A Meta-Analysis of Dropout Prevention Outcomes and Strategies

A Technical Report by:

Shanan L. Chappell, PhD • Old Dominion University

Patrick O’Connor, PhD • Kent State University

Cairen Withington, MA • National Dropout Prevention Center/Network

Dolores A. Stegelin, PhD • Clemson University

April 2015

Please direct all correspondence regarding this meta-analysis to:

Dr. Shanan L. Chappell

Research Assistant Professor

The Center for Educational Partnerships

Darden College of Education

Old Dominion University

Norfolk, VA 23508



For additional information regarding the NDPC/N and its work contact:

Dr. Sandy Addis

National Dropout Prevention Center/Network

Eugene T. Moore School of Education

Clemson University

209 Martin Street

Clemson, SC 29631



The authors would like to acknowledge the contributions of Terry Pickeral, Eurmon Hervey, Jr., Jay Smink, Nicholas Benson, Monique Colclough, Anna Veprinsky, and Julia Zaharieva.

Recommended Citation:
Chappell, S. L., O’Connor, P., Withington, C., & Stegelin, D. A. (2015, April). A meta-analysis of dropout prevention outcomes and strategies (A Technical Report in Collaboration with The Center for Educational Partnerships at Old Dominion University). Clemson, SC: National Dropout Prevention Center/Network at Clemson University.

©2015, The Center for Educational Partnerships at Old Dominion University and the National Dropout Prevention Center/Network at Clemson University.

The opinions expressed herein are those of the authors, and do not necessarily reflect the views, position, or policy of Old Dominion University or Clemson University.


A History of the Discovery of 15 Effective Strategies for Dropout Prevention

Jay Smink, PhD, Retired Executive Director, National Dropout Prevention Center/Network

Almost from the start of the public schools system in America, we have had students leaving school without high school diplomas. However, the dropout issue did not rise to the level of significance it has today until the early 1980’s when social pressures along with business leaders leveraged their influence on educators to address the dropout issue in an attempt to find an acceptable workforce. The business and corporate world began to recognize how America’s graduates were poorly prepared for the workplace compared to graduates from other nations. This realization sparked many different initiatives to target the world of school dropouts including two developmental approaches with implications to the solutions and strategies now used to address school dropouts.

One initiative was driven by a group of business leaders to form a national center focused on the dropout crisis. Concurrently, there was a similar initiative to form a national network of practitioners and researchers to address the dropout issue. Both initiatives were birthed almost simultaneously and eventually combined forces to form The National Dropout Prevention Center/Network (the Center), located at Clemson University in South Carolina. Begun in 1986, the Center was developed to serve as a clearinghouse on issues related to dropout prevention and to offer effective strategies designed to increase the graduation rate in America's schools.

From the beginning of the Center, one major mission was to develop a database of model dropout prevention programs. In addition, the initial database had numerous other components including professional development activities, active research projects, and a listing of organizations providing professional services related to the dropout issue. Separate lists of leading experts in the field of dropouts were also included and eventually the database was refined to focus mostly on model dropout prevention programs.

Throughout the refinement of the database it became clear that practitioners, business and community leaders, and legislators were focused on securing research and statistics related to the dropout issue. However, the dominant request from practitioners was to identify and implement dropout prevention programs that would “fix” the dropout crisis. These needs were clearly met by the database because it grew to include more than 500 model dropout prevention programs. These programs were self-reported from local education agencies, state education agencies, and national organizations that nominated their successful lighthouse programs to be included in the prestigious and user-friendly database.

The database provided an excellent pool of successful model programs to be promoted by the Center and then to be reviewed by local districts and schools searching for proven dropout prevention programs. However, as practitioners were gaining experiences in matching programs to meet their needs, they began to request additional information about the programs. This request for additional information about programs prompted the Center to initiate a more thorough and comprehensive analysis of the model programs.

In the early 1990s, I initiated a systematic review focused on all of the programs in the database. The intent of the research was to review each program to see the multiple design patterns in the programs such as objectives, instructional activities and tasks, staff training and engagement activities, management practices, and evaluation methods and findings. This analysis attempted to find common ground in the programs that could be used to guide potential users in their planning efforts to reduce the school dropout rates. The result of this thorough review of each program identified about 50-55 categories of descriptors that were most common in the database of programs. Some of the early findings confirmed the three structures used in the database, which were characterized as prevention, intervention, and recovery programs. Other structures that became obvious were community settings such as rural, suburban, and urban. Another main focus was the location of the high-risk students, which was either elementary, middle, or high school.

The Center began to promote these 50-55 categories of additional information and quickly found the information was useful but it was too broad and very difficult to use effectively at the school level or in community-based programs. Therefore, a more systematic and in-depth analysis was initiated to gain easier access to each of the programs and make a more accurate match between the needs of schools and communities and strategies used by successful programs. Completed in 2000, this second review of the model programs and their related categories provided a higher level of descriptive information and provided the first glimmer at a smaller number, approximately 25 categories, and a different structure for the commonalties found in the model programs.

Continued research of this information about models programs along with feedback from practitioners led directly to the 15 Effective Strategies for Dropout Prevention now promoted by the Center and used in many state agencies and school districts across America. However, there remained one last refinement of the 15 strategies and that was the final structure used to present the strategies to a wider audience of school and community leaders and to promote their use in a comprehensive dropout prevention plan at the school district level. These final 15 strategies are examined in this report, bringing the work on the NDPC to a more scientific level to further promote the effectiveness of the strategies in combating the national dropout crisis.

A Meta-Analysis of Dropout Prevention Outcomes and Strategies

The impact of dropping out of high school has implications on individuals as well as society overall. Research tells us that about one million students drop out every year, and nearly half of all African Americans, Hispanics, and Native Americans fail to graduate (Bridgeland, DiIulio, & Morison, 2006). These alarming statistics have far-reaching consequences for these individuals and the country’s economic and civic health, as dropouts are more likely than high school graduates to be unemployed, in poor health, living in poverty, in prison, on public assistance, and to have children who also drop out of high school (Bridgeland et al., 2006). On average, a high school dropout earns $9,200 less per year than a high school graduate, and about $1 million less over a lifetime than a college graduate.

Over the past three decades, researchers and organizations have sought to identify those factors that put students at risk of dropping out of high school. The research indicates that dropout risk factors are many and are multifaceted (Hammond, Linton, Smink, & Drew, 2007; Smink & Schargel, 2004). As we become familiar with dropout risk factors and the impact of dropout on individuals and our society, it has also become clear that we need to determine the best solutions for each risk factor to reduce or eliminate the risks. The ultimate goal is to help schools create the most engaging environments to successfully prepare kids to graduate with clear pathways to rewarding postsecondary lives.

