Preventing Disruptive Behavior in the Urban Classroom:Effects of the Good Behavior Game on Student and Teacher Behavior
Teachers are often ill-prepared to manage classrooms in urban schools. In the present study, an empirically-based behavioral management strategy, the Good Behavior Game (Game), was investigated. The effects of the Game on student behavior and teacher response statements, including praise, were examined. A teacher with 22 students in a first grade classroom of an urban elementary school participated in implementation of the Game. Using a withdrawal design, results showed that student on-task behavior increased while disruptive behavior decreased, replicating previous findings. The number of teacher praise statements remained at near zero levels across conditions. Frequency of teacher neutral and negative statements varied with the level of student disruptive behavior. Teacher praise and limitations are discussed.
group contingency, Good Behavior Game, urban schools, disruptive behavior, teacher praise
In this age of NCLB, schools are struggling to close the achievement gap that exists between groups of students in American schools. No where has this issue become more apparent than in urban schools which serve the highest percentage of poor, minority, non-English speaking and special needs students. For these schools, closing the gap has been achieved with a great deal of effort and sacrifice on behalf of teachers, administrators and even students and their families, including increased time in after school tutoring sessions, increased time devoted to testing, and greater emphasis on content covered. This leaves the question, why are these outcomes so difficult to achieve in urban schools?
Many researchers point to the fact that urban schools are in a state of deterioration (Lippman, Burns, & McArthur, 1996; Noguera, 2003). Although urban schools are typically defined by high concentrations of poverty, a particularly robust risk factor for several undesirable conditions, a comparison of urban and non-urban schools [End Page 85] serving student populations with similar poverty levels reveals that urban schools are further distinguished by, among other concerns, (a) higher rates of student mobility, (b) difficulty hiring teachers, and (c) a greater percentage of students evidencing classroom discipline problems (Lippman et al., 1996). It is clear that teachers electing to teach in urban schools must come prepared with equally refined skill arsenals of pedagogy and behavior management.
Students in urban schools enter the classroom with diverse academic and behavioral needs. School readiness studies of children who grow up in urban areas show that exposure to several critical risk factors related to early health and caretaking, including birth to a single parent, birth to a teenage mother, low birth weight, child maltreatment, and exposure to lead all significantly and negatively impact school adjustment (Weiss & Fantuzzo, 2001). Urban school classrooms taught by poorly-prepared, or even novice, professionals with little or no city school experience place students at further risk for impairment, including chronic patterns of antisocial behavior and even conduct disorder (Kellam, Ling, Merisca, Brown, & Ialongo, 1998; Walker, Ramsey, & Gresham, 2003/04). These teachers over-rely on reactive and aversive strategies in the absence of planned preventive and educational approaches to address classroom discipline despite mounting evidence supporting the ineffectiveness of such an approach (Van Acker, Grant, & Henry, 1996). When the situation becomes so intolerable, the teacher leaves (i.e., walks out or formally resigns).
Graziano (2005) indicates that teacher preparation is the key to reducing attrition rates. Specifically, she suggests that training in learning theory alone can reduce teacher attrition rates by as much as 16 percent. Although professional development is a necessary step, it is often not sufficient to improve teacher management strategies (Sawka, McCurdy, & Mannella, 2002; Shapiro, Miller, Sawka, Gardill, & Handler, 1999). For example, Sutherland, Wehby, and Copeland (2000) demonstrated a functional relationship between a teacher's use of behavior-specific praise and performance feedback by an observer trained to assess the rate of praise in a class for students with emotional/behavioral disorders. Results demonstrated that higher rates of behavior-specific praise occurred only in response to observer feedback. Other research has indicated that teacher praise for students at risk for aggressive behavior, while generally low overall, occurs mostly in response to correct academic responses (Sutherland et al., 2000; Sutherland, Wehby, & Yoder, 2002; Van Acker et al., 1996; Wehby, Symons, & Shores, 1995). Furthermore, attempts to increase correct academic responding have resulted in concomitant increases in teacher praise (Sutherland et al., 2002). However, there has not been an investigation of teacher praise with disruptive behavior. [End Page 86]
Given that teachers are often ill-prepared to effectively manage urban classrooms (Graziano, 2005) and often require repeated follow-up to improve skills (Sutherland et al., 2000), it is essential to provide teachers with easy-to-implement interventions with demonstrated effectiveness in decreasing problem behavior. One such strategy is the Good Behavior Game (i.e., the Game). The Game is an interdependent group contingency that is user-friendly and is applied class-wide. It has benefited from multiple studies demonstrating its effectiveness in improving appropriate student behavior and decreasing disruptive behavior. The Game was first applied successfully in a classroom of fourth grade students with high rates of talk-outs and out-of-seat behavior (Barrish, Saunders & Wolf, 1969). Since then, numerous investigations have been implemented to evaluate the impact of the Game on disruptive behavior in various school settings, including regular and special education classrooms (Darveaux, 1984; Salend, Reynolds, & Coyle, 1989), the school library (Fishbein & Wasik, 1981) and with a variety of age levels including elementary- as well as secondary-age students (Salend et al., 1989; Werthamer-Larsson, Kellam, & Wheeler, 1991).
