I want to find the impact of both perpetration of abusive violence and observation of abusive violence on PTSD symptoms, controlling for combat exposure, age, and clinical status. Do I have to apply hierarchical regression analysis?
If I were doing this analysis, I'd enter combat exposure, age, and clinical status as predictors in the first step of a regression, then add your other two predictors at a second step. That allows you to see how much variance your two predictors of interest account for (R-squared change) after you have taken into account the variance already accounted for by your control variables). You'll also be able to find out whether both or only one of your predictors of interest accounts for unique variance (whether the predictor itself is significant).
See this link for additional help:
https://www.youtube.com/watch?v=xgA8qY63dX0
Simon
No, this is a simple multiple regression, where combat exposure, age, and clinical status enter as additional regressors/predictors. If you want to have p-values for the effects of the target predictors make sure to use type-III sum-of-squares (best option is to compare the model[s] with and withou the predictor[s] of interest).
- Badges
- Science topic