Psychology in the Schools, Vol. 41(3), 2004 Published online in Wiley InterScience (www.interscience.wiley.com).

© 2004 Wiley Periodicals, Inc. DOI: 10.1002/pits.10168

USING A WHOLE-CLASS TOKEN ECONOMY AND COACHING OF TEACHER SKILLS IN A PRESCHOOL CLASSROOM TO MANAGE DISRUPTIVE BEHAVIOR HOLLY A. FILCHECK, CHERYL B. MCNEIL, LAURIE A. GRECO, AND REBECCA S. BERNARD

West Virginia University The Level System is a whole-classroom approach for managing disruptive behavior that utilizes behavioral management strategies such as a token economy, response cost, stimulating rewards, and strategic attention. Using an ABACC⬘ treatment comparison design with follow-up, this study evaluated the effectiveness of the Level System in a preschool classroom compared to (a) strategies already employed by the teacher, and (b) coaching the teacher in the Child-Directed Interaction (CDI) and Parent-Directed Interaction (PDI) phases of Parent–Child Interaction Therapy (PCIT). Teacher- and parent-report measures were administered, and behavioral observation data were collected for child and teacher behavior using videotapes. Results suggested that the amount of inappropriate behavior exhibited by children decreased when the Level System was implemented. Additionally, inappropriate behavior decreased further during the CDI and PDI conditions. © 2004 Wiley Periodicals, Inc.

The number of young children with behavior problems has increased, with current estimates ranging from 2 to 17% (Lavigne et al., 1998). This increase results in more children per classroom with behavior problems than ever before, making it imperative for preschool teachers to implement effective behavioral management techniques. Previous research has indicated that behavioral techniques (e.g., token economies, response costs, positive reinforcement) are effective in reducing the amount of inappropriate behavior exhibited by disruptive children in the elementary classroom setting (e.g., McGoey & DuPaul, 2000; Wolfe, Boyd, & Wolfe, 1983). Most often, these strategies are used individually with children exhibiting behavior problems. Traditionally, behavioral school consultation has involved behavioral observation, functional analysis, didactic instruction, coaching, feedback to implementers, and booster sessions (e.g., Noell, Witt, Gilbertson, Ranier, & Freeland, 1997). Watson and Robinson (1996) have suggested that parent-training techniques may be beneficial when conducting behavioral school consultations. Because the number of children with behavior problems is increasing, whole-class interventions may be more time and cost efficient than developing specific behavioral plans for individual children. The current study examined the effectiveness of a whole-class token economy (i.e., Level System) versus parent-training techniques (i.e., Parent–Child Interaction Therapy; PCIT) for managing problem behaviors in the classroom. Both interventions were implemented to compare the effects of a tangible reinforcement program (i.e., Level System) to the effects of an exclusively social program based on behavioral modification principles (i.e., PCIT components). Additionally, parent and teacher satisfaction with the interventions was assessed. Token Economies Because some common interventions (e.g., individual token economies) may not be time and cost effective, using whole-class token economies may be a practical option for classroom management problems. Yet, whole-class interventions for modifying behavior problems have not been widely used. Anhalt, McNeil, and Bahl (1998), however, implemented a whole-class token economy (i.e., The AD/HD Classroom Kit, also called, “The Tough Class Discipline Kit”; KIT; McNeil,

Correspondence to: Cheryl B. McNeil, Department of Psychology, West Virginia University, P.O. Box 6040, Morgantown, WV 26506-6040. E-mail: [email protected]

351

352

Filcheck et al.

