Effective classroom management systems are essential for creating an optimal learning environment (Heering & Wilder, 2006), and therefore, it is critical that teachers be equipped with the necessary tools to manage and prevent challenging behaviors in general education classrooms. While the majority of classroom management research has focused on using behavior management interventions with individuals in special education classrooms or those diagnosed with intellectual or learning disabilities (McKissick et al., 2010), public school teachers report that the occurrence of disruptive behaviors in general education is prevalent (Naylor et al., 2018). This high prevalence of challenging behaviors leads to increased stress and discontent among teachers, ultimately leading to early departures from the field. It was reported that in the USA, half of new teachers resigned from their positions within their first 5 years, with a major reason being high rates of challenging behaviors (Aloe et al., 2014). This concerning statistic highlights the need for more research on classroom management and training for teachers to effectively apply interventions in their classroom.

Research is crucial for designing and employing interventions for general education teachers to easily apply in their classrooms. Group contingencies, which are rules and expectations applied to a group, are often suggested as a classroom management contingency due to the relative ease with which they can be implemented and the impact the strategies have across a number of students (Heering & Wilder, 2006). The use of group contingencies within classrooms may be more feasible for one teacher to implement in a classroom up to 30 students than are individualized contingencies (Heering & Wilder, 2006).

Group contingencies are defined as the application of operant behavior procedures to manage the behavior of a group (Litow & Pumroy, 1975). There are three categories of group contingencies: independent, interdependent, and dependent. In an independent group contingency, the same response contingencies are in place for all group members; however, reinforcement is delivered based on individual performance. These interventions are used more often in special education classrooms compared to general education, which is most likely because the performance of one individual does not affect reinforcement for other group members (Litow & Pumroy, 1975).

Within general education classrooms, independent group contingencies have been effective for increasing academic engagement behaviors (Dart et al., 2016), decreasing disruptive behaviors (Lum et al., 2019), and increasing physical activity (Wahl-Alexander et al., 2020). In one study, Dart et al., (2016) used a novel form of an independent group contingency, titled the classroom password, with middle school students to increase academic engagement and decrease disruptive behaviors. During this game, the teacher said the specified word a certain number of times while the students tallied the frequency. Up to five students, of those who scored the correct frequency, were randomly selected to receive the reward at the end of the session. Results demonstrated an increase in academic engagement behavior and decrease in disruptive behavior, with variable data for decreasing off-task behavior. Social validity results indicated that two of the three teachers rated the procedure as acceptable and one teacher agreed with its immediate positive effect. Findings suggest this intervention may be valuable and that additional research is needed to confirm results and increase teacher acceptability.

An interdependent group contingency requires all members to achieve a certain level of performance for the group to receive access to the reinforcer. This type of contingency can help students work together and encourage one another (Helton & Alber-Morgan, 2020). On the other hand, if one or a few students prevent the class from receiving the reinforcer, those students may experience negative social consequences from peers (Helton & Alber-Morgan, 2020). Studies have used interdependent group contingencies to increase homework completion (Chafouleas et al., 2012), frequency of teacher praise (Clair et al., 2018), physical activity during recess (Foote et al., 2017), and on-task behavior (Kamps et al., 2015). Research has also evaluated interdependent group contingencies for decreasing disruptive behavior (Lee et al., 2017) and off-task behavior (Ling et al., 2011).

One form of an interdependent group contingency that has received attention in the literature is the good behavior game (GBG), in which the class is separated into teams, rules are enforced for all students, and points are delivered for either appropriate behaviors or challenging behaviors (Pennington & McComas, 2017). A potential obstacle in the design of this system is determining a reinforcer that is preferred by all students (Lo & Cartledge, 2004). Methods to increase motivation include conducting frequent preference assessments, varying the reinforcer often, and using a mystery box (Silva & Wiskow, 2020). GBG has presented positive results for decreasing challenging behaviors including disruptive behavior (Bohan et al., 2021; Dadakhodjaeva et al., 2020), off-task behavior (Hernan et al., 2019), and out of seat behavior (Elswick et al., 2016). Studies have also demonstrated success with increasing teacher attention (Donaldson et al., 2015), teacher’s use of behavior specific praise (Lastrapes et al., 2018), academic engagement behavior (Lynne et al., 2017), reading performance (Weis et al., 2015), and on-task behavior (Pennington & McComas, 2017).

