Secondary prevention efforts at the middle school level: an application of the Behavior Education Program.
Article Type:
Report
Subject:
Junior high school students (Behavior)
Behavior modification (Research)
Authors:
Lane, Kathleen Lynne
Capizzi, Andrea M.
Fisher, Marisa H.
Ennis, Robin Parks
Pub Date:
02/01/2012
Publication:
Name: Education & Treatment of Children Publisher: West Virginia University Press, University of West Virginia Audience: Professional Format: Magazine/Journal Subject: Education; Family and marriage; Social sciences Copyright: COPYRIGHT 2012 West Virginia University Press, University of West Virginia ISSN: 0748-8491
Issue:
Date: Feb, 2012 Source Volume: 35 Source Issue: 1
Topic:
Event Code: 310 Science & research Canadian Subject Form: Behaviour modification
Product:
Product Code: E197300 Students, Junior High
Geographic:
Geographic Scope: United States Geographic Code: 1USA United States

Accession Number:
285089960
Full Text:
Abstract

In this study we examine the impact of the Behavior Education Program (BEP; Hawken, MacLeod, & Rawlings, 2007) with four middle school students who were not responsive to a comprehensive primary prevention program including academic, behavioral and social components. To extend this line of inquiry we (a) conducted a functional behavioral assessment prior to implementing the BEP and (b) employed a changing criterion design to determine if a functional relation could be established between the introduction of the BEP program and changes in student performance. Results suggest this intervention may be effective for students whose challenging behaviors are maintained by attention and escape. However, the variability in scores raises questions and concerns regarding the nature of the behavior change that may have occurred. Additional limitations and future direction are discussed.

Keywords: Behavior Education Program, three-tiered models of prevention, secondary supports

Positive behavior intervention and supports (PBIS; Sugai & Horner, 2002) is a three-tiered model of prevention and intervention dedicated to (a) preventing the development of challenging behavior and (b) responding to students with existing concerns. The PBIS model involves a systematic, data-driven continuum of support with primary, secondary, and tertiary levels of prevention (Turnbull, et al., 2002). In PBIS, primary prevention supports are available to all students to encourage desired behavior in key school settings (e.g., cafeteria, classroom, hallways). Schools establish 3-to-5 positively stated expectations to clarify desired behaviors and provide consequences for rule-violations (Crone, Horner, & Haw ken, 2004). Students are taught these expectations explicitly, given opportunities to practice expectations, and receive reinforcement for meeting them. In short, it involves an instructional approach to teaching expectations, similar to the way one would teach academic content (Lane, Kalberg, & Menzies, 2009).

School-wide data collected as part of regular school practices are analyzed by school-site leadership teams to identify students who do not respond to the primary level of prevention/ intervention in social, behavioral, and/or academic domains (Hawken, MacLeod, & Rawlings, 2007; Lane, 2007). Identified students are offered secondary supports to address their specific needs, provided their parents agree to these additional programs or practices. Secondary supports include targeted interventions provided to groups of students who show signs of similar acquisition or performance deficits (Elliott & Gresham, 2007). Approximately 10-15% of the student body is expected to require secondary prevention efforts (Sugai & Horner, 2002).

Finally, students exhibiting serious behavior problems are provided tertiary supports such as individualized behavior support plans based on results of a functional behavioral assessment (FBA; Lane, Kalberg, & Shepcaro, 2009; Turnbull et al., 2002) or other individual supports such as mental health supports. Approximately 3-5% of the student body may require tertiary prevention efforts (Sugai & Homer, 2002).

While considerable research has examined the effectiveness of primary and tertiary levels of behavior support, fewer studies have been conducted to evaluate secondary prevention efforts (Hawken, 2006; Kalberg, Lane, & Lambert, in press; Robertson & Lane, 2007). The Behavior Education Program (BEP; Hawken et al., 2007) is a secondary prevention support to assist middle school students identified as at-risk for developing behavior problems due to poor peer relations, low academic achievement, and/or chaotic home environments.

The Behavior Education Program: A Promising Secondary Prevention Effort

The BEP, also referred to as check-in/check-out procedure (McCurdy, Kunsch, & Reibstein, 2007), is designed to provide targeted support to students at-risk for developing serious or chronic behavior problems (Crone et al., 2004). Students involved in the BEP receive frequent feedback regarding daily behavior, with an emphasis on providing positive reinforcement for appropriate behavior. Students start each day and class period with a positive adult interaction, thus increasing the rate of effective adult-student relationships and providing antecedent conditions to prompt positive behavior. Generally, a group of 10-30 students can participate in the BEP at a given time. Participating students are those who have consistently violated school-wide expectations, yet whose behavior does not warrant an individualized behavior support plan. Given the increased behavior support provided by the BEP, the rationale is it should be effective in reducing the problem behavior of those at-risk for school failure. Decreases in problem behavior may, in turn, lead to increases in academic engagement (Fairbanks, Sugai, Guardino, & Lathrop, 2007).

Hawken (2006) and March and Horner (2002) evaluated the efficacy of the BEP on reducing problem behavior for middle school students attending a school of approximately 500 students. Effectiveness of the BEP was determined by comparing the mean rate of each student's weekly office discipline referrals (ODRs) prior to the intervention (pre-BEP) to referrals following BEP implementation (post-BEP). The intervention was considered effective if the rate of discipline contacts post-BEP was lower than pre-BEP. March and Horner found the BEP to be effective for 12 of 24 students and Hawken found 7 of 10 students showed reductions in office discipline referrals. Fidelity of implementation was high for the majority of components in both studies.

After introducing the BEP, March and Horner (2002) completed functional behavioral assessment interviews and examined results of the BEP based on behavioral functions. The BEP resulted in decreased office discipline contacts for four out of five students whose problem behavior was maintained by adult attention. Five of eight students whose problem behavior was maintained by peer attention showed a decrease in ODRs post-BEP. The BEP was least effective for students whose problem behavior was maintained by escape from academic demands, with only 3 of 11 students demonstrating decreased ODRs post-BEP. March and Horner concluded the effectiveness of the BEP was associated with the student's behavior, indicating behavior maintained by adult and peer attention is most likely to be responsive to the BEP. Whereas, behavior maintained by escape from academic demands may require an alternate method of secondary support or tertiary support.

Hawken (2006) drew similar conclusions concerning the effectiveness of the BEP, finding office contacts actually increased following implementation of the BEP for students who did not find adult attention reinforcing. In short, students attempted to escape the BEP intervention. Further, Hawken (2006) stated the BEP did not provide enough support for some students who needed a more intensive, individualized behavioral intervention.

In response to the need to consider the function of the student's behavior prior to implementing the BEP, Hawken and Horner (2003) used FBA to identify four middle school students whose problem behavior was determined to be maintained by peer and/or adult attention. Hawken and Horner found that after implementing the BEP, the variability and mean level of problem behavior was reduced for all four students. Furthermore, two students' mean level of problem behavior was similar to typical classroom peers during the intervention phase. While academic engagement was highly variable during baseline, variability decreased and the mean level of academic engagement increased for all four students during the intervention. Because the authors implemented the BEP with students whose behavior was maintained by adult and/or peer attention, students were more likely to demonstrate a reduction in problem behavior while on the program (March & Horner, 2002).

Treatment-outcome studies of the BEP suggest the program is effective in reducing problem behavior for some students, but is ineffective for other students (Hawken, 2006; Hawken & Horner, 2003; March & Homer, 2002). Furthermore, when the BEP is implemented without consideration of the behavior's maintaining function, problem behaviors can be exacerbated. While these studies present strong preliminary evidence to support the utility of the BEP, certain limitations in this body of work must be addressed in future research. For example, March and Horner (2002) and Hawken (2006) did not determine the function of the problem behavior prior to implementing the BEP. As their results clearly indicate, the function of the student's behavior may be a strong determinant of the effectiveness of the intervention and the BEP may be contraindicated for students who find adult attention aversive. To address this limitation, Hawken and Horner (2003) completed the Functional Assessment Checklist for Teachers and Students (FACTS; March, et al., 2000) with all participants at the start of their study. Yet, the results were used for descriptive purposes only and were not used to design the intervention. A more comprehensive functional behavioral assessment, including direct observations of antecedent-behavior-consequence (ABC) patterns as well as a systematic analysis of data collected from interviews and direct observation (e.g., Umbreit, Ferro, Liaupsin, & Lane, 2007) may have been more informative to determine if the BEP would be the most efficient and effective intervention to use with the student participants.

Another limitation of this work was the absence of a true experimental design and emphasis on limited outcomes measures. Both the March and Horner (2002) and Hawken (2006) studies employed a quasi-experimental design, focusing solely on the extent to which the BEP was associated with reductions in ODRs. Although a logical starting point, additional studies using rigorous single case designs with sufficient demonstrations to establish internal validity (see Horner et al., 2005) and randomized control designs (see Gersten et al., 2005) are needed to confirm the findings of these preliminary studies and extend this line of inquiry.

