Given that there were no a priori hypotheses with regard to emotional eating and dietary restraint,
α was adjusted to .008 (for analyses on expectancy and desire) or .016 (for all other analyses) to correct for the number of interactions tested among these variables.