The effects of the GBG on young adult outcomes suggest that directing a universal intervention at the first grade teacher and classroom to improve socializing children into the role of student and classroom behavior management has both immediate and long-term benefits. This is particularly true for males and even more so for those at higher levels of aggressive, disruptive behavior early in first grade. The GBG is directed at improving early mastery of appropriate student behavior, thereby improving the developmental trajectories and preventing a set of problem outcomes, mainly those that could be thought of as externalizing outcomes. summarizes our significant findings, with fitted proportions calculated in each intervention condition based on individual-and classroom-level analytical models. We emphasize these particular two-level model results because they incorporate our theory of individual-level maladaptive aggressive, disruptive behavior embedded in the classroom social adaptational process involving teacher and classmates, thus providing the clearest test of our theory.
Comparison of Rates of Young Adult Outcomes for Cohort 1 GBG and Internal GBG Controls
Life course/social field theory
postulates that aggressive, disruptive maladaptive behavior in the social field of the classroom will lead to long term poor outcomes in later social fields, and that directing an intervention, such as the GBG, at the social adaptational classroom process will ameliorate such outcomes. This central concept of the theory suggests a hypothetical mechanism for how the GBG, a two-year intervention in first grade, yielded such long-term outcomes in young adulthood. Children who are successfully taught how to master the social task demands of the first grade classroom may carry with them to higher age and grade levels the experience of mastery (“I can do it”). In addition, the GBG involves teachers determining the classmate team membership and making certain that the teams are mixed in regard to gender and behavior. Van Lier et al. (2005)
reported the GBG decreased antisocial behavioral outcomes among high-risk youth coinciding with decreased affiliations with deviant peers and lower rates of peer rejection, indicating these factors may mediate the beneficial effect of the GBG. It is also helpful to look back at the immediate earlier effects of the GBG to assess potential mechanisms for these long-term effects. Besides the obvious advantages of improved behavior in first grade and elementary school, the GBG also reduced off-task behavior, thus allowing teachers more time/opportunity to teach (Brown, 1993
) and the beneficial consequences of mastery of academic subjects such as reading skills required for further mastery later on. The results reported here are consistent with life course/social field theory with regard to the importance of early mastery of social task demands of the classroom in promoting later successful social adaptation not only in school but also in other main social fields.
The present trial had a limited ability to carry out more detailed mediational models to examine specific mechanisms leading to improved child outcomes. We plan to report on the role of early responses to the GBG in mediating long-term outcomes, but other analyses will be required to more fully test these and other mediational processes. For example, the current third-generation trial in Baltimore incorporates independent observation of children’s behavior, child outcomes, and the quality and fidelity of implementation of teachers’ classroom behavioral management.
By design the intent-to-treat analyses reported here examine two single measurement times separated by 14 years. Thus, the malleability of the relationship between first grade and young adulthood is summarized in these analyses without a developmental examination. In this way, these analyses are notably blind to detecting impact through changes in growth trajectories and the known interrelationships of SAS and PWB over time, especially those involving the co-occurrence of substance use and mental disorders. Other analyses reported previously (Muthén et al., 2002
; Wang et al., 2005
) address developmental questions through different forms of growth mixture analyses. Other papers in this special issue report survival analyses and GGMM to examine not only baseline aggressive, disruptive behavior but also developmental trajectories leading to the young adult outcomes of ASPD, violent and criminal behavior, and suicide ideation and attempts (Petras et al., in press, this issue
; Wilcox et al., in press, this issue
). In the case of ASPD and violent and criminal behavior, the highly aggressive, disruptive first graders had trajectories in both Cohorts 1 and 2 leading to high rates of outcomes and significant impact of the GBG in reducing the rates of juvenile delinquency measured by court records and violent and criminal behavior measured by records of incarceration. Baseline aggressive, disruptive behavior coupled with the developmental trajectories reveal GBG impact clearly among those children with high initial levels and persistent trajectories of aggressive, disruptive behavior (Petras et al., in press, this issue
For Cohort 1 in this paper there are clear, consistent findings in the results across these young adult outcomes, although findings differ somewhat in the details. We found that GBG had a significant and substantial impact among all of the externalizing behaviors. The impact of GBG was overall highest among higher aggressive, disruptive first-grade males compared with less or moderately aggressive, disruptive males. In addition, the effects of GBG on all categories of substance abuse/dependence disorders were consistently strong, whereas the effects on psychological disorders in these intent-to-treat analyses were consistently non-significant. In general, risk did not increase with baseline aggressive, disruptive behavior for females, and for all outcomes, the effects appeared stronger for males than for females. Interestingly, however, the outcomes where the gender differences were the smallest, lifetime regular smoking and alcohol abuse/dependence disorders may be considered somewhat less antisocial outcomes.