In addition, while all three studies reported high fidelity of implementation for the majority of intervention components, data suggest parents were inconsistent in their participation. One goal of the BEP was to increase school-home collaboration (Crone et al., 2004). Yet, there is relatively limited information regarding parent involvement in the interventions, an area in need of additional inquiry.

Despite the modest findings and limitations of these studies, social validity data suggest teachers and students found the intervention helpful and easy to implement, and would recommend it for other students (March & Homer, 2002; Hawken, 2006; Hawken & Homer, 2003). Given perceptions of social validity may be predictive of treatment fidelity (Lane, Kalberg, et ah 2009; Lane, Little, et al., 2009), it is encouraging to see the parties involved viewed this intervention as feasible and effective.

Purpose

In this study we extend this line of inquiry by examining the impact of the BEP with four middle school students who were not responsive to a comprehensive primary prevention program. Specifically, we (a) conducted a functional behavioral assessment (including interviews and direct observation data) prior to implementing the BEP and (b) employed a changing criterion design (Hartmarm & Hall, 1976) to determine if a functional relation could be established between the introduction of the BEP program and changes in student performance.

Method

Participants

Students. Student participants were 4 eighth-grade boys: Jacob, Curtis, Andy, and John, all of whom were Caucasian. Jacob was 14 years old, whereas, the other three boys were 13 years of age (see Table 1). They attended an inclusive, public middle school in a rural Southern state implementing a three-tiered model of prevention that included academic, behavioral, and social components. One student (Andy) was receiving special education services under the category of Other Health Impaired as determined by a multidisciplinary team (Individuals with Disabilities Education Improvement Act, 2004). All students participated in the standard middle school curriculum for the majority of their 7-period day.

These 4 students were identified by the assistant principal for possible participation in this secondary support due to (a) scoring in the moderate and/or high risk categories on the Student Risk Screening Scale (SRSS; Drummond, 1994; description to follow) which was implemented as part of regular school practices or (b) low levels of work-completion and poor classroom behavior according to teacher reports.

Mentor. One on-site mentor served as the site coordinator for this intervention. He was an African-American, male paraprofessional who was also an athletic coach. He worked with the eighth-grade students in classrooms throughout the day and had a desk in a resource room close to the eighth-grade hallway. The primary role of the mentor was to provide monitoring sheets to students each day, tell each student his daily goal, and monitor student points obtained on their Daily Progress Report (DPR) at the end of the day.

Setting

The school, Foster Garden Middle School served 505 students and employed 34 teachers and two administrators. The school was predominantly Caucasian (96.24%), with 1.78% African American, 1.78% Hispanic, and.20% from other racial backgrounds.

At the time of this study, the middle school was beginning the third year of implementation of their three-tiered model of prevention. The primary plan included academic, behavioral, and social skills supports. The plan specified behavioral expectations for students and teachers in each domain; procedures for teaching the expectations; procedures for reinforcing expectations; as well as an assessment plan for monitoring implementation and student outcomes. Academic and behavioral expectations were skills rated by the majority (> 50%) of teachers as essential for students' success. Social skills expectations included skills and constructs delineated in the Character Under Construction program (Forrest, 2000), a district-wide social skills program. Students were taught school-wide expectations using student-developed instructional videos, assemblies, and posters. Students received PBIS tickets (a small piece of paper with a listing of school-wide expectations) paired with verbal praise on an intermittent schedule of reinforcement for meeting expectations. PBIS tickets could be exchanged for reinforcers (e.g., school supplies, front of the lunch line passes) or entered into lottery type drawings for prizes of greater value (e.g., football tickets or iPod shuffles). This middle school monitored a range of data as part of regular school practices to (a) monitor the overall level of risk evident in the building and (b) identify students who did not respond to the primary prevention plan who might benefit from secondary and tertiary supports. For example, the team monitored grades, ODRs, attendance, referrals for supplemental support (e.g., counseling, alternative learning center) and at-risk status using the Student Risk Screening Scale (SRSS). The SRSS is a mass-screening tool used to identify students who are at moderate or high risk for antisocial behavior. Homeroom teachers rated each of 7 items (a) steal; (b) lie, cheat, sneak; (c) behavior problems; (d) peer rejection; (e) low academic achievement; (f) negative attitude; and (g) aggressive behavior on a 4-point Likert-type scale ranging from never (0), occasionally (1), sometimes (2) to frequently (3). Total scores range from 0 to 21, with high scores indicating higher risk for antisocial behavior patterns. The developer specified three categories: 0-3 low, 4-8 moderate, and 9-21 high risk. Although initially developed for use with elementary school students, recent studies support the utility of this instrument for use in rural and urban middle schools with alpha coefficients exceeding.80 and predictive validity established with ODRs, suspensions, grade point averages (GPA), and course failures (Lane, Bruhn, Eisner, & Kalberg, 2010; Lane, Parks, Kalberg, & Carter, 2007). Participants scoring in the moderate to high risk-range on the SRSS or those with low academic achievement were considered for participation.

Procedures

After securing the necessary approvals from the Institutional Review Board (IRB) and the district, the vice principal identified four eighth-grade students who were not responsive to primary prevention efforts as defined by the previously stated inclusion criteria. Specifically, the vice principal selected students for possible participation in this project if they (a) scored in the moderate or high risk categories on the SRSS or (b) demonstrated low levels of work-completion and poor classroom behavior according to teacher reports to the vice principal. The vice principal met with the teachers of these four students to see if they were potentially interested in Project Function, one targeted support this school offered as part of their integrated model. Three teachers volunteered to participate, with one teacher serving as the point person for two students.

Project staff contacted these volunteer teachers to secure teacher consent. In each case the teacher consented, and sent home the request for parental consent. Once parental consent was returned to the university (100% consented), the research assistants (RAs), all of whom were interns in a Board Certified Behavior Analyst (BCBA) program, sought the assent of each student. The interns and the primary investigator (first author) were not informed of the students' names until after teacher and parental consents were secured. All four students assented. Next, a formal functional assessment was conducted.

Functional Assessment

RAs worked with teachers to conduct a functional assessment prior to beginning the intervention. The assessment involved four components: preliminary information gathering, functional assessment interviews, direct observations, and behavioral rating scales. First, an informal meeting was held with each teacher to (a) identify the students' specific instances of behavior impeding academic success, (b) conduct informal observations of the student during the setting of concern, and (c) schedule the formal functional assessment interview.

Second, a series of interviews were conducted. RAs conducted a functional assessment interview, the Preliminary Functional Assessment Survey (PFAS; Dunlap et al., 1993), with each teacher before or after school or during a planning period. The PFAS is a 22-item survey designed to identify and define the primary target behaviors of concern as well as determine the antecedents prompting the problem behavior and the consequences maintaining the behavior. Next, the PFAS was also conducted with the parents to determine antecedent conditions and maintaining consequences from the parent perspective. Lastly, RAs completed the Student-Assisted Functional Assessment Interview (SAFAI; Kern, Dunlap, Clarke, & Childs, 1994) with each student to obtain their perspectives. The SAFAI contains four sections: initial information, open-ended items, feedback on core content areas, and Likert-type items, focusing on identifying conditions prompting the behavior, specific facets of the target behavior, maintaining consequences, and reinforcers. Third, teachers and parents completed the Social Skills Rating System (SSRS; Gresham & Elliott, 1990). The teacher version (SSRS-T) of the secondary scale contains three sections: social skills, problem behavior, and academic competence. The social skills section contains 30 items constituting three factor analytically derived domains (cooperation, assertion, self-control). Teachers rated the frequency (0 = never, 1 = sometimes, 2 = very often) and importance (0 = not important, 1 = important, 2 = critical for success) of each item on a 3-point Likert-type scale. The problem behavior section contains 12 items equally distributed across two domains (internalizing, externalizing), rated only on the frequency dimension. The academic competence section contains 9 items rated on a 5-point Likert-type scale (1 = lowest 10% to 5 = highest 10%; internal consistency estimates:.78 to.95). The parent version (SSRS-P) contains two sections: social skills and problem behavior and involves the same Likert-type scales as the SSRS-T version. The SSRS-P offers social skills (40 items) and problem behavior (12 items) statements similar to those on the teacher version but places the context in the home setting (internal consistency estimates: 0.65 to 0.87). We used information from both measures to (a) discriminate between skill and performance deficits and (b) confirm behavioral excesses or deficits.