2000) in an elementary classroom. The KIT resulted in increasing the frequency of appropriate and on-task behaviors. Additionally, Bahl, McNeil, Cleavenger, Blanc, and Bennett (2000) evaluated the effectiveness of the KIT compared to a school-wide discipline program (i.e., red-, yellow-, green-light) in two first-grade classrooms. Results indicated that the KIT was effective in increasing the frequency of appropriate and on-task behavior in children exhibiting disruptive behavior and children exhibiting typical behavior. Furthermore, the teachers involved in each of these studies reported high satisfaction with the KIT (Anhalt et al., 1998; Bahl et al., 2000). Even though there is literature to support the effectiveness of whole-class token economies in elementary classrooms, this type of intervention previously has not been evaluated in preschool classrooms. Individual token economies, however, have been found effective in reducing the amount of disruptive behavior exhibited by children in preschool settings (e.g., McGoey & DuPaul, 2000; Wolfe et al., 1983). Specifically, Wolfe et al. examined the effects of a token economy on cooperative play of three children with behavior problems in the preschool classroom. Results suggested that the children’s level of cooperative play increased by 50% during the treatment phase (i.e., receiving stickers for 1 min of cooperative play), and the number of time-outs decreased for two children. McGoey and DuPaul (2000) investigated the effects of a token economy (i.e., receiving buttons for following classroom rules) plus a response cost (i.e., losing buttons for breaking classroom rules) in decreasing the disruptive behavior of four preschool children with Attention-Deficit/ Hyperactivity Disorder (ADHD). The authors found that both phases (i.e., token economy and response cost) were associated with decreased disruptive behavior. Because of the effectiveness of individual token economies with preschool children with behavior, and the effectiveness of a whole-class token economy with elementary school children, it was hypothesized that a wholeclass token economy for preschool children would be effective in managing disruptive classroom behavior. The Level System The Level System (McNeil & Filcheck, in press) was designed as a whole-classroom token economy approach for several reasons. First, it was developed to address the issue of increased numbers of children with behavior problems in the preschool setting. If individual systems were used, the teacher would have to manage several systems all with different targeted behaviors, consequences for behaviors, and monitoring techniques. Thus, using a whole-class token economy is more time and cost efficient for the teacher. Second, parents and teachers may disapprove of using individual behavioral management systems because they believe that these systems are unfair to the other children in the class. Specifically, the children without a management system would not have the opportunity to receive rewards for exhibiting the same behaviors as the children on the program. Lastly, some may believe that the children with individual behavioral management systems are “singled out.” In other words, other children and parents would recognize that these children have problems because of the management system (Kohn, 2000), and this may cause the children to feel humiliated or ashamed (Corrigan, 1995). Using a whole-class approach would allow all of the children in the classroom to participate in the token economy. Therefore, each child would have the opportunity to receive rewards and no child would be “singled out.” Parent–Child Interaction Therapy Although there is research to support the use of whole-classroom token economy systems to manage disruptive behavior, some researchers (e.g., Eyberg, 1988) have purported that social approaches, such as Parent–Child Interaction Therapy (PCIT), should be used. Parent–Child Interaction Therapy is a 12-week treatment program for children aged 2 to 7 with disruptive behavior

