JOURNALOFAPPLIEDBEHAVIORANALYSIS 2010, 43, 565–567 NUMBER3 (FALL2010) INFLUENCE OF ASSESSMENT SETTING ON THE RESULTS OF FUNCTIONAL ANALYSES OF PROBLEM BEHAVIOR RUSSELL LANG UNIVERSITYOFCALIFORNIA,SANTABARBARA JEFF SIGAFOOS VICTORIAUNIVERSITYOFWELLINGTON,NEWZEALAND GIULIO LANCIONI UNIVERSITYOFBARI,ITALY ROBERT DIDDEN RADBOUDUNIVERSITYNIJMEGEN,THENETHERLANDS AND MANDY RISPOLI TEXASA&MUNIVERSITY Analogue functional analyses are widely used to identify the operant function of problem behaviorinindividualswithdevelopmentaldisabilities.Becauseproblembehavioroftenoccurs across multiple settings (e.g., homes, schools, outpatient clinics), it is important to determine whethertheresultsoffunctionalanalysesvaryacrosssettings.Thisbriefreviewcovers3recent studiesthatexaminedtheinfluenceofdifferentsettingsontheresultsoffunctionalanalysesand identifies directions for future research. Keywords: autism, assessment, developmental disability, functional analysis, setting _______________________________________________________________________________ Analogue functional analyses are designed to (Hanley, Iwata, & McCord, 2003), and the identify functional relations between problem person who implements the assessment (Ring- behavior and reinforcement contingencies dahl&Sellers,2000)haveallbeenidentifiedas through systematic environmental manipula- variables that can influence the results of the tions (Iwata, Dorsey, Slifer, Bauman, & Rich- assessment. Given this evidence, it is plausible man, 1982/1994). Several idiosyncratic and that the setting in which the functional analysis contextual variables have been shown to is conducted might also influence results. influence the results of analog functional Functional analyses to identify the operant analyses (Carr, Yarbrough, & Langdon, 1997). function of problem behavior have been For example, presession attention (McComas, conducted in a variety of settings, ranging from Thompson, & Johnson, 2003; O’Reilly, clinical settings (e.g., hospitals, outpatient Edrishina,&Sigafoos,2007),thecharacteristics clinics, unoccupied rooms in schools) to of task demands (Smith, Iwata, Goh, & Shore, relatively natural contexts (e.g., bedrooms in 1995), the order of assessment conditions children’s homes and classrooms with other children present). It is not uncommon for the AddresscorrespondencetoRussellLang,whoisnowat resultsofafunctionalanalysisconductedinone the Department of Curriculum and Instruction, Texas settingtobeusedtodesignanintervention that StateUniversity,601UniversityDrive,SanMarcos,Texas is implemented in another setting (Hanley et 78666(e-mail: [email protected]). doi:10.1901/jaba.2010.43-565 al.,2003).Whenthisisdone,itisassumedthat 565 566 RUSSELL LANG et al. the variables that influence behavior in the the whether the discrepancy in the results across functional analysis setting are the same as those settings was important in the design of an that control the behavior in the intervention intervention by comparing an attention-based setting. intervention and a tangible-based intervention Thepurposeofthisbriefreviewistodescribe in both settings. Results of the comparison three recent studies (i.e., Lang et al., 2008, validated the assessment results (i.e., the 2009, 2010) in which this assumption was attention-based intervention was more effective tested by examining the influence of setting on on the playground and the tangible-based functional analysis results and subsequent intervention was more effective in the class- function-based interventions. Based on this room) and demonstrated the importance of brief review, we offer practical suggestions for considering potential variations in the function practitioners who implement functional analy- of problem behavior across settings in the ses and elucidate directions for future research. design of function-based interventions. Lang et al. (2008) conducted functional In the thirdstudy (Lang et al., 2010),a child analyses using a procedure similar to that of with Asperger syndrome was assessed in two Iwata et al. (1982/1994) with two school-aged different rooms (i.e., resource room and girls with autism who engaged in problem classroom).Functionalanalysisresultsindicated behavior. In the Lang et al. study, both children that elopement was maintained by access to participatedinfunctionalanalysesintwosettings attention in the resource room and by access to (atherapyroomandaclassroom).Toisolatethe tangible items in the classroom. Two interven- potential influence of setting, other variables tions (an attention-based intervention and a knowntoaffecttheresults offunctional analyses tangible-based intervention) were then com- were held constant (i.e., the sequence of pared in an alternating treatments design in conditions, materials, task demands, and thera- both settings. The attention-based intervention pist). For both children, escape from task was more effective in the resource room, and demands and attention from the therapist were the tangible-based intervention was more effec- identifiedas maintainingconsequencesforprob- tive in the classroom. As in the 2009 study, lem behavior in both settings. However, for the results of this comparison again validated the second child, levels of problem behavior were findings of the functional analyses and demon- higher in the escape condition in the therapy strated