the effect that occurs when a third variable changes the nature of the relationship between a predictor and an outcome, particularly in analyses such as multiple regression. For example, structural equation modeling can be used to assess whether a predicted association between quantitative skill and performance fits equally well across different teaching style groups (e.g., lecture based vs. hands-on learning). If the prediction is different across the two groups, then teaching style is said to have produced a moderating effect. Also called moderator effect.