Teacher praise and the Game are two empirically-based effective classroom management tools. While the Game has been shown to be effective in reducing rates of disruptive behavior, there has not been an investigation of teacher behavior with the Game. Specifically, teacher praise has not been measured within the context of the Game. The purpose of the present study was to evaluate the impact of the Game on student behavior when implemented at the class-wide level in an urban elementary school. It was hypothesized that disruptive behavior would decrease while on-task behavior would increase with implementation of the Game. This investigation was designed to extend the research on the Game by evaluating its impact in an urban classroom serving a population of students characterized by a high level of poverty and also evaluating the collateral effects on teacher behavior, specifically the teacher's use of praise. Similar to other investigations designed to examine the rate of teacher praise, it was hypothesized that improved student behavior, including more frequent opportunities to respond, would lead to higher rates of teacher praise.
Participants and Setting
The study was conducted in a general education first-grade classroom of an elementary school in a large urban area in the Northeastern U.S. The school is comprised of 462 students in grades K to 5 with approximately 92% of students receiving free or reduced price [End Page 87] lunch. Most recent school achievement data indicate that 16% of the students are proficient in reading and math. The classroom in which the study was conducted was comprised of 22 students, 11 males and 11 females. The intervention agent was a female teacher who was identified by administration as experiencing difficulty with classroom management. Moreover, parents of her students were requesting classroom transfers due to poor classroom management.
Behavior Definition and Measurement
The dependent measures included student on-task and disruptive behaviors as well as teacher response statements. Student on-task behavior was defined as the student attending to the assigned work or teacher (i.e., having eyes oriented to work or teacher). On-task behavior included both active and passive forms (e.g., looking at teacher during lecture, writing answer to math worksheet). Student disruptive behavior was defined as any behavior that is not included in the on-task category such as academically unrelated verbal (e.g., call outs, talk to other students) or motoric (e.g., out-of-seat, throwing objects) behaviors. A broad category of disruptive behavior was chosen, so as to represent the host of behaviors that would typically elicit reprimands by a teacher during instruction or independent seatwork (e.g., staring around room, out-of-seat, tapping pencil, talking to a peer). On-task behavior was measured through momentary time sampling, and disruptive behavior through partial interval recording. Momentary time sampling requires the observer to record whether the student is engaging in the defined on-task behavior at the moment the cue is heard through the earphone. Partial interval recording requires the observer to record whether the student engages in the defined behavior at any time during the interval. The observation procedure involved observing one student for one 10-second interval in a rotating fashion across the teams for three consecutive intervals. This resulted in 45 student observations for each 10-minute observation.
Teacher response rate was defined along 3 categories: (1) positive, consisting of a praise statement following a student behavior (behavior-specific praise was not required); (2) neutral, consisting of statements that do not have a positive, negative, or instructional connotation; and (3) negative, consisting of a warning or negative response to a student behavior. Teacher response statements were recorded with a frequency count. Teacher behavior was recorded every fourth interval, resulting in 15 observations of teacher behavior per 10-minute observation. [End Page 88]
The study employed an ABAB withdrawal design. Phase changes were initiated following a stable rate of student behavior or trend in an undesired direction.
