IT should be notified now!

Association for Behavior Analysis International

The Association for Behavior Analysis International® (ABAI) is a nonprofit membership organization with the mission to contribute to the well-being of society by developing, enhancing, and supporting the growth and vitality of the science of behavior analysis through research, education, and practice.

Search
Donate to SABA Capital Campaign
Portal Access Behavior Analysis Training Directory Contact the Hotline View Frequently Asked Question
ABAI Facebook Page Follow us on Twitter LinkedIn LinkedIn

43rd Annual Convention; Denver, CO; 2017

Event Details

Previous Page

 

Symposium #32
CE Offered: BACB
Quantifying Effects, Identifying Relations, and Extending the Generality of Behavior Analytic Research on Problem Behavior
Saturday, May 27, 2017
10:00 AM–11:50 AM
Convention Center Four Seasons Ballroom 1
Area: DDA/AUT; Domain: Applied Research
CE Instructor: SungWoo Kahng, Ph.D.
Chair: SungWoo Kahng (University of Missouri)
Discussant: Gregory P. Hanley (Western New England University)
Abstract: Studies employing single-case experimental designs are perfectly suited to examine behavior-environment interactions at the level of the individual participant; however, a body of literature that is comprised mostly of such studies has its limitations. Because data are analyzed using visual analysis and studies report on small numbers of participants, the objectivity of our data analysis methods and the generality of findings of individual studies may be limited. Although some have called for the use of randomized clinical trials and parametric statistics, those methods are conceptually inconsistent with applied behavior analysis, and are practically incompatible with the individualized response-guided approach to assessment and treatment that is a hallmark of our field. The current presentations will describe methods aimed at improving the objectivity of data analysis, and enhancing the generality of behavior analytic findings. A critical feature of these methods is that they preserve the analysis of individual behavior by either quantifying behavior change within the individual participant, or by analyzing accumulated datasets from multiple participants within or across studies. These methods have the potential to identify relations not otherwise evident when examining individual datasets in isolation, and extend the generality of findings within and across studies.
Instruction Level: Advanced
Keyword(s): generality, problem behavior, Single-case designs
Empirically Supported Treatments in Applied Behavior Analysis
MICHELLE A. FRANK-CRAWFORD (Kennedy Krieger Institute), Patricia F. Kurtz (Kennedy Krieger Institute), Louis P. Hagopian (Kennedy Krieger Institute)
Abstract: A number of procedures describing criteria for systematically reviewing a body of literature have been developed. For example, the American Psychological Association (Task Force Promoting Dissemination of Psychological Procedures, 1995) described a process of evaluating whether treatments have been sufficiently researched to characterize them as “empirically supported treatments” (EST). Those interventions with the highest level of support are characterized as “well-established” (Chambless et al., 1996). This and similar efforts have been undertaken for the purposes of guiding clinical practice, influencing regulations and standards, providing priorities for funding (for both research and treatment), and guiding professional training. The presentation will review these methods and discuss modifications to make them more suitable for evaluating behavior analytic research. EST studies on behavior analytic research will be summarized, followed by a discussion of the value of quantitatively synthesizing research finding across studies with regard to documenting the reliability of effects of interventions, as well as establishing the generality of those effects across individuals, settings, and researchers.
The Use and Utility of Consecutive Controlled Case Series in Applied Behavior Analysis
GRIFFIN ROOKER (Kennedy Krieger Institute), Clare Liddon (Kennedy Krieger Institute), Christopher M Dillon (Kennedy Krieger Institute), Louis P. Hagopian (Kennedy Krieger Institute)
Abstract: In a Consecutive Case Series, all patients with a particular condition are identified and outcomes of a procedure (s) with these patients are reported. The use of single subject design in Applied Behavior Analysis (ABA) allows this methodology to be extended and produces Consecutive Controlled Case Series (CCCS). The use of CCCS in ABA increases the generality of findings, while retaining appropriate focus on the individual and his or her behavior. The purpose of this study was to review ABA research where CCCS has been used in relation to problem behavior to highlight the applicability of this methodology, as well as to demonstrate the utility of conducting such studies. Review of the research indicates that in the past 25 years more than 20 studies met our criteria as a CCCS. These CCCS detailed over 1000 behavioral assessments and/or treatments for individuals with a wide range of behavioral disorders. Specific examples of how CCCS research has produced novel findings and increased the generality of behavioral assessments and treatments will be discussed. Finally, the overall quality of this literature, as well as specific future directions for CCCS research will be discussed.
Examining Correspondence Between Statistical Modeling and Visual Analysis of Behavioral Assessment and Treatment Data
NICOLE LYNN HAUSMAN (Kennedy Krieger Institute), Gayane Yenokyan (Johns Hopkins Bloomberg School of Public Health), Julia Iannaccone (Kennedy Krieger Institute), Louis P. Hagopian (Kennedy Krieger Institute)
Abstract: The current presentation will discuss potential limitations and benefits of using statistical modeling to quantify effects observed during assessment and treatment. We employed a model for statistical analysis that mirrors visual analysis for magnitude of effects, stability, and trend. Generalized Linear Models (GLM) with a distribution for the outcome and a link function that describes how average response depends on treatment was used. The simplest distribution for the outcome is normal; however, others can be selected, by checking the correspondence of observed versus predicted values. To estimate “percent change” in response, a logarithmic link function is used. Robust, or “model- agnostic” variance can be specified to calculate 95% confidence interval for the treatment effect. We examined the correspondence between the GLM model to visual analysis of published functional analysis and treatment datasets and a high degree of correspondence was observed. Statistical methods need not obscure the individual analysis of behavior, replace visual analysis, or eliminate the pursuit of producing socially meaningful change. Rather, quantitative methods build on visual analysis to adequately model the data, provide a highly objective and reliable means to evaluate outcomes, and would make findings of behavior analytic studies more interpretable by the broader scientific community.
Identifying Predictive Behavioral Markers: Implications for Advancing Practice and Research
LOUIS P. HAGOPIAN (Kennedy Krieger Institute), Griffin Rooker (Kennedy Krieger Institute), Gayane Yenokyan (Johns Hopkins Bloomberg School of Public Health)
Abstract: Recent research on automatically reinforced self-injurious behavior (SIB) has identified subtypes based on distinct patterns of responding in the functional analysis (FA). Subtypes were shown to differ greatly in terms of their resistance to first line treatment (reinforcement). Those findings were largely replicated in a subsequent analysis of published datasets. The current study combined data from these two studies (n = 78) and examined with the quantitative methods used to evaluate predictive biomarkers – biological measures that predict response to treatment. This probabilistic analysis preserves the individual dataset by first classifying cases based on whether treatment targets were achieved (80% reduction in SIB), and then determining how accurately the predictor distinguishes those groups. Findings revealed that both the level of SIB differentiation in the FA and subtype classification to be “good to excellent” predictive behavioral markers (PBM). These PBMs identified sensitivity of SIB to disruption by alternative reinforcement as a critical dimension for automatically reinforced SIB. The potential utility of this approach for applied behavior analysis research and practice is discussed. Identifying other PBMs could help inform individualized treatment selection, identify classes of problem behavior that are responsive and non-responsive to treatment, and advance knowledge about how treatments exert their effects.
 

BACK TO THE TOP

Modifed by Eddie Soh
SABA DONATE ABAI HOTLINE