Using a Whole-Class Token Economy

353

problems and their families, which contains two phases: Child-Directed Interaction (CDI) and Parent-Directed Interaction (PDI) (see Hembree-Kigin & McNeil, 1995 for detailed description). Parent–Child Interaction Therapy (Hembree-Kigin & McNeil, 1995; Herschell, Calzada, Eyberg, & McNeil, 2002) has received substantial empirical support as an effective method to train parents to manage disruptive behavior of young children (e.g., Eisenstadt, Eyberg, McNeil, Newcomb, & Funderburk, 1993). Specifically, research indicates that desirable parent behavior (e.g., praise, reflection, imitation, description, enthusiasm) increases, and inappropriate child behavior, such as noncompliance, decreases after mastery of PCIT skills. Because PCIT has been shown to be extremely effective in clinical settings, it is not surprising that PCIT recently has been adapted to the classroom as a technique to manage preschool disruptive behavior (McIntosh, Rizza, & Bliss, 2000). Specifically, McIntosh et al. described a case study in which they used Teacher–Child Interaction Therapy (TCIT) to decrease a child’s disruptive behavior in a preschool classroom. The protocol and goals for TCIT were the same as those described for PCIT, except the intervention was conducted with the teacher in the classroom instead of with the parent in the clinic. In the McIntosh et al. study, five CDI sessions and seven Teacher-Directed Interaction (TDI) sessions were conducted with a teacher and a 2-year-old female child outside of the classroom once a week for 20 minutes. During CDI, the teacher increased her descriptions, reflections, and labeled praises, and decreased her questions and commands with the child compared to baseline. During TDI, the amount of disruptive behavior exhibited by the child decreased, and the amount of compliance displayed by the child increased compared to CDI. In addition, the number of commands given by the teacher decreased. Thus, the results of this study provide preliminary support for using PCIT skills (i.e., TCIT) in a preschool setting. Method Setting and Participants This study was conducted in a preschool classroom described as “out of control.” The classroom included 17 children (M ⫽ 2.9 years, SD ⫽ .33), a female teacher, and one teacher’s aide who changed frequently (i.e., many different individuals were hired for this position). Parental consent was obtained from the entire class. Most of the children were girls (65%) and Caucasian (88.2%), and all children were from two-parent homes. The entire class was videotaped for approximately 1 hour per day during a structured circle time. The circle time typically consisted of learning numbers, letters, and shapes, listening to stories, and singing songs. Measures School observation coding system. One behavioral category (i.e., inappropriate) of the School Observation Coding System (SOCS; McNeil, Eyberg, Eisenstadt, Newcomb, & Funderburk, 1991) was used to help determine the effects of each intervention on child behavior. Specifically, each instance of inappropriate child behavior (e.g., whining) was coded via videotape (see Appendix for definitions). Therefore, coders observed videotapes of circle time and coded each instance of inappropriate behavior exhibited by any child in the classroom. The resulting frequency of inappropriate behavior exhibited across participants was divided by the number of children in the class during the observation and then divided by the length of the observation. Thus, the frequency of inappropriate behavior per child per minute was obtained. Teacher behavior coding system. The frequencies of three teacher behaviors were coded: labeled praise, unlabeled praise, and criticism (see Appendix for definitions). The definitions of these behaviors were adapted from the Dyadic Parent–Child Interaction Coding System (DPICS; Robinson & Eyberg, 1981). Coding occurred via videotapes of circle time that lasted approximately

354

Filcheck et al.

1 hour. Specifically, coders observed the videotapes and coded each instance of labeled praise, unlabeled praise, and criticism exhibited by the teacher to assess the effects of each intervention on positive and negative verbal teacher behavior. Treatment integrity checklist. A 16-item treatment integrity checklist was developed for this study to provide an index of teacher adherence to the Level System protocol (e.g., Labeled praise is used when moving child up a level). A trained observer completed the integrity checklist for each observation session during the Level System condition. This measure was completed based on the teacher’s behavior and the structure of the classroom (e.g., each child had a shape of the chart). The teacher was provided with daily feedback regarding her integrity. A treatment integrity checklist was not developed for the CDI and PDI conditions. Conners’ Global Index. The Conners’ Global Index (CGI; Conners, 1997) is a 10-item measure of disruptive behaviors (e.g., temper outbursts) rated on a 4-point Likert-type scale ranging from “not at all” (0) to “very much” (3). This index has been shown to be sensitive to detecting behavioral changes due to treatment, and it requires minimal administration time (Conners, 1997). This measure results in a total raw score that is converted to a T-score (M ⫽ 50, SD ⫽ 10). The teacher completed the CGI for each child at the end of each condition to determine if these scores changed. Classroom manageability. The teacher completed a single-item, daily classroom management scale throughout the study. The teacher rated the degree to which her classroom was manageable using a 5-point Likert-type scale from “completely unmanageable” (1) to “completely manageable” (5). Time-out log. The teacher completed a daily time-out log indicating which children received time-outs each day, and how many time-outs were given. Teacher satisfaction. A teacher satisfaction measure was developed for this study and was completed at the end of each condition to assess satisfaction of the different management programs. This measure contains 10 items such as: “How likely are you to use these management strategies in your classroom in the future?” Ratings ranged from: 1 ⫽ “poor” to 5 ⫽ “excellent” on a 5-point Likert scale. High total scores indicate high satisfaction with the intervention. Parent interview. The parents completed a 15-question phone interview concerning their views of different classroom management strategies (e.g., time-out, token economies). For questions 1–10, parents were asked to rate the level of acceptability (1 ⫽ “very unacceptable” to 6 ⫽“very acceptable”) of various classroom management strategies. Items 11–15 were open-ended questions regarding parents’ views on certain management strategies (e.g., “List any negative things regarding time out.”). These items were asked to assist in future program development. Only items 1–10 were used in scoring. Higher total scores indicate more acceptability of behavioral management strategies for the classroom. Parents completed this measure once at the beginning of the study, and once at the end of the study. Procedure Teacher Training Overview of the level system. The Level System (McNeil & Filcheck, in press) is a chart with seven levels. The top three levels consist of sunny faces that get happier and brighter as the levels ascend. The bottom three levels consist of cloudy faces that get sadder and grayer as the levels descend. The center level (i.e., starting point) is blank and white. At the beginning of the class, each child’s shape (e.g., dinosaur, kite) is placed on the center level of the Level System. Approximately two to four times daily, the teacher provides a reward to children whose shapes are in the sunny levels.