Reinforcer Preference Assessment
A reinforcer survey listing seven tangible rewards, that were approved by the teacher, was created on 8½" x 11" paper.
A script was devised for the teacher to train the students. The script outlined and explained the procedures, rules for the Game, examples and nonexamples of appropriate "Game" behaviors, and a brief role play of the Game.
Good Behavior Game
The following materials were required for daily implementation of the Game: the Game integrity checklist, the Game rules, a recording sheet, weekly chart, and kitchen timer. The Game integrity checklist listed the 10 steps involved in daily implementation of the Game, from posting the recording sheet to distributing rewards to winning teams. The Game rules were listed on 8½" x 11" paper. The recording sheet was on 11" x 17" paper and contained four boxes with the team names (blue, white, black, and gold) with a space for each team's tally, and entry spaces for the date, start time, and stop time. For the weekly chart, the days of the week were printed on laminated sheets of 8½" x 11" paper with four boxes below for each team. The teacher initiated the use of a kitchen timer to ensure the game only occurred for 30 minutes.
Materials for classroom observations of student and teacher behavior by the experimenter included an audio cueing tape, earphone, and a recording sheet. The audio cueing tape contained cues at 10-sec fixed-time intervals. The recording sheet was divided into 60, 10-sec intervals.
For the reinforcer preference assessment, a reinforcer survey was distributed to the students prior to commencement of the study. Directions were provided for each student to rank order their choices for rewards. The pool of reinforcers included pencils, erasers, and candy.
During baseline, data were collected on student on-task and disruptive behaviors as well as teacher response statements during a 30-minute math period while students were instructed in math. [End Page 89]
Following baseline, the teacher was trained by the experimenter in the procedures of implementation of the Game such as reviewing the rules with the class, recording occurrences of disruptive behavior, and identifying winning teams. The training involved reviewing the purpose of the Game, outlining the procedures of the Game, and practicing the procedures or steps of the Game.
Good Behavior Game
The Game was implemented once daily during math for 30 minutes. It should be noted that the duration for conducting the Game was chosen based on the length of the academic period. Previous investigations have implemented the Game for entire academic periods, with Salend et al. (1989) reporting the use of 30 minutes for Game duration. The 10-minute observation was conducted at various times throughout the 30-minute math period, so as to sample the experimental session. Prior to implementation, the teacher divided the class into four teams and trained the class on the Game. Before daily implementation of the Game, the teacher placed a recording sheet on the bulletin board adjacent to the blackboard and a large envelope containing the numerical criterion for the period. The criterion remained a mystery to the students until the end of the period. The teacher reviewed the Game rules (e.g., Raise your hand and wait to be called on; Keep hands, feet, and objects to self; Remain in seat; Follow directions the first time given), announced the start of the Game to the students, and set the timer. During each occasion of disruptive behavior, the teacher made a tick mark on the recording sheet under the team to which the student belonged. At the end of the period, the teacher tallied the tick marks and revealed the criterion. The team(s) with tick marks falling at or below the criterion earned a reward. The criterion was selected by the experimenter based on recent classroom behavior. It ranged from 7 to 15 and was adjusted accordingly. The teacher announced the winning team(s) and recorded each team's tick marks on a weekly chart posted in the classroom. In addition to the daily rewards, team(s) that fell at or below a weekly criterion earned a reward (Medland & Stachnik, 1972).
Following the first three sessions of Game implementation (sessions 5 to 7), the experimenter provided the teacher with feedback on her performance with the intention of improving intervention integrity. The feedback specifically addressed the teacher's completion of the ten steps of the Game integrity checklist (e.g., (a) post recording sheet, (b) announce that the Game will begin, (c) review the Game rules with the students and remind teams to not exceed the criterion of tick marks, (d) start the game, (e) identify and record occurrences of disruptive behavior on sheet, (f) announce end of game, (g) tally tick marks for each team, (h) announce the criterion and winning teams, [End Page 90] (i) post each team's tick marks on the progress chart, and (j) provide winning teams with reward).
Treatment Integrity, Interobserver Agreement, and Treatment Acceptability
The primary investigator served as experimenter and collected all data. Two graduate-level students served as observers who were blind to the purpose of the study. The observers were trained on the observation system and achieved proficiency before conducting interobserver agreement measures.
The primary investigator developed scripted protocols to measure adherence to the Game procedures. The protocols listed the primary steps of the Game (e.g., post recording sheet, identify occurrences of disruptive behavior, record occurrences on sheet, determine and announce winning teams). During 29% of the sessions, the primary investigator recorded whether the teacher followed the procedures of Game implementation according to the Game integrity checklist. Across all sessions sampled, treatment integrity was assessed to be at 88%.
Interobserver agreement (IOA) was calculated during 29% of the sessions. Percentage of agreement for student on-task and disruptive behaviors was calculated on an interval by interval basis by dividing the number of agreements by the number of agreements plus disagreements and multiplying by 100%. Mean IOA for on-task behavior was 81.4% (range = 71 to 93%) and 82.8% (range = 73 to 90%) for disruptive behavior.
Although teacher behavior was recorded with a frequency count, it was observed within intervals, or every fourth interval. Therefore, IOA for teacher response statements was calculated by event recording on an interval by interval basis. This is a more stringent method of IOA for event recording. Typical procedures for event recording require dividing the smaller number of observed behaviors by the larger across an entire observation, and multiplying by 100%. Mean IOA for teacher response statements for positive, neutral, and negative statements were 96%, 94.8%, and 90.8%, respectively.
The teacher completed the Intervention Rating Profile (IRP; Martens, Witt, Elliott, & Darveaux, 1985) at the conclusion of the study and found the Game acceptable with a total rating of 75 (possible range = 15 - 90). The students completed a modified version of the Children's Intervention Rating Profile (CIRP; Turco & Elliott, 1986) with a 3-point rating scale (1 = agree a lot, 2 = agree a little, 3 = agree not at all). With a possible range of 6 to 18, student responses on the CIRP suggested the Game was moderately acceptable (M = 9.28, SD = 2.24). [End Page 91]
Figure 1 displays student on-task and disruptive behaviors as percentage of intervals observed across conditions. During initial baseline, student on-task behavior was stable (M = 53.25%) while disruptive behavior showed an increasing trend (M = 36.5%). Implementation of the Game resulted in an increasing trend for on-task behavior (M = 68%), with the exception of session 7, and a decreasing trend for student disruptive behavior (M = 22.33%). Further analysis of the data during implementation of the Game revealed a small increase in on-task behavior during the first three data points (M = 56.33%) when the teacher was provided feedback by the primary investigator around intervention implementation. However, on-task behavior evidenced a change in level and trend (M = 79.67%) during the final three data points of the condition. Note the dotted line delineating the feedback and no feedback phases during the intervention. During the withdrawal phase, a level change was evident for both on-task and disruptive behaviors (M = 47% and 43%, respectively). With the re-implementation of the Game, some initial variability (sessions 13 – 14) preceded an increasing trend for on-task behavior and a concomitant decreasing trend for disruptive behavior (M = 75.6% and 25%, respectively). [End Page 92]
Figure 2 displays teacher behavior as frequency counts across conditions. Results of teacher observations showed little change for teacher positive statements across conditions. Teacher praise did not exceed 2 statements and remained at 0 for 14 consecutive sessions despite increases in student on-task behavior and decreases in disruptive behavior across conditions. In contrast, following initial variability during baseline, collateral increases in negative and neutral responses were evidenced with increases in disruptive behavior, regardless of condition.
The results of the study suggest that the Game was effective in increasing student on-task behavior and reducing disruptive behavior. The results are consistent with previous findings suggesting that the Game is an effective intervention (Barrish, Saunders, & Wolf, 1969; Tankersley, 1995). A poignant remark made by the teacher reflected the strength of the intervention and attested to its social validity, "It was actually quiet in here for a couple minutes."
