Classroom-Based Functional Behavioral Assessment

Empirical Base of FBA Use in Schools

The professional literature uses various terms that reflect components of FBA use in the schools, including functional analysis, functional assessment, descriptive analysis, and structural analysis (Lewis & Sugai, 1996). (See Table 1 for definitions.) Each of these terms represents a slightly different aspect of behavior analysis; however, they all support components of FBA. Use of FBA in schools is widely supported in the literature. For example, there are practical application articles that provide suggested directions on how to complete FBA (Shippen, Simpson, & Crites, 2003) and how FBA can promote desirable student behaviors (Ryan, Halsey, & Matthews, 2003).

OPEN IN VIEWER

Table 1.

Behavior Assessment Terminology and Definitions

Term	Definition
Functional behavioral assessment	Systematic method of using behavioral data to create hypotheses to determine antecedents and maintaining consequences of problem behavior
Functional analysis	Systematic manipulation of antecedents and consequences to test hypothesized behavioral function
Descriptive analysis	Utilizing direct observation data (e.g., scatterplots, ABC analysis) to describe problem behavior in terms of environmental events
Structural analysis	Analyzes behavior in terms of behavioral topography (e.g., frequency, rate, severity) to determine behavioral constraints and parameters
Behavior intervention plan	Plan derived from functional behavioral assessment data that focuses on changes to the problem behavior through positive interventions

Beyond practical articles, there exists a base of knowledge that suggests teachers should be trained in conducting FBA. Within this literature, much of the suggested teacher training is for implementation of intervention. For example, teachers have received training on how to implement different functional analysis conditions (Lewis & Sugai, 1996; Moore et al., 2002). Asmus, Vollmer, and Borrero (2002) provided teacher training in conducting FBA; however, no empirical evidence was presented that shows training improved student outcomes. Witt, Noell, LaFleur, and Mortenson (1997) provided empirical evidence of the importance of performance feedback following teacher training in implementing behavioral interventions; however, data related to student outcomes were not presented.

Reviews of literature have been published that focused on the implementation of FBA in school-based settings. Solnick and Ardoin (2010) reviewed 39 data-based articles and noted several important findings related to classroom-based FBA. First, the authors reported that neither teachers nor paraprofessionals were involved in data collection. Second, they reported classroom teachers were responsible for only one-half of the conditions (i.e., interventions) implemented. The authors noted that more than one-third of classroom-based FBA were conducted with students without disabilities, whereas less than 10% were conducted with students with emotional and behavioral disorders (EBD). These findings support a legitimate concern that students who need FBA (i.e., students with EBD and students with disabilities) and the people who need to implement FBA procedures (i.e., teachers) did not play a significant role in the research process.

Ervin, Radford, Bertsch, et al. (2001) conducted a review of 100 published articles related to FBA development in schools. Results of their findings were similar to those disseminated by Solnick and Ardoin (2010). Specifically, the authors reported that school staff implemented 23% of interventions without researcher assistance, whereas the researcher implemented 17% of interventions. In addition, Ervin, Radford, Bertsch et al. noted that a majority of participants had psychiatric diagnoses, with more than 70% having a cognitive disorder (e.g., differing severity levels of intellectual disabilities) and 18% of participants having a school-based diagnosis of EBD. The authors did not note who collected behavioral data, but they did report that observations were conducted in 100% of the reviewed articles.

Other reviews of empirical literature have returned similar results and confirm the need to ensure high fidelity FBA development and implementation by teachers in public schools.For instance, the use of FBA procedures with students with more severe disabilities (see Gable, 1996; Gresham, Quinn, & Restori, 1999; Nelson, Roberts, Mathur, & Rutherford, 1999) and the scant use of classroom teachers as active participants (see Reid & Nelson, 2002) have shown to be a common theme in FBA research. A third theme is the use of analog settings—a setting that resembles the natural context. Analog settings can increase validity of functional assessments by controlling extraneous variables (Ellis & Magee, 2004); however, implementation of FBA in public schools require teachers to use natural settings and contexts (Conroy, Fox, Crain, & Jenkins, 1996). Within these natural contexts, there are certain steps that are generally suggested for teachers to follow. The following section addresses some key components of high-fidelity FBA implementation.