Using a Whole-Class Token Economy

355

The rewards are designed to be time-efficient, novel, inexpensive (or free), and stimulating. Specifically, each child is given the same reward, all rewards are designed to take less than 3 minutes to dispense, and most of the rewards are activity-based (e.g., act like your favorite animal). The Level System includes 30 reward cards from which the teacher may choose a reward. The teacher is encouraged to use each reward card once before repeating a reward to ensure that the children receive novel rewards. The children who are not involved in the reward simply begin the next class activity (e.g., stations) while the other children receive the reward. Children’s shapes are moved up for engaging in appropriate behavior (e.g., following instructions, sharing with other children). When moving a child’s shape up a level, the teacher is required to provide a labeled praise for the appropriate behavior (e.g., “You are doing a great job of sharing, so I’m going to move your shape up a level!”). When children exhibit mild inappropriate behavior, they are given a visual (i.e., V-shaped hand signal) and verbal warning (e.g., “If you continue to throw the toys, I will have to move your shape down a level”). If they continue to engage in the inappropriate behavior, their shape is moved down a level. For high-level misbehavior (e.g., hitting), the child’s shape is moved down a level with no warning. Children’s shapes can be moved up and down several times throughout the period. At the end of a certain period of time (e.g., after morning routines), all of the children whose shapes are in the sunny area receive a reward (e.g., sticker, activity with teacher). Immediately prior to the reward, all of the children’s shapes are moved back to the center level (i.e., “clean slate approach”) for the next period in which they can earn rewards. The Level System allows for taking individual differences into account. For example, children that lack appropriate play skills could be reinforced for exhibiting these behaviors, while children that have mastered these skills would be reinforced for other behaviors. Training method. Teacher training for the Level System occurred on multiple levels. Specifically, didactic training, in-room coaching, modeling, and immediate feedback were used. A 1-hour didactic was conducted with the teacher at the beginning of condition B. During the first 2 days of the B condition, the teacher was coached in the classroom and given immediate feedback on her use of the System (total of 15 min coaching and 15 min feedback each day). Also, on the 7th and 10th days of condition B, use of the Level System was modeled by the authors during circle time (15 min each day) because the teacher’s integrity with the system was low. After modeling occurred, the teacher was instructed to increase her level of involvement in using the Level System, while the authors gradually decreased (i.e., “faded out”) their level of involvement. By the end of the session, the teacher assumed full responsibility for conducting circle time while using the Level System. The authors provided corrective feedback regarding her use of the Level System following this training session (1 hr each day for 2 days). Total training time for the Level System was 4 hours 30 minutes. When treatment integrity was below 80%, the teacher was provided with verbal feedback by the authors concerning the skills that she needed to improve. This occurred seven times throughout this condition. Parent–Child Interaction Therapy. During the CDI phase, parents are taught specific skills to increase their positive parenting skills and enhance their relationship with their child (HembreeKigin & McNeil, 1995). Specifically, parents are taught to use labeled praise, reflection, imitation, description, and enthusiasm (i.e., PRIDE skills). During PDI, parents are taught to give effective instructions, use two-choice statements, and implement an effective time-out procedure (HembreeKigin & McNeil, 1995). In PCIT, parents learn new skills by being coached and receiving immediate feedback from the therapist while they practice the skills with their children. The teacher was trained in the CDI and PDI phases of PCIT in a similar manner. Specifically, a 1-hour didactic session was conducted for CDI and a 1.5-hour didactic session was conducted for PDI on the day preceding the conditions. After the teacher was taught the skills, she was