Perhaps even more important than the replication of the Game's effects are the implications of using the Game with a young, inner city population of students. Researchers have repeatedly demonstrated the negative impact of poorly managed urban school environments [End Page 93] on the behavior of students at risk for aggression (Guerra, Huesmann, Tolan, Van Acker, & Eron, 1995; Van Acker & Talbott, 1999). In particular, the work of Kellam and his colleagues (Kellam et al., 1998) has provided evidence of a link between poorly managed first grade classrooms and the continuation of severe aggression and related academic problems in middle school among boys who were aggressive at school entry. Making matters worse, the link is stronger for boys coming from impoverished families. Clearly, effective classroom management is essential to the socialization of young urban students making it critical to continue to identify effective behavioral management practices that teachers can easily utilize.
Classroom management is an important component of effective teaching. Just as important are academic activities incorporating such teaching practices as opportunities to respond and corrective feedback (Sugai, Horner, & Gresham, 2002). During sessions 7 and 14, there was a notable level drop in on-task behavior and increase in disruptive behavior for session 14. During these sessions, the students were completing workbook pages for an extended period of time. As some students completed their work, additional work was not assigned, allowing for more occasions of negative behaviors. Although the Good Behavior Game can be instrumental in establishing on-task behavior, high-quality instructional practices are critical for maintenance.
The hypothesis regarding teacher praise was not supported. Although the level of classroom disruption decreased with a concomitant increase in on-task behavior, allowing for more occurrences of positive behavior for the teacher to praise, there was not an increase in teacher praise statements. Alternatively, the teacher neutral and negative statements (largely, negative) did vary with the level of disruptive behavior. As the level of disruptive behavior decreased, the teacher neutral and negative responses decreased. Similarly, increases in student disruptive behavior seemed to occasion an increase in the level of teacher neutral and negative responses. These results suggest that more occasions of appropriate student behavior, facilitated by the implementation of the Game, do not necessarily precipitate increased teacher praise. However, student disruptive behavior may set the occasion for an increase in teacher neutral-to-negative responding (i.e., increased teacher talk with an emphasis on negative comments). Future studies should continue to examine a wider use of the Good Behavior Game and, as well, the reciprocal impact of student behavior on teacher behavior in an effort to improve the practice of classroom management, particularly for novice teachers operating in urban schools. Present results, and those of other investigations (e.g., Sutherland et al., 2002; Van Acker et al., 1996; Wehby et al., 1995), [End Page 94] suggest that only overt student behavior (e.g., calling out/correct academic responding) is likely to occasion a teacher response (e.g., neutral, negative/praise) and that increased teacher responsiveness to more passive student behavior (e.g., on-task behavior) will require direct intervention via feedback (Sutherland, Wehby, & Copeland, 2000), self-evaluation (Sutherland & Wehby, 2001), goal setting (Martens, Hiralall, & Bradley, 1997), or student recruitment (Craft, Alber, & Heward, 1998).
Several limitations should be noted when reviewing this study. First, the instructional activities were not controlled during the study. As discussed previously, the instructional activity can have a significant impact on student behavior. Although its negative impact was discussed, it may also have a positive impact, accounting for the improvement in student behavior. Future studies should consider controlling or tightly monitoring instructional activities in place during the investigation.
The second limitation was around time, particularly the length of the study and the length of the observations. Time constraints imposed by nearing the end of the school year resulted in condition changes that were initiated based on the stabilization of only one student variable (i.e., on-task or disruptive behavior). Longer conditions may have produced more stable data series for both student variables. The length of observations presented a concern as student behavior was observed for only 10 minutes during the 30-minute Game period and therefore, may not have accurately captured the behavior exhibited throughout the period while the Game was in effect.
Another limitation was in regard to teacher praise. Behavior-specific praise was not required for a positive statement to be included in the frequency count of teacher behavior. During pre-assessment of teacher behavior, low levels of behavior-specific praise were observed. A looser definition of praise was adopted to include a larger response class of praise. However, many of the studies examining praise have intervened with behavior-specific praise (Sutherland et al., 2000), producing positive results. It may be meaningful to only include behavior-specific praise in future research.
Finally, the current study reflects the behavior of one teacher. Including multiple teachers in a study will allow for more definitive conclusions regarding teacher behavior within the context of the Game. Moreover, a more rigorous assessment method of teacher behavior (e.g., more frequent measures of behavior) is warranted to directly examine the link between student and teacher behavior. Thus, the findings are preliminary and need to be investigated further. [End Page 95]