PROTOCOL: Prevention,Intervention,and Compensation Programs to Tackle School Dropout: An Evidence and Gap Map

The Problem, Condition or Issue

The problem of students dropping out before completing secondary education is a matter of global concern. Early School Leaving (ESL), Early Leaving from Education and Training (ELET), and school dropout are the terms most used by international and national organizations to define this phenomenon. These terms and the related indicators reflect different perspectives and policy focuses. For instance, in 2003, the European Union adopted the term “Early School Leavers” to refer to individuals “aged 18 to 24 with only lower secondary education or less and not in education or training” (Council Conclusion 2003/C 134/02), only to later define them as “Early Leavers from Education and Training” (ELET), because this concept more clearly includes those who have dropped out of vocational training tracks (European Commission et al. 2019). Similarly, in the United States., the school status dropout rate is defined as “the percentage of 16- to 24-year-olds who are not enrolled in school and have not earned a high school credential” (National Center for Education Statistics [NCES] 2023). The NCES also tracks event dropout rates, status completion rates, and adjusted cohort graduation rates (i.e., students who graduate on time with a regular diploma). Several primary studies evaluate the impact of interventions at the end of secondary school by measuring dropout or graduation rates, as they are typically unable to track the attainment of qualifications years later (e.g. when participants reach the age of 24). Thus, we use more the definition by De Witte et al. (2013, 14) for the scope of our EGM: “school dropout has been defined as leaving education without obtaining a minimal credential (most often a higher secondary education diploma).”

International organizations and national policies consistently focus on establishing new targets and priorities to tackle dropout. The Incheon Declaration for Education 2030 – held by UNESCO in 2015 and signed by 160 countries – proposed Sustainable Development Goals to be achieved by 2030. Among them, Goal 4 aims to “ensure inclusive and equitable quality education and promote lifelong learning opportunities for all” (UNESCO 2016, 29). This target is monitored in each participating country through multiple indicators, including school-level completion rate and out-of-school rate. As another example, the 2021 European Council Resolution set the target of reducing the school dropout rate to below 9% by 2030 (Council Resolution 2021/C 66/01). According to the latest estimate from UNESCO, approximately 244 million children and youths between the ages of 6 and 18 were out of school in 2021, with different distributions according to age group and world region. The highest percentage of dropouts was in upper secondary school, with 30% of youths aged 15–18 out of school. At lower secondary and primary school levels, 14% and 9% of children were out of school, respectively. Europe and North America had the lowest rates for each age group, while Sub-Saharan Africa had the highest rates (Global Education Monitoring Report Team and UNESCO Institute for Statistics 2022). In Europe, the average rate of ELET in 2022 was 9.6%, which is 0.6 percentage points higher than the 2030 target (Eurostat 2023). Despite an overall declining trend in Europe (European Commission 2023), progress has not been equally distributed. The proportion of early school leavers ranged from 2.3% in Croatia to 15.6% in Romania, with high variations within and between EU Member States. These variations highlighted significant disparities among Europe, with 18 countries already meeting the 2030 target and nine countries experiencing an increase in the percentage of early school leavers (Eurostat 2023).

Dropping out of school before achieving a qualification has wide-ranging implications for individuals and society. At an individual level, school dropout is associated with elevated rates of substance abuse, antisocial behaviors, criminal activity, and restricted career prospects (Audit Commission 2010; OECD 2023). It is also correlated with higher rates of poor mental and physical health (European Education and Culture Executive Agency, Eurydice 2014). At a societal level, individuals who drop out of school are less likely to engage in democratic processes and more likely to experience marginalization and unemployment (Dale 2010; OECD 2023), which have enormous economic implications. In the long run, consequences at a national level include increased welfare expenditures in healthcare and social services (Audit Commission 2010; International Labour Organization 2019; OECD 2023).

Initiatives at global, regional, and national levels have been implemented to address school dropout. For instance, “Safe Back to School,” created by Save the Children in 2020, aimed to support the return of over 150 million of the most marginalized children safely back into learning after the COVID-19 pandemic (Save the Children 2023). As another example, the Nordic Council of Ministers, involving Norway, Sweden, Iceland, Finland, and Denmark, commissioned the Nordic Web Resource on Dropout from Upper Secondary Education project. Its goal is to collect effective initiatives to increase the rate of upper high school completion (Tägtström and Olsen 2016). Structured interventions have also been developed with the aim of reducing school dropouts, especially in Anglo-Saxon countries. For example, the National Dropout Prevention Center in the United States offers a database of programs designed to address school dropout. Additionally, the What Works Clearinghouse regularly updates reviews of studies focusing on the effects of educational programs to address dropout, rating the magnitude of their impact and the strength of their evidence.

Evaluating the effectiveness of school dropout programs is challenging, as it demands longitudinal assessments to gauge their long-term impact on dropout rates. If a program is implemented in elementary or lower secondary school, it is unlikely that a study will track students until the end of secondary school, due to the high costs and challenges involved in following students for a long time. Consequently, many studies focus on precursors or predictors of dropout, such as student academic achievement, attitudes, and behaviors. Rumberger (2011) provides a conceptual framework aimed at understanding the path that leads to school dropout and identifies the most important individual, family, and school risk factors. A description of the logic model is provided in Section 3.4. Our Evidence and Gap Map (EGM) addresses these challenges by identifying areas where substantial evidence exists on the impact of school dropout prevention, intervention, and compensation programs on measures of long-term outcomes (i.e., school dropout and completion/graduation), along with intermediate outcomes, which include risk factors or precursors of dropout (e.g., educational performance, behaviors).

The Intervention

This EGM focuses on all interventions aimed at reducing school dropout or supporting school completion/graduation. Based on the 2011 European Council Recommendations, updated in 2022, we will categorize the programs into three broad domains: prevention, intervention, and compensation programs (Council Recommendation 2022/C 469/01 2022).

Prevention programs aim to reduce the risk of school dropout before the problem starts, optimizing education opportunities offered to all students and removing any obstacles to school success throughout grade levels. These measures include the early identification of risk factors, and the use of learner-centered and inclusive curricula, including academic, social, and emotional education. Examples of programs in this category are high-quality instructional strategies, teacher professional development, family engagement programs, and schoolwide approaches. These programs address all students to promote attendance and a culture of the importance of going to school (Balfanz et al. 2007). Intervention programs aim to avoid students dropping out of school by reacting to early warning signs and providing specific support for students at risk. This category of programs may address all grade levels, from early childhood to secondary school. However, it includes support programs targeting only students identified as at risk of underachievement, facing learning challenges, or broader multifaceted difficulties. Examples are mentoring practices, behavioral interventions, and supplemental services. The intensity of these programs varies according to students' needs. Targeted programs, such as those utilizing data tracking, typically involve 15%–20% of the student population. For students who do not respond to targeted interventions, more intensive programs, such as one-to-one mentoring, are implemented and involve approximately 5%–10% of the students (Balfanz et al. 2007). Compensation programs aim to address and re-engage individuals who have already dropped out of school without earning a secondary school qualification. These programs offer them alternative ways to reach a qualification. Examples of this category of programs include Vocational Education and Training and alternative schools, such as General Equivalency Diploma preparation programs, for students who have left regular schools.

Within the three domains, eligible interventions for our EGM are school-based, school-affiliated, and community-based programs that perform actions with the expectation of having a beneficial impact on staying in or returning to school. School-based programs are those that take place under the control of school authorities. School-affiliated programs are those that take place before or after school hours with the support of external professionals. Community-based programs are services that are implemented at a local level, outside of the schools (Wilson et al. 2011).

Why It Is Important to Develop the EGM

Developing this EGM is important for several reasons. First, EGMs are valuable resources for policymakers, program developers, and researchers to inform the design, implementation, and evaluation of effective programs. By using a visual and accessible tool to show the extent of the evidence base, EGMs can support the discoverability of evidence and its use to make informed decisions in practice and policy. Our EGM will provide a comprehensive and up-to-date overview of studies that evaluated the impact of dropout prevention, intervention, and compensation programs on reducing school dropout rates and improving intermediate outcomes (see Section 3.4). Second, no previous EGMs on interventions to tackle school dropout have been published. Several EGMs exploring the impact of interventions related to student well-being (Saran et al. 2020a), violence (Pundir et al. 2020), and inclusion (Saran et al. 2020b) considered school dropout or school enrollment as outcomes. However, the primary focus was on interventions other than programs directly aimed at reducing school dropout. The most recent systematic reviews with meta-analysis on the impact of school dropout interventions were published in the 2010s (Chappell et al. 2015; Freeman and Simonsen 2015; Tanner-Smith and Wilson 2013; Wilson et al. 2011). Wilson et al. (2011) included 167 impact studies that evaluated the effect of general prevention and intervention programs, as well as those targeting teen parents, for increased school completion. Tanner-Smith and Wilson (2013) examined the effect of dropout programs on student absenteeism across 74 randomized and quasi-experimental studies. Chappell et al. (2015) evaluated the effect of 61 dropout prevention programs on dropout rates and graduation rates. Freeman and Simonsen (2015) address research questions similar to the ones of our EGM. The recent meta-analysis by Wang et al. (2024) included 26 studies conducted between 2010 and 2022, using similar eligibility criteria to Wilson et al. (2011). Although this is a relevant up-to-date resource, the overall lack of recent reviews on the impact of school dropout programs shows the need to conduct a comprehensive exploration of the evidence available through an EGM. Thirdly, our EGM will have a broad scope, including dropout compensation programs, which have not been considered in previous syntheses.