Although numerous evaluations have been conducted on the impact of dropout prevention programs, there is little evidence that relates particular prevention strategies to dropout and graduation rates. In fact, there have been no rigorous syntheses of dropout prevention strategies to help identify the most successful features of these efforts and assist educators and policymakers categorize programs based on scientific evidence. We initiated this study to provide sound, empirical support of the most effective strategies for preventing dropout and improving graduation rates for all students.

Theoretical Framework

Since 1986, the National Dropout Prevention Center (NDPC) has focused on its mission to reduce the national dropout rate by meeting the needs of youth in at-risk situations. The Center has partnered with school districts and statewide departments of education, sponsored workshops and national conferences for researchers and practitioners, and collaborated with policymakers to help meet their respective goals. During this time, researchers at the NDPC identified 15 research-based strategies that have the most positive impact on the high school graduation rate. Though these strategies are independent of one another, in practice they work well together and frequently overlap. Further, the NDPC has gathered evidence that these strategies can be successful in all school levels from K-12 and in rural, suburban, or urban settings (Smink & Schargel, 2004). However, while the literature identifies methods that may in part alleviate or eliminate the at-risk population, no synthesis of the literature has been conducted to establish which program strategies are most successful in reducing dropout rates and improving graduation rates.

Based on the body of research on dropout prevention, many schools, districts, and states are making appreciable gains in graduation rates. Over the most recent seven years, 42 states have, on average, increased their averaged freshman graduation rates by up to 2.4% per year. However, not all schools, nor all states, have shown such progress. In addition, funding for education is often the first on the chopping block in many areas of the nation (see Trends in AFGR, Dropout, and Funding, NDPC, 2013a). Cuts in funding and continued struggles to address the dropout issue mean that efficiency and effectiveness are perhaps even more crucial today than ever.

Purpose of the Study and Research Questions

The purpose of this study was to provide an overall estimated effect size of dropout prevention efforts on dropout and graduation rates and to provide estimated effect sizes of research-based strategies on these two outcome variables. However, our search of the literature rendered less than 30 evaluations that (a) included graduation rate as the dependent variable and (b) met our inclusion criteria. This small number of evaluations did not allow for examination of the strategies as predictors of graduation rates. However, we foresee that changing in the near future as states have begun to include graduation rate as a measure of student and school success, which should translate into more studies/evaluations that use graduation rate as the dependent variable.

Our final analyses focused on the following four research questions:

  1. 1. What are the mean effect size estimates of dropout prevention programs on dropout rates?

  2. 2. What are the mean effect size estimates of dropout prevention programs on 4-year graduation rates?

  3. 3. How do dropout rate effect sizes vary as a function of prevention strategies?

  4. 4. Which dropout prevention strategies are significant predictors of the overall dropout rate?


We employed a meta-analysis research design (Borenstein, Hedges, Higgins, & Rothstein, 2009) to estimate a mean effect size of dropout prevention efforts on dropout and graduation rates. The meta-analytic approach requires a number of steps from determining inclusion criteria, identifying variables and predictors to examine, conducting a search for appropriate studies, and coding studies according to the chosen predictors. Once all data were extracted and the studies were coded for predictors, we calculated overall estimates of effect size for the two dependent variables (separate analyses were conducted for each variable) and then for each strategy’s impact on dropout rates. We also conducted a meta-regression to determine which dropout prevention strategies were statistically significant predictors of the dropout rate effect size. Institutional Review Board approval was not necessary because this study entailed a review of previously published research and the analyses included secondary data.

Selection Criteria

To be included in this meta-analysis, dropout prevention program evaluations [1]

were required to meet the following criteria:

Data Sources and Sample

All available dropout prevention program evaluations were examined for inclusion in this study. First, we reviewed the Model Programs Database established by the National Dropout Prevention Center. The NDPC established this database to identify programs according to the level of evidence of effectiveness. All those programs categorized as having a strong evidence of success were reviewed to gather evaluations conducted using experimental designs.

We also accessed the What Works Clearinghouse (WWC), a resource provided by the U.S. Department of Education’s Institute of Education Sciences. The WWC presents some evidence of selected dropout prevention programs that have been submitted for inclusion in the WWC. Many of these did not meet WWC standards, but 19 dropout prevention program evaluations presented small or medium to large effect sizes (U.S. Department of Education, 2013).

Additionally, other database searches were conducted seeking dropout prevention program evaluations. ERIC (Education Resources Information Center), a source for full-text articles and abstracts from educational journals; and PsycINFO, a database containing full-text and summarized scholarly articles, books, and dissertations; were accessed in this search. Reference sections and bibliographies from evaluations found through these search methods were also reviewed for additional studies, which were included in this analysis. Any evaluation that met the inclusion criteria was included.

We reviewed over 500 studies that evaluated dropout prevention efforts. Overall, we included 67 effect sizes [2]

from 35 programs in the dropout rate analysis and 26 effect sizes from 16 programs in the graduation rate analysis that met our inclusion criteria. Sample sizes were n = 42,291 (23,535 treatment and 18,756 comparison) for the dropout rate analysis and n = 310,257 (43,111 treatment and 267,146 comparison) for the graduation rate analysis. Technical Appendix A lists the studies included in this report along with their corresponding numbers of dropout and/or graduation rate effect sizes.

Data extracted from each evaluation included sample sizes, means, and standard deviations for treatment and comparison groups, and effect sizes. We used the standardized mean difference value as effect size to allow for a standardized measure of impact across studies. Where standardized mean difference effect sizes were not reported, we gathered the reported outcome statistics and transformed them into standardized mean difference effect sizes using the Comprehensive Meta-Analysis software (Borenstein, Hedges, Higgins, & Rothstein, 2013). Program information was also gathered from the studies, including program name, program strategies (these were included as predictors in the analyses), and year the study was conducted.

Model diagnostics. Diagnostic procedures indicated that six of the effect sizes for the dropout rate analyses violated assumptions of the model. These six effect sizes were removed from all subsequent dropout rate analyses. All of the graduation rate effect sizes fell within the acceptable diagnostic ranges and were retained in the appropriate analysis. The final analysis sample included 61 effect sizes from 30 programs, with n = 38,155 (21,255 treatment and 16,900 comparison) for the dropout rate analyses and 26 effect sizes from 16 programs, with n = 310,257 (43,111 treatment and 267,146 comparison) for the graduation rate analysis.

Analytic Approach

The meta-analytic approach allows us to either determine a fixed effect or calculate an overall mean effect. The fixed effect approach is used when all the included studies have very similar individual effect sizes. When that is not the case, the random effects model is appropriate (Borenstein et al., 2009). The random effects model allows us to calculate an overall mean effect across the varying effect sizes using a weighting procedure. We used the random effects model and applied the weighted linear combination approach to allow studies with larger sample sizes to carry more weight. Because studies with large sample sizes are considered to have higher statistical power, this provided a more accurate and representative overall mean effect size for each of our outcome variables and each of the strategies we examined as predictors of dropout rates (Hedges & Olkin, 1985).

Predictors/Strategy Variables

In our analysis, we included program strategies as predictor variables to examine if the dropout rate varied as a function of these program strategies and to determine if any of the strategies statistically significantly predicted the overall effect size for the dropout rate. First, we needed to establish effective strategies for dropout prevention and graduation rate improvement. To do this, we accessed the current Effective Strategies for Dropout Prevention Web page (2013b) on the National Dropout Prevention Center’s Web site. The NDPC has identified 15 strategies that have been shown in the literature to have positive impacts on dropout rates. These strategies have been implemented successfully at all education levels and environments throughout the nation. The strategies are: Systemic Renewal, School-Community Collaboration, Safe Learning Environments, Family Engagement, Early Childhood Education, Early Literacy Development, Mentoring/Tutoring, Service-Learning, Alternative Schooling, After-School Opportunities, Professional Development, Active Learning, Educational Technology, Individualized Instruction, and Career and Technology Education (CTE; see Appendix B for strategy definitions and descriptions).

These overarching strategies were further partitioned into definable, tangible strategies to be used as predictors in the mixed effects models. To do this, we relied upon the 2007 report by Hammond et al. titled Dropout Risk Factors and Exemplary Programs: A Technical Report. The strategies we examined in our analysis are:

We also intended to include Adult Education, Credit Recovery, and Gang Prevention/Intervention as predictors of the dropout rate effect size in our analyses but these three strategies were not widely represented in the included evaluations and therefore we were unable to calculate effect sizes for each. See the attached Appendix for complete strategy definitions and example of services provided under each strategy.

Coding Procedures

Each of the four authors served as raters for the strategy coding procedures. The list of programs to be included in the final analyses was provided to each rater along with an electronic coding spreadsheet which included the programs to be coded and complete strategy definitions, descriptions, and services that applied to each strategy. Additionally, coders attended a training session that provided hands-on coding practice. We employed a dummy-coding approach, where a study received a code of “1” if it employed a given strategy and “0” if it did not. A total of 34 programs were coded by 2 independent coders each, with each coder rating 17 programs (not all 34 programs ended up in the final analyses as indicated in the Diagnostics section above). To determine interrater reliability, Pearson product moment correlations were conducted between the two coders for each program (Multon, 2010). [3]

Correlations were calculated for each rating team (coders 1 and 2, coders 1 and 3, coders 1 and 4, coders 2 and 3, coders 2 and 4, and coders 3 and 4) and then averaged to determine an overall interrater reliability coefficient of .71, which exceeded the acceptable rate of .70 (Multon, 2010).


Research Questions One and Two

Standardized mean difference effects and sample sizes for each included study were used to calculate weights to be used in the random effects model for the dropout rate analysis. Using Borenstein et al.’s (2009) estimation of effect sizes, the overall weighted mean effect on dropout rates was estimated at d+ = .15 with an overall variance of s2(d+) = 0.001. The standard error of the mean was .03. To test for statistical significance of the overall effect size, 95% confidence intervals were calculated using the standard error of the mean and standard normal distribution values of +/- 1.96. The results indicated that the overall mean effect was statistically significant, with δL = 0.08 to δU = 0.21. Table 1 reports model statistics for the dropout and graduation rate analyses.

Similarly, standardized mean difference effects and sample sizes for each included study were used to calculated weights to be used in the random effects model for the graduation rate analysis. The overall weighted mean effect on dropout rates was estimated at d+ = .36 with an overall variance of s2(d+) = 0.005. The standard error of the mean was .07. To test for statistical significance of the overall effect size, 95% confidence intervals were calculated using the standard error of the mean and standard normal distribution values of +/- 1.96. The results indicated that the overall mean effect was statistically significant, with δL = 0.23 to δU = 0.49.

Sensitivity Analysis

To determine if the results of the two main analyses would change if one study was removed, we conducted sensitivity analyses using the Comprehensive Meta-Analysis software. For the dropout rate effect sizes, the overall values with one study removed ranged from .136 to .157 (overall effect size was .149) and were all statistically significant (p < .001). This indicated that the results from the overall dropout rate analysis would not be different if any of the 61 included effect sizes were removed. Results from the graduation rate sensitivity analysis revealed effect sizes of .313 to .391 (overall effect size was .36) and were all statistically significant at the .001 alpha level. Again, this indicated that the results of the overall graduation rate effect size would not differ if any of the 26 effect sizes were removed from the analysis.

Publication Bias

We used Rosenthal’s Fail Safe N (Rosenthal, 1979; Begg & Mazumdar, 1994) approach to examine publication bias for the dropout rate and graduation rate analyses. Assuming a nil effect, we would need 1,677 studies for the dropout rate analysis to bring the significance level above the .05 alpha level. For the graduation rate analysis, we would need 1,223 studies to bring the overall effect to a level of nonsignificance.

Research Question Three

We also conducted group analyses to determine how the dropout rate varied as a function of prevention strategies employed by the programs evaluated in the included studies. We used Borenstein et al.’s (2009) mixed effects model to estimate overall effects of programs that employed each of the 12 strategies identified above. In these analyses, we were only measuring the effect of each strategy individually on the dropout rate. Research question four addresses the impact of all of the strategies together as determined by the meta-regression procedure.

Four strategies revealed larger effect sizes than the overall effect of .15: Behavioral Intervention, Career Development/Job Training, Family Engagement, and Literacy Development. This means that programs that employ these strategies have larger dropout rate effect sizes than those who do not employ the strategy. Four others have effect sizes that are about the same as the overall effect: Academic Support, Health and Wellness, Life Skills Development, and Mentoring. However, the effect size for those studies that did not include these strategies is equivalent or larger than those that did. Four strategies had smaller effect sizes for those that included the strategy: Afterschool, School/Classroom Environment, Service-Learning, and Work-Based Learning. However, effect size by subgroup were only statistically significantly different for four of the strategies: Career Development/Job Training and Family Engagement, where the programs that employed these strategies had statistically significantly higher effect sizes than those who did not, and School and Class Environment and Service-Learning, where the programs that employed these strategies had statistically significantly smaller effect sizes than those who did not.

