Reed Phil
Swansea University.
J Exp Psychol Anim Learn Cogn. 2015 Oct;41(4):419-31. doi: 10.1037/xan0000081. Epub 2015 Aug 31.
Four experiments explored the rate and structure of human responding on random ratio (RR) schedules of reinforcement, using 3 different methods of analysis. All experiments demonstrated that responding on RR schedules was higher with intermediate ratio values-rates being higher on an RR-30 schedule than on either a RR-10 or RR-60 schedule. All experiments identified 2 distinct types of responding generated by RR schedules: "response-initiation" responding and "within-burst" responding. In contrast to overall rates, response-initiation responding was greatest on RR-10 schedules, and less pronounced as the ratio value increased, but response rates did not vary greatly within the burst (being higher on longer ratio schedules when there was a difference). These findings are consistent with previously obtained data from nonhumans, and suggest human schedule performance should be considered as a joint product of 2 different aspects of responding: response-initiation responding which is sensitive to rates of reinforcement; and within-burst responding which is less sensitive to the molar characteristics of the schedule in operation. The presence of 2 distinct forms of responding may require reinterpretation of many human schedule-controlled phenomena that rely on assumptions about the similarity of all responses to one another.
四项实验运用三种不同的分析方法,探究了人类在随机比率(RR)强化程序下的反应速率及结构。所有实验均表明,在RR程序中,中等比率值时的反应率更高——RR-30程序下的反应率高于RR-10或RR-60程序。所有实验都识别出RR程序产生的两种不同类型的反应:“反应起始”反应和“爆发内”反应。与总体反应率不同,“反应起始”反应在RR-10程序中最为强烈,且随着比率值增加而减弱,但在爆发内反应率变化不大(当存在差异时,较长比率程序下的反应率更高)。这些发现与先前从非人类实验中获得的数据一致,并表明人类程序表现应被视为两种不同反应方面的联合产物:对强化率敏感的“反应起始”反应;以及对当前运行程序的总体特征不太敏感的“爆发内”反应。两种不同形式反应的存在可能需要重新解释许多依赖于所有反应彼此相似性假设的人类程序控制现象。