Methodological artifacts in moderated multiple regression and their effects on statistical power. | Synapse