Taking Stock of Programs to Develop Socioemotional Skills A Systematic Review of Program Evidence

Possessing a range of cognitive, socioemotional, and technical skills is important for individuals to maximize their chances of success in many aspects of life. In particular, a growing body of research highlights the effects that socioemotional skills have on a variety of outcomes, from wages and academic performance to health. Programs to help participants develop such skills continue to expand in both high-income and lower-income countries, targeting individuals of almost all ages and life stages. Socioemotional skills development is embedded in programs as diverse as early-childhood nutrition programs for adolescent mothers, K-12 academic curricula, and workforce training programs for vulnerable adults. However, the characteristics that make some programs more successful than others, or even what types of outcomes programs use to measure ‘success’, are less clear in the literature. This analysis seeks to fill this knowledge gap through a systemic review of socioemotional skills development programs. To categorize the wide range of socioemotional program objectives, this analysis breaks down programs by where they occur, program characteristics, participant characteristics, and outcomes measured. Only programs with randomized or quasi-randomized evaluations are included. The programs in this review are divided into three categories, on the basis of when they are implemented in the life cycle: (a) before-school programs (infants and young children), (b) school-based programs (preschool to secondary education), and (c) out-of-school programs (usually targeted, vulnerable populations). Next, the review assesses programs by their specific characteristics: objectives, components and complementary elements, setting, and quality.

relevance Possessing a range of cognitive, socioemotional, and technical skills is important for individuals to maximize their chances of success in many aspects of life. In particular, a growing body of research highlights the effects that socioemotional skills have on a variety of outcomes, from wages and academic performance to health. Programs to help participants develop such skills continue to expand in both high-income and lower-income countries, targeting individuals of almost all ages and life stages. Socioