Instrumental Conditioning

A typical independent variable in instrumental conditioning procedures is the number of response-contingent events ( e.g., the number of food-reinforced responses).  Another independent variable that is often manipulated is a discriminative stimulus (i.e. the presence and absence of a discriminative stimulus).

Typical dependent variables:

Discrete trial procedures: response latency; response time; # of errors

Free-Operant procedures: cumulative responses; response rate

Contiguity theories of behavior-event learning

(Contingent relationships between behavior and events are needed to ensure contiguous pairings of the behavior and event; learning cause-effect relationships is not necessary)

  1. Thorndike: animal intelligence and the law of effect
  2. B.F. Skinner
      1. Operant conditioning and the principle of reinforcement
      2. superstition experiment (contiguity is sufficient for behavior-event learning)
        3. Staddon & Semmilhag's reinterpretation of Skinner's superstition experiment (event-event learning is a better explanation of the superstition experiment than behavior-event learning).

     

    Biological constraints on behavior-event learning

    e.g., Breland & Breland's Instinctive drift)

    More on Reinforcement theory

    1. Skinner's 3-term contingency:

    Discriminative stimulus S, operant (response; r), reinforcer (event; s*)

    2. What associations are learned during behavior-event learning

    R-S*, S-R,  S-S* and S(R-S*)

    The R-S* association is emphasized by cognitive and teleological explanations of instrumental conditioning.  Evidence of  R-S* associations come from the Reinforcer devaluation experiments conducted by Colwill and Rescorla.