The Past, Present, and Future of the Good Behavior Game

The Good Behavior Game (GBG) is an effective behavior-management intervention used in classrooms worldwide (Barrish, Saunders, & Wolf, 1969; Donaldson, Vollmer, Krous, Downs, & Berard, 2011; Joslyn, Donaldson, Austin, & Vollmer, 2019; Saigh & Umar, 1983) and has been likened to a “behavioral vaccine” (Embry, 2002, p. 1) in that it may prevent problem behavior and reduce the likelihood of problem behaviors occurring later in life (e.g., aggression and drug abuse; Embry, 2002; Kellam et al., 2008; Kellam et al., 2011). The GBG is an interdependent group contingency, in which all members of a group are subject to the same response contingencies and the entire group must meet a specified criterion before reinforcement is delivered (Litlow & Pumroy, 1975). Interdependent group contingencies often require only limited time and resources to implement and increase cooperative behaviors and positive social interactions among peers, which makes the GBG an efficient intervention for classroom teachers (Skinner, Cashwell, & Dunn, 1996). Specific components of the GBG include rules the players must follow, the delivery of feedback (e.g., tallies), a criterion for winning the game, and delivery of reinforcement to the winning team (Barrish, Saunders, & Wolf, 1969).

The First 20 Years: 1969-1989

In the first use of the GBG, Barrish et al. (1969) evaluated the effects of an interdependent group contingency with specific components on out-of-seat and talking behaviors in a fourth-grade classroom. Shortly after, Medland and Stachnik (1972) conducted a replication and systematic analysis of the GBG in a fifth-grade classroom, specifically evaluating the rules and feedback components of the GBG. Results were consistent with those of Barrish et al. (1969) and expanded on the literature by showing that certain components of the game (i.e., stating the rules and providing feedback when the rules were upheld or violated) continued to exert control over student behavior when other aspects (i.e., contingency management and rewards) were withdrawn. While Barrish et al. (1969) and Medland and Stachnik (1972) provided feedback (i.e., tallies) for rule violations, other researchers began providing feedback for behaviors they wanted to increase (i.e., instead of a tally for being out of one’s seat, tallies were awarded for on-task behavior; Darch & Thorpe, 1977; Robertshaw & Hiebert, 1973; Wolf, Hanley, King, Lachowicz, & Giles, 1970). Additional modifications were introduced in the early years of the GBG, including the Astronaut Game, in which feedback was displayed on a bulletin board where each group of first graders had a “spaceship” that moved one step closer to the moon when a group won the game (Robertshaw & Hiebert, 1977), and winners received  attention from the principal of the school (e.g., classroom visits) (Darch & Thorpe, 1977). In the Library Game, the school librarian conducted the game in the library, but the classroom teacher delivered reinforcement, due to the nature of the reinforcers (i.e., extra recess could not be provided by the librarian; Fishbein & Wasik, 1981). The GBG was also successfully implemented internationally during this time (Nolan, Houlihan, Wanzek, & Jenson, 2014), in elementary classrooms in Germany (Huber, 1979), Canada (Kosiec, Czernicki, & McLaughlin, 1986), and Sudan (Saigh & Umar, 1983) and in a secondary-school classroom in the United Kingdom (Phillips & Christie, 1986). The first 20 years of the GBG provided the basis for an effective behavior management intervention used in classrooms across the globe (Nolan et al., 2014).

The Next 40 Years: Different Populations

            The next 40 years of GBG research saw a shift from modifying the GBG itself to expanding the settings and target populations with which the GBG was used, including low-income students (Dion, Roux, Landry, Fuchs, Wehby, & Dupéré, 2011; Lannie & McCurdy, 2007), individuals diagnosed with emotional and behavior disorders (EBD; Joslyn, Vollmer, & Kronfli, 2019) and intellectual and developmental disabilities (I/DD; Wiskow, Ruiz-Olivares, Matter, & Donaldson, 2018), and preschoolers (Ling & Barnett, 2013; Murphy, Theodore, Aloiso, Alric-Edwards, & Hughes, 2007; Swiezy, Matson, & Box, 1992; Wiskow, Matter, & Donaldson, 2019). Research also expanded internationally, with positive effects seen in elementary classrooms in Belgium (Leflot, van Lier, Onghena, & Colpin, 2010), the Netherlands (van Lier, van der Sar, Muthen, & Crijnen, 2004), Chile (Nolan et al., 2014), and Canada (Dion et al.,  2011) and in secondary-school classrooms in Spain (Ruiz-Olivares, Pino, & Herruzo, 2010).

Research on the GBG has also extended more recently to preschoolers, an important population, as intervening on problem behaviors as early as possible is considered best practice (Jolivette, Gallagher, Morrier, & Lambert, 2008; Leaf et al., 2016; Ling & Barnett, 2013; Stevenson & Goodman, 2001; Wiskow et al., 2019). While this body of work is still limited, what exists is promising (Donaldson, Lozy, & Galjour, 2019; Foley, Dozier, & Lessor, 2019; Ling & Barnett, 2013; Murphy et al., 2007; Swiezy et al., 1992; Wiskow et al., 2019). Research on the implementation of the GBG with low-incidence disability populations is also limited but promising (Axelrod, 1973; Breeman et al., 2016; Groves & Austin, 2019; Joslyn et al., 2019; Popkin & Skinner, 2003; Salend, Reynolds, & Coyle, 1989; Wiskow et al., 2018). For example, Wiskow et al. (2018) examined the effects of the GBG on the disruptive behavior of a 4-year-old diagnosed with fetal alcohol syndrome.  Disruptive behavior was reduced to near-zero levels, similar to that of his peers playing the GBG (Wiskow et al., 2018).

Researchers have investigated several concerns about generalization and maintenance of the GBG and found that its effects appear to extend beyond the periods in which the game is conducted (Donaldson, Wiskow, & Soto, 2015). One such concern was behavioral contrast (Cooper, Heron, & Heward, 2014), which could result in undesired target behaviors increasing when the game is not being played. Donaldson et al. (2015) found that while the intervention does not result in a decrease in disruptive behavior during non-intervention activities, disruptive behaviors during those activities did not exceed baseline levels. Another concern was the removal of the GBG: Behaviors often revert to baseline levels following the termination of the game (Donaldson et al., 2015; Donaldson et al., 2019). To target this concern, Donaldson et al. (2019) systematically removed components and found that all could be faded without an increase in disruptive behavior except for vocal corrective feedback when rules were broken. Despite conflicting short-term maintenance results, preliminary long-term results suggest the GBG could have effects reaching into young adulthood under conditions providing naturally occurring reinforcement of target behaviors, although additional longitudinal research is needed (Donaldson et al., 2019; Groves & Austin, 2019; Kellam et al., 2008; Kellam et al., 2014).

What’s Next: A Model for Dissemination

 In order to be widely disseminated, an intervention should be effective, reproducible, and affordable (Fixsen & Blase, 1993). Geller (1989) suggests using social marketing (i.e., advocating socially beneficial ideas) as a “bridge” (p. 27) between our knowledge of human behavior and the wide-scale social implementation of that knowledge. The GBG has made its way into some 650 schools across 32 states (Weis, Osborne, & Dean, 2015) in the form of an adaptation known as the PAX Good Behavior Game (PAX GBG). The disseminators of this trademarked intervention, whose website does not explicitly mention the game’s behavior analytic origins, market benefits of the PAX GBG (“The Science,” 2018) that have not been examined in the behavior-analytic literature (e.g., decreased suicidal ideation, decreased illicit drug abuse) or are not typically considered behavioral in nature, (e.g., changes in neural connectivity and epigenetic makeup). The PAX GBG uses the elements of the traditional GBG to build self-regulation skills and promote social and emotional learning (“What is PAX?,” 2018).