356

Filcheck et al.

coached (2 hr per condition) while using the skills out of the classroom with selected children to increase her skill level. During the CDI phase, the teacher was coached to use the PRIDE skills. A mastery criterion was set such that she had to engage in 15 or more praises (at least eight labeled praises), 25–50 descriptions and reflections, and a high level of enthusiasm (i.e., 4 or 5 on a 5-point scale) during a 5-minute observation. Additionally, she had to make no criticisms, give no commands, and ask no questions during the observation. The teacher was coached with one, two, and then three children, each to mastery. During the PDI phase, the teacher was taught the PDI skills (see Hembree-Kigin & McNeil, 1995 for a detailed description), and was coached to use these skills with one to three children at a time. Before coaching began, the time-out procedure was rehearsed with the children so that the consequence for noncompliance with commands was clear. During the coaching sessions, the teacher was instructed to use the skills learned in the CDI phase, and then to provide instructions. If a child was noncompliant with an instruction, the teacher was coached through the time-out procedure. These practices resulted in five time-outs for noncompliance. After these coaching sessions were complete, the teacher began using the skills in the classroom. Additionally, during the first 2 days of each condition, the teacher was observed and provided with immediate feedback on her skills (2 hr and 30 min per condition). Total time spent on the CDI and PDI interventions were 5 hours 30 minutes and 6 hours, respectively. Classroom Coding Classroom observations were coded from videotapes of circle time (approximately 1 hr) on a daily basis using the inappropriate behavior definition of the SOCS (McNeil et al., 1991). The frequency of inappropriate behavior exhibited by all children for the observation period was coded. Teacher behavior also was coded from the videotape. Specifically, the frequency of labeled praise, unlabeled praise, and criticism for each observation period was coded. Two advanced undergraduate students and one graduate student were trained to code the behaviors through pilot videotapes. Concurrent coding of videotapes occurred until 80% interrater agreement was met. Then, independent coding was conducted. Interrater agreement was assessed for each behavioral category for 25% of the observations. These observations were chosen randomly across raters and conditions. Mean percentages for each category were as follows: 80% for inappropriate behavior, 86.9% for labeled praise, 84.2% for unlabeled praise, 75.1% for criticism. Experimental Conditions An ABACC⬘ treatment comparison design with follow-up was used to determine the effects of the Level System as well as the CDI and PDI phases of PCIT on the amount of inappropriate behavior exhibited in the classroom during circle time. The baseline (eight observations) and withdrawal phases (six observations) (conditions A) were the teacher’s typical classroom management strategies (e.g., time-out). Condition B was the Level System (28 observations). Conditions C and C⬘ were the CDI (seven observations) and PDI (four observations) phases of PCIT respectively. Finally, a 4.5-month follow-up was conducted for two observations. During the followup, the teacher had the choice of whether or not to use the Level System; she chose to use it. Results School Observation Coding System The frequency of inappropriate behavior exhibited by the children in the classroom is displayed in Figure 1. The amount of inappropriate behavior exhibited steadily decreased throughout

Using a Whole-Class Token Economy

Figure 1.

357

Frequency of inappropriate behavior exhibited per child per minute with horizontal mean lines.

the study. Specifically, mean frequencies of inappropriate behavior per child per minute were .45 during baseline, .29 while using the Level System, .21 during the withdrawal phase, .12 during CDI, .06 during PDI, and .05 during the follow-up phase. Teacher Behavior Coding System Teacher behavior also was examined (See Figures 2, 3, and 4). Figure 2 indicates the frequency of labeled praises exhibited per minute. The number of labeled praises increased from baseline to the Level System condition and decreased again during the withdrawal condition. During the CDI phase, labeled praises increased again. However, they decreased during the PDI and follow-up conditions. Mean labeled praises for each condition were as follows: .07 baseline, .50 Level System, .27 withdrawal, .47 CDI, .22 PDI, and .11 follow-up. The same general trend occurred concerning the teacher’s unlabeled praise (See Figure 3). Specifically, the mean frequency

Figure 2.