A dependent group contingency delivers reinforcement to the entire group based on the performance of one or more selected individuals in the group (Litow & Pumroy, 1975). A potential undesirable outcome of this contingency is negative social consequences from peers to the selected individual (Heering & Wilder, 2006). Alternatively, this can be a benefit as it allows one student or a small group of students to receive appreciation from peers if reinforcers are earned. Due to this potential limitation, the individual’s anonymity should be maintained until the contingency is met (Helton & Alber-Morgan, 2020). Another possibility is for the teacher to target several behaviors simultaneously to reduce the likelihood that the selected individual will be identified prematurely (Helton & Alber-Morgan, 2020).

Dependent group contingencies have been used for a variety of targets including reducing disruptive behavior (Deshais et al., 2018), increasing on-task behavior (Bulla & Frieder, 2018), and increasing engagement in physical activities (Vidoni et al., 2012). One study successfully evaluated the use of a randomized dependent group contingency for reducing disruptive behaviors during hallway transitions with first grade students in a public elementary school (Deshais et al., 2018). A dependent group contingency may also be applicable in physical education classes as Vidoni et al., (2012) successfully increased students’ heart rates and physical activity.

There are numerous advantages to using group contingencies in classrooms to manage challenging behaviors and increase desirable behaviors. These systems are also economical and practical as it is possible for one teacher to implement. Group contingencies are also flexible as they allow teachers to modify the intervention to fit the classroom needs, address the behavior of one student or the entire class, and be used alone or in combination with additional interventions. These contingencies can also accommodate the varying needs and abilities of students in general education classrooms. While the recent research in school settings demonstrates their effectiveness for classroom management, generalization is a key component to extending the use of these strategies as classroom or school wide interventions. Further research is vital for determining how group contingencies can optimally be applied within general education classrooms and generalized across students and teachers.

Purpose

The results of previous literature reviews (Bowman-Perrott et al., 2016; Little et al., 2015; Maggin et al., 2017; Pokorski et al., 2017) suggest group contingencies are an appropriate classroom management strategy for modifying behaviors on a group level. An updated review is necessary to evaluate how group contingencies have been implemented in general education classrooms, specifically the targeted behaviors, ages of students, social validity, and identified areas for future research. The purpose of the current paper is to explore the research on group contingencies within general education classroom settings. The evaluation of these studies aims to identify the effectiveness of group contingencies, the limitations within the current literature, and directions for future research. Previous literature reviews have been conducted on various uses of group contingencies including studies set in preschools (Pokorski et al., 2017), those with students who present with challenging behaviors (Maggin et al., 2017), and implementation of the good behavior game (Bowman-Perrott et al., 2016). Little et al., (2015) reviewed the application of group contingencies with school aged children from 1980 to 2010 across various settings and included 50 studies. There has yet to be a literature review to focus solely on the application of group contingencies in general education classrooms. Given the substantial amount of research in the past 10 years, the use of group contingencies within general education settings needs an updated comprehensive review to synthesize the current research and evaluate trends in the literature.

Method

Selection Criteria

A review of empirical literature on group contingencies within educational settings was conducted through the identification and evaluation of single-subject studies. A search of the literature was conducted through electronic databases, including Wiley Online Library Database, ProQuest Central, and SAGE Online. A combination of the following terms was used: group contingency, general education, school aged children, independent group contingency, interdependent group contingency, dependent group contingency. Inclusion criteria consisted of studies with an independent variable of a group contingency or a group contingency within the intervention package, single subject design with at least one participant, students in kindergarten through 12th grade general education classrooms, students receiving special education within a general education classroom, and studies set in any school location, including academic classrooms, non-academic classes (i.e., physical education class, art class), and hallways. Studies were excluded if they were conducted in an alternative school, resource classroom, small group instruction (i.e., students pulled out of the classroom), special education classroom, self-contained classroom, or if adult participants were included. Additional parameters included the necessity for publication in peer-reviewed journals, publication date between 2011 and 2021, and publication in the English language. The original search yielded 24,388 articles. A review of the title and abstracts resulted in 107 studies meeting criteria for inclusion in this review. Articles were then further screened by reading the methods section for final selection yielding 53 studies in total.

