Undermatching and overmatching as deviations from the matching law

Area of Science:

• Behavioral Psychology
• Operant Conditioning
• Animal Behavior

Background:

• The matching law accurately predicts behavior under concurrent reinforcement schedules.
• Deviations like undermatching and overmatching are observed but not fully explained by the simple matching law.

Purpose of the Study:

• To present a novel model of operant performance under concurrent variable-interval reinforcement schedules.
• To explain undermatching and overmatching as deviations from the matching law.

Main Methods:

• Developed a theoretical model based on the hypothetical "burst" structure of operant responding.
• Incorporated two opposing tendencies: random allocation of response bursts and prolongation of high-probability bursts.

Main Results:

• The model predicts undermatching is more common than overmatching.
• It also predicts that overmatching, when it occurs, is generally limited in extent.
• These predictions align with existing empirical data.

Conclusions:

• The proposed model accounts for undermatching and overmatching by adding specific processes to behavior allocation governed by the simple matching relation.
• This framework allows reinforcement rate or probability optimization to remain explanatory mechanisms even when the simple matching law is not strictly followed.