These GBG outcome differences among males and females suggest that early developmental processes that are salient for males may be different than those for females. The differences may be critical in understanding the gender differences in developmental and in intervention outcomes, and in directing preventive interventions at those that are more germane to females. Early aggressive, disruptive behavior may not have the same developmental salience for females as for males. We reported in 1983 on gender differences in the Woodlawn longitudinal study that among girls, more developmental continuity from first grade through adolescence was found in internalizing symptoms such as depression and anxiety as compared with boys. Among boys, early aggressive, disruptive behavior was more developmentally predictive of later teenage aggression and drug use from first grade to ages 16–17 (Kellam et al., 1983
; Ensminger et al., 1983
). We have noted some gender differences in these data; for example, the predictability of later external behavior problems as a function of first-grade aggressive, disruptive behavior is far stronger for males than females. Such differences in predictability do not seem to be explained by the generally lower level of aggressive, disruptive behavior in girls compared with boys, although the problem may lie in our aggressive, disruptive measures being more fitting for males than females.
There were some other differences in impact across outcomes as well. The impact of the GBG was the greatest among those with the highest aggressive, disruptive behavior in first grade in the case of the two outcomes most closely related to illegal behavior—drug abuse/dependence disorders and ASPD. These outcomes are also where we found the greatest gender differences. We infer from these results that the GBG has the greatest impact on the highest levels of aggressive, disruptive behavior (highly aggressive, disruptive males compared to lower aggressive, disruptive males and females) and where the outcome is the more illegal. This suggests that early aggressive, disruptive behavior plays a specific and important role in the development and etiology of the profile of outcomes addressed in this paper—the higher the level of early aggressive, disruptive behavior and the more illegal the outcome, the more GBG impact.
The effects on males’ high school graduation resembled that for ASPD, but the effect was not significant due to significant random effects at the school/community level. An additional point is high school graduation may depend more on the course of aggressive, disruptive behavior, not merely the rating of it in the fall of the first grade. In any case, more research is needed to explain the variation at the school/community level combined with the impact of GBG on high school graduation rates.
We found null effects of the GBG on anxiety and depressive disorders. Similar to findings in our previous studies, we note that at a population level, early antecedents of depressive and anxious symptoms are distinct from those of externalizing behaviors (Kellam et al., 1983
). We note that the pattern of impact on GAD, although non-significant, is intriguing. Because the statistical power in these models depends on the prevalence of the outcome (as well as the strength of association between baseline aggressive, disruptive behavior and the outcome), the low frequency of GAD may partly explain why we failed to find effects for this outcome. By using symptom- and/or syndrome-level measures, we may well find significant impacts of intervention on internalizing psychiatric disorders. Furthermore, GBG may have an impact on depression or anxiety disorders through a reduction in co-occurrence and co-morbidity. It is noteworthy that GBG had an impact on suicide ideation and suicide attempts among both genders (see Wilcox et al., 2008 for discussion of the role of early depressive symptoms).
For Cohort 2, we concluded that there was a somewhat dissimilar pattern of replication of beneficial GBG findings for drug abuse/dependence disorders. In the second cohort the impact was stronger at the lower levels of aggressive, disruptive behavior while in Cohort 1 the effect was apparent across all levels and strongest at the higher levels of aggressive, disruptive behavior. Furthermore, there was a similar pattern of GBG benefit for regular smoking, but it was non-significant. There is a suggestion of beneficial impact on male MDD rates. The second cohort analyses reported in this paper for other outcomes yielded non-significant findings but in the same direction as Cohort 1. There are three major hypotheses for these diminished findings in the second cohort. First, the GBG showed a much reduced effect on short-term aggressive, disruptive behavior in the second cohort compared with the first, and our earlier finding of the impact of the GBG on adolescent smoking appeared not to be mediated through aggressive, disruptive behavior (Kellam and Anthony, 1998
). Second, we hypothesize that GBG was implemented in the second cohort with less precision than it was in the first cohort because we did not have in place sufficient mentoring and monitoring procedures, resulting in two types of consequences: a diminished impact and a shift in for whom impact occurred. If this reasoning is correct, it would explain the less significant impact and/or a shift in which children were benefited from more highly aggressive, disruptive to less aggressive, disruptive in the second cohort. Third, another important area for further study is the somewhat weaker relationship in the second cohort between aggressive, disruptive behavior in first grade and adult behavioral outcomes. Classroom variability and heterogeneity across the three control conditions were markedly higher in the second cohort compared to the first so that statistical power was reduced. Despite the reduced mentoring and monitoring, the GBG effect remained for the highly aggressive, disruptive males for ASPD and violent and criminal behavior (Petras et al., in press, this issue
) and drug abuse/dependence disorders for Cohort 2, although in the case of the drug abuse/dependence disorders there was a shift in impact towards the lesser aggressive, disruptive males.
Although we did not measure the quality of GBG implementation in this first generation of trials, we strongly hypothesize that the GBG must be carried out with precision including continuing mentoring and monitoring. In the second generation of trials in the early 1990s where we did measure implementation, the results revealed marked reduction in impact when that intervention (combined classroom behavior management and enhanced curriculum/instruction) was done with less precision (Ialongo et al., 1999
). This problem of low sustainability without continuing mentoring and monitoring is being reported by other investigators and is the new research frontier as we move through the phases of prevention research from efficacy through effectiveness into the problem of sustainability of new practices and ultimately into system-wide fidelity as programs are disseminated (Elliott and Mihalic, 2004
; Hallfors and Godette, 2002
). Olds and colleagues (2003)
have reported success using an ongoing monitoring system that provides information on fidelity to mentors as well as to the researchers who remain in close partnership with agencies carrying out the Nurse-Family Partnership program. In the current, third-generation GBG trial in Baltimore, we are testing a multilevel mentoring and monitoring structure while we move the program toward dissemination as the data warrant.
From a prevention science and policy perspective there are important lessons if these results are replicated in places other than Baltimore. First, the overall strategy of directing a universal intervention at a shared antecedent of a set of later problem outcomes appears to have been successful. This success provides grounds for real optimism that universal prevention strategies such as the GBG can be done early and economically and can address a set of outcomes, not just one at a time. The importance of this finding cannot be overemphasized. It suggests that each problem outcome may not require a separate early preventive intervention, but that a set of outcomes can respond to an intervention aimed at a single shared antecedent risk factor.
Second, the GBG generally had its strongest effect on the highest-risk youth. We have reported elsewhere that such children, particularly males in highly aggressive, disruptive, disrupted classrooms, are the children most at risk for continuing and later problem behavior (Kellam et al., 1998a
). Providing the teacher with tools for socializing children into the role of student and managing his or her classroom appears to reduce the high risk for these early aggressive, disruptive, disruptive children, and demonstrates the utility of this universal intervention for maintaining such children in the mainstream classroom and helping them to develop successfully.
Third, the results point to new areas for training new teachers and in-service training for more experienced ones. Teachers are not as often trained in classroom behavior management as these results strongly suggest they should be. Our data indicate that up to half of the teachers are not prepared to manage their first-grade classrooms effectively, and effective tools such as the GBG can and must be provided to teachers (Kellam et al., 1998a
Fourth, the results underline the vital importance of the first-grade classroom as a social field where developmental trajectories are displayed and further shaped beyond the earlier social fields of family or preschool settings. This is in no way meant to deny the importance of earlier interactions between parents and children and preschool teacher and students. Social adaptation in first grade left to itself without intervention is strongly predictive of later social adaptation, but by first grade there is still considerable malleability—room for improvement—particularly among higher-risk males. Universal interventions in the first-grade classroom can be decisive in setting the direction for life course in school and beyond. The strength and clarity of the impact of this precisely directed early intervention emphasizes the validity of early universal interventions in this critical social field. Finding the most salient early antecedents for girls will allow extending the testing of this strategy.
Fifth, continual monitoring and mentoring at multiple levels of a school district are likely requirements of moving to high levels of sustained effectiveness in interventions in classrooms. The No Child Left Behind Act of 2001 (Public Law 107–110) emphasizes information systems that can play a vital role in monitoring and mentoring. The current high level of attention being paid to measuring school building levels of achievement should also focus on classrooms and individual children at the micro level, as well as to city- and statewide information systems at the more macro levels. A multilevel monitoring system can provide a framework for understanding what problems exist and where. Randomized field trials can provide answers to what works, for which children, under what circumstances. Such trials should be conducted within a multistage epidemiological framework, however. Statewide information systems allow for demographic epidemiology of where the nonrandom distribution of problems is occurring. A second-stage representative sample can provide a sampling frame for causal modeling to examine potential early risk factors that are then potential targets for preventive interventions. The third stage, the randomized field trial, can then be conducted on a representative third-stage sample directed at the early antecedent.
In the end, there is still a great need for replicating the GBG and other universal prevention programs in other school districts with similar and different social, ethnic, and economic characteristics. Even as we produce evidence of effective programs, we face new challenges: how to sustain teacher practices over consecutive cohorts of children and how to go to scale within and across school districts. Such questions need to be addressed at the same time that we are replicating trials of the GBG and other programs in different contexts. We have already learned in this first-generation trial that merely training teachers for 1 year is not sufficient support to sustain new practices. The third-generation trial under way in Baltimore is testing a multilevel mentoring and monitoring structure, and similar and different designs for dissemination are needed. How to test these emerging designs is a challenge to the public health prevention and education research fields.
Ultimately, prevention research in the public education and public health fields needs to be integrated. The risk factors for outcomes in each field overlap and are in many cases the same. Both require basic partnerships between scientists and school districts. This particular study is the product of such a partnership. To develop a design and carry out randomization at multiple levels as described here required a continuing partnership with the BCPSS, just as was required in the early Woodlawn studies beginning in 1964 (Kellam, 2000
). The studies reported here and elsewhere from the Baltimore partnership have been based on strong mutual self-interests of the BCPSS, the families and their children, and the researchers. We have functioned over the 21 years of our work together as a research and development arm of BCPSS, while carrying out basic and applied prevention research. The ongoing third-generation trial in Baltimore is part of the BCPSS Master Plan, and it represents the kind of real-world mutual benefit that can occur with such a partnership.