Failure to Replicate Matching in Conversations

In an attempt to replicate and extend a 1974 demonstration of matching in conversation by Conger & Killeen, nine German subjects, aged 20-63, each participated in three hours of discussion with two confederates offering different rates of verbal agreement. In condition A, approval was given contingent upon and contiguous with the subjects’ statements made while they had eye-contact with the approving confederate. In condition B, approval was uttered following a subject’s statement whenever scheduled and non-contingent upon eye-contact. Sessions were video-recorded and coded to obtain the duration of the subjects’ looking at and talking to each confederate, as well as the actual reinforcement rates delivered by each confederate. The equation for the generalized matching law was fitted to the data. Contrary to Conger and Killeen’s findings, data were not described well by the generalized matching law. What are possible reasons for the data not yielding slopes suggesting that the subjects’ behavior were sensitive to the approval delivered? In natural interactions, Germans do not express their approval as vividly as Americans do, which might have derogated the reinforcers’ effects. Additionally, continued interaction with the confederates may have precipitated biases as opinions were stated. Moreover, although statements of approval were standardized as much as possible, merely due to its sparseness, approval delivered by the confederate currently on the lean schedule might have had a larger reinforcement value than approval given by the confederate on the rich schedule. These possibilities are to be reviewed.