Key Components in FBA

There are several components of FBA that must be addressed; however, O’Neill et al. (1997) suggested three key components of the FBA process: (a) hypothesis development, (b) data collection, and (c) BIP development. Each of these components is utilized throughout the FBA process, which requires teachers to understand and be able to complete each component with high fidelity. (See Figure 1 for schematic of FBA process.) For instance, hypothesis development requires that teachers be able to identify and operationally define the problem behavior. This is followed by collecting baseline data through interviews, reviews of record (e.g., discipline referrals) or direct observation (e.g., systematic observation, frequency counts, ABC charts) to document the severity of the behavior. Third, teachers must have the ability to combine and interpret behavioral data. Interpretation of behavioral data includes communication with the team as well as utilization of data to determine hypothesis of behavioral function. Finally, considering the data collected, analysis of the data and team consultation, the hypothesis can be developed.

OPEN IN VIEWER

Figure 1.

Flowchart of functional behavioral assessment process

Data collection continues throughout hypothesis development and testing, which sets the stage for the BIP development process. In the BIP development component, data collected through hypothesis testing lead to the creation of a BIP that addresses the function of the challenging behavior and attempts to replace the problem behavior with a socially appropriate behavior (Sugai, Lewis-Palmer, & Hagan-Burke, 2000). For example, brief functional analysis (BFA; Northup et al., 1991) can be conducted to test a hypothesis. Based on the results of BFA testing, teachers can develop interventions that address the function of the behavior. Evaluation of the BIP, through data collection, helps to ensure the success of the behavior change.

The Current Review

It is clear that high-fidelity implementation of FBA requires skill and expertise in understanding the complex components of behavior. As previously noted, much of the professional literature reports that teachers have little involvement in data collection or implementation of interventions. Furthermore, much of the professional literature presents FBA procedures being conducted in analog settings. Because of the low involvement of teachers in the FBA process and the high rate of unnatural settings (i.e., out of the classroom), it is important to study what aspects of the FBA process are being reported in the literature. The current review analyzes FBA literature to determine if classroom teachers can implement FBA with high fidelity. This review focuses on data-based and practical articles published in peer-reviewed journals. Data-based articles are analyzed to determine the level of training and involvement of school personnel, student outcomes, and use of FBA in actual classrooms, as opposed to analog settings. Practitioner and data-based articles are analyzed to determine if each of the key components of FBA are reported. Identifying content and outcomes of this literature base will provide a basis for the practicality of high-fidelity FBA implementation by classroom teachers.

Methods

Coding and Summarizing Studies

Selected data-based and practical articles were coded for inclusion of key components of effective FBA development (i.e., data collection methods, hypothesis development, brief functional analysis, BIP development, and plans for evaluation). Data were coded for the data-based studies to determine school personnel involvement in FBA procedures including: (a) participant selection, (b) problem behavior identification, (c) active data collection (e.g., personnel conducting observations), (d) passive data collection (e.g., personnel providing interview data), (e) intervention implementation, (f) inclusion/exclusion and amount of training received, (g) involvement in implementation of brief functional analysis or treatment conditions, and (h) evaluation of intervention (e.g., student outcome and social validity).

Active data collection refers to the inclusivity of school personnel, primarily teachers, in taking an active role in the data collection process through either direct observation or conducting interviews. Passive data collection implies that personnel were providing researchers with data collected either by being interviewed by the researchers or providing feedback through rating scales. Implementation of BFA or treatment conditions refers to school personnel delivering the intervention to the target participants. Social validity measures were included because they provide basic evidence of the usability of procedures used in the studies. For this review, the classification of social validity measures included the type of measure (i.e., checklist, interview) used by the authors. Results of the web-based searches produced a total of 28 articles met the inclusion criteria.

Practical articles were coded for the key components of FBA suggested by O’Neill et al. (1997). Data collection methods were identified based on suggested methods for FBA completion. Data collection was coded by type of data (e.g., interview, frequency/rate, systematic observation, ABC chart, records review, or rating scales) and whether or not the article provided practical data collection methods. That is, articles were examined to determine if the authors provided readers with practical applications of the data collection method. Second, hypothesis development was coded by inclusion or exclusion in the article and the focus on the use of data collection to drive hypothesis development. Third, BFA were coded to determine inclusion or exclusion as a means to test the hypothesis. Finally, BIP development coding focused on transfer of data from the hypothesis-development and testing phases to the intervention and BIP development. Evaluation of BIP was coded for inclusion or exclusion as well as if data collection was suggested during evaluation. A total of 17 articles met the inclusion criteria.