Fourth, RAs conducted three hours of direct observation for each student using A-B-C data collection to confirm the presence of the target behavior and determine the antecedents (A) occurring before the target behavior (B) and the consequences (C) maintaining the behavior. During observation sessions, two RAs, trained in ABC data collection, noted instances of target behaviors as determined by the teachers during the PFAS interview (i.e. off-task, out of seat, profanity, inappropriate volume or tone of voice, refusal to work, and talking out). RAs were seated in the back of the classroom out of the student's direct line of sight. RAs wrote down each instance of the target behavior, the context (e.g., seatwork, small group) and time of occurrence. They recorded the antecedent events occurring prior to the behavior as well as consequent events occurring following the behavior. Each instance of the behavior was analyzed to determine the hypothesized function of the occurrence. This method is based on the assumption that all operant behavior occurs either to seek (positive reinforcement) or avoid (negative reinforcement) certain consequences, typically taking the form of attention, activities/ tangibles, or sensory reinforcement (Umbreit et al, 2007). For each occurrence, the function was recorded and all behavioral occurrences were organized by hypothesized function to determine which function(s) was (were) most salient. For example, reviewing all data may result in a hypothesis such as: the student's noncompliance was maintained by peer attention and escape from difficult tasks.

For each student, data from the informal observations, interviews, rating scales, and direct observations, were placed into a Function Matrix (Umbreit et al., 2007) to determine the hypothesized function(s) (see Table 2). The Function Matrix is a six-celled grid used to analyze data from multiple sources. Specifically, it provides a structure to determine if the behavior occurs to obtain (positive reinforcement) or avoid (negative reinforcement) attention, tangibles/activities, or sensory stimulation. After reviewing the data in the Function Matrix, the RA and teacher developed a statement describing the function(s) of the target behavior (results are depicted in Table 2 and explained in the Results section). After completing the functional assessment process and identifying the hypothesized function of each student's target behaviors, the intervention (described below) was implemented.

Intervention: Behavior Education Program

Materials. A daily progress monitoring form (DPR) was developed based on the information provided from Crone and colleagues (2004) to rate student performance of school rules and work completion in all class periods during the school day (see Figure 1). These forms listed each of the school rules and work completion followed by a 3 point Likert-type scale: 0 = Did not meet expectations at all; 1= Partially met expectations; 2= Always met expectations. A column for ratings and comments was provided for each class period of the school day. Students could earn up to 84 points per day, which were converted to a daily percentage of points earned (range 0 to 100%). Other components included the daily goal, student signature, check-out checklist for the mentor, and a line for parent signature. Parents were asked to sign the DPR each night and were encouraged to provide positive feedback to their child on positive scores and comments on the daily sheet. DPRs were used daily with all students during the intervention.

Procedures. Throughout the study, students checked-in and checked-out with the mentor on each day of school. Every morning, each student met individually with the mentor, received his DPR, and was given his point goal for the day. The RA and mentor worked together to establish point goals for each intervention, with the actual goals established based on the student's initial performance levels (details to follow).

Then, the student carried the DPR to each of his class periods. At the end of each period, teachers rated on the DPR how well the student met each expectation. At the end of the school day, the student met with the mentor to turn in the DPR. The mentor totaled the points earned; computed and graphed the percentage of points earned for the day; and provided positive feedback for good behavior and corrective feedback to improve misbehavior. The DPR was photocopied and sent home for a parent signature and returned the following day.

All participants started the intervention with a goal of 60%. At the outset of the study, the intent was to conduct phase changes by increasing a students7 daily goal after the students met their goal for three consecutive days. However, two weeks into the study, RAs noted changing goals midweek was difficult to manage for the mentor and students. For this reason, goal increase guidelines were modified, and goal increases were made at the beginning of a school week based on students achieving their goal for at least 80% of trials during the previous week. Goal changes were individualized for students based on their response to the intervention.

Mentor training. Prior to the start of the intervention, RAs taught the intervention procedures to the mentor at the school site. RAs explained the goals of the BEP, reviewed the DPR, modeled how to complete the form, and addressed questions. Training continued until the mentor was able to complete all components with 100% fidelity. A fidelity checklist was created as a guide for the mentor to remind him to complete all components of the BEP.

Treatment integrity. Treatment integrity was evaluated in two ways. First, session integrity for check-in and check-out mentor-student meetings was computed using a treatment integrity checklist. The list contained a behavioral listing of all essential intervention components. RAs marked the presence or absence of each component listed. The percentage of treatment integrity was computed by dividing the number of items completed correctly by the total number of components, multiplying the quantity by 100. RAs completed treatment integrity checklists for 20% of mentor and student meetings. Second, the DPRs were evaluated. Specifically, approximately 50% of DPRs were evaluated for treatment integrity across phases and students. Treatment integrity ranged from 52-100% (M = 93%). Treatment integrity data for check-in meetings, check-out meetings, and DPR form completion is reported in Table 3.

Measures

Student outcomes: Compliance. The dependent measure was the percentage of possible points obtained on the daily DPR (range 0 to 100%), referred to as compliance or cooperation.

Social validity. Social validity was evaluated prior to intervention and again following completion of the intervention from the teacher and student perspectives using standardized instruments. Teachers completed the Intervention Rating Profile (IRP-15; Witt & Elliott, 1985) to measure their perceptions of intervention goals, procedures, and outcomes. Teachers rated 15 statements about procedures and outcomes (e.g., "I liked the procedures used in this intervention.") on a 6-point Likert-type scale (strongly disagree = 1 to strongly agree = 6). Total scores range from 15-90, with high scores indicating high acceptability (internal consistency:.88 to.98). The primary teacher prior to implementation completed the IRP-15 after the BEP was explained to them and following the conclusion of implementation. All classroom teachers responsible for completing the DPR post-implementation also completed the IRP-15.

Students completed the Children's Intervention Rating Profile (CIRP, Witt & Elliott, 1985) to measure their perceptions of intervention goals, procedures, and outcomes. Students rate 7-items on a 6-point Likert-type scale (I do not agree = 1 to I agree =6). Total scores range from 7-42, with high scores suggesting high acceptability (internal consistency estimates:.75 -.89). Students completed the CIRP prior to implementation, immediately after the BEP was explained to them and following the conclusion of implementation.

Experimental Design and Statistical Analysis

A changing-criterion design was used to shape students' behavioral performance as measured by the DPR, via gradual increases in goals. As previously mentioned, phase changes initially occurred when students met their respective goals for three consecutive days. However, to address feasibility concerns raised by the mentor during the second week, phase changes occurred when students' met their goals four out of five days each week. We used this design rather than a withdrawal to avoid having to withdraw the intervention. Although traditional applications of the changing-criterion design begin with a baseline phase, we began immediately with the first goal phase (60%) to avoid delaying supports for students, a concern expressed by all teachers. Instead, we included a minimum of four phases for each student allowing at least the three shifts in goals necessary to determine if a functional relation was established between the introduction of a new goal and changes in students' behavior (Horner et al., 2005).

Phase changes were guided by student performance. Jacob's goals changed from 60% to 70% to 80% to 90%. Based on Jacob's difficulty with the 90% goal, his goal was moved to 85%. Curtis' goals changed from 60% to 70% to 80% to 85%. Based on Andy's past experience with interventions, goals were modified to make smaller goal percentage changes for him as compared to other students. His goals changed from 60% to 70% to 75% to 80% to 85%. John's goals changed from 60% to 70% to 75% to 85%. He was given a smaller increase in goal after 70% because he had difficulty meeting the 70% goal consistently during the initial phase. Maintenance data were collected for all students for two weeks following discontinuation of RA support of the intervention. Data were analyzed using traditional visual inspection techniques by examining stability, level, and trend. Also, we report mean and slope changes by phase (Table 3).

Results

Participant #1: Jacob

Functional assessment findings. During the PFAS, Jacob's math teacher stated she was primarily concerned about Jacob's off-task behavior. She also noted Jacob was disorganized and often came to class late without all of his materials. The teacher stated Jacob was a very likeable student who was very social and had a lot of friends. The teacher identified attention from peers as the main function of his off task behavior.

Jacob's mother and father completed the PFAS together regarding Jacob's behavior. They noted they were most concerned about Jacob's off task behavior and his poor attitude. Jacob's parents stated they dealt with Jacob's difficulty remaining on task by providing him with extra reminders, making lists for him, and removing privileges. They stated his off task behavior occurred frequently at home when Jacob was to be completing homework and chores.

When Jacob was asked about his off-task behavior at school during the student interview, the SAFAI, Jacob stated he, "Like(d) to goof around a lot," indicating the, "work is ALWAYS boring so sometimes I don't do it." Also, he stated he "like(d) to make his friends laugh".