Frequency of labeled praises exhibited by the teacher per minute with horizontal mean lines.

358

Filcheck et al.

Figure 3.

Frequency of unlabeled praises exhibited by the teacher per minute with horizontal mean lines.

was .25 during the baseline phase, .68 during the Level System phase, .63 during the withdrawal phase, .93 during the CDI phase, .80 during the PDI phase, and .49 during follow-up. The amount of teacher criticism per minute for each phase is shown in Figure 4. This behavior decreased slightly from the baseline phase (1.09) to the Level System phase (.96), and then increased during the withdrawal phase (1.58). Criticisms then decreased and remained stable for the last three conditions (i.e., CDI, PDI, follow-up). Mean number of criticisms for each of these conditions were: .27, .37, and .26, respectively. Treatment Integrity Checklist The mean teacher treatment integrity for the Level System ranged from 56.3–100% (M ⫽ 67.8%). Seven booster sessions were provided to the teacher to enhance treatment integrity over the 2 months that she used the Level System.

Figure 4.

Frequency of criticisms exhibited by the teacher per minute with horizontal mean lines.

Using a Whole-Class Token Economy

359

Conners’ Global Index The means and ranges of CGI T-scores were calculated for each condition. Conners’ Global Index scores ranged from 42– 62 during baseline (M ⫽ 48.3, SD ⫽ 5.5), and ranged from 40–58 during the Level System condition (M ⫽ 48.7, SD ⫽ 4.9). Scores were between 43 and 68 (M ⫽ 50.6, SD ⫽ 6.9) during withdrawal, and between 44 and 63 (M ⫽ 50.1, SD ⫽ 4.9) at the end of the study. Finally, at follow-up, scores were between 44 and 57 (M ⫽ 50, SD ⫽ 3.9). Classroom Management Classroom management ratings remained relatively stable throughout the study. However, this rating decreased during CDI compared to the other conditions. Mean ratings were as follows: 2.7 (baseline) (SD ⫽ .48), 2.68 (Level System) (SD ⫽ .55), 2.83 (withdrawal) (SD ⫽ .41), 2.29 (CDI) (SD ⫽ .49), 2.5 (PDI) (SD ⫽ .58), and 2.67 (follow-up) (SD ⫽ .58). Time-out Log The mean number of time-outs given per day steadily increased throughout the study. However, time outs decreased to below baseline levels at follow-up. Specifically, mean time outs per condition were as follows: 2.22 for baseline (SD ⫽ 1.48), 2.68 for the Level System (SD ⫽ 1.63), 2.83 for withdrawal (SD ⫽ 1.47), 3.71 for CDI (SD ⫽ .95), 4.00 for PDI (SD ⫽ 1.82), and 1.67 for follow-up (SD ⫽ 1.53). Treatment Satisfaction The teacher was more satisfied with her typical classroom management strategies (total score ⫽ 26) than with the Level System (23). However, she was more satisfied with using CDI and PDI skills (31 and 31 respectively) than with the typical classroom management strategies. Despite the teacher’s reported level of satisfaction with the Level System, she chose to use the Level System at follow-up rather than her typical classroom management strategies. Parent Interview Results from the parent interview indicated that the parents were very accepting of the Level System, and CDI and PDI before and after the study. Specifically, the mean total score at the beginning of the study was 47.9 out of 60 (SD ⫽ 5.2), and the mean increased to 49.9 at the end of the study (SD ⫽ 5.6). Discussion In the current study, we examined the effectiveness of the Level System compared to skills learned in PCIT in a preschool classroom with high levels of disruptive behavior. Results provided preliminary support for the use of both of these interventions to manage disruptive behavior in preschool classrooms. Specifically, the amount of inappropriate behavior exhibited in the classroom decreased with implementation of the Level System. However, using CDI and PDI skills further decreased inappropriate behavior. Results for the use of PCIT skills with the entire class were consistent with those found by McIntosh et al. (2000) when using these skills with one child outside of the classroom. The Level System and PCIT skills both resulted in a more positive classroom environment. For example, the teacher used more labeled and unlabeled praises while these interventions were in place. Also, the teacher used less critical statements when she used the Level System and PCIT skills. Critical statements decreased substantially when the teacher was taught to decrease their use during the CDI and PDI phase. The teacher used criticism during these two conditions at a