Research involving implementation of the PAX GBG is limited (Domitrovich et al., 2016; Smith, Osgood, Oh, & Caldwell, 2018). PAX GBG has been found to increase prosocial behavior (i.e., “caring,” sharing, listening) and decrease hyperactivity (compared to a control group) when implemented during an afterschool program (Smith, Osgood, Oh, & Caldwell, 2018). Domitrovich et al. (2016) found the PAX GBG had effects on teacher behavior as well, decreasing teacher “burnout.” Much more research is needed in the area of teacher behavior (Domitrovich et al., 2016; Smith et al., 2018), as well as on the effects of GBG on school building climate (Domitrovich et al., 2016), family life (Smith et al., 2018), and skill acquisition (e.g., “agency,” “delay gratification and impulsivity,” “pro-social behavior,” “self-regulation”; Weis et al., 2015; “What is PAX?”, 2018). The PAX GBG has also had important effects on academic outcomes, an understudied target compared to disruptive behavior (Weis et al., 2015): Following a countywide rollout of PAX GBG across six school districts, small but significant effects were found on participants’ end-of-year reading and math scores (Weis et al., 2015).

While, again, there is a notable absence of any mention of behavior analysis on the PAX GBG website, behavior analysts could cautiously view the PAX intervention and subsequent research as a potential model for dissemination. Important components of the PAX GBG package, which make it reproducible and possibly cost effective, include on-site training, training materials (e.g., manuals), and materials necessary for implementation (e.g., posters, timers, and suggested rewards), which might be useful to implementers who do not have a behavior analytic background (Weis et al., 2015).

Future research on the GBG should continue to evaluate the effects of the intervention on novel populations, settings, and target behaviors (i.e., behaviors beyond disruptive behavior), as well as variations in procedures and materials that could have an effect on the efficiency and acceptability of the intervention (Joslyn et al., 2019). Tingstrom, Sterling-Turner, & Wilczynski (2006) suggested two areas of future research in their literature review of the GBG: negative group peer pressure (Groves & Austin, 2019) and effects on social behavior within populations where limited research has been done, such as preschoolers. Research on the GBG has in fact begun to focus on both preschoolers, as previously noted (Donaldson, Lozy, & Galjour, 2019; Foley, Dozier, & Lessor, 2019; Ling & Barnett, 2013; Murphy et al., 2007; Swiezy et al., 1992; Wiskow et al., 2019) as well as social behavior (Groves & Austin, 2019).

Perhaps most important, we might focus empirically on large-scale behavior analytic dissemination of the GBG. What, if anything, does behavior analytic technology gain or lose when disseminated by non-behavior analysts, as in the case of the PAX GBG? What will be required for the GBG to be implemented with maximum fidelity across the widest possible number of appropriate settings, behaviors, and populations? Acceptance by policymakers, a treatment package that can be implemented by nonspecialists (such as that the PAX GBG has provided), and sufficient funding will almost certainly be on the list. How to accomplish these things is a question that has not gone unasked in behavior analysis, but that needs more empirical research. The GBG, with its relatively long history and large body of literature, may be a good case study in dissemination.

Detrich (2018) proposed that when functionally, rather than topographically defined, dissemination has been accomplished “only when a practice has been adopted” (p. 2), and likened what he termed passive dissemination, such as publishing and presenting, to “train and hope” generalization (Stokes & Baer, 1977). Active dissemination, by contrast, involves working directly with target audiences (Detrich, 2018).  Pennypacker (1986), writing about his experience in disseminating a behavior analytic technology for teaching breast self-exams, explained that he and his colleagues ultimately decided to take it to the marketplace themselves, “setting as our objective the delivery of the benefits of the research to the maximum number of women, while allowing the minimum possible dilution in quality” (p. 152) and “using the experimental method to teach ourselves how to transfer a behavioral technology to the marketplace without destroying it” (p. 152). It may be necessary for behavior analysts to step into the marketplace once again to apply these lessons to the GBG and learn new ones about how to scale up their technologies.

https://pixabay.com/photos/classroom-school-education-learning-2093744/

References