Results of the Social Skills Rating System suggested Jacob had average levels of social skills and problem behaviors according to both teacher and parent ratings, with standard scoring in the 85 to 115 range. The academic competence subscale suggested Jacob's performance was within the average level as evidenced by a standard score of 86. Thus, it was determined Jacob had the requisite skills to be engaged, suggesting a performance deficit.

Direct observation data confirmed the information from the interviews. Namely, 23 instances of the target behavior suggested Jacob's off task behavior was maintained by attention and 6 instances were maintained by escape. Data from interviews and direct observation were placed into a Function Matrix (See Table 2) and reviewed by the teacher and two RAs. They collectively hypothesized: when in class, Jacob engaged in off-task behavior to gain peer attention and escape non-preferred tasks.

Outcomes. During the first phase of the intervention, Jacob had a goal of 60% of points on the DPR form (see Figure 2). Jacob exceeded his goal on each of the 3 data points. Data were stable with an average of 72.33% (SD = 1.16) of points earned. There was little variability in scores (71-73%) during the initial phase. Treatment fidelity of the check in and DPR forms were adequate, with mean scores of 83% and 95.70% respectively (see Table 3).

After meeting the criterion for 3 days, Jacob's goal was increased to 70%. Jacob again exceeded his goal on each of the four days, although data were variable with daily DPR scores ranging from 75-100%.

As previously mentioned, phase change decisions were shifted from three consecutive days to four out of five days per week due to difficulties in implementing phase changes reported by the mentor. Jacob's criterion was increased to 80% after meeting the criterion on 4 consecutive days in the previous phase. Jacob exceeded his goal on each day (M = 89.00%; range = 86-95%) and Jacob's DPR percentage scores became more stable (SD = 4.08). There was a clear increase in level and decreased variability relative to the previous phase.

As Jacob had shown consistent achievement of his goals, the RAs and mentor decided to increase Jacob's goal to 90%. Upon this goal change, Jacob's data began to show great variability (range = 26-95%; SD = 21.49), a slight deceleration in trend and a decreased level from the previous phase (M = 76.98%). These changes occurred despite adequate treatment fidelity of the check-in (87.25%), check-out (80%), and DPR forms (89.61%) components raising questions regarding the value of reinforcement (positive reinforcement in the form of attention) relative to the effort required to attention this goal.

In response to the substantial decrease in performance and inability to meet criterion for 3 consecutive days, along with Jacob's comments he did not feel he could reach 90%, the RAs and the mentor decided to decrease Jacob's goal to 85% (the final intervention phase) to allow him a better chance to access reinforcement. Jacob's data at the beginning of the criterion showed an increase in trend, but data became highly variable as the phase progressed (again raising questions regarding relative cost-benefit ratio of effort to reinforcement). Data points in the phase ranged from 0-100% with an average of 81.06% and a slight increase in level from the previous phase.

During the three data points in the 85% criterion maintenance phase, Jacob's data remained variable with a range of 41-96%. His average DPR point percentage was 73% (SD = 28.58). The first two data points approximated the goal, yet the final data point displayed a sharp decline in the percentage of DPR points acquired (reportedly a difficult day).

Social validity. Despite the modest outcomes, Jacob rated the intervention favorably and consistently prior to intervention onset and during the final intervention phase (85% goal), with CIRP values of 36 at both time points. However, the intervention did not meet the math teachers' expectations, with IRP-15 declining from 79 prior to intervention onset to 68 during the final intervention phase. During the post intervention phase, 7 teachers serving Jacob completed IRP-15 forms to obtain their opinions of intervention outcomes. There was tremendous variability in their total score ranging from 15 to 90, with a mean score of 64.29 (SD = 25.52; see Table 3).

Participant #2: Curtis

Functional assessment findings. During the PFAS Curtis's science teacher stated Curtis seemed apathetic about completing work in class and had a negative attitude about class. The teacher noted this behavior was most common during note taking activities, class discussion, and labs. The teacher prioritized work completion as a target behavior and felt Curtis avoided completing work, noting he seemed to enjoy her attention and praise. She typically responded to his incomplete work by moving him to a new seat or sending him to the assistant principal.

Curtis's father completed the PFAS with the RA. Curtis's father noted concerns regarding Curtis's incomplete homework and general attitude. He was most concerned about Curtis's large amount of incomplete schoolwork. In the past, Curtis's father had responded by requiring Curtis to complete the incomplete work at home. He felt Curtis might be neglecting to complete classwork because he was bored, wanted attention from others, or was avoiding work.

During the student interview, Curtis stated, "Work is always boring so sometimes I don't do it." During classroom observations, there were several instances of Curtis stating he had "forgotten" assignments and materials. On one instance he commented to the whole class he completed his work incorrectly because he was "retarded" and then laughed it off.

Results of the Social Skills Rating System suggested, like Jacob, Curtis had average levels of social skills and problem behaviors according to both teacher and parent ratings, with most standard scores in the 85 to 115 range. The academic competence subscale suggested Curtis's performance was within the average level as evidenced by a standard score of 108. Thus, it was determined Curtis had the requisite skills to be engaged, suggesting a performance deficit.

Direct observation data confirmed the information from the interviews. Namely, during the three hours of direct observation, there were 8 instances of the target behavior. All 8 instances appeared to be maintained by attention. Data from interviews and direct observation were placed into the Function Matrix and reviewed by the teacher and two RAs. They hypothesized the following: when in class, Curtis failed to complete work to gain peer and teacher attention and avoid classwork.

Outcomes. During the first phase of the intervention, Curtis's initial goal was set at 60% of DPR points (see Figure 3). During this phase, data showed an accelerating trend (slope 6.50), with all three data points clearly above the goal line. Treatment fidelity of the check in and DPR forms were adequate, with respective mean scores of 83% and 91%.

After meeting criterion for 3 days, Curtis's goal was increased to 70% and an immediate change in level occurred. He exceeded his goal each day (range 82 to 94%), with an average of 86.60% of points earned during this phase. While there was a clear decelerating trend (-3.30) during this phase despite adequate treatment integrity, the final three data points were very stable and above the 70% goal criteria.

Given Curtis exceeded the 70% criteria for all five days, a phase change occurred and his goal increased to 80%. Yet, Curtis was absent for several days, resulting in missing data. When Curtis returned from school, he met criteria on the first day, but not the next two days. All but two of the next data points indicated Curtis met or exceeded the 80% goal. Collectively, data were variable across the 13 data points in this phase (range 56-99%; SD = 12.28). Curtis remained in this phase for 4 weeks, until such time he met criteria for 4 out of 5 days during an instructional week. Curtis's goal was increased to 85% during the final intervention phase.

During this final intervention phase, Curtis met or exceeded the goal required for 10 of the 11 data points. There was a decrease in variability in the data (SD = 7.83; range: 70 to 99%) relative to the previous phase as well as an increase in mean level of performance (M = 89.64%) providing additional evidence of a functional relation.

Upon introduction of the maintenance phase, Curtis showed a rapid drop in DPR point percentage. Performance was variable with a range of 69-100%. His mean DPR point percentage showed a slight decrease in mean level of performance from the previous phase with an average of 81% (SD = 13.70) of points earned across the 6 data points. It should be noted Curtis was absent for several days during this maintenance phase (which occurred just prior to state academic assessments) and as the school year drew to a close.

Social validity. Like Jacob, Curtis rated the intervention favorably and consistently prior to intervention onset and during the final intervention phase (85% goal), with CIRP values of 33 at both time points. The intervention exceeded the science teacher's expectations, with IRP-15 ratings increasing from 76 prior to intervention onset to 87 during the final intervention phase. During the post intervention phase, all 7 teachers serving Curtis completed IRP-15 forms with a mean score of 66.29 and limited variability in their ratings (SD = 1.70).

Participant #3; Andy

Functional assessment findings. During the PFAS Andy's social studies teacher expressed concerns regarding Andy's lack of work completion and tendency to be disrespectful towards adults. She stated Andy "love(d) to talk" and "(did) not like to be told to do his work".

Andy's mother completed the PFAS with the RA. She also expressed concerns about his poor attitude and rebellion. She wondered if these were related to puberty changes. Andy's mother indicated Andy preferred to be outside and it was very difficult to get Andy to work on homework after a full day of school. Andy's mother also noted Andy related better to some of his teachers than others.

During the student interview, Andy stated it "(took) a while" for him to get started on work and some teachers "(didn't) like (him) at all". He said he "(got) mad" when teachers start "saying stuff (he) need(ed) to work on".

Neither the teacher nor the mother completed the Social Skills Rating System. Thus, it was not possible to determine if Andy's social skills, problem behaviors, and academic performance levels were within typical ranges.