These findings should be taken into context with the fact that each strategy was analyzed independently of the others. In other words, these results do not taken into account any confounding effects of other variables on the effect size outcomes. Table 2 reports the model statistics by subgroup for each strategy, including effect sizes and between-group statistics for each strategy.

Research Question Four

Finally, we conducted a meta-regression to determine if any of the strategy variables (covariates) were able to statistically significantly predict the overall dropout rate effect size. First, a model was converged that included all 12 predictor variables to determine the variance (R2) accounted for by the set. The application output includes several scenarios to assist in determining the most appropriate model for the given task, including correlation and increment matrices. The correlation matrix identifies correlations among covariates; strong correlations indicate that two or more covariates are highly confounded and should not be analyzed independent of one another. The increments matrix reports results from an automated series of analyses that incrementally introduce each of the covariates into the model, collating the changes in the variance accounted for at each iteration, or increment. This allows the researcher to easily identify any covariates that have a nil or negative impact on R2. An examination of these two matrices can provide the researcher with knowledge regarding any covariates that may be able to be removed to develop the most appropriate model for explaining the variance of an effect.

Examination of the correlation matrix in our analysis revealed no strong correlations (correlations were all < +/- .50) between any of the 12 covariates, indicating that covariates were not highly confounded and that covariates could be removed from the model if desired. Additionally, an examination of the increments matrix indicated that the Afterschool and Life Skills Development covariates negatively influenced R2. These two covariates were removed from the model and the meta-regression was conducted with just 10 covariates: Academic Support, Behavioral Intervention, Career Development/Job Training, Family Engagement, Health and Wellness, Literacy Development, Mentoring, School/Classroom Environment, Service-Learning, and Work-Based Learning.

The results of the test of the model revealed a Q value of 101.59 with df = 10 and p < .001, indicating that at least one of the covariates is likely related to the overall dropout rate effect size. However, the goodness of fit test indicated that the model is still incomplete ( Q = 103.42, df = 50, p < .001), suggesting that there are other (unknown) covariates that could help explain the variance in the dropout rate effect. The total variance in true effects (between-study T2) was .042 and the variance not explained by the model (null model T2) was .01, with an R2 value of .76. This means that though the model explains 76% of the between-study variance in effects, only 3.2% of the overall variance is explained by the model.

Though the model is not complete in terms of explaining the variance within the overall dropout rate effect size, the model did identify eight of the 10 variables as being statistically significant predictors of the dropout rate effect. Further, effect sizes for each of the 10 covariates can be calculated from the coefficient values. These effect sizes differ from the effects reported above in that they are the effect sizes for each strategy holding all other covariates constant. This allows us to calculate an effect size that accounts for any confounding effects of other covariates. The model revealed effect sizes for the 10 covariates ranging from .11 to .81, all of which are statistically significant predictors of the overall dropout effect except Career Development/Job Training and Mentoring. Table 3 reports the complete model statistics by covariate for the dropout rate meta-regression and Table 4 reports the effect sizes derived from the meta-regression for the 10 included covariates in order of effect size.

Discussion and Implications

Improving dropout rates and graduation rates are actually comprehensive K-12 issues, although much of the prevention work occurs at the high school level. To appropriately interpret the effect sizes estimated here, we need to view them in light of other high school reform efforts. Research indicates that mean effects for high school reform efforts are estimated at about .24 (Hill, Bloom, Black, & Lipsey, 2007). However, these are effect sizes that represent school-wide populations, not just the at-risk or very at-risk populations that are included in most of the studies we have included here. In that context, we interpret our overall effect sizes of .15 for dropout rate as promising and .36 for graduation rate as excellent. These effects provide us with an indication that dropout efforts are resulting in positive impacts on students, which is especially important considering the cost of dropouts on society and individually.

This is good news for a couple of reasons. First, many of these programs are implemented using taxpayer dollars; others are funded through private sources. Second, substantial resources other than finances go into these programs, including time and research.

There is obviously still substantial work to be done in the dropout prevention field, but our results indicate that we are making progress in the area. Hopefully, the positive results will allow us to more smoothly build or extend relationships with schools and program developers as we now have empirical evidence to identify those strategies that are impactful on dropout rates. Programs that employ these strategies have larger effects than those who do not, particularly those programs that include Family Engagement, Behavioral Intervention, and Literacy Development strategies. Other strategies, however, such as Work-Based Learning, School/Classroom Environment, Service-Learning, and Health and Wellness are also significant predictors of positive dropout prevention outcomes, as is Academic Support.

Costs involved with implementing prevention programs are often quite large, and can inhibit broad scale implementation of programs. Not surprisingly, some strategies are more costly to implement than others, some are more appropriate to specific populations and settings than others, and some are more easily modified to fit within existing programs and processes than others. Those needing to make decisions about program implementation usually know these types of constraints but until now have not been aware of the varying strengths of strategies.

Ultimately, we foresee our findings as being influential in policy and practice as it will allow us to help schools “personalize” programs that have the best chances of promoting students beyond high school onto meaningful college or career pathways.

Implications for Practitioners

For those who work directly in schools and community-based organizations as well as key stakeholders who support their efforts, the findings here provide compelling empirical data which points to family engagement, behavioral intervention, and literacy development as the most significant evidence-based solutions to address the dropout problem in America. Given that programming for school and community-based projects can be quite costly and many districts, especially those serving students in at-risk situations, are woefully underfunded, these findings enable youth leaders and advocates to narrow the focus of resources. Financial, human, and material resources can be redirected toward strategies that ultimately improve efficiency and effectiveness and therefore should result in better and more meaningful student outcomes.

School boards, school and district administrators, teachers, and community service organizations should reflect on the findings here and review current policies, practices, and programs for alignment. Should there be dissonance between current practices and those aligned with the results, programming modification should be considered. The following are suggestions for specific areas to target:

Implications for school boards and district administrators. While policy development related to the findings of the meta-regression is highly recommended, school boards and district administrators should conduct a self-assessment to determine all programming both central to and tangential to dropout prevention efforts. Where feasible, ineffective policies, practices, and programs should be reformed to conform to the findings emergent in the meta-analysis. Broad scale implementation of effective programming should be pursued versus continuation of multiple programs deemed ineffective.

