A Comparison of Functional Behavioral Assessment and Functional Analysis Methodology among Students with Mild Disabilities

Functional behavioral assessment (FBA) and functional analyses (FA) are grounded in the applied behavior analysis principle that posits problem behavior is functionally related to the environment in which it occurs and is maintained by either providing access to reinforcing outcomes or allowing the individual to avoid or escape that which they find aversive. Previous research has pointed to the limitations across FBA methodologies in comparison to the direct experimental trials used in FA. The purpose of this study was to evaluate the degree to which hypotheses generated by common FBA strategies (i.e., interview, rating scale, and direct observation) match hypotheses generated through FA trials. Results of a multiphase descriptive study indicated that traditional school personnel with behavioral expertise were able to generate FBA hypotheses that were later confirmed by independent review and largely aligned with FA outcomes. The impact of the findings for the field and implications for future research and practice are discussed. ▪ In their seminal article, Baer, Wolf, and Risley (1968) outlined the interconnectedness and functional relationships between an individual’s behavior and environmental events that precede and follow. However, at that time, behavioral interventions for significant challenging behavior largely remained focused on the “form” of the problem behavior and intervention strategies consisted of reinforcing appropriate behavior and punishing inappropriate behavior. In the late 1970s and early 1980s, researchers began to shift their focus from the form of problem behavior to the “function” the problem behavior served for the individual (Carr, 1977; Carr & Durand, 1985; Iwata, Dorsey, Slifer, Bauman, & Richman, 1982). Early research demonstrated that by (a) identifying what occasions and/or what maintains problem behavior (i.e., the function of the problem behavior) and then teaching an appropriate replacement behavior that more efficiently accesses the same or similar outcomes as the problem behavior and (b) altering the environment to withhold reinforcement that was maintaining the problem behavior, problem behavior can be significantly reduced (Carr & Durand, 1985). Three possible functions of problem behavior have been empirically validated: (a) access external reinforcing stimuli, (b) avoid aversive stimuli, and (c) access automatic or sensory reinforcing stimuli (Conroy & Stichter, 2003). Throughout the 1980s and early into the 1990s, the majority of research on functional behavioral assessment (FBA), functional analysis (FA), and functional communication training was conducted within clinical settings with young adults with significant disabilities (Blakeslee, Sugai, & Gruba, 1994; Nelson, Roberts, Mathur, & Rutherford, 1999). As the evidence base supporting FBA/FA to determine function-based interventions grew within the literature on low-incidence disabilities, research on FBA/FA extended to school, home, and other settings (Lalli, Browder, Mace, & Brown, 1993; Northup et al., 1995), as well as to students with and at risk for highincidence disabilities (Kamps et al., 1995; Kern, Childs, Dunlap, Clarke, & Falk, 1994; Lewis & Sugai, 1996a, 1996b; Umbreit, 1995) in the early 1990s. The reauthorization of the Individuals with Disabilities Education Act (IDEA) in 1997, which included regulatory language mandating an FBA be conducted to develop a behavior intervention plan (BIP) if a child with a disability is suspended from school beyond 10 days, brought forth a rapid increase in research within school settings among students with mild disabilities (Dunlap et al., 2006; Fox & Davis, 2005; Scott & Kamps, 2007). Shortly after the 1997 IDEA regulations, Behavioral Disorders, 41 (1), 5–20 November 2015 / 5 the literature base supporting FA to guide intervention development within clinical settings was fairly robust; however, several within the field of emotional and behavioral disorders were quick to point out that the knowledge base at that time was not mature enough to provide a clear consensus on essential elements of applied FBAs within school settings (Nelson et al., 1999; Sasso, Conroy, Stichter, & Fox, 2001). Over the past 20 years, a plethora of FBA/ FA research conducted within educational settings using a range of assessment strategies, targeting a range of problem behaviors, and including students with varying disabilities has emerged (Gage, Lewis, & Stichter, 2012; Solnick & Ardoin, 2010; Wood, Blair, & Ferro, 2009). Equally important, research has shown that interventions guided by FBA/FA are more efficacious in reducing problem behavior in comparison with non–function-based strategies (Gage et al., 2012; Ingram, Lewis-Palmer, & Sugai, 2005; Liaupsin, Umbreit, Ferro, Urso, & Upreti, 2006; Newcomer & Lewis, 2004; Park & Scott, 2009; Payne, Scott, & Conroy, 2007; Stichter, Lewis, Johnson, & Trussell, 2004). Although the evidence supporting the use of FBA/FA to guide interventions for students with mild disabilities and those at risk continues to grow, the field has not reached consensus on the minimal features of an FBA/FA and the requisite skill sets school personnel need to competently complete an FBA/FA within school settings (Scott & Kamps, 2007). Functional analyses typically involve the direct manipulation of antecedent and consequent variables paired with the direct measurement of student behavior (Neef & Peterson, 2007). Because FA involves the direct experimental manipulation of variables under controlled trial conditions, it is typically viewed as essential to empirically identify functional relationships but less feasible within