Direct observation data confirmed the information from the interviews. During the three hours of direct observations, 41 instances of the target behaviors were determined to be maintained by attention, with 40 also maintained by escape from tasks. Data from interviews and direct observation were placed into a Function Matrix (see Table 2) and reviewed by the teacher and two RAs. They hypothesized the following: when in class, Andy failed to complete work and engaged in disrespectful behavior to gain peer and teacher attention and escape classwork.

Outcomes. During the first phase of the intervention, Andy had a goal of 60% of points on the DPR form (Figure 4). Although data were quite variable, Andy exceeded his goal on each day. Mean level of performance was 82.50% (SD = 13.82), with the percentage of points attained each day ranging from 65% to 98% during this first phase. Treatment integrity was acceptable for the check in components, check out components, and DPR form completion, with respective mean scores of 83%, 80%, and 81.72%, respectively.

After meeting the criterion for 4 days, Andy's goal was increased to 70%. He exceeded his goal on each day, with a mean level score of 82.33% points (SD = 9.87). Performance revealed an increasing trend, with a slope of 9.00 across the three data points in this phase.

Following the third day of meeting criterion, Andy's criterion was increased to 75%. As in the previous phase, he exceeded his goal on each day (M = 80.5%; SD = 0.71; range = 80-81%). Although there were only two data points in this phase, the mentor decided to increase Andy's goal to 80% during the following week as (a) Andy had exceeded the benchmark consistently during both the 70% and 75% performance goals for the last two phases and (b) researchers had decided it would be much easier for the mentor to make goal changes for all students at the beginning of the week rather than midweek.

During the 80% target goal phase, Andy's DPR performances remained relatively consistent with the performance objective, with 4 of the 11 data points falling below the goal line. Mean performance during this phase was 80.73% (SD = 6.83). Andy spent three weeks in this phase before meeting criteria for a phase change.

During the final intervention phase, Andy's goal was increased to 85%. His DPR data remained quite variable during the phase (range = 71-99%), but the average DPR points rose to 89% (SD = 8.85%) suggesting an improving level from the previous phase yet again raising questions about the balance between required effort and value of the existing reinforcers. Only three data points fell below the 85% objective during this phase all of which occurred on Mondays. Treatment fidelity remained high throughout the intervention process, including check in components (100%), check out components (80%), and DPR form completion (89.86%).

During the maintenance phase, Andy's criterion remained at 85%. His mean DPR point percentage showed a slight drop in level with an average of 80% and a range of 69-87% as Andy approached the end of the academic year.

Social validity. Andy's pre- and post-intervention CIRP scores were quite consistent (33 and 31, respectively). Andy's social studies teacher's IRP-15 scores increased slightly from 86 prior to intervention onset to 89 following intervention completion. As with Jacob's case, all 7 teachers serving Andy completed IRP-15 forms to obtain their opinions of intervention outcomes. Their mean score was 62.94, with high level of variability in ratings (SD = 22.94).

Participant #4; John

Functional assessment findings. During the PFAS John's science teacher stated John was slow to start work during each class period, was frequently off-task, and frequently made inappropriate comments in class (e.g., off topic, "whatever") "to make the other kids laugh." The teacher prioritized work completion as a target behavior. The teacher noted John's behavior was probably due to wanting attention from his peers and getting out of work. The teacher's primary method of dealing with his behavior was to verbally reprimand John for failure to get to work.

John's mother completed the PFAS with the RA. John's mother noted concerns regarding John's "forgetfulness" about turning in his work. She stated John does complete his work at home. John's mother was also concerned about verbal aggression towards her as well as his female teachers. She indicated he might be exhibiting aggression to escape work.

During the student interview, John stated he had the most trouble in his science class. He said he forgot to write things down in his journal and his forgetfulness was due to "genetics". He also stated he got angry at teachers and other students when they said "mean things".

Results of the Social Skills Rating System suggested John had average levels of social skills and problem behaviors according to both teacher and parent ratings. However, the teacher rated John as having fewer social skills (standard score [SS] = 83) than did his mother (SS = 102). The academic competence subscale suggested John's performance was within the average level (SS = 93). Thus, it was determined John had the requisite skills to complete the requisite activities, suggesting a performance deficit.

Direct observation data confirmed information from the interviews. Namely, 22 instances of the target behavior were maintained by attention, with 16 instances suggested incomplete work was also maintained by escape. Data from interviews and direct observation were placed into a Function Matrix and reviewed by the teacher and two RAs. They collectively hypothesized: when in class, John failed to complete work to gain peer and teacher attention and escape class-work.

Outcomes. During the first phase of intervention, John's goal was set at 60% of points on the DPR form (Figure 5). He exceeded the goal each day, with a mean level of performance 79.60% (SD = 7.89; range 70% to 88%) during this first phase. Treatment integrity was adequate for check in components, check out components, and DPR form completion with respective scores of 83%, 80%, and 92.90%.

After meeting the criterion for 5 days, John's goal was increased to 70%. DPR percentages during this criterion phase were highly variable ranging from 50% to 100% (M - 73.75%; SD = 20.75). Immediately following the criterion change, there was a sharp decline in John's DPR point percentage. After 5 days below criterion, John's DPR points rapidly increased to percentages well above criterion with the last 3 points far exceeding the 70% goal. Thus, John was informed a phase change would occur the following week, increasing his criteria to 75%.

John was absent during the first three days of this phase. However, he exceeded his goal on the two days for which he was in attendance. Although the RAs intended for John to remain in the same goal phase during the next week, the mentor inadvertently increased the student's goal to 85% the following week.

During this last intervention phase, John's DPR point scores were fairly stable with one data point decreasing below criterion. John's DPR percentages averaged 91.17% with a range of 71-100%, suggesting an increase in level from the previous phase. Treatment integrity remained high during this final phase with check in components at 100% fidelity and check out at 90%.

Despite John's improvements, he elected to withdraw from the study prior to the maintenance phase and before he completed the post social validity form. John stated he no longer wanted to complete the forms and did not want to meet with the mentor.

Social validity. While John's initial CIRP score was positive (35), the fact he withdrew from the study was a clear indication the intervention was not acceptable to John. However, the intervention exceeded the science (target) teacher's expectations, with IRP-15 increasing from 74 to 78. During the post intervention phase, all 7 teachers serving John completed IRP-15 with a mean score of 64.71. There was variability in ratings, with totals ranging from 35-90.

Summary

Results support a functional relation between the introduction of the BEP and changes in students' behavior for Curtis, Andy, and John. Their performance increased to match the reinforcement criterion established in each phase. Yet, maintenance data (collected during the last month before the academic year concluded) were less promising, indicating the behavior change did not consistently maintain. Yet, interventions were rated favorably by both students (save for John) and teachers involved in the initial assessment process. To a lesser extent, post-intervention ratings were relatively favorable for the other teachers serving these students.

Discussion

Many schools across the country are embracing three-tiered models of prevention to (a) prevent the development of learning and behavior problems and (b) respond more effectively to students requiring more intensive supports (Lane, 2007; Sugai & Horner, 2002). While a substantial amount of research has examined the effectiveness of primary and tertiary levels of prevention, less attention has been devoted to exploring the utility of secondary prevention efforts--particularly at the middle school level (Hawken, 2006; Kalberg et al., in press; Robertson & Lane, 2007). The BEP (Hawken et al, 2007) is a secondary prevention program that has met with demonstrated success in supporting middle school students identified as nonresponsive to primary prevention efforts (e.g., Crone et al., 2004).

In this study, we attempted to extend the knowledge base by examining the impact of the BEP with four middle school students who were not responsive to a comprehensive, integrated, primary prevention program. We conducted a functional behavioral assessment prior to implementing the BEP, as there is initial evidence to suggest the BEP may be more effective for students whose target behaviors are maintained by attention (positive reinforcement) rather than escape (negative reinforcement; Hawken, 2006; March & Horner, 2002).

Also, given many practitioners are resistant to withdrawing an intervention once behavior change is observed, we elected to employ a changing-criterion design (Hartmann & Hall, 1976) rather than a more traditional withdrawal design (Kennedy, 2005). A changing criterion design allowed us to evaluate if a functional relation existed between the introduction of the BEP program with the corresponding changes in behavioral goals and changes in student performance. This design, which was modified to eliminate the baseline condition, also enabled all students to begin intervention procedures after teacher consent, parental consent, and student assent were secured. However, given the consenting process took longer for Andy and John, their initial intervention phase occurred approximately 1 (Andy) to 2 (John) weeks after Jacob and Curtis. The changing criterion design allowed us the opportunity to avoid withdrawing the intervention to establish a functional relation as in the case of an ABAB design or delaying intervention to some students as in a multiple baseline design. In short, our goal was to enhance social validity by selecting an amenable intervention design as well as a design that might allow school-site personnel the opportunity to explore the value of existing reinforcers. Given this was their third year of implementation, the school-site team was not only seeking interventions they could implemented with limited university support, but also methodology they also could use in a more independent fashion.