Because Family Engagement emerged as the most significant factor and families have an indelible impact on both the performance and the behavior of their children, districts should examine current practices related to family engagement. Central to this examination is the common understanding of the term family. As many youth in at-risk situations may see their biological/nuclear family as an impediment to success, districts must recognize that the term family is relative to each student. Caring influential adults who offer real-time guidance to youth in at-risk situations could very well be regarded as extended family. Nonetheless, once family is defined for individual students, districts can deliberately include them in their programming through regular and meaningful two-way communication.

Implications for schools and community-based organizations. Practitioners working directly with students in at-risk situations who are on the verge of leaving school before graduation should understand the findings and the triad nature of the most significant factors: Family Engagement, Behavioral Intervention and Literacy Development. As Family Engagement is understood to be the practices that engage and inform parents and families of the performance and behavior of their child(ren)/family member, strategies that include the family in behavioral intervention and/or literacy development will undoubtedly have a positive compounded effect.

Family Engagement coupled with Behavioral Intervention may offer the most useful and scalable model. Whereas some families may see literacy development as academic (school’s role), behavior is different. Families see behavior as a home issue and will more readily address this (in collaboration with school officials) as they feel more empowered and capable to intervene. Thus, families should be considered full partners in decision making that impacts their children, especially regarding behavior modification.

In short, our findings are particularly beneficial to district, school and community-based programs. From this body of work, practitioners are well-positioned to adopt new relevant activities that reduce the likelihood that students will leave school before graduation.

Implications for Policymakers

Our findings would be incomplete without the inclusion of thoughtful, tangible recommendations for policy decisions related to dropout prevention. This section aligns each of the strategies with a set of school policy options.

Academic Support

(Policies that provide academic support for students at risk of dropping out)

Behavior Intervention

(Policies that identify and address behaviors that predictably lead to dropping out of school)

Career Development and Job Training

(Policies that support career and job training as essential component of school and demonstrate the relevance of education)

Family Engagement

(Policies that engage and inform parents and families of the performance and behavior of their child(ren)/family member)

Health and Wellness

(Policies that support health and wellness in each student to be adequately prepared for school and to fully develop)

Literacy Development

(Policies that support literacy development for students to be successful in school and in life)


(Policies that support school-based mentoring activities and programs that effectively engage students as mentors and mentees building their academic and social-emotional knowledge and skills)

School/Classroom Environment

(Policies that identify, measure and improve the school and classroom environment most conducive to positive student development)


(Policies that support high-quality service-learning that combine classroom instruction with community service)

Work-Based Learning

(Policies that support high-quality work-based learning that engage students in career orientation)


We hope our findings here will provide a mechanism by which decision makers can weigh the effectiveness and feasibility of particular strategies so they choose the strategies that will give them the most bang-for-the-buck for their education environments. As we finalize this study and extend our research efforts in this area, we believe we are on the verge of having evidence that can be directly applied to practical, real-time decisions made by local school officials. But we also should be able to extend our findings to legislators, policymakers, and program developers at state and national levels to help develop evidence-based solutions to the dropout problem.

Ultimately, it seems appropriate to conclude that our findings should help us achieve our goal of increasing efficiency and effectiveness of dropout prevention programs so we can impact real change. Conducting this meta-analysis of dropout prevention strategies has provided further insight into what we already knew. Previous research identified dropout risk factors, which has helped educators establish methods that may help alleviate or eliminate these risk factors. Unfortunately, many prevention programs require substantial resources (time, personnel, financing), but current economic conditions often prohibit the implementation of such programs on a broad scale. However, by identifying those program strategies that have the strongest probability of positively impacting student outcomes, we can assist schools in selecting those strategies that address their particular school’s issues. We plan to extend our research to include additional evaluations and additional predictors to help explain the overall variance in the dropout rate, and to clearly discuss the context in which these strategies are successful.


1.   The terms “study” and “evaluation” are used interchangeably in this report.

2.   Some evaluations contained more than one effect size, so overall n’s are reported in effect sizes rather than number of studies included.

3.   The kappa statistic is often used to calculate interrater reliabilities but it was not appropriate here because (a) only two coders rated each program and (b) programs and coders were not fully crossed. In this case, the Pearson product moment correlation was an appropriate measure of interrater reliability (Multon, 2010).


(References marked with an asterisk indicate studies included in the analyses.)

*Abele Mac Iver, M. (2009). Improving educational opportunities: A randomized evaluation study of a high school dropout prevention program. Paper presented at the American Educational Research Association, San Diego, CA

*Adler, L., Cragin, J., & Searls, P. (1995). A study of the impact of a community based school to work program for high risk youth. West Covina, CA: East San Gabriel Regional Occupational Program.

*Allen, J. & Allen, C.W. (2009). Escaping the endless adolescence. New York: Ballantine Books.

*Barkely, R., Guevremont, D., Anastopoulos, A., & Fletcher, K. (1992). A comparison of three family therapy programs for treating family conflicts in adolescents with attention deficit hyperactivity disorder. Journal of Consulting and Clinical Psychology, 60, 450-462.

Begg, C., & Mazumdar, M. (1994). Operating characteristics of a rank correlation for publication bias. Biometrics, 50, 1088-1101.

Borenstein, M., Hedges, L., Higgins, J., & Rothstein, H. (2009). Introduction to meta-analysis. West Sussex, UK: John Wiley & Sons, Ltd.

Borenstein, M., Hedges, L., Higgins, J., & Rothstein, H. (2013). Comprehensive Meta-Analysis software (Version 3.2). Englewood, NJ: Biostat.

Bridgeland, J., DiIulio Jr., J., & Morison, K. (2006). The silent epidemic: Perspectives of high school dropouts. Civic Enterprises and Peter D. Hart Research Associates for the Bill and Melinda Gates Foundation.

*Byrne, P. J. (2001). The educational effectiveness of the Albany I Have a Dream program (Unpublished doctoral dissertation). University at Albany, New York.

*Clarke, G., Lewinsohn, P., & Hops, H. (1990). Instructor’s Manual for the Adolescent Coping with Depression course. Eugene, OR: Castalia Press.

*Clawson, H., Lamb, Y., Porowski, A., O’Conner, R., Passa, K., Burden, F. et al. (2008). Evaluation of Communities In Schools (CIS) of Texas. Fairfax, VA: ICF International.

*Cobb, B., Sample, P., Alwell, M., & Johns, N. (2005). The effects of cognitive- behavioral interventions on dropout for youth with disabilities. Clemson, SC: National Dropout Prevention Center for Students with Disabilities.

