A model that combines variables (cues) to form a prediction. It is compensatory because high values for some cues can compensate for low values in other cues. Adding and averaging are compensatory models.