Coding

For each study, the participants, methods, results, limitations, and future research directions were coded to synthesize this information. Variables were selected, operationally defined, and coded by the first author. Results were then summarized as the percentage of articles by category and displayed in Table 1. Interobserver agreement data were collected for 35% of the included articles by one additional graduate student. This observer was trained by the first author on how to code each variable including how to calculate effect size and how to report results on the coding form. IOA results demonstrate 98% agreement across all coded variables.

Table 1 Participants and setting
Full size table

Participants and Setting

Participants of the 53 studies were coded by gender, number, race/ethnicity, and whether any participants were receiving special education services. The setting of the study was coded to identify whether the intervention was implemented in an academic classroom, non-academic classroom (e.g., physical education, art classroom), an alternative location within the school (e.g., recess, hallways), and the number classrooms involved in the study.

Independent Variable

The independent variable was coded by type of group contingency (dependent, independent, interdependent, or comparison), and whether there was an intervention in addition to the group contingency. The majority of studies clearly specified the type of group contingencies, but for any that did not, an inference was made based on the description of the intervention provided.

Dependent Variable

Each study’s dependent variables were coded by the following categories: on-task behavior, off-task behavior, academic engagement behavior, disruptive behavior, average number of steps, heart rates, teacher statements, academic performance, or other. The other category included variables such as transitioning between classrooms, intervention acceptability, teacher accuracy with data collection, percentage of potential points earned, and mobile device presence.

Findings

To provide information on the magnitude of the effect of the intervention on the dependent variable, effect sizes were calculated on all dependent variables and participants in the study. Nonoverlap of all pairs (NAP) is an index used to determine the effect between phases by comparing all data points in baseline to all data points in intervention (Parker & Vannest, 2009). The amount of change between baseline and intervention can be measured by the extent to which baseline and intervention data points do not overlap, which can be quantified by NAP values (Parker & Vannest, 2009). If the study reported NAP values as part of their results, those scores were used for coding. For the remaining studies that did not report NAP values, data were extracted using the digitization program, WebPlotDigitizer (Moeyaert et al., 2016), and inserted into the NAP calculator. Moeyaert et al., (2016) evaluated the usability of data extraction programs and found that WebPlotDigitizer was the best selection based on usability, time to obtain data, and cost.

Limitations

The most frequent limitations as reported by the author were coded across all studies. These limitations included short implementation period, limited generalizability, applicability of rewards, multiple components, lack of maintenance, small sample size, low or minimal IOA and treatment integrity, function of behavior not identified, reinforcers may not function as reinforcers, no individuals data collection, procedural deviations from research methodology, constant researcher presence, minimal social validity, time constraints, academic performance not assessed, or other (e.g., teacher withdrawal, no preference assessment, possibility of observer bias, staff or student absences).

Future Research

Directions for future research as reported by the authors were coded across studies. The most common areas were defined and coded as assessing social validity, conducting a component analysis, analyzing effects on individual students, assessing generalization, evaluating long-term effects, assessing academic performance, conducting preference assessments, assessing feasibility on teacher’s part, adjusting the schedule of reinforcement, assessing the intervention without outside support, and other (e.g., blind research assistants, frequency with which intervention is implemented, economical prize rewards).

Results

Participants and Setting

The majority studies included participants in kindergarten through third grade with 19% of studies in kindergarten, 25% in first grade, 6% for second grade, and 21% in third grade. In regard to gender, 2% of studies used only male students, 81% were coeducational, and 17% of studies did not specify gender. Race or ethnicity data were reported in 75% of studies with 64% specifying the percentage of students per race/ethnicity and 11% categorized students generally. Studies also varied with respect to number of students and classrooms as 25% included one classroom, 49% assessed multiple classrooms, 9% included one classroom with target students, 15% included multiple classrooms and target students, while 2% of studies collected data only on target students. Students also varied in diagnoses with 13% all typical development, 4% with at least one student with autism, 8% with at least one student with a learning disability, 11% with at least one student receiving special education, 6% with least one student with other health impairment, 8% with at least one student with EBD, 4% with at least one student with ADHD, 8% categorized as other, and 51% studies did not specify. The settings of the studies were less varied with 2% in the hallway of the school, 4% in art class, 4% in recess, 8% in physical education classes, 81% in the classroom, and 2% did not specify the location.