*Coleman, M. (1992). Aggression replacement training with behaviorally disordered adolescents. Behavioral Disorders, 18, 54-66.

*Crean, H., Hightower, A., & Allan, M. (2001). School-based child care for children of teen parents: Evaluation of an urban program designed to keep young mothers in school. Evaluation and Program Planning, 24, 267-275.

*Dynarski, M., & Wood, R. (1997). Helping high-risk youth: Results from the alternative schools demonstration program. Princeton, NJ: Mathematica Policy Research.

*Dynarski, M., Gleason, P., Rangarajan, A., & Wood, R. (1998). Impacts of dropout prevention programs: Final report. Princeton, NJ: Mathematica Policy Research, Inc.

*Eckenrode, J., Campa, M., Luckey, D., Henderson, C. R., Cole, R., Kitzman, H., et al. (2010). Long-term effects of prenatal and infancy nurse home visitation on the life course of youths: 19-year follow-up of a randomized trial. Archives of Pediatrics & Adolescent Medicine, 164 (1), 9-15.

*Edmunds, J.A., Bernstein, L., Unlu, F., Glennie, E. & Smith A. (2013). Graduating on-time: The impact of an innovative high school reform model on high school graduation rates. Paper presented at the Annual Meeting of the American Educational Research Association, San Francisco, CA.

*Etscheidt, S. (1991). Reducing aggressive behavior and improving self-control: A cognitive-behavioral training program for behaviorally disordered adolescents. Behavioral Disorders, 16, 107-115.

*Feiner, R., Brand, S., Adan, A., Mulhall, P., Flowes, N., Sartain, B., & Dubois, D. (1994). Restructuring the ecology of the school as an approach to prevention during school transition: Longitudinal follow-ups and extensions of the School Transitional Environment Program. Prevention in Human Services,10(2), 103-136.

*Gambone, M. A., Klem, A. M., Summers, J. A., Akey, T. A., & Sipe, C. L. (2004). Turning the tide: The achievements of the First Things First education reform in the Kansas City, Kansas public school district [Executive Summary] . Philadelphia, PA: Youth Development Strategies, Inc.

*Gandara, P., Larson, K., Mehan, H., & Rumberger, R. (1998). Capturing Latino students in the academic pipeline. Berkley, CA: Chicano/Latino Policy Project.

*Garvie, J. L. (1992). The effect of a teacher-as-advisor program on factors related to students’ dropout rate (Unpublished doctoral dissertation). University of Miami, Florida.

*Goerge, R., Cusick, G. R., Wasserman, M., & Gladden, R. M. (2007). After-school programs and academic impact: A study of Chicago’s After School Matters (Issue Brief No. 112). Chicago, IL: Chapin Hall Center for Children.

*Grossman, J. & Sipe, C. (1992). Summer Training and Education Programs (STEP): Report on the 1985 summer experience. Philadelphia, PA: Public/Private Ventures.

*Haithcock, F. O. (1996). The effects of a research based high school dropout prevention program on multiple student outcomes (Unpublished doctoral dissertation). University of Miami, Florida.

Hammond, C., Linton, D., Smink, J., & Drew, S. (2007). Dropout risk factors and exemplary programs: A technical report . Clemson, SC: National Dropout Prevention Center and Communities In Schools, Inc. Retrieved July 17, 2013, from www.dropoutprevention.org/resource/major_reports/communities_in_schools.htm

Hedges, L., & Olkin, I. (1985). Statistical methods for meta-analysis. San Diego, CA: Academic Press.

Hill, C., Bloom, H., Black, A., & Lipsey, M. (2007). Empirical benchmarks for interpreting effect sizes in research. Working paper. New York: MDRC.

*Huang, D., Gribbons, B., Kim, K. S., Lee, C., & Baker, E. L. (2000). A decade of results: The impact of the LA’s Best after school enrichment program on subsequent student achievement and performance. Los Angeles, CA: University of California

*Huang, D. Kim, K. S., Marshall, A., & Pérez, P. (2005). Keeping kids in school: An LA’s Best example: A study examining the long-term impact of LA’s Best on students’ dropout rates. Los Angeles, CA: University of California.

*ICF International. (2009). Evaluation of Georgia’s Performance Learning Centers Final Report: Leveraging results/research to practice. Retrieved from http://www.cisga.org/cisgawpress/wp-content/uploads/2011/09/PLC_Evaluation_Summary.pdf

*ICF International. (2010). Communities In Schools national evaluation five year summary report. Retrieved from http://www.communitiesinschools.org/about/publications/publication/five-year-national-evaluation-summary-report

*Kahne, J., Sporte, S., de la Torre, M., & Easton, J. (2006). Small high schools on a larger scale: The first three years of the Chicago High School Redesign Initiative. Chicago, IL: Consortium on Chicago School Research at the University of Chicago.

*Karp, M. M., Calcagno, J. C., Jeong, D. W., & Bailey, T. R. (2007). The postsecondary achievement of participants in dual enrollment: An analysis of student outcomes in two states. Minneapolis, MN: National Research Center for Career and Technical Education.

*Kemple, J. J., & Snipes, J. C. (2000). Career Academies: Impacts on students’ engagement and performance in high school. New York, NY: Manpower Demonstration Research Corporation.

*Langley, K. (1993). A program evaluation of Project Grads: A dropout prevention program. Dissertation Abstracts International, 54(05-A).

*Larson, K. & Gerber, M. (1987). Effects of social metacognitive training for enhancing overt behavior in learning disabled and low achieving delinquents. Exceptional Children, 54, 201-211.

*Larson, K. A., & Rumberger, R. W. (1995). ALAS: Achievement for Latinos through Academic Success. In H. Thornton (Ed.), Staying in school. A technical report of three dropout prevention projects for junior high school students with learning and emotional disabilities (pp.1-60). Minneapolis, MN: University of Minnesota, Institute on Community Integration.

*Lecroy, C. (1988). Anger management or anger expression: Which is most effective? Residential Treatment for Children & Youth, 5, 29-39.

*Lentz, F. E., (2011). Evaluation of 2010-2011 student cohort of the Educational Options for Success (EOS) program. Cincinnati, OH: University of Cincinnati.

*Lentz, F. E., (2012). Evaluation of 2010-2011 student cohort of the Educational Options for Success (EOS) program: Executive summary. Cincinnati, OH: University of Cincinnati.