typical educational settings (Gage et al., 2012). The limited feasibility of FA within schools is related to (a) the lack of personnel with expertise to design and implement trials, (b) the time involved in conducting multiple trial configurations, and (c) the need for sophisticated direct observation data collection across trials. Functional behavior assessments are more commonly conducted in school settings and typically include both indirect measures— such as interviews, rating scales, and archival reviews—and direct observations through descriptive summaries of behavioral patterns (Borgmeier, Loman, Hara, & Rodriguez, 2015; Scott & Kamps, 2007). Because environmental events are not directly manipulated as part of the assessment, FBAs rely on personnel with expertise in applied behavior analysis to review multiple measures to draw summary hypotheses about what is occasioning and maintaining problem behavior (Borgmeier et al., 2015). Unfortunately, reviews of FBAs and BIPs conducted by school personnel within the decade following the 1997 IDEA regulations showed a lack of fluency with the process. For example, Blood and Neel (2007) in a review of FBAs and BIPs found little correspondence between the two. Scott and colleagues (2004) found less than a quarter of behavior plans were guided by the FBA. Similarly, Barton-Atwood, Wehby, Gunter, and Lane (2003) found that common rating scales used in FBAs had inconsistencies across educators’ ratings. Recent work on establishing the essential features of efficacious FBAs shows strong promise (Gage et al., 2012; Lane, Kalberg, & Shepcaro, 2009). For example, Scott and colleagues (2004) demonstrated that, through a schoolbased team process that examined multiple data sources and included school personnel with behavioral expertise, educators were able to accurately identify functional relationships between problem behavior and environmental events (Scott et al., 2004; Payne et al., 2007). Lane, Weisenbach, Phillips, and Wehby (2007) further expanded on the teaming approach advocated by Scott and colleagues (2004) and created a problem-solving rubric to assist educators in determining the function of problem behavior with minimal assistance. Focusing on the essential skills needed to interpret typical FBA measures, Borgmeier et al. (2015) demonstrated that with minimal targeted professional development, the majority of educators participating in their study were able to identify the same function of problem behavior as experts. Other research has demonstrated that teachers (e.g., Kamps, Wendland, & Culpepper, 2006) and school support personnel such as paraprofessionals (e.g., Bessette & Wills, 2007) can accurately complete FBAs with appropriate supports and guidance. In one of the few studies to directly compare FBA and FA methodologies within educational settings among students with mild disabilities, Cunningham and O’Neil (2007) examined the degree to which traditional indirect FBA hypotheses were confirmed through direct FA manipulations. Using two rating 6 / November 2015 Behavioral Disorders, 41 (1), 5–20 scales, student interviews, and classroom observations, school personnel (e.g., classroom teachers, school psychologists) developed independent hypotheses regarding the function of student behavior, along with summaries across measures, for 20 students with emotional/behavioral disorders. The FBA results were then compared with brief FA trials conducted by researchers using Northrup and colleagues’ (1991) FA trial conditions and procedures. Results indicated some agreement among the FBA measures but little convergence with the FA trials. In a similar study, Alter, Conroy, Mancil, and Haydon (2008) examined the alignment of resulting hypotheses between two rating scales, the Functional Analysis Interview (O’Neill et al., 1997) and the Motivation Assessment Scale (Durand & Crimmins, 1992), direct observations using an Antecedent-Behavior-Consequence (ABC) format, and FA trials among four young children at risk for emotional/behavioral disorders. They reported little convergence between the rating scale hypotheses and the ABC observation or FA hypotheses but complete agreement between the ABC observations and the FA hypotheses. In the 18 years since the IDEA mandate advocating, and in some instances requiring, an FBA to guide intervention development, a substantial amount of progress has been made. Significant effects on improved outcomes have been documented across studies for students with mild disabilities in educational settings (Gage et al., 2012). Building on the logic of multidisciplinary teams found within special education, research has demonstrated that educators can, with sufficient technical assistance and professional development, accurately identify the function of problem behavior (Borgmeier et al., 2015; Lane et al., 2007; Payne et al., 2007). And yet, when compared with the gold standard of FA, FBA outcomes, especially using indirect methods, remain inconsistent in hypothesis development (Alter et al., 2008; Cunningham & O’Neill, 2007). The purpose of this study was to ascertain the degree to which the current FBA technology aligns with FA trials using school personnel with behavioral expertise. Specifically, the study was designed to answer the following research questions: 1. Towhat degree do FBA hypotheses generated by school personnel with behavioral expertise align with FA generated hypotheses? 2. To what degree do FBA methods (i.e., rating scale, interview, and ABC observations) triangulate to identify function of problem behavior? 3. If disagreements in hypothesized functions generated from FBA measures and FA are found, are there specific functions or contexts that are common across the agreements and/or disagreements?