EGM: Definition and Purpose

An EGM is defined as a “systematic evidence synthesis product which displays the available evidence relevant to a specific research question” and is “used to identify gaps requiring filling with new evidence, collections of studies for review and increase the discoverability and use of studies by decision-makers, research commissioners and researchers” (White et al. 2020, 1). Based on the EGM results, decision-makers and commissioners may be facilitated in making evidence-informed policies, as well as commissioning relevant research, based on the identified gaps. Program developers may be oriented to develop new approaches that have the best chance of being effective, based on the available evidence. Researchers may direct their efforts to investigate areas that need further development. It is worth noting that, while EGMs offer a visual overview of the available evidence, providing information on the impact of the interventions studied is outside of their scope. Further analyses are needed to code the study results and synthesize the magnitude of the effects.

Framework Development and Scope

Creating the EGM framework is recognized as the initial and crucial step in developing an evidence map (White et al. 2020). We will follow the typical framework of a matrix with interventions/programs in rows and outcomes in columns to develop our EGM. After drafting the framework based on the previous literature, we will review it through consultation with stakeholders. The process is outlined as follows. First, we will draft the framework, with the program and outcome domains and categories based on previous policy and research literature. Programs will be grouped into three broad domains—prevention, intervention, and compensation—based on the framework outlined by European Council Recommendation 2022/C 469/01. Each domain will be further distinguished into program categories, which we will identify through consulting the practice guide on preventing school dropout by the Institute of Education Sciences (Rumberger et al. 2017), the database of programs by the National Dropout Prevention Center (https://dropoutprevention.org/), and the Campbell Collaboration review conducted by Wilson et al. (2011). These three resources include a wide range of programs aimed at tackling school dropout. The categories will be updated according to the programs evaluated in the studies included in our review and their characteristics. Outcomes will be grouped into two broad domains (i.e., long-term outcomes and intermediate outcomes) to distinguish between studies that directly evaluated the effects on school completion or dropout—long-term outcomes—and studies that evaluated the effects on precursors of school dropout—intermediate outcomes (Battin-Pearson et al. 2000; Rumberger 2011). Intermediate outcomes will be further distinguished into categories based on Rumberger's (2011) conceptual framework of individual risk factors, including areas of educational performance, problem behaviors, attendance, attitudes, and relational and social factors. The categories will be updated according to the outcomes considered in the studies. Program and outcome domains and categories are drafted in Annexes 1 and 2.

Second, the draft framework will be revised through a consultative process involving the authors of the map, members of the advisory group, school stakeholders, and local administrators. A workshop with stakeholders and local administrators will be organized to review the framework, with the aim of making the EGM easy to use for practitioners and policymakers. We will also ask for adjustments to the labels used to categorize programs and outcomes and their definitions to ensure that clear and easily understandable language is used.

Stakeholder Engagement

We have engaged research and school experts who agreed to contribute to the project as members of the Advisory Group. The Advisory Group has national and international members with different areas of expertise. We will also engage with teachers, principals, and local administrators in a workshop to advise on the EGM framework.

Therese Deocampo Pigott is a Professor at Georgia State University and an expert in methods for evidence synthesis and meta-analysis.

Daniel Princiotta is an Assistant Research Scientist at Johns Hopkins University and an expert in school dropout prevention and recovery policies, practices, and interventions.

Veronica Mobilio is the Director of the Research Unit at Fondazione per la Scuola (FpS), an Italian training and research organization strongly committed to improving the connection between research and practice by making better use of available evidence.

The members of the Advisory Group will advise and review the EGM protocol and the final paper, supporting the quality of the process, the development of the EGM framework, and the interpretation of the results. Once we have a draft of the map, we will engage with teachers, principals, and local administrators in a workshop to advise on the EGM framework and share suggestions on how to make the EGM usable for stakeholders.

Conceptual Framework

Research has shed light on how school dropout is often the long-term manifestation of an individual and academic discomfort, evident much earlier through dropout risk factors (Battin-Pearson et al. 2000; European Agency for Special Needs and Inclusive Education [EASNIE] 2019). We adopted a broad logic model to portray the general pathway expected for the impact of the programs on precursors of school dropout and, ultimately, on dropout rates. Figure 1 shows our logic model, which is based on the following: Rumberger's (2011) conceptual framework, which identifies the most important individual and institutional risk factors of school dropout; Dupéré et al.'s (2015) review and the EASNIE (2019) eco-systemic approach to preventing school failure, which suggests a set of risks and protective factors at national, school, and individual levels; and Battin-Pearson et al. (2000), who tested five models to predict school dropout and identify precursors, such as academic achievement and attendance.

OPEN IN VIEWER

Figure 1

Logic model.