360

Filcheck et al.

remarkably lower rate than the other conditions (i.e., from approximately 1 per min to approximately .3 per min). Because criticisms did not decrease substantially during the Level System, CDI training appears to be necessary to assist the teacher in noticing and decreasing such statements. Teacher’s ratings of the Level System were inconsistent. Specifically, compared to conditions using PCIT skills, the teacher rated her class as more manageable when she used the Level System. However, classroom manageability ratings with the Level System were similar to phases without the Level System. Additionally, the teacher reported that she was more satisfied with using CDI and PDI skills than the Level System. It is important to note, however, that the teacher chose to use the Level System during follow-up and after the end of the study. Surprisingly, the mean number of time-outs steadily increased throughout the study, but decreased during follow-up. A possible reason for this effect might be that the training involved with the Level System and the PCIT phases enabled the teacher to feel more confident in using behavioral management strategies, and she therefore used time outs more for misbehaviors that occurred in the classroom. Another possible explanation involves the potential importance of PDI. The PDI phase emphasized time out, encouraging the teacher to use an appropriate and effective time out procedure for defiance and aggression. Therefore, the increase in time-outs during this phase, where it is the skill being taught, is expected and desirable. Perhaps, once the teacher used time-out appropriately for a while, the need for time-out decreased during follow-up. Although both behavioral interventions were effective, when implementing these programs clinically, it is essential to consider the amount of time involved in training teachers and consultation. Specifically, implementing the PCIT phases took 11 hours 30 minutes of training whereas implementing the Level System took only 4 hours 30 minutes including all consultation and feedback time. The choice of intervention should involve a consideration of the degree of classroom disruption and an analysis of the function of the problem. When a teacher skill deficit is present, PCIT may be the best approach. Specifically, if the teacher is not providing positive reinforcement for appropriate behavior (e.g., praise, attention) and managing inappropriate behavior with negative consequences (e.g., ignoring, time-out), then learning PCIT skills will provide the teacher with these basic behavior management skills. Skills deficits can be assessed by direct behavioral observation of the class. When consulting with a teacher who has excellent skills but has a particularly challenging constellation of children with behavior problems, the Level System might be the intervention of choice. Finally, in cases involving both teacher skill deficits and a very challenging group of children with behavior problems, a combination of a token economy and skill coaching may be required to intervene effectively. Because of methodological limitations (e.g., lack of reversal of inappropriate behavior following the Level System condition, low number of observations in PDI and follow-up, use of only one classroom, no treatment integrity check for PCIT skills), the current results must be considered with caution. Specifically, due to low levels of treatment integrity with the Level System (i.e., treatment integrity fell below 80% seven times) as well as the other methodological limitations, it is unclear whether decreased child inappropriate behavior was a result of implementation of the interventions or something else. Despite these concerns, this study does provide preliminary support for the effectiveness of two new approaches to management of behavior problems in preschool classrooms: (a) a whole-classroom token economy, and (b) the coaching of preschool teachers in skills found effective in parent training programs (e.g., labeled praise, ignoring, effective instructions, time-out). Yet, further research is needed to determine the applicability of these findings to diverse Head Start and preschool programs across the country. References Anhalt, K., McNeil, C.B., & Bahl, A.B. (1998). The ADHD Classroom Kit: A whole-classroom approach for managing disruptive behavior. Psychology in the Schools, 35, 67–79.