*Long, D., Gueron, J. M., Wood, R. G., Fisher, R., & Fellerath, V. (1996). LEAP: Three-year impacts of Ohio’s Welfare Initiative to improve school attendance among teenage parents. New York, NY: Manpower Demonstration Research Corporation.

*Mahoney, J. L. & Cairns, R. B. (1997). Do extracurricular activities protect against early school dropout? Developmental Psychology, 33(2), 241-253.

*Mauldon, J., Malvin, J., Stiles, J., Nicosia, N., & Seto, E. (2000). Impact of California’s Cal-Learn demonstration project: Final report. Berkley, CA: University of California.

*Maxfield, M., Schirm, A., & Rodriguez-Planas, N. (2003). The Quantum Opportunity Program demonstration: Implementation and short-term impacts . Washington, DC: Mathematica Policy Research, Inc.

*McGlumphy, W. P. (1994). An analysis of an at-risk dropout prevention program in Seaford high school (Unpublished doctoral dissertation). University of Delaware, Newark.

*McLaughlin, N. L. (2000). Developmentally appropriate practices: Interventions to help at-risk high school students meet math graduation requirements. Baylor University, Texas.

*Millenky, M., Bloom, D., & Dillon, C. (2010). Making the transition: Interim results of the National Guard Youth ChalleNge evaluation. New York: MDRC.

Multon, K. (2010). Interrater reliability. In N. Salkind (Ed.), Encyclopedia of research design. (pp. 627-629). Thousand Oaks, CA: SAGE Publications, Inc.

National Dropout Prevention Center. (2013a). Trends in AFGR, dropout, and funding. Clemson, SC: National Dropout Prevention Center at Clemson University.

National Dropout Prevention Center. (2013b). The 15 effective strategies for dropout prevention. Clemson, SC: National Dropout Prevention Center at Clemson University.

*Nowicki, S., Duke. M. P., Sisney, S., Stricker, B., & Tyler, M. A. (2004). Reducing the drop-out rates of at-risk high school students: The Effective Learning Program (ELP). Genetic, Social, and General Psychology Monographs, 130(3), 225-239.

*Pearson, L.C. & Banerji, M. (1993). Effects of a ninth-grade dropout prevention program on student academic achievement, school attendance, and dropout rate. The Journal of Experimental Education, 61(3), 247-256.

*Phillips, G. (n.d.). Career Choices curriculum positively impacts high school transition to small learning communities. Retrieved from http://www.academicinnovations.com/dataproof/silveradoslc.pdf

*Quint, J., Bos, J., & Polit, D. (1997). New Chance: Final report on a comprehensive program for young mothers in poverty and their children. New York: MDRC.

*Ravenell, M. (1998). The effects of cohort ealignment on achievement, self-concept, and dropout rates of at-risk students. South Carolina State University, South Carolina.

*Robinson, T., Smith, S., & Miller, M. (2002). Effect of a cognitive-behavioral intervention on responses to anger by middle school students with chronic behavior problems. Behavioral Disorders, 27, 256-271.

*Rodriguez, N. M. (2003). Evaluation of a 4-year dropout prevention program: Comparison of high school Hispanic and African American students (Unpublished doctoral dissertation). Temple University, Pennsylvania.

Rosenthal, R. (1979). The file drawer problem and tolerance for null results. Psychological Bulletin, 86 , 638-641.

*Royer, M. L. (2007). Investigating the success of a school-within-a-school model for dropout prevention: Goshen high school’s freshman academy for at-risk students (Unpublished doctoral dissertation). Ball State University, Indiana.

*Sinclair, M. F., Christenson, S. L., Evelo, D. L., & Hurley, C. M. (1998). Dropout prevention for youth with disabilities: Efficacy of a sustained school engagement procedure. Exceptional Children, 65 (1), 7–21.

*Sinclair, M. F., Christenson, S. L., & Thurlow, M. L. (2005). Promoting school completion of urban secondary youth with emotional or behavioral disabilities. Exceptional Children, 71(4), 465-482.

Smink, J., & Schargel, F. P. (Eds.) (2004). Helping students graduate: A strategic approach to dropout prevention . Larchmont, NY: Eye on Education.

*Solomon, R., & Liefeld, C. P. (1998). Effectiveness of a family support center approach to adolescent mothers: Repeat pregnancy and school drop-out rates. Family Relations, 47, 139-144.

*Thorton, H. (1995). Staying in school: A technical report of three dropout prevention projects for middle school students with learning and emotional disabilities. Minneapolis, MN: Institute on Community Integration.

U.S. Department of Education, Institute of Education Sciences, National Center for Education Evaluation and Regional Assistance, What Works Clearinghouse. (2013).


Table A1

Effective Strategies for Dropout Prevention.



Systemic Renewal

A continuing process of evaluating goals and objectives related to school policies, practices, and organizational structures as they impact a diverse group of learners.

School-Community Collaboration

When all groups in a community provide collective support to the school, a strong infrastructure sustains a caring supportive environment where youth can thrive and achieve.

Safe Learning Environments

A comprehensive violence prevention plan, including conflict resolution, must deal with potential violence as well as crisis management. A safe learning environment provides daily experiences, at all grade levels, that enhance positive social attitudes and effective interpersonal skills in all students.

Family Engagement

Research consistently finds that family engagement has a direct, positive effect on children's achievement and is the most accurate predictor of a student's success in school.

Early Childhood Education

Birth-to-five interventions demonstrate that providing a child additional enrichment can enhance brain development. The most effective way to reduce the number of children who will ultimately drop out is to provide the best possible classroom instruction from the beginning of their school experience through the primary grades.

Early Literacy Development

Early interventions to help low-achieving students improve their reading and writing skills establish the necessary foundation for effective learning in all other subjects.


Mentoring is a one-to-one caring, supportive relationship between a mentor and a mentee that is based on trust. Tutoring, also a one-to-one activity, focuses on academics and is an effective practice when addressing specific needs such as reading, writing, or math competencies.


Service-learning connects meaningful community service experiences with academic learning. This teaching/learning method promotes personal and social growth, career development, and civic responsibility and can be a powerful vehicle for effective school reform at all grade levels.

Alternative Schooling

Alternative schooling provides potential dropouts a variety of options that can lead to graduation, with programs paying special attention to the student's individual social needs and academic requirements for a high school diploma.

After-School Opportunities

Many schools provide after-school and summer enhancement programs that eliminate information loss and inspire interest in a variety of areas. Such experiences are especially important for students at risk of school failure because these programs fill the afternoon "gap time" with constructive and engaging activities.