Functional assessment results showed all four students had target behaviors maintained primarily by positive reinforcement in the form of attention and negative reinforcement in the form of escaping non-preferred tasks. However, Curtis (the only student to score a zero on the Student Risk Screening Scale) exhibited the fewest instances of the target behavior (n = 8) and only interview data indicated his poor work completion was maintained by escape as well as attention. The other three students (Jacob, Andy, and John) exhibited their respective target behaviors with higher frequency relative to Curtis. Furthermore, these remaining three boys exhibited behaviors serving a clear dual function: access to attention and escape from tasks/activities. Not surprisingly, all three of these young men scored in the moderate-risk (Andy, 7) or high-risk (Jacob, 10; and John, 16) range on the SRSS during the fall administration.

In terms of intervention outcomes, for Curtis, Andy, and John, there was modest evidence of a functional relation due to the variability in responding evident when criterion levels were increased. This warrants consideration when evaluating the functional relation. Yet, in general students' performance increased to match or exceed the established goals in each phase. Curtis was probably the most responsive student. As you may recall, Curtis exhibited the fewest instances of the target behavior during the functional assessment process and had the least evidence supporting escape motivated behavior. The other three students, who exhibited higher frequencies of their respective target behaviors during the assessment phase and whose target behavior was more equally maintained by attention and escape, were somewhat less responsive to intervention efforts than Curtis. However, partial evidence of a functional relation was still established for Andy and John, although Andy struggled to meet criterion each Monday during the 85% goal phases (perhaps due challenges associated with returning to school after being home for the weekend). Yet, Andy was able to consistently meet criterion the remaining days of the week. The exception was Jacob.

In looking at Jacob's performance between the 80% goal phase and the 90% goal phase, it is quite possible the 90% goal was perceived as too challenging or requiring too much effort, a concern Jacob raised himself during one of the interactions with the mentor. This perceived challenge may have prompted variability in performance, with levels far below the first goal phase 60%. It is possible had a goal of 85% been the final target following 80%, Jacob's behavior may not have been so variable. Moving forward it will be important to establish reasonable goals for the students and to make gradual changes in criterion.

Despite the generally favorable outcomes during each intervention phase, maintenance of effects was limited for all students. This may have been due to the end of the school year, bringing with it many competing reinforcers (e.g., time spent socializing with friends before summer break), the pressures associated with the pending state assessments to determine academic performance, or the loss of the novelty effect (Cooper, Heron, & Heward, 2007).

Clearly variability existed in terms of student performance over time. Based on the pattern of student responding, it may be that the increased criterion made the effort not consistently "worth it" for the students resulting in variability of student performance. Another potential source of variability in scores could be the result of potential variability in teachers' judgments regarding student performance. This cautious interpretation leaves unanswered questions. For example, was this an intervention producing changes in students' behavior? Or, is it possible, there were changes in teachers' behavior? Or were there changes in student and teachers' behavior? While the design of the current study does not allow for full exploration of these questions, we encourage future inquiry to explore the potential transactional relations that might have existed with respect to behavior changes.

Yet, despite these unanswered questions, interventions were rated favorably by most students--except for John who elected to withdraw from the study just prior to the maintenance phase. Also, the teachers most directly involved in the initial assessment process rated the intervention favorably.

With the exception of Jacob's case, the intervention met teachers' expectations according to the IRP-15 ratings for teachers most directly involved with each student (e.g., completing the intervention). To a lesser extent, post-intervention social ratings were relatively favorable for the other teachers serving these students. While the overall teachers' ratings were comparable with mean scores ranging from 62.94 (Andy) to 66.29 (Curtis), there was variability in teachers' perceptions for most students.

Limitations and Future Directions

We encourage the readers to interpret the findings presented in light of the following limitations. First, only one teacher--a volunteer--participated in the functional assessment process to determine the reason why each student engaged in his respective undesirable behaviors (e.g., off-task, disrespectful behavior, and incomplete work). Had additional teachers been invited and willing to participate in the process, a more accurate picture of maintaining consequences might have been developed. When intervening in middle and high schools, we encourage future studies to solicit participation from all teachers serving the students of interest. In this way, a more thorough assessment could be conducted and plans for promoting generalization can be better developed (Cooper et al., 2007).

Second, the dependent variable in this study was the percentage of points earned on the DPR form. While this is an adequate variable, had resources permitted, it would have been beneficial to also collect data on each student's respective target behaviors. Clearly, this would have been very labor intensive, requiring direct observations for multiple students across multiple periods within each school day. Given the situational specificity of behavior (Kazdin, 1987), which is clearly reflected in the collective IRP-15 ratings completed by teachers serving each of the four students, such information would provide a more detailed view of intervention outcomes across the school day. We recommend experts in the BEP process secure external funding to conduct more comprehensive investigations addressing this limitation and explore the potential shift that may have occurred in terms of teachers' behavior.

Third, absenteeism appears to have impacted intervention outcomes. For example, consider Curtis's performance. At the end of the 70% goal phase, the mentor informed Curtis he would be changing goals (increased to 80%) at the beginning of the following week. Thus, the phase change was decided at the end of the week as part of intervention procedures. Unfortunately, Curtis was absent for several days the next week. This may have contributed to the initial variability in the 80% goal phase. The lack of absenteeism may partially explain why there was less variability and a higher mean level performance in the 85% goal phase relative to the 80% goal phase. Curtis was also absent for several days during the maintenance phase which may have contributed to the variability in performance during this phase. Future study may be improved by attending to issues associated with absenteeism. It may be better to inform students of their intervention goals at the onset of the week--when the mentor can be certain the student was present. When conducting randomized control trials of BEP, it would be interesting to build attendance into the statistical model as a moderating variable. This would provide additional information regarding the implications of absenteeism with respect to intervention outcomes.

Fourth, a related limitation pertains to goal setting. As we mentioned previously, we speculate some of the variability in Jacob's performance may have been due to his perception that a 90% goal was too stringent. In other words, his perception of his ability to attain this goal may have negatively impacted his motivation to succeed (Lane, Kalberg et al., 2009). To this end, we recommend future studies consider assessing students' perception of the feasibility of goals before instituting phase changes. It may be performance will be enhanced if students play a role in the goal setting process (Lee, Palmer, & Wehmeyer, 2009). Future inquiry is needed to explore the issue of balance between effort required to access reinforcers and the student's perception of reinforcer value. At what point does a given reinforcer simply lack sufficient value? And, at what point should a secondary support such as the BEP be determined insufficient and a tertiary support be introduced?

Fifth, and related to the value of reinforcers, the potential for novelty effects on the outcome of the invention is a possible limitation. For example, in Jacob's case, we question his declining performance during the last 4 data points during the final intervention phase. All students were informed when they began the final intervention phase. It is possible finding out the intervention was drawing to a close and there would not be any additional phase changes led to the variability and declining performance for Jacob at the end of the final intervention phase.

Finally, the extreme variability in the data across phases for most students may indicate relatively weak reinforcement contingencies relative to the existing reinforcement contingencies for engaging in the less than desirable target behaviors (Cooper et al., 2007). Couple this with the fact that some participants' target behaviors were dual-functioned behaviors maintained by attention and escape; it is not surprising a clear, unambiguous functional relation was not established between the intervention (with varying goals criteria) and the resulting DPR scores for all students. Previous research indicated BEP intervention efforts are perhaps less effective for participants for whom adult attention is aversive (March & Horner, 2002). The escape maintained behavior noted in this study may suggest that some of the adult attention (e.g., giving instructions) may have been aversive to student participants. If so, future research will need to examine how the BEP could be modified to better support students whose behavior is also maintained by escape. For example, it might be possible to include a bonus clause that allows a student to escape a non-preferred activity (e.g., homework pass) or interaction (e.g., meeting with the mentor or teacher for one day the following week) if they meet DPR score criteria for a specified amount of time (e.g., for the entire 5 day week). In the current study it is possible the DPR monitoring procedures may have assured that the participants could not escape the impact of the behavior on the DPR results. Their behavior eventually conformed to the expectations or improved even more. This might have been a sign of the potentially reinforcing effects of engaging in behavior matching teachers' expectations that resulted in receiving positive feedback for meeting expectations.

Summary

Despite the limitations noted above, this study offers additional evidence for the utility of the BEP program with middle school students who are not responsive to primary prevention efforts. Results suggest this intervention is generally effective for students whose challenging behaviors are maintained by attention and escape. Findings also suggest the intervention is particularly effective for students with lower frequency behaviors and when goal changes are implemented incrementally.

