Association for Behavior Analysis International

This symposium includes talks addressing variables that affect response hierarchies and verbal behavior. The variables analyzed include: a) comparing the use of a conditional- versus simple-discrimination task in assessing intraverbal control, b) evaluating the effects of an alternative response on the resurgence of terminal members of a response-class hierarchy, and c) manipulating reinforcement schedules, discriminative stimuli, and motivating operations in producing mands in multioperant environments.

Intraverbal control plays an important role in the study of complex verbal behavior (Skinner, 1957). Palmer and Katz (2005) used a lexical-decision task, which requires participants to press a specific button depending on whether two English words were related or unrelated, in their assessment, and response latency between the related and unrelated pairs of words served as a measure of intraverbal control. A potential limitation of this task is that it confounds the assessment of intraverbal control by including a conditional discrimination, which may contribute to the variability observed within and across participants. We evaluated (n = 15) and compared (n = 15) a textual-response task (simple discrimination) to a lexical-decision task (conditional discrimination) with adults of typical development. Shorter response latencies were observed with related English words compared to unrelated English words with a textual-response task, supporting a systematic replication of the results obtained by Palmer and Katz with a lexical-decision task. Within-subject comparison results show that the use of a textual-response task resulted in reduced differences in response latencies between the types of word pairs in comparison to a lexical-decision task. Future research should focus on identifying the boundary conditions for measuring intraverbal control using both tasks.

In the assessment and treatment of response-class hierarchies of problem behavior, researchers have found that directly altering one member of the hierarchy can indirectly affect other members. One of the most problematic forms of these indirect effects is resurgence of previously extinguished, more severe, members when initial, less-severe members contact extinction. Shabani, Carr, and Petursdottir (2009) described a method for developing an arbitrary response-class hierarchy, which provides a method for evaluating parameters of resurgence without requiring researchers to evoke severe problem behavior. Three typically-developing children participated in the study. In Experiment 1, Shabani et al.s method was replicated and extended. In Experiment 2, an alternative response was taught in isolation and then placed under extinction in the presence of the button hierarchy. When the alternative response was present but under extinction, there was no significant delay to resurgence as compared to when the card was absent and the first member of the hierarchy was under extinction. However, when the alternative response was first reinforced in the presence of the hierarchy and then placed on extinction, there was a delay to resurgence of terminal members of the hierarchy, which may have significant treatment implications.

Individuals with an autism spectrum disorder (ASD) emit repetitive and restricted responding. Limited research on treatment of repetitive or restricted manding exists. Across 4 experiments and for 8 individuals diagnosed with an ASD or Multiplex Developmental Disorder, we evaluated the effects of procedures designed to increase the rate of targeted mands in a multioperant environment. In Experiments 1 through 3, relative reinforcement schedules, discriminative stimuli, and motivating operations were manipulated. In Experiment 4, we combined multiple manipulations to produce targeted manding. Interobserver agreement data were collected across more than 30% of sessions with an average agreement score of greater than 90%. All manipulations were effective in increasing target mands for at least one participant and effects were reliable within participants. Results were idiosyncratic across participants.