Date of Award


Degree Name

Doctor of Philosophy



First Advisor

Dr. Cynthia J. Pietras

Second Advisor

Dr. Alan D. Poling

Third Advisor

Dr. Scott T. Gaynor

Fourth Advisor

Dr. Eric A. Jacobs


log-survivor, bout, rats, human behavior


Operant psychologists typically use response rate as a primary measure of behavior. Although response rate has proven a useful dependent measure resulting in the identification of many important behavioral regularities, many researchers have argued that the measure has significant limitations. Primarily, response rate treats all responses in the measured response class as functionally equivalent and distributed uniformly across time. This conceptualization of behavior is useful as long as all responses are affected similarly by different experimental manipulations. Research has shown, however, that certain manipulations differentially affect responses with relatively short or long interresponse times. This has led to a new conceptualization of responding in terms of periods of engagement and disengagement, or response bouts. In this approach, responses are characterized as either initiating bouts of responding (i.e., responses with relatively long inter-response times), or comprising bouts of responding (i.e., responses with relatively short inter-response times). Initiation responses are thought to be affected by motivational or reinforcement variables, whereas within-bout responses are thought to be affected by variables affecting the capability to respond. Most studies investigating the bout nature of responding have used rats responding for food pellets. Few studies have explored the generality of approach with other reinforcers or species. Thus, Experiment 1 attempted to assess the generality of a response-bout conceptualization of behavior by investigating rat lever pressing for milk reinforcers across manipulations shown in prior studies to differentially affect within-bout and bout-initiation responses. Experiment 2 sought to extend the generality of the bout analysis to human button and footswitch pressing for hypothetical monetary reinforcers. Data were analyzed with log-survivor functions, which have previously been used to assess the two-state nature of responding. Only results from some of the manipulations in Experiment 1 provided compelling support for the two-state conceptualization of responding. Therefore, the generality of the bout conceptualization of responding appears limited. The bout analysis has been posited as a possible method for elucidating behavioral mechanisms of drug action. The potential utility of the bout analysis for investigating drug effects is discussed in light of the present findings. Other limitations of the bout analysis are also discussed.

Access Setting

Dissertation-Open Access