The purpose of the present study was to evaluate Herrnstein's law of effect as a description of socially significant behavior in an applied setting. The subject was an 18‐year‐old retarded girl with a history of autistic‐like and aggressive behavior. Using a baseline design for two response classes and stimulus conditions, eight categories of subject and staff behavior were monitored over a 3‐week period. A computerized observation system, developed for use in the present investigation, was used to obtain real‐time durations of the behavior categories alone and in combination. Overlapping durations of teacher and subject behavior were then correlated to yield approximations to a functional definition of reinforcement. Plots of behavior by contingent reinforcement revealed a hyperbolic relationship for each response class, the shape of which varied as a function of extraneous reinforcementr0. In addition, estimated parameters in Herrnstein's equation did not differ significantly from those obtained through independent observation. Finally, Herrnstein's equation accounted for an average 63% of variance in response allocation. Results are discussed in terms of the relevance of matching‐law theory to behavior in appli