|Patterns Everywhere: Sequential Analysis and What It Can Do in Applied and Basic Behavior Analysis
|Monday, May 26, 2008
|9:00 AM–10:20 AM
|Chicago & Alton
|Area: EAB/AUT; Domain: Basic Research
|Chair: Susan M. Schneider (Florida International University)
|Discussant: Alliston K. Reid (Wofford College)
|Abstract: Sequential analysis enables temporally-ordered patterns in data to be identified and evaluated for statistical significance. Behavior analyst Ernst Moerk used it to demonstrate the prevalence of the three-term operant contingency in language acquisition (note his well-known 1990 publication in the Journal of the Experimental Analysis of Behavior). The classic reference is Bakeman and Gottman (Observing Interaction), and the method has been of considerable value in a wide variety of areas in psychology. This symposium will show how sequential analysis is currently being applied across a range of behavior-analytic specialty areas, from applied research in developmental disabilities and autism, to basic matching law research with pigeons.
|Sequential Analysis and the Matching Law: Molecular Patterns in Three Species.
|SUSAN M. SCHNEIDER (Florida International University)
|Abstract: When demarcated two-response sequences were concurrently reinforced, pigeons’ sequences rather than individual keypecks obeyed the generalized matching law. At a molecular level, a sequential analysis revealed that the sequences themselves were emitted in nonrandom patterns that were consistent across individuals. After reinforcer delivery, different patterns occurred, again consistent across individuals. A sequential analysis performed on similar data from rats produced similar patterns, even when minimum inter-response times were required between the sequence elements. Higher-order organizing principles for these patterns were identified. Finally, quail chicks pecked touchscreen targets for heat in a standard concurrent variable-interval schedule paradigm. Their responses obeyed the generalized matching law. A sequential analysis found patterns that were different from those for the concurrent-sequences paradigm for the rats and pigeons. Some patterns were present immediately upon exposure to the concurrent schedules, when the chicks were only 24 to 48 hours old. Such results can advance our understanding of behavior-environment causal relations at a molecular level. They also have applied implications.
|Sequential Analysis within Single Case Experimental Designs in Home and School: Connecting Behavior and Physiological Arousal.
|JOHN D HOCH (University of Minnesota), Jennifer J. McComas (University of Minnesota), Timothy R. Moore (University of Minnesota), Frank J. Symons (University of Minnesota)
|Abstract: Theoretical accounts of autism have long hypothesized a link between arousal and behavior. In applied settings, parents and teachers often attribute problem behavior to internal levels of arousal. Sequential analysis of data collected during a home-based concurrent operant choice procedure and a school-based functional behavioral analysis (FBA) were combined with real-time heart rate monitoring to determine whether arousal preceded or followed choice among activities for one participant, or occurrences of problem behavior for the other participant. In the concurrent operant procedure, sequential analyses showed that activity choice was significantly related to the preceding level of arousal (p < 0.05). For the participant in the school based functional analysis, changes in arousal did not reliably proceed or follow instances of self injury or aggression. Results from the sequential analysis are compared to summary level ANOVA results. Sequential analysis is discussed as a method to examine the link between behavior and arousal within the context of single case experimental designs and socially significant problem behavior.
|Sequencing the Behavioral Unit: Conditional Probabilities in Basic and Applied Settings.
|SCOTT S. HALL (Stanford University)
|Abstract: Sequential (conditional probability) analysis can be used to uncover a variety of fundamental and important dynamic behavioral processes operating in a variety of settings. For example, in an experimental setting, a conditional probability analysis was employed to examine the dynamic shifts between choices of 'cooperate' and 'defect' in an iterated Prisoner's Dilemma Game. Two pigeons responded sub-optimally when playing against a simulated opponent playing Tit-for-Tit. A conditional probability analysis (i.e., an examination of the probability that a pigeon received a particular outcome on a trial given that he had received one of these outcomes on a preceding trial) indicated that local reinforcement contingencies determined suboptimal performance. In an applied setting, the problem behaviors (e.g., aggression, self-injury, pica) of four individuals with developmental disabilities were observed in naturalistic settings. An examination of the conditional probability of attention, demands, and tangibles (prior to and following problem behavior) indicated that these events were presented contingently on thin schedules. These examples, and others, demonstrate the utility of conducting sequential analyses of behavioral data streams. Possible reasons for the non-adoption of sequential analysis methodology by behavior analysts, including the base-rate problem, specifying time windows, statistical issues and interpretation, will be explored.