Interrater Reliability

The first and third authors collected interrater reliability data to check the accuracy of the coding procedures. The first author provided a written coding sheet with the aforementioned inclusion criteria, explained the criteria and terminology, and provided an example of a coded article. The third author independently coded 20% of the selected articles. Interrater reliability was determined by dividing the total number of codes by the number of codes agreed upon by the two raters. Results indicated 87% (range = 82%–94%) reliability in coding. Although an acceptable reliability coefficient was attained, coding errors were magnified by the small denominator, and in many cases, the error reflects a disagreement in only 2 codes (which accounted for up to 15% of the coding reliability).

Results

Key Components of FBA

Inclusion of key components in both data-based and practical articles are reported in Table 2. It is evident from inspection of Table 2 that three of the key components were suggested in more than 75% of the articles. The BFA component was not widely suggested within practical articles (29%). In addition, it should be noted that data-based articles reported constructing a hypothesis (94%), but a much smaller percentage of the articles (57%) reported testing the hypotheses.

OPEN IN VIEWER

Table 2.

Overview of Inclusion of Key Components of Functional Behavioral Assessment by Percentage

Key Component	Data-Based Articles(Percentage Included)	Difference	Practical Articles(Percentage Included)
Data collection	100	=	100
Hypothesis development	93	< +1	94
Brief functional analysis	57	> +28	29
Behavior intervention plan development	78	> +2	76

The following sections provide a detailed analysis of the results.

Data collection methods

Results of the data collection measures are shown in Figure 2. Structured interviews were the most common form of data collection for both types of articles, being included in 93% of data-based and 76% of practical articles. Results suggested disparity between the two types of articles for three data collection procedures. Record reviews were utilized in only 3% of data-based articles, but 47% of practical articles suggested the collection of these data. Systematic observation (i.e., whole interval, partial interval, or momentary time sampling procedures) were used in 71% of data-based articles, but were suggested for use in only 35% of practical articles. A third obvious disparity was in the use of ABC analysis, with 46% use in data-based articles and 76% use in practical articles. Both intensity (n = 2) and task analysis (n = 1) were suggested as data collection measures in the practical articles, but were not utilized in data-based articles. There was also a practical article that did not mention specific data collection methods; rather, it stated that data should be collected.

OPEN IN VIEWER

Figure 2.

Percentage of data collection methods reported based on type of article

Hypothesis development

This key component of FBA procedures was included in 93% of data-based articles. Of the included articles, 14% vaguely mentioned the use of hypothesis development by stating that data were used to determine the function of the behavior; however, the verbiage “hypothesis” was not used. For practical articles, 94% included the use of hypothesis development in the creation of FBA.

Brief functional analysis

Data-based articles were more likely to note BFA procedures to test the hypotheses than were practical articles. A total of 57% of data-based articles noted BFA procedures, whereas only 29% of practical articles suggested that teachers use BFA. The majority of data-based articles provided clear descriptions of how to replicate the procedure in future research. Within the data-based articles containing this component, only two explicitly included teachers in developing and implementing BFA procedures. Remaining articles utilized the teacher in varying roles of the procedure, from assisting with development to implementation of the procedure. Most practical articles mentioned this key component, but the description of how to complete BFA did not contain sufficient instructions to replicate the procedure.

Behavior intervention plan development

Data-based articles reported the use of BIP in 78% of articles. The specific type of intervention or treatment varied among the participants. The most common form of intervention was teacher application of differential reinforcement of other behaviors/differential reinforcement of alternative behaviors (DRO/DRA) (18% of articles) conditions. Other types of academic (e.g., audiotaped notes, multiple choice assessments) and behavioral (e.g., contingent reinforcement, fixed interval reinforcement) interventions were reported. All articles noted that BIP development was driven by data collection. Evaluation of BIP was difficult to measure because the research design (e.g., ABAB, alternating treatment) served as an evaluation of the BIP intervention; however, evaluation procedures were clearly stated in 21% of articles.

Practical articles were analyzed to determine the use of BIP evaluation in the FBA process. Furthermore, the practical articles were analyzed for data collection suggestions within the evaluation process. A total of 76% of articles suggested that evaluation measures should be in place to determine the effectiveness of the BIP on problem behavior. Of the 17 selected articles, three specific types of data collection methods were suggested: structured interviews (n = 3), frequency measures (n = 1), records review (n = 1). Almost one-half of the articles suggested that data be collected, but did not provide readers with specifics on data collection methods to use to evaluate BIP effectiveness.

Teacher Involvement in FBA Procedures

Results from data-based articles will be reported within this section.