Independent Variables

While all three types of group contingencies are expressed in the literature, 80% used interdependent, 6% used an independent, 6% used a dependent, and 9% compared at least two group contingencies. Of the interdependent studies, 36% of studies implemented the good behavior game and 2% used the caterpillar game (a novel modification of an interdependent group contingency) (Table 2).

Table 2 Methodological variables
Full size table

Dependent Variables

The dependent variables assessed in each study varied somewhat with 58% measuring disruptive behavior, 30% measuring academic engagement behavior, 23% assessing on-task behavior, 11% including teacher statements, and 9% collecting the average number of steps taken.

Findings

To synthesize the results of each study, NAP scores were calculated for each dependent variable and the effect was categorized as weak, moderate, or strong. Scores between 0 and 0.65 are considered weak effects, scores of 0.66–0.92 are considered moderate effects, and scores of 0.93–1.0 are considered strong effects (Parker & Vannest, 2009). Of the 53 studies, 68% demonstrated a strong effect for at least one dependent variable, 43% demonstrated a moderate effect for at least one dependent variable, and 6% demonstrated a weak effect for at least one dependent variable (Table 3).

Table 3 Results and discussion variables
Full size table

Limitations

The limitations as identified by the authors are also widespread with 28% mentioning limited generalizability, 25% indicating a small sample size, 15% reporting not collecting individual data, 15% stating low or minimal treatment integrity data, 13% reporting minimal social validity, 13% identifying time constraints, 8% describing low or minimal IOA data, 8% acknowledging the reinforcer delivered may not have truly functioned as reinforcers, 11% indicating there were multiple components to the intervention, 13% reporting maintenance data were not collected, 8% indicating there may have been procedural deviations from the research methodology, 8% suggesting the function of the behavior was not identified, 8% mentioning the applicability of rewards may be limited, 6% discussing the researcher presence, and 8% considering not assessing academic performance a limitation.

Identified Directions for Future Research

Authors of the included studies also mentioned numerous directions for future research. Of the 53 studies, 40% suggested generalization assessments, 15% suggested component analyses, 15% recommended assessment of social validity, 11% indicated the need for evaluating additional or long-term effects of the intervention, 9% suggested academic performance be measured, 9% mentioned the need for measuring the effect on target students, 6% indicated the schedule of reinforcement be adjusted, and 4% mentioned the need for assessing feasibility on the part of the teacher, and 8% suggested the intervention be implemented without outside support.

Discussion

This review evaluated the use of group contingencies from 2011 to 2021 and identified 53 studies that met inclusionary criteria. The settings were narrowed down to any general education classroom within the school, including academic and non-academic classrooms. Overall, results of this literature demonstrate the use of group contingencies across grades, behaviors, and school locations. While the majority of studies included elementary aged students, research supports their use in grades from kindergarten through 12th grade. Interdependent group contingencies are also the most commonly implemented, suggesting higher feasibility compared to dependent and independent. Additionally, the results of this review highlight the limitations that have been addressed, limitations that still exist, and directions for future research.

Participants and Setting

A common limitation and area for future research was the use of a small sample size and the need to address a range of ages. While each study used a specific age range, the effectiveness of this intervention has been demonstrated across ages within the literature. While 25% discussed small sample size as a limitation and 28% mentioned limited generalizability, classrooms from kindergarten to 12th grade are represented in the literature, suggesting this limitation has been addressed. However, the majority of studies (71%) used participants in kindergarten through third grade. Future research should focus on evaluating group contingencies with upper elementary, middle school, and high school students.

Based on the literature, a current limitation is the application of group contingencies to general education classrooms also consisting of students with developmental or learning disabilities. While 51% of studies did not specify the development of the students, 13% included students all of typical development and 11% included at least one student receiving special education services. More research is warranted on the effectiveness of the intervention with students with disabilities in a general education classroom.

Findings of this literature review demonstrate the limited number of studies that took place in non-academic classrooms. The majority of studies were set in the student’s academic classroom even though classroom management strategies are still necessary in other settings within the school. Students who receive special education are more likely to attend general education non-academic classes than academic classes, suggesting a possible need for classroom management strategies in these classrooms. Of the studies included in this review, two studies took place in art class, four studies were set in a physical education class, and two studies occurred during recess. Future research is necessary to determine if and how group contingencies can be applied in these types of settings to decrease challenging behavior and increase desirable behaviors.