In our logic model, programs designed to address school dropout are expected to have an impact on intermediate outcomes and long-term outcomes. These programs operate through distinct pathways based on the three domains and their corresponding program categories, with direct effects on the long-term outcomes or effects mediated by intermediate outcomes. The former outcome category is distinguished in the model as school dropout and school completion/graduation. The latter is organized according to the following target areas of student development: (a) educational performance, including academic achievement, educational persistence, and attainment (e.g., grade retention); (b) problem behaviors, including factors that reflect what students do in and outside of school (e.g., school suspension, school expulsion, deviant behaviors, substance use); (c) attendance, defined as a student being present in an accredited organized learning environment at any time during the school year; (d) attitudes, including a wide range of psychological factors, such as self-efficacy, self-esteem, learning motivation, and engagement; (e) relational and social factors, related to student–teacher interactions, peer relationships, and family relationships.

As mentioned above, we expect different pathways for the three program domains. Based on our logic model, prevention programs targeting all students have an impact on the intermediate outcomes, which, in turn, have a long-term effect on the long-term outcomes. This pathway finds support in the nature of this type of program, which aims to reduce the risk of school dropout before the problem manifests. Intervention programs target students identified as at risk of school dropout. The identification of such students typically involves using data from several variables, including individual, family, school, and community factors, along with intermediate outcomes, such as student attendance and educational performance. The dashed arrow connecting intermediate outcomes and other factors to intervention programs illustrates this pathway. Intervention programs have a direct impact on long-term outcomes, as well as a mediated impact through intermediate outcomes. The intensity of intervention programs for the identified at-risk students varies according to their individual needs. Targeted programs, typically delivered in small groups, reach a larger section of student populations, while intensive one-to-one programs involve a smaller section.

In contrast to prevention and intervention domains, compensation programs take a distinct trajectory, as they address individuals who have already dropped out of school. The objective of these programs is to re-engage them by providing ways to earn a qualification. Similarly to intervention programs, we expect a direct impact of compensation programs on school completion/graduation and a mediated impact through intermediate outcomes. In Figure 1, this trajectory is depicted by an arrow connecting the long-term outcome of school dropout to compensation programs, which, in turn, are linked to intermediate outcomes and school completion/graduation.

The characteristics of students who drop out of high school, along with their associated risk factors, have been extensively researched (e.g., Freeman and Simonsen 2015; Rumberger 2011). Although no individual risk factor accurately predicts dropout, the prediction becomes more accurate when multiple risk factors are considered. Thus, the presence of multiple risk factors significantly increases the likelihood that a student will leave school before achieving a secondary qualification (Freeman and Simonsen 2015). Our logic model considers these factors—either protective or risk factors—influencing the pathway to school dropout, including the above-mentioned intermediate outcomes alongside individual and family factors, such as socioeconomic status (SES), gender, parents' level of education, and family structure, as well as school and community factors, such as school structure, school resources, urbanicity level, and employment opportunities (Rumberger 2011). We expect that a limited number of studies will provide information about these factors. Thus, we will code the ones that are commonly reported in studies to describe the population of interest (e.g., SES). We will use the filter function of the interactive EGM to show the results for different populations of students based on these factors.

The model also incorporates the contribution of Dupéré et al. (2015), which integrates the traditional life course framework with the stress process framework, emphasizing the interplay between long-term risks and proximal life events and circumstances in the dropout pathway. School-aged youths may face stressors— external events and circumstances that challenge their adaptive capacities through either disruptive isolated events or prolonged difficulties. Additionally, they may encounter turning points—significant departures from their past development brought on by events powerful enough to reshape their lives. Finally, the historical, geographical, and socioeconomic context frames the entire logical model. Lives are embedded in specific contexts that determine the conditions under which schools operate, and these contexts may influence the immediate environment in which the decision to drop out is made (Dupéré et al. 2015).

Dimensions

The EGM will be presented as an online matrix of intersections between programs (in rows) and outcomes (in columns). Programs and outcomes will be grouped into broad domains and categories. Each study will be allocated to the appropriate cell according to the evidence it presents, resulting in numerous appearances of the studies on the map. Bubbles, representing groups of studies of the same type, will have the following: (a) a different color based on the study design (systematic reviews with meta-analysis, Randomized Controlled Trials [RCTs], Quasi-Experimental Designs [QEDs]); and (b) a different size based on the number of studies. In addition to interventions and outcomes, several variables related to the PICOS framework will be coded and included in the EGM as filters to support the selection of interventions matching certain criteria of interest. We plan to include the population and study design characteristics described in the Filters for presentation section. EPPI Mapper will be used to generate an online interactive EGM.

Other Eligibility Criteria

Search Methods and Sources

We propose a comprehensive search strategy to identify published and unpublished studies that potentially meet the eligibility criteria, following the guidelines provided by MacDonald et al. (2024). The following strategies will be used to search for studies.

Analysis and Presentation

Data Collection and Analysis