Using a Whole-Class Token Economy

361

Bahl, A.B., McNeil, C.B., Cleavenger, C.J., Blanc, H.M., Bennett, G.M. (2000). Evaluation of a whole-classroom approach for the management of disruptive behavior. Proven Practice, 2, 62–71. Conners, C.K. (1997). Conners’ Rating Scales—Revised. New York: Multi-Health Systems Inc. Corrigan, P.W. (1995). Use of token economy with seriously mentally ill patients: Criticisms and misconceptions. Psychiatric Services, 46, 1258–1263. Eisenstadt, T.H., Eyberg, S.M., McNeil, C.B., Newcomb, K., & Funderburk, B. (1993). Parent–Child Interaction Therapy with behavior problem children: Relative effectiveness of two stages and overall treatment outcome. Journal of Clinical Child Psychology, 22, 42–51. Eyberg, S.M. (1988). Parent–Child Interaction Therapy: Integration of traditional and behavioral concerns. Child and Family Behavior Therapy, 10, 33– 46. Hembree-Kigin, T.L., & McNeil, C.B. (1995). Parent–Child Interaction Therapy. New York: Plenum Press. Herschell, A.D., Calzada, E.J., Eyberg, S.M., & McNeil, C.B. (2002). Parent–Child Interaction Therapy: New Directions in Research. Cognitive and Behavioral Practice, 9, 9–16. Kohn, A. (2000). The schools our children deserve: Moving beyond traditional classrooms and “tougher standards.” Boston, MA: Houghton Mifflin Company. Lavigne, J.V., Gibbons, R.D., Christoffel, K.K., Arend, R., Rosenbaum, D., Binns, H., Dawson, N., Sobel, H., & Isaacs, C. (1998). Prevalence rates and correlates of psychiatric disorders among preschool children. Journal of the American Academy of Child and Adolescent Psychiatry, 35, 204–214. McGoey, K.E., & DuPaul, G.J. (2000). Token reinforcement and response cost procedures: Reducing the disruptive behavior of preschool children with Attention-Deficit/Hyperactivity Disorder. School Psychology Quarterly, 15, 330–343. McIntosh, D.E., Rizza, M.G., & Bliss, L. (2000). Implementing empirically supported interventions: Teacher–Child Interaction Therapy. Psychology in the Schools, 37, 453– 462. McNeil, C.B. (2000). The Tough Class Discipline Kit. Congmont, CO: Sopris West. McNeil, C.B., Eyberg, S., Eisenstadt, T.H., Newcomb, K., & Funderburk, B. (1991). Parent–Child Interaction Therapy with behavior problem children: Generalization of treatment effects to the school setting. Journal of Clinical Child Psychology, 20, 140–151. McNeil, C.B., & Filcheck, H.A. (in press). The Level System. Longmont, CO: Sopris West. Noell, G.H., Witt, J.C., Gilbertson, D.N., Ranier, D.D., & Freeland, J.T. (1997). Increasing teacher intervention implementation in general education settings through consultation and performance feedback. School Psychology Quarterly, 12, 77–88. Robinson, E.A., & Eyberg, S.M. (1981). The dyadic parent–child interaction coding system: Standardization and validation. Journal of Consulting and Clinical Psychology, 42, 471– 481. Watson, T.S., & Robinson, S.L. (1996). Direct behavioral consultation: An alternative to traditional behavioral consultation. School Psychology Quarterly, 11, 267–278. Wolfe, V.V., Boyd, L.A., & Wolfe, D.A. (1983). Teaching cooperative play to behavior problem preschool children. Education and Treatment of Children, 6, 1–9.

Appendix Inappropriate behavior. Behavior that is considered disruptive or annoying to yourself or others in the classroom environment. Examples of inappropriate behaviors include: “whining, crying, temper tantrums, yelling, destructiveness, negativism, pathological self-stimulation, demanding attention, high-rate behavior, talking out of order, being out of area or cheating” (McNeil et al., 1991, p. 143). Labeled praise. A labeled praise lets the child know exactly what is liked about his or her behavior because it is specific in that it describes the behavior being praised (e.g., Nice job sharing the toys!) (Hembree-Kigin & McNeil, 1995). Unlabeled praise. Any positive statement directed toward the child that does not include a specific behavior being praised (e.g., Good job! Terrific!) (Hembree-Kigin & McNeil, 1995). Criticism. A critical statement conveying a negative tone toward the child (e.g., “You’re bad. That was a stupid choice.”), and any statement including the words: no, don’t, stop, quit, or not (Hembree-Kigin & McNeil, 1995).

Copyright of Psychology in the Schools is the property of John Wiley & Sons, Inc. and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use.