Dependent Variables

The majority of studies (58%) included disruptive behavior as at least one of the dependent variables. Other common dependent variables include academic engagement behavior, off-task behavior, and on-task behavior. Only four studies assessed academic performance, which should be addressed in future research. A primary purpose for decreasing off-task behaviors is to optimize instructional time and facilitate learning. While measuring off-task behavior may be an important indicator, decreasing off-task behavior may not directly translate to improved academic performance (Heering & Wilder, 2006). Perhaps additional interventions should be used in conjunction with group contingencies to increase appropriate classroom behaviors and create academic success.

Interventions

Overall, the most common group contingency was the interdependent with 80% of studies using some form of an interdependent group contingency. Of those 80%, 36% used the good behavior game and 2% used the caterpillar game. The quantity of studies using the good behavior game and their results indicate its feasibility and effectiveness. Of these studies, only three mentioned minimal or low social validity, signifying teacher and student acceptance of this intervention. Additionally, only one study included low or minimal treatment integrity and two studies included the need for researcher presence throughout implementation as a limitation. The few studies that included minimal social validity or treatment integrity as limitations suggest the good behavior game may be easily implemented by a classroom teacher and rated favorably by teachers and students.

The majority of these studies used participants in kindergarten (7) and first grade (5), although grades through eleventh are represented in the literature. Similarly, the majority of studies assessed academic engagement behavior (7) and disruptive behavior (13) as at least one of their dependent variables. Off-task behavior, on-task behavior, and frequency of teacher statements were each represented in two studies and academic behavior was only measured in one study. Future studies should measure additional variables to determine the generalizability of the good behavior game.

Another limitation within the current literature is minimal studies conducted with upper elementary, middle, and high school students. An advantage of the good behavior game is the ability to modify procedures to meet the needs of the learners (Kleinman & Saigh, 2011). The delivery of rules and terminology used should be altered based on the age and population of the classroom. Future research should explore the good behavior game with older students to ascertain appropriate modifications of the intervention. Future research could also compare the number of teams in which the class is divided while playing the good behavior game. While the majority of research has used two teams (Fallon et al., 2020; Lynne et al., 2017; Pennington & McComas, 2017), a comparison study to determine the optimal number of teams could be beneficial.

There were significantly fewer studies that used an independent or dependent group contingency with three studies of each represented in this review, implying these are potentially less feasible than interdependent. Research has suggested potential obstacles when designing dependent group contingencies which may contribute to the limited number of studies. A dependent group contingency delivers reinforcement to the entire group based on the performance of one or more selected individuals in the group. A potential undesirable outcome of this contingency is negative social consequences from peers to the selected individual; however, this limitation can be minimized by maintaining the individual’s anonymity (Heering & Wilder, 2006).

Another limitation within dependent group contingencies is the potential for reinforcing disruptive behavior (Bulla & Frieder, 2018). Bulla & Frieder, (2018) discussed the possibility that students may be less likely to engage in the appropriate classroom behaviors if their team earns a reward despite some students engaging in disruptive behavior. In this study, one student was selected randomly from each team yielding a low probability for a student being selected and coming into contact with the reinforcement contingency. Due to this implementation, a slow treatment effect was observed. A possible solution could be to select students who exhibit higher rates of disruptive behaviors compared to students who are engaging in desirable behaviors.

Additionally, there may be ethical concerns with dependent group contingencies including negative peer pressure towards the target student if criteria is not met, unwanted peer attention towards the target student if the criteria is met, unachievable criteria for some students if using class average, desire for target student to sabotage peers, and lack of prerequisite skills in the target students’ repertoires (Vidoni et al., 2012). Based on these concerns, caution should be taken when creating this type of intervention and individual skills of the students need to be assessed to determine the appropriate performance criteria and the potential for negative social consequences.

Two studies that implemented independent group contingencies mentioned minimal teacher acceptability of the procedures or reinforcement schedule (Dart et al., 2016; Lum et al., 2019). Independent group contingencies require that students earn rewards based on independent performance. Dart et al., (2016) reported low acceptability of the intervention by the classroom teachers, even though data represented a decrease in off-task behavior. Qualitative feedback from teachers regarding potential modifications to the intervention could be beneficial for creating a more feasible procedure. Another possibility for low teacher acceptability ratings is the delivery of reinforcement including the time and cost to obtain the rewards (Lum et al., 2019). Rewards in the form of bonus points and homework passes were also offered; however, teachers expressed concerns that these would artificially inflate the students’ grades (Lum et al., 2019).