Participant selection and problem behavior identification

Each data-based article reported consulting teachers to identify participants for the research, as well as to identify the problem behavior in which the participant was engaging. There were researchers that conferred with other school personnel (e.g., administrators) to verify participant selection and problem behavior identification. Problem behavior identification most often included the use of structured interviews and anecdotal data collected from teachers. Several problem behaviors were identified and received some form of intervention. Specifically, off-task (43%) and disruptive (43%) behaviors were the most common problem behaviors noted. Aggression (32%; verbal, physical, self-injurious), out of seat (7%), defiance (3%) and noncompliance (3%) made up the remaining problem behaviors addressed in the FBA.

Teacher involvement in data collection

Active and passive data collection methods were coded for teacher involvement in the data collection process. Figure 3 shows data collection methods used by teachers in the articles. Results suggested that teachers participated in passive data collection through providing data to researchers in the form of interviews (78% of articles) or rating scales (25% of articles). Active data collection by teachers was rarely utilized. Only 2 of the 28 reviewed articles included teachers actively collecting multiple forms of data. For example, one article reported teachers collecting interview and ABC analysis data, whereas the other reported teachers gathering frequency and duration measures. Other forms of reported active data collection were systematic observation (n = 1) and scatterplot (n = 1). There was no specified teacher involvement in data collection for 18% of the reviewed articles.

OPEN IN VIEWER

Figure 3.

Involvement in data collection by teachers (percentages)

Teacher-directed intervention implementation

Interventions reported in the articles included the implementation of BFA conditions (e.g., escape, attention), treatment conditions (e.g., DRA/DRO), or BIP conditions (e.g., teaching expectations). A total of 86% of articles reported that teachers were involved in some level of intervention implementation. Teacher involvement was not reported or specified in the remaining 14% of articles.

Teacher training

Reported amounts of teacher training varied across articles. Only 43% of articles reported training teachers in the study’s FBA procedures. Time spent in training and procedural integrity of intervention implementation varied among the articles. For example, one article reported a 10-min training, whereas another reported a total of 12 hr. An average of 241 min of training was received across the six articles that reported time in training. Four articles reported that teachers received training, but did not report the amount of time spent in training, while one article simply reported that two sessions of training were conducted. Articles reporting teacher training also reported providing either verbal or written feedback to the teacher.

Evaluation of FBA procedures

Two sources of evaluation were noted in the data-based articles: Student outcome and social validity. Student outcomes were measured by the problem behavior being changed in the desired direction (e.g., disruptive behaviors decreasing in frequency). Reports of students’ outcomes were overwhelmingly positive (89% of articles). Surveys were the primary source of social validity measures (61% of articles). Indirect measures of social validity (i.e., conversation with parents and teachers) were reported in 3% of articles, whereas 3% did not report the type of social validity measure that was used.

Discussion

Functional behavior analysis enjoys a rich research base covering more than half a century (Dunlap, 2006; Skinner, 1953) and have been shown to be useful tools in understanding behavioral function and providing function-based interventions (Scott & Kamps, 2007). With the research-based history of FBA and efficacy of FBA use in changing behavior, it is difficult to understand how school-based implementation can be so difficult (Lalli, Browder, Mace, & Brown, 1993). Reasons have been identified and ranged from requiring too much time (Northrup et al., 1991) to requiring too much specialized expertise (Scott & Kamps, 2007); however, with the federal mandate, teachers and schools must be able to conduct FBA with high fidelity.

The purpose of this review was to reflect on professional literature that has addressed FBA completion in the public schools since the inception of the federal mandate. Specifically, key components of FBA development (O’Neill et al., 1997) were coded for both data-based and practical articles to determine the inclusion of these components. Inclusion of both data-based and practical articles was intentional for two primary reasons. First, data-based articles provide clear methodologies needed for replication research, whereas practical articles address a form of dissemination (i.e., less technical terminology) that many classroom teachers are more comfortable reading. Second, both types of articles represent the breadth of dissemination in the field of education (i.e., practitioners and researchers) and must be compared to ensure consistency of message delivery.

There are several important findings that deserve attention. First, all reviewed articles noted the need for data collection in the FBA process. (See Table 2.) The means of data collection varied greatly between the types of articles, with the exception of structured interviews. Figure 2 shows that less than 50% of practical articles suggested direct observation methods, with the exclusion of ABC analyses. Specifically, frequency counts were suggested in about one-third of practical articles, whereas systematic observation was suggested in less than 40% of articles. Perhaps these methods were not suggested because of the perceived difficulty in collecting data while implementing instruction or intervention (e.g., systematic observation; Alberto & Troutman, 2009). ABC analyses were suggested in about 75% of practical articles. Although ABC analysis can and should be used as a direct measure of behavior, they can be used to summarize interview data (e.g., Kern, Dunlap, Clarke, & Childs, 1994). Another concern with ABC analysis is that direct observation could lead to highly inaccurate results (Lerman, Hovanetz, Strobel, & Tetreault, 2009), which suggests that teachers should only collect these data after receiving training. Because of the variability of information being presented to teachers regarding data collection, it would behoove researchers and practitioners to come to a consensus of empirically acceptable and practical data collection methods that allow for high fidelity FBA.