the importance of conducting the room than in the classroom. functional analyses in the setting in which the Lang et al. (2009) replicated the procedures intervention will be implemented. of Lang et al. (2008) with a child with autism Overall, these three studies suggest that at who engaged in aggression. Lang et al. (2009) least two patterns of results are possible when conducted functional analyses in the child’s functionalanalysesarecomparedacrosssettings. classroom and on the school playground. The The first pattern is one in which the functional playground assessment results indicated that analyses identify the same behavioral function problem behavior was maintained by adult or functions. In this pattern, the same conclu- attention as reinforcement, whereas the class- sion and subsequent treatment decisions would room assessment results indicated that problem be made regardless of the assessment setting. behavior was maintained by access to toys. ThiswasthecaseforParticipant1inLangetal. These results suggested that the controlling (2008). The second pattern is one in which the variablesforproblembehaviormaydifferacross results of the functional analyses differ across settings. Lang et al. (2009) then validated the settings; this was the case in Lang et al. (2009, results of the functional analyses and evaluated 2010).Itisthissecondpatternthatisimportant BRIEF REVIEW 567 topractitionersbecauseitmayleadtothedesign REFERENCES of an ineffective function-based intervention. Carr,E.G.,Yarbrough,S.C.,&Langdon,N.A.(1997). Specifically, if a functional analysis is conducted Effects of idiosyncratic stimulus variables on func- in one setting and the results of that analysis are tional analysis outcomes. Journal of Applied Behavior Analysis,30,673–686. usedtodesignaninterventionthatisimplement- Hanley, G. P., Iwata, B. A., & McCord, B. E. (2003). ed across settings, the intervention may fail in Functional analysis of problem behavior: A review. settings in which the maintaining consequences Journal ofAppliedBehavior Analysis, 36,147–185. aredifferentfromthoseoftheassessmentsetting. Iwata,B.A.,Dorsey,M.F.,Slifer,K.J.,Bauman,K.E., & Richman, G. S. (1994). Toward a functional For example, a functional analysis conducted at analysis of self-injury. Journal of Applied Behavior schoolmayresultinthedesignofanintervention Analysis, 27, 197–209. (Reprinted from Analysis and that is effective at school but is ineffective in the Intervention in Developmental Disabilities, 2, 3–20, home. It is perhaps not uncommon for imple- 1982) Lang, R., Davis, T., O’Reilly, M. F., Machalicek, W., mentation fidelity to be questioned following Rispoli, M. J., Sigafoos, J., et al. (2010). Functional such intervention failures. The results of these analysisandtreatmentofelopementacrosstwoschool studies suggest that, in addition to implementa- settings. Journal of Applied Behavior Analysis, 43, 113–118. tion fidelity, practitioners should also consider Lang, R., O’Reilly, M., Lancioni, G., Rispoli, M., the possibility of ungeneralizable functional Machalicek, W., Chan, J. M., et al. (2009). analysis results. Discrepancy in functional analysis results across two The results of these studies raise several appliedsettings:Implicationsforinterventiondesign. Journal ofAppliedBehavior Analysis, 42, 393–397. questions that warrant future research. First, it Lang, R., O’Reilly, M., Machalicek, W., Lancioni, G., is not always practical to implement functional Rispoli, M., & Chan, J. M. (2008). A preliminary assessments in every setting in which problem comparison of functional analysis results when behavior occurs. Therefore, future research conducted in contrived versus naturalistic settings. Journal ofAppliedBehavior Analysis, 41, 441–445. should determine what factors contribute to the McComas, J. J., Thompson, A., & Johnson, L. (2003). development of setting-specific functions, with The effects of presession attention on problem the aim of creating a procedure for identifying behavior maintained by different reinforcers. Journal of AppliedBehaviorAnalysis,36,297–307. when functional reinforcers will differ across O’Reilly,M.F.,Edrishina,C.,&Sigafoos,J.(2007).The settings. Second, conducting functional analyses effects of presession attention on subsequent atten- in classrooms, onplaygrounds,andother similar tion-extinction and alone conditions. Journal of settings often requires extensive planning to Applied BehaviorAnalysis,40,731–735. Ringdahl, J. E., & Sellers, J. A. (2000). The effects of control possible confounding variables inherent different adults as therapists during functional in such settings (i.e., other students may deliver analyses. Journal of Applied Behavior Analysis, 33, attention when attention is not a programmed 247–250. consequence for problem behavior). To the Smith, R. G., Iwata, B. A., Goh, H., & Shore, B. A. (1995). Analysis of establishing operations for self- extent that the procedural integrity of these injury maintained by escape. Journal of Applied assessments is compromised, functional analyses Behavior Analysis,28,515–535. may yield inaccurate results. Therefore, future researchshouldalsofocus on identifying adapta- Received February 17,2010 tionstofunctionalanalysesthatcontrolpotential Final acceptanceMay 4,2010 confounding variables in applied settings. Action Editor,Cathleen C.Piazza