It is also worth noting that all three studies that implemented an independent group contingency used participants in middle (Dart et al., 2016; Wahl-Alexander et al., 2020) or high school (Lum et al., 2019). Generalization of the effectiveness of this type of contingency has not been demonstrated across grades, suggesting more research is necessary for determining if an independent group contingency is more accepted by elementary school teachers.

Identified Directions for Future Research

Based on the current literature, many of the limitations discussed have been addressed in other studies including generalization across populations and settings, individual data collection, and limited number of participants. Future research should address the remaining limitations which include determining procedures that do not require the researcher to be present for the duration of implementation, measuring academic performance as an additional dependent variable, conducting follow-up phases, and assessing methods to increase feasibility.

Future researchers may also consider the method for collecting social validity data. Collecting data on the continued use of the interventions may be more indicative of teachers’ opinions and acceptability compared to a standard questionnaire (Wahl et al., 2016). While in one study social validity results reported high teacher acceptability, follow-up sessions indicated teachers discontinued the use of the good behavior game (Wahl et al., 2016). Treatment integrity data also needs to be included in future research to ensure classroom teachers are able to implement procedures accordingly. Acceptable rates of treatment integrity are essential to the success and long-term use of an intervention (Lee et al., 2017). Of the studies included, 15% reported low or minimal treatment integrity, which is a concern for generalizability and replication of results. Similarly, 8% of studies reported the classroom teacher deviated from the method and 6% of studies reported the need for the researcher to be present for the duration of implementation. Future research on training teachers efficiently is necessary to ensure successful implementation with high treatment fidelity without the need for outside individuals.

An important direction for future research is to continuing developing the most feasible strategies for implementing group contingencies. There is little research that takes advantage of technology for implementing classroom management strategies. Lynne et al., (2017) discussed how electronic devices have received little attention, but should be evaluated more as they could increase the ease with which teachers implement group contingencies. The use of the latest electronic devices, such as iPads and smartboards, as vehicles for implementing group contingencies should be explored. Elswick et al., (2016) compared paper and pencil data collection to computer-based data collection when implementing the good behavior game. Results demonstrated that teachers collected more accurate data with the computer-based method compared to the data collected by hand. These findings further demonstrate the need for more research with technology and group contingencies. Additional research can determine the most efficient and feasible means of implementation to create the optimal learning environment for all students.

Conclusion

The effectiveness and practicality of group contingencies on managing behaviors in general education classrooms has been demonstrated in the current research. The results of this review further demonstrate that group contingencies are an evidence-based practice that has been successfully implemented across multiple participants, general education settings, and dependent variables. One limitation of this review is the confined setting and population parameters as general education classrooms was the focus. Studies that implemented these procedures in other settings such as special education classes, alternative schools, employment settings, and residential settings were excluded. A second limitation is the time frame as this review only included studies published from 2011 to 2021, and therefore the history of applications of group contingencies within general educations is not covered. Based on these limitations, findings cannot be generalized to other settings and only provide clinical implications for use of group contingencies in general education settings from kindergarten through 12th grade. Future research should continue to evaluate group contingencies across various settings and populations to generalize these findings and expand the use of these procedures. The evidence supports the use of group contingencies to modify behavior in the classroom and additional research can determine the most successful and feasible means of implementation. When educators employ evidence-based practices, students will benefit from an enhanced learning environment.

There are several implications of this review that can be useful for teachers and administrators. Target populations included students in general education classrooms in kindergarten through 12th grade, suggesting educators of these grades may be able to apply these interventions to their classrooms. Similarly, several common challenging behaviors were also identified that may be applicable for other classrooms. Results of this review synthesized the settings and for what types of challenging behaviors group contingencies have been effective. Teachers and administrators can review these results and use the outlined evidence-based interventions in their classrooms of similar needs. Due to the occurrence of challenging behaviors in all classrooms, it is important that teachers have access to behavior supports and classroom management systems. The increased use of classroom management systems will assist educators in creating the optimal learning environment that supports the students’ needs and fosters their success.