In addition to being a relatively effective intervention, findings support the feasibility of using scientifically rigorous, single case designs within the context of three-tiered models of prevention. In this case a changing criterion design was used to enhance feasibility for the school site by allowing all students to participate in the BEP simultaneously and avoid withdrawal. Furthermore, the changing criterion design may be of value to school-site personnel, serving as a vehicle for identifying the value of the reinforcers involved, determining when a given criterion may be too stringent, and potentially determining when secondary supports are insufficient. For example, in terms of the value of the reinforcers, the changing criterion design may be particularly useful in examining existing contingencies (e.g., in the maintenance phase) that fall short of desired performance patterns. Furthermore, it appears in some instances in the current study students began under-performing when the criterion was perhaps set too high. This may have indicated positive reinforcement from the mentor was insufficient and a more powerful reinforcer necessary. Or, another possibility is students whose behavior does not maintain or respond to the established reinforcers (positive reinforcement in the form of attention from the mentor) may require more intensive supports such as tertiary intervention efforts (Lane, Kalberg, & Menzies, 2009).

We suggest the experiments presented in this paper were methodologically and pragmatically successful as defined by the Institute of Education Sciences. The methodology included recommended quality indicators such as (a) described participants and selection criteria in sufficient detail to allow replication; (b) precise definitions of the dependent and independent variables; (c) reliability of the dependent and independent variables; (d) established experimental control for three experiments; and (e) assessed social validity from multiple perspectives (Horner et al., 2005). In addition, the study was pragmatically successful in that modest evidence for experimental control was established for three of the four students - with the fourth student later being referred to and placed in special education services as specified in the Individuals with Disabilities Education and Improvement Act (IDEA, 2004).

We hope this article will be another piece of evidence to support the utility of the BEP within the context of three-tiered models of prevention. Furthermore, we hope our suggestions for future investigation will be considered by those seeking to conduct scientifically rigorous and feasible secondary prevention programs within these models (Lane, 2007).

References

Cooper, J. O., Heron, T. E., & Heward, W. L. (2007). Applied behavior analysis. Upper Saddle River, N.J.: Pearson Education, Inc.

Crone, D. A., Horner, R. H., & Hawken, L. S. (2004). Responding to problem behaviors in schools: The behavior education program. New York: The Guildford Press.

Drummond, T. (1994). The Student Risk Screening Scale (SRSS). Grants Pass, OR: Josephine County Mental Health Program.

Dunlap, G., Kern, L., dePerczel, M., Clarke, S., Wilson, D., Childs, K., ... Falk, G. (1993). Functional analysis of classroom variables for students with emotional and behavioral disorders. Behavioral Disorders, 18, 275-291.

Elliott, S. N., & Gresham, F. M. (2007). Social Skills Improvement System: Classwide intervention program guide. Bloomington, MN: Pearson Assessments.

Fairbanks, S., Sugai, G., Guardino, D., & Lathrop, M. (2007). Response to intervention: Examining classroom behavior support in second grade. Exceptional Children, 73, 288-310.

Forrest, D. B. (2000). Character under construction. Chapin, SC. Youth-light.

Gersten, R., Fuchs, L. S., Compton, D., Coyne, M, Greenwood, C, & Innocenti, M. S. (2005). Quality indicators of group experimental and quasi-experimental research in special education. Exceptional Children, 71, 149-164.

Gresham, F. M., & Elliott, S. N. (1990). Social Skills Rating System. Circle Pines, MN: American Guidance Service.

Hartmann, D. P., & Hall, R. V. (1976). The changing criterion design. Journal of Applied Behavior Analysis, 4, 527-532.

Hawken, L. S. (2006). School psychologists as leaders in the implementation of a targeted intervention: The Behavior Education Program. School Psychology Quarterly, 21, 91-111.

Hawken, L. S., & Horner, R. H. (2003). Evaluation of a targeted intervention within a schoolwide system of behavior support. Journal of Behavioral Education, 12, 225-240.

Hawken, L. S., MacLeod, K. S., Rawlings, L. (2007). Effects of the Behavior Education Program (BEP) on office discipline referrals of elementary school students. Journal of Positive Behavior Interventions, 9, 94-101.

Horner, R. H., Carr, E. G., Halle, J., McGee, G., Odom, S. L., & Wolery, M. (2005). The use of single-subject research to identify evidence-based practices in special education. Exceptional Children, 71,165-179.

Individuals with Disabilities Education Improvement Act of 2004, 20 U.S.C. 1400 et esq. (2004). (reauthorization of Individuals with Disabilities Act 1990).

Kalberg, J. R., Lane, K. L., & Lambert, W. (in press). The utility of conflict resolution and social skills interventions with middle school students at risk for antisocial behavior: A methodological illustration. Remedial and Special Education. Advance online publication, doi: 10.1177/0741932510362514

Kazdin, A. E. (1987). Treatment of antisocial behavior in children: Current status and future directions. Psychological Bulletin, 102,187-203.

Kennedy, C. H. (2005). Single-case designs for educational research. Boston, MA: Allyn & Bacon.

Kern, L., Dunlap, G., Clarke, S., & Childs, K. E. (1994). Student-assisted functional assessment interview. Diagnostique, 19, 20-39.

Lane, K. L. (2007). Identifying and supporting students at risk for emotional and behavioral disorders within multi-level models: Data driven approaches to conducting secondary interventions with an academic emphasis. Education and Treatment of Children, 30,135-164.

Lane, K. L., Bruhn, A. L., Eisner, S. L., & Kalberg, J. R. (2010). Score reliability and validity of the Student Risk Screening Scale: A psychometrically-sound, feasible tool for use in urban middle schools, journal of Emotional and Behavioral Disorders, 18, 211-224.

Lane, K. L., Kalberg, J. R., Bruhn, A. L., Driscoll, S. A., Wehby, J. H., & Elliott, S. (2009). Assessing social validity of school-wide positive behavior support plans: Evidence for the reliability and structure of the Primary Intervention Rating Scale. School Psychology Review, 38, 135-144.

Lane, K. L., Kalberg, J. R., & Menzies, H. M. (2009). Developing school-wide programs to prevent and manage problem behaviors: A step-by-step approach. New York, N.Y.: Guilford Press.

Lane, K. L., Kalberg, J. R., & Shepcaro, J. C. (2009). An examination of quality indicators of function-based interventions for students with emotional or behavioral disorders attending middle and high schools. Exceptional Children, 75, 321-340.

Lane, K. L., Little, M. A., Casey, A. ML, Lambert, W., Wehby, J. H., Weisenbach, J. L., & Phillips, A., (2009). A comparison of systematic screening tools for emotional and behavioral disorders: How do they compare? Journal of Emotional and Behavioral Disorders, 93-105.

Lane, K. L., Parks, R. J., Kalberg, J. R., & Carter, E. W. (2007). Systematic screening at the middle school level: Score reliability and validity of the students risk screening scale. Journal of Emotional and Behavioral Disorders, 15, 209-222.

Lee, S., Palmer, S., & Wehmeyer, M. (2009). Goal setting and self-monitoring for students with disabilities: Practical tips and ideas for teachers. Intervention in School and Clinic, 44, 139-145. Retrieved from ERIC database.

March, R.E. & Horner, R.H. (2002). Feasibility and contributions of functional behavioral assessment in schools. Journal of Emotional and Behavioral Disorders, 10,158-170.

March, R., Lewis-Palmer, L., Brown, D., Crone, D., Todd, A. W., & Carr, E. (2000). Functional assessment checklist for teachers and staff (FACTS). Education and Community Supports. Eugene, Oregon: University of Oregon.

McCurdy, B. L., Kunsch, C, & Reibstein, S. (2007). Secondary prevention in the urban school: Implementing the Behavior Education Program. Preventing School Failure, 51,12-19.

Robertson, E. J., & Lane, K. L. (2007). Supporting middle school students with academic and behavioral concerns within the context of a three-tiered model of support: Findings of a secondary prevention program. Behavioral Disorders, 33, 5-22.

Sugai, G. & Horner, R. (2002). The evolution of discipline practices: School-wide positive behavior supports. Child and Family Behavior Therapy, 24, 23-50.

Turnbull, A., Edmonson, H-, Griggs, P., Wickham, D., Sailor, W., ... Freeman, R. (2002). A blueprint for schoolwide positive behavior support: Implementation of three components. Exceptional Children, 68, 377-40.

Umbreit, J., Ferro, J., Liaupsin, C, & Lane, K. (2007). Functional behavioral assessment and function-based intervention: An effective, practical approach. Upper Saddle River, N. J.: Prentice-Hall.

Witt, J. C., & Elliott, S. N. (1985). Acceptability of classroom intervention strategies. In T. R. Kratochwill (Ed.), Advances in schoo psychology (Vol. 4., pp. 251-288). Mahwah, NJ: Earlbaum.

