The effects of response contingent probabilities which favor response change.

Probability learning of interpersonal events1

“…In the smiplest, subjects were instructed to predict the occurrence of one of two mutually exclusive events (El or E2) over a senes of tnals, with either Ei or E2 occuning after each prediction The relative frequency of Ei and E2 were predetermmed and therefore noncontmgent on the S's responses The results of studies m this category have been generally m support of the "probability matchmg hypothesis " ) have pomted out deviations m obtamed results from the hypothesis.A second category of studies has employed the same basic procedures and m addition has mtroduced a stimulus prior to the S's prediction on each tnal which served as a partially vahd cue to either Ei or E2 Vanous event signalmg stimuh have been utihzed notably, the event on the previous tnal (Anderson, i960), the response made on the previous trial (Shelly, 1958), or • …”

“…

This paper reports three expenments which attempt to assess the degree to which the "probability matching hypothesis" apphes to the prediction of mterpersonal events The large body of literature purportmg to demonstrate that Ss leam to adjust their predictions of events to match the actual probability of occurrence of the altemative events, has been restncted to such stunuh as the Hashing of one or another hght, or similar innocuous events It was felt that an mvestigation of whether comparable probabihty matcthmg behavior occurs m the mterpersonal reahn would have direct relevance to the processes mvolved m social learmng and mterpersonal perceptionPubhshed reports m this area can be grouped mto three related paradigms. In the smiplest, subjects were instructed to predict the occurrence of one of two mutually exclusive events (El or E2) over a senes of tnals, with either Ei or E2 occuning after each prediction The relative frequency of Ei and E2 were predetermmed and therefore noncontmgent on the S's responses The results of studies m this category have been generally m support of the "probability matchmg hypothesis " ) have pomted out deviations m obtamed results from the hypothesis.A second category of studies has employed the same basic procedures and m addition has mtroduced a stimulus prior to the S's prediction on each tnal which served as a partially vahd cue to either Ei or E2 Vanous event signalmg stimuh have been utihzed notably, the event on the previous tnal (Anderson, i960), the response made on the previous trial (Shelly, 1958), or •

…”

“…A second category of studies has employed the same basic procedures and m addition has mtroduced a stimulus prior to the S's prediction on each tnal which served as a partially vahd cue to either Ei or E2 Vanous event signalmg stimuh have been utihzed notably, the event on the previous tnal (Anderson, i960), the response made on the previous trial (Shelly, 1958), or •…”