Association for Behavior Analysis International

Functional analysis (FA) is the gold standard for identifying the function of problem behavior, but it is sometimes impractical or not feasible to implement. Therefore, it is important to develop alternative and adapted methods that can be implemented in settings where FAs are not possible. The current symposium will present recent research evaluating alternative methods such as indirect assessments, use of mand and preference assessments, and trial-based FAs. Data presented will compare alternative methods to FAs to determine the correspondence and validity of these alternative methods. Finding show improved correspondence when indirect assessments were completed by experts when compared to those completed by caregivers, good correspondence between concurrent operant preference assessment using the same reinforcers provided during the FA and positive treatment outcomes for interventions developed based on trial-based FAs. Taken together these results provide preliminary support for these alternative methods. Data will also be presented on the use of mand assessments as an alternative method for identifying the function of problem behavior. Results of the mand assessment were compared to those of FAs, showing low correspondence between mand assessments and FAs. Results from this study indicate mands may not be a viable alternative to FAs.

Functional analysis (e.g., Iwata, Dorsey, Slifer, Bauman, & Richman, 1982/1994) is the most effective methodology for identifying the function of problem behavior. However, skills and resources needed to conduct functional analyses are often not available in many settings, which has resulted in the use of indirect assessments to predict the function of problem behavior. To date, researchers have found that caregiver-completed indirect assessments are not valid (i.e., they do not correspond with functional analysis outcomes; Smith et al., 2012), but it is possible that "experts" may be better at accurately completing indirect assessments. The purpose of the current study was to compare the outcomes of an indirect assessment (Functional Analysis Screening Tool; Iwata, DeLeon, & Roscoe, 2012) completed by two caregivers and two experts in functional behavioral assessment and the outcome of a functional analysis. Five children with autism who engaged in problem behavior, their caregivers, and several experts participated. Comparison of the outcomes of the indirect assessments and functional analysis outcomes suggested that experts were more likely than caregivers to identify all functions of problem behavior via indirect assessments, but sometimes identified additional functions. The use of experts for completing indirect assessments could have significant impact on their utility.

Although functional analyses (FAs) are the gold standard in identifying the function of problem behavior, they may not always be feasible. One alternative may be to assess the function of appropriate requesting (i.e., mands) under the same circumstances that problem behavior is assessed during traditional FAs. The assumption behind this type of assessment is that the motivating variables that evoke target problem behavior would also evoke functionally equivalent mands. Two recent studies have evaluated correspondence between FAs of problem behavior and mand assessments, however, results conflicted: Scheiltz et al. (2010) found correspondence for 20% of participants, whereas LaRue et al. (2011) found correspondence for 75%. The studies differed in whether or not they incorporated mand prompting. This study sought to determine whether procedural variations accounted for the difference in correspondence. Thus, we conducted three assessments with four children: a mand assessment with prompts, a mand assessment without prompts, and an FA of problem behavior. The results indicate poor correspondence between FAs of problem behavior and mand assessments with prompts (0%), as well as mand assessments without prompts (25%), suggesting mand assessments should not be used as a basis for identifying function of problem behavior.