Kathleen Lynne Lane, Andrea M. Capizzi, Vanderbilt University

Marisa H. Fisher Vanderbilt University, Kennedy Center

Robin Parks Ennis Georgia State University

Correspondence to Kathleen Lynne Lane, School of Education, 201-E Peabody Hall, Campus Box 3500, University of North Carolina at Chapel Hill, Chapel Hill, NC 27599-3500; e-mail: Kathleen.Lane@unc.edu.
Table 1

Characteristics of Students with At-Risk Behavior

                           Student

Variable         Jacob      Curtis         Andy        John

Demographics

Age                   14         13            13         13

Gender              Male       Male          Male       Male

Ethnicity      Caucasian  Caucasian     Caucasian  Caucasian

Special               No         No           Yes         No
Education                 ADHD (a),  Other Health
& Special                   but not    Impairment
Needs                     receiving      for ADHD
                           services

SRSS (b)

Fall                  10          0             7         16

Winter                 5          0             7         14

Spring                 3          0            13          0

SSRS-T (c)

Social Skills         91         84             *         83

Problem               86        113             *        115
Behavior

Academic              86        108             *         93
Competence

SSRS-P (d)

Social Skills         88        112             *        102

Problem               86         89             *        106
Behavior

Notes. (a) ADHD refers to Attention Deficit Hyperactivity Disorder.
(b) Student Risk Screening Scale (Drummond, 1994): raw scores range
from 0 to 21, with higher scores indicating greater risk for
antisocial behavior. Categories are as follows: 0-3 = low risk,
4 - 8 = moderate risk; and 9-21 - high risk;
(c) Social Skills Rating System-Teacher version (Gresham & Elliott,
1990) - standard scores (M = 100; SD = 115);
(d) Social Skills Rating System-Parent version (Gresham & Elliott,
1990) - standard scores (M = 100; SD = 115).


Table 2

Function Matrices

             Jacob's Target Behavior and Function Matrix

Off-Task     Any behavior during the instructional period
             outside the scope of instructional requirements and not
             related to completing assignments and/or listening to
             directions and instruction from the teacher.

Function     Positive Reinforcement      Negative Reinforcement

Attention    ABC Data Instances: 23
             Teacher interview:
             Identified peer attention
             as a motivator.
             "He likes to talk to his
             friends in class"; "He is
             very social"
             Parent interview: "He
             likes to talk to his
             friends on the phone"
             Student interview:
             "like(d) to make his
             friends laugh"

Activities/                              ABC Data Instances: 6
Tangibles                                Parent interview: "He
                                         doesn't bring his homework
                                         home after school"
                                         Student interview:
                                         "like(d) to goof around a
                                         lot"

Sensory

             Curtis's Target Behavior and Function Matrix

Incomplete   Any class-work that is finished beyond the
Work         allotted time or never completed

Function     Positive Reinforcement      Negative Reinforcement

Attention    ABC Data Instances: 8
             Teacher interview:
             identified attention and
             teacher praise as
             effective reinforcers
             Parent interview: "Curtis
             likes attention"; "Curtis
             likes doing things with
             me"

Activities/                              Teacher interview:
Tangibles                                identified escape from
                                         work as a negative
                                         reinforcer, "He
                                         consistently avoids his
                                         work"
                                         Parent interview: "He is
                                         not very motivated to do
                                         his school-work"
                                         Student interview: "Work
                                         is always boring, so
                                         sometimes I don't do it"

Sensory

             Andy's Target Behavior and Function Matrix

Incomplete   Any class-work that is finished beyond the allotted
Work         time or never completed.

Disrespect   Language or actions toward a teacher that are
             impolite or discourteous, such as profanity,
             derogatory, language, talking back, direct
             defiance, and/or refusal.

Function     Positive Reinforcement      Negative Reinforcement

Attention    ABC Data Instances: 41
             Teacher and student
             interviews: identified
             attention as an effective
             reinforcer,
             Teacher interview:
             "love(d) to talk"; will
             return to work after
             redirection from the
             teacher. Parent interview:
             "likes being with older
             kids"; " has always been
             rowdy"
             Student interview:
             "Friends are in my classes
             [and I] talk to them"

Tangibles/                               ABC Data Instances: 40
Activities                               Teacher and student
                                         interview: identified
                                         escape from work as a
                                         negative reinforcer.
                                         Teacher interview: "(did)
                                         not like to be told to do
                                         his work".
                                         Parent interview: "After a
                                         full day of school, he
                                         can't sit still to do
                                         work."
                                         Student interview: "(took)
                                         a while" for him to get
                                         started on work and that
                                         some teachers "(didn't)
                                         like (him) at all.

Sensory

             John's Target Behavior and Function Matrix

Incomplete   Any class-work that is finished beyond the allotted
Work         time or never completed.

Function     Positive Reinforcement      Negative Reinforcement

Attention    ABC Data Instances: 22
             Teacher interview:
             identified attention as a
             reinforces "He likes to
             make the other kids
             laugh"; "I reprimand him
             right away"
             Parent interview: "We talk
             about why he is upset
             right away"
             Student interview: likes
             "extra time with friends"

Activities/                              ABC Data Instances: 16
Tangibles                                Teacher interview:
                                         identified escape from
                                         work as a negative
                                         reinforcer; he "tries to
                                         escape work"
                                         Parent interview:
                                         "Sometimes he doesn't want
                                         to do work" Student
                                         interview: "Likes breaks"

Sensory


Table 3

Intervention Outcomes and Treatment Integrity

Student  Phase      Outcomes:                    Treatment Integrity
         Criterion  Percentage of DPR Points     Components
         (No.       Earned
         Points)

                                                 Check       Check
                                                 In          Out
                    M(SD)          Slope         M           M
                                   (S yx)

Jacob      60% (3)   72.33 (1.16)  -1.00 (0.82)          83
           70% (4)  86.00 (10.74)         -0.40         100         80
                                        (13.14)
           80% (4)   89.00 (4.08)   -1.00(4.74)                     80
          90% (10)  76.98 (21.49)         -1.89       87.25         80
                                        (21.97)
          85% (13)   81.06(25.72)         -2.33       93.33
                                        (25.13)
         Main. (3)  73.00 (28.58)        -20.50
                                        (28.17)

Curtis     60% (3)   78.33 (7.23)   6.50 (4.49)          83
           70% (5)   86.60 (5.90)  -3.30 (3.17)         100
          80% (13)  84.39 (12.28)  0.76 (12.44)          81         80
          85% (11)   89.64 (7.83)   -1.04(7.41)         100         80
         Main. (6)   81.00(13.70)  0.40 (15.29)

Andy       60% (4)   82.50(13.82)  3.40 (16.05)          83         80
           70% (3)   82.33 (9.87)   9.00 (5.72)                     80
           75% (2)   80.50 (0.71)         1.00*          83
          80% (11)   80.73 (6.83)    1.06(6.18)       87.67         90
          85% (13)   89.00 (8.85)   0.31 (9.16)         100         80
         Main. (5)   80.00 (6.78)  -0.50 (7.79)

John       60% (5)   79.60 (7.89)  -2.70 (7.67)          83         80
           70% (8)  73.75 (20.75)  6.93 (13.28)        91.5         80
           75% (2)    86.00(1.41)       -2.00 *
           85% (6)   91.17(10.57)   0.77(11.71)         100         90

Student                       Social Validity

         DPR           CIRP             IRP-15
                       (a)              (b)
         Form

                              Target    All (c)
         M(SD)         Total  Total     M(SD)

Jacob    95.70 (4.93)     36        79
         94.62 (4.93)
          98.39(1.86)
                89.61
              (14.68)
                88.27     36        68    64.29
              (13.94)                   (25.52)
         98.39 (2.28)

Curtis   91.40 (3.72)     33        76
         97.85 (3.72)
          92.8 (4.40)
                96.48     33        87    66.29
              (11.67)                    (1.70)
               100.00
               (0.00)

Andy     81.72 (6.72)     33        86
          92.47(1.86)
          91.94(2.28)
         92.96 (8.51)
         89.86 (7.55)     31        89    62.94
                                        (22.94)
         93.55 (0.00)

John     92.90 (5.77)     35        74
         91.94 (5.68)
              97.00 *
              94.00 *     **        78    64.71
                                        (18.56)

Note. DPR refers to Daily Progress Report Forms. * only two data
points in this phase, therefore, there was no variability about
the regression line. (a) Children's Intervention Rating Profile
(CIRP; Witt & Elliott, 1985) (b) Intervention Rating Profile
(IRP-15; Witt & Elliott, 1985) (c) For each student, the IRP-15
was given post-intervention to all of the teachers with whom the
student had a class. ** student withdrew.
Gale Copyright:
Copyright 2012 Gale, Cengage Learning. All rights reserved.