decline?



Technical: How does one calculate residualized gain scores in Stata?
simply?:
Code:
 reg Motivation_Time2  Motivation_Time1
predict New_var, residuals



Best,
Mads