A second important finding related to data collection was the lack of active data collection by teachers. Quinn, Gable, Rutherford, Howell, and Hoffman (2000) suggested that many teachers face challenges in knowing which type of data collection measure to use. Only 3% of the reviewed data-based articles had teachers actively involved in the data collection process, which suggests that teachers were either not given opportunities to collect data or expressed hesitation about collecting data. If special education teachers are held accountable for their expertise in completing FBA/BIP, they will need to be provided with practical means to collect data accurately.

The IDEA requires that FBA lead to BIP that provide positive intervention for problem behavior (Sugai et al., 2000), suggesting that teachers should develop BIP that address behavior function. This leads to a question of fidelity of hypothesis development and testing by teachers. It is encouraging that 93% of articles suggested the use of data to develop hypotheses for possible behavioral functions and 86% of articles utilized teachers, in some capacity, to implement interventions. A discouraging result was that only 29% of practical articles and 57% of data- based articles suggested testing the hypothesis (e.g., BFA) prior to implementation of behavior intervention or BIP. It would seem plausible that hypotheses should be tested prior to BIP implementation to insure success.

Research has reported that hypothesis testing is conducted either by researchers or by teachers under their guidance (Heckarnan, Conroy, Fox, & Chait, 2000). Furthermore, hypothesis testing in natural classroom settings can be difficult to accomplish, which is why it has been suggested that it may be more efficient to utilize analog settings (Repp & Karsh, 1994). Unfortunately, many teachers do not have access to analog settings or the personnel to commit to testing the hypothesis in this setting. Effective collaboration between researchers and practitioners to establish a teacher-led, practical methodology for testing hypotheses is important. Research has shown that BFA use in natural settings can provide effective hypothesis testing (e.g., Cihak, Alberto, & Frederick, 2007; Wallace & Iwata, 1999); therefore, supplying teachers with training and a methodology could lead to better FBA/BIP.

Limitations

Utilizing a computer-based search of the Academic Search Premier and ERIC databases may not have addressed all articles that focus on school-based FBA. Although these databases contain a large sample of relevant literature, it is likely that an article may have been inadvertently omitted from the study. A second limitation may be the inclusion of a specified date of publication. Much research exists related to school-based FBA prior to the 1997 amendments. This review focused on the literature from the time of the FBA mandate to the present; however, it may be beneficial to review the years prior to 1997 to determine teacher involvement.

Conclusions

Functional behavioral assessments were mandated to address problem behaviors that result in excessive removal from the child’s LRE (Asmus et al., 2002) and are intended to be completed by school teams (e.g., IEP teams) that are usually headed by a special education teacher. Because the special education teacher is responsible for IEP and FBA/BIP implementation, it is vital for this person to be well trained in understanding FBA components (Nelson et al., 1999) and have the collaborative skills to lead a team through problem solving and BIP development (Quinn et al., 2000).

Ten years ago, Quinn et al. (2001) suggested 10 gaps in research and practice in FBA development (p. 268). Based on results of the current review of literature, there remain gaps in both research and practice. For instance, five of the gaps in practice that Quinn et al. (2001) suggested were related to teacher knowledge ranging from selecting behaviors to implementing BIP. Results of the current review show that many teachers involved in data-based research were not provided training in any aspect of the research methodology or training in active data collection. Research has shown that teachers have the ability to conduct FBA with high fidelity (Doggett, Edwards, Moore, Tingstrom, & Wilczynski, 2001; Mueller, Sterling-Turner, & Moore, 2005; Wallace, Doney, Mintz-Resudek, & Tarbox, 2004). As a field, if we are to fulfill our mandate, we must provide teachers with the skills to successfully complete the tasks assigned to them. Researchers should begin to utilize teachers more in their active data collection and provide them with opportunities to test hypotheses. Practitioners should begin to translate research-validated methods of data collection, analysis, and implementation into practical (perhaps task analyzed) steps that teachers can follow to develop FBA and implement BIP effectively.