Professional Development

Teachers who work with youth at high risk of academic failure need to feel supported and have an avenue by which they can continue to develop skills, techniques, and learn about innovative strategies.

Active Learning

Active learning embraces teaching and learning strategies that engage and involve students in the learning process. Students find new and creative ways to solve problems, achieve success, and become lifelong learners when educators show them that there are different ways to learn.

Educational Technology

Technology offers some of the best opportunities for delivering instruction to engage students in authentic learning, addressing multiple intelligences, and adapting to students' learning styles.

Individualized Instruction

Each student has unique interests and past learning experiences. An individualized instructional program for each student allows for flexibility in teaching methods and motivational strategies to consider these individual differences.

Career and Technology Education (CTE)

A quality CTE program and a related guidance program are essential for all students. School-to-work programs recognize that youth need specific skills to prepare them to measure up to the larger demands of today's workplace.

Table A2

Service/Strategy Categories.

Service/Strategy Category


Mapping to 15 Effective Strategies (described below)

Academic Support

Help with remediation, support learning, other than tutoring, such as computer labs; academic skills enhancement programs that use instructional methods designed to increase student engagement in the learning process and hence increase their academic performance and bonding to the school (e.g., cooperative learning techniques and “experiential learning” strategies); includes homework assistance and tutoring.

  • Mentoring/Tutoring
  • Active Learning
  • Individualized Instruction
  • Professional Development

Adult Education

Educate adults through a variety of means, such as continuing education courses or online courses; adult secondary education, including GED preparation; English-as-a-Second-Language programs; adult basic education, literacy; work skills or work-based education; lifelong learning/opportunities for adult growth and development.

  • Alternative Schooling


Rewarding, challenging, and age-appropriate activities in a safe, structured, and positive environment after regular school hours. They may reduce delinquency by way of a socializing effect through which youth learn positive virtues such as discipline or simply reduce the opportunity for youth to engage in delinquency.

This category also includes:

Structured Extracurricular Activities

Recreation, sports, creative, performing arts, academic, or social activities usually in afterschool programs; community service opportunities.

  • After-School Opportunities

Behavioral Interventions

Individualized interventions designed to decrease a specific behavior, by shaping and reinforcing a desired alternative replacement behavior, while

tracking changes over time; designed to improve the individual’s overall quality of life (i.e., student development).

This category also includes:

Conflict Resolution/Anger Management

Encourage nonviolent dispute resolution through a wide range of processes; teach decision-making skills to better manage conflict; learn to identify interests, express own views, and seek mutually acceptable solutions to disputes. Common forms of conflict resolution include: negotiation, mediation, arbitration, community conferencing, and peer mediation.

Court Advocacy/Probation/Transition

Individuals who serve as advocates for youth with social services, the juvenile justice, or school system to make sure they receive appropriate services; provision of resources and support during transition and reintegration after being released; probation services, monitoring, and support through intensive supervision programs or school-based probation.

Substance Abuse Prevention

Reduce the use or abuse of illegal drugs, alcohol, or steroids by educating youth about the effects of drugs/alcohol/steroids.

Truancy Prevention

Promotes regular school attendance through one or more strategies including an increase in parental involvement, the participation of law enforcement, the use of mentors, court alternatives, or other related strategies.

  • Safe Learning Environments

Career Development/Job Training

Provision of social, personal, and vocational skills and employment opportunities to help youth achieve economic success, avoid involvement in criminal activity, and subsequently increase social and educational functioning.

  • Career & Technology Education

Credit Recovery

An alternative to repeating a course for students who have failed required courses for graduation. Services may be offered online or face-to-face using traditional or technology-based instruction.

  • Individualized Instruction

Family Engagement

Encompasses a broad range of events from picnics and field trips to activities that involve families in their children’s education.

This category also includes:

Family Strengthening

Educating parents on specific parenting skills, management skills, and communication skills; providing education on various topics such as abuse and sexuality; training on ways to assist child academically.

Family Therapy

Focuses on improving maladaptive patterns of family interaction and communication.

Teen Parent Support

Parenting skills training; financial management; other types of training and/or services to assist teen parents in staying in school and developing family life; includes pre-post natal care; and provision of child care for children of teen parents while they attend programs, schools, etc.

  • Family Engagement

Gang Prevention/Intervention

Prevent youth from joining gangs; intercede with existing gang members during crisis conflict situations.

  • Safe Learning Environment
  • School-Community Collaboration

*Health and Wellness

Health issues are known to affect a student’s risk of dropout and should be addressed to reduce the impact on school experience. These issues may include obesity, mental and physical health as well as the following:

This category also includes:

Mental Health Services.

Substance abuse treatment such as 12-step programs such as Alcoholics Anonymous or Narcotics Anonymous; counseling related to substance use; counseling related to suicide prevention; counseling related to other mental health syndromes or issues.

Pregnancy Prevention

Aims to reduce the incidence of teen pregnancy through education and provision of comprehensive information.

*Life Skills Development

Communication skills; the ability to cope effectively with relationships; problem solving/decision making; critical thinking; assertiveness; peer selection; low-risk choice making; self-improvement; stress reduction; consumer awareness; peer resistance; recognize and appropriately respond to risky or potentially harmful situations; appreciation for diversity; social influences on behavior; overviews of conflict resolution skills and social skills; leadership skills/training; and health education.

Literacy Development

Early interventions to help low-achieving students improve their reading and writing skills establish the necessary foundation for effective learning in all other subjects.

  • Literacy Development


Relationship over a prolonged period of time between two or more people where an older, caring, more experienced individual provides help to the younger person as he or she goes through life.

  • Mentoring/Tutoring

School/Classroom Environment

Reducing or eliminating problem behaviors by changing the overall context in which they occur; interventions to change the decision-making processes or authority structures; redefining norms for behavior and signaling appropriate behavior through the use of rules; reorganizing classes or grades to create smaller units, continuing interaction, or different mixes of students, or to provide greater flexibility in instruction; and the use of rewards and punishments and the reduction of down time.

  • Safe Learning Environments


Community service with integration of service experience into classroom curricula.

  • Service-Learning

Work-based Learning

Consists of a variety of learning experiences designed to narrow the gap between theory and practice. Experiences include apprenticeships, career fairs, field studies, mentoring, guest speakers, job shadowing and student internships. WBL can be a component of Career-Technical Education programming or offered to all students usually at the secondary level.

  • Active Learning

*Not currently listed as an effective strategy by the NDPC but identified as program characteristic in one or more programs included in the study.