In the experimental analysis of behavior, response allocation is typically studied under concurrent interval schedules, with two response alternatives, in a static environment. The natural environment of the unfettered organism, however, is dynamic insofar as even frequently visited environments are rarely identical from encounter to encounter. Additionally, natural environments usually offer more than two concurrently available behaviors that are often scheduled for reinforcement contingent on rate of responding. The purpose of this study was to determine how the addition or removal of a third response alternative affected response allocation between two topographically dissimilar operants on independent concurrent ratio schedules in a dynamic environment. Results indicate that the addition of the third operant served to temporarily suppress response rates of the first two operants but had no or only minor and inconsistent effects on relative allocation. The reintroduction of the third operant in a new location again suppressed response rates in three of four subjects and slightly shifted response allocation for one subject. Highly individual anecdotal patterns could be seen in tangential observations of each of 4 subjects. The results suggest the possibility that new alternatives have temporary effects on response bias, and that these effects may be variable and dependent on details of individual history and stimulus control.

The study of choice and allocation has often focused on certain parameters of reinforcement, but rarely on historical variables. The goal of this study was to investigate the potential effects of gradual vs. abrupt teaching methods on future response allocation. A secondary goal was to see if the results of teaching-method manipulations might be correlated with the parameters that the teaching method produced, specifically unit costs, rate of reinforcement, or error rates. Results indicate that these teaching procedures can produce transient shifts in allocation, but not in a consistent direction. Neither unit cost nor rate or reinforcement alone can account for observed response allocation shifts after training. Researchers saw that subjects reliably shifted towards a manipulanda that produced higher rates of errors, therefore investigating the influence that error rate (or types of errors) may have on response allocation may aid in general teaching method preferences. Future research could focus on the combination of historical procedural variables and current variables that could determine response allocation.

Perseverative conversation is a problem vocal behavior that is unique to individuals with Asperger's disorder. It is characterized by long-winded monologues revolving around circumscribed interests. The current research literature suggests that perseverative conversation may exacerbate already poor social relations, impede productivity at work and home, and lead to disruptive behavior when it is restricted. Despite the negative repercussions there is a lack of research related to the assessment and treatment of this behavior. In the field of applied behavior analysis, concurrent operants assessments have been used effectively to identify relative preference for concurrently available reinforcers. In the current study, choice assessments using concurrent operants arrangements were used to evaluate the reinforcing properties of perseverative conversation in children and adolescents with Asperger's disorder. Five participants, all with Asperger's disorder and reported difficulties with perseverative conversation, were assessed in three phases: A preference assessment for conversation topics; an assessment of preference for reinforcer dimensions; and an assessment of preference for competing dimensions. In the first phase a two-stage preference assessment separated high-preferred from less-preferred conversation topics and in the second and third phases participants were asked to make choices regarding their preference for conversation content, conversational style, and conversation duration. Phase II results suggested that participants preferred conversing about their respective circumscribed interests over neutral topics, actively conversing rather than listening, and conversing for longer duration rather than shorter duration. Phase III, which assessed relative preference for the three reinforcer dimensions, resulted in a hierarchy of preference for all but one participant. Two participants demonstrated the strongest preference for conversation content and two participants demonstrated the strongest preference for conversational style. Conversation duration was least preferred by four of the five participants. Perseverative conversation occurred at a relatively high rate across all assessment phases. The results are discussed in terms of current theoretical explanations for perseverative thinking and behavior, implications for treatment, and future research of this problem vocal behavior.

The present study investigated the effects of a concurrent schedules arrangement, in which three dimensions of reinforcement (duration, attention, and stimuli) were manipulated, on choice responding, appropriate behavior, and problem behavior for three participants with disabilities who had escape-maintained problem behavior. Three experiments were conducted in which participants could choose between work, break, or problem behavior. In the first experiment, the choice analysis, three reinforcement dimensions were varied simultaneously for choice responses. In the second experiment, the component choice analysis, reinforcement dimensions were evaluated in isolation. In the third experiment, the effort analysis, increasing task demand requirements and how they affected response allocation were investigated. Results of the first experiment were consistent across all participants, and suggested that participants allocated their choices in favor of reinforcement contingencies that resulted in breaks with the longest duration, high preference stimuli, and high quality attention. Results of the second and third experiments, however, were idiosyncratic across participants. Component choice analysis results suggested that only specific reinforcement dimensions maintained responding for some participants, whereas all reinforcement dimensions maintained response allocation for others. Results of the third experiment suggested that as task demands increased, reinforcement contingencies that previously maintained responding in the second experiment did not always continue to maintain responding for all participants. This study contributes to and extends the literature on choice making in several ways. The majority of previously published investigations evaluated different dimensions of reinforcement when only two response options were concurrently available, and many of these studies only examined one or two reinforcement dimensions. The present study used a concurrent schedules arrangement in which three concurrently available response options existed. Like previous research the present study suggests that quality of reinforcement can be manipulated to effectively bias individuals' responding in favor of adaptive responses, and the quality variables that impact choice responding may or may not be related to the function of problem behavior. However, further research is needed to understand how choice responding is impacted by increasing demand requirements, as this study demonstrated that choice responding was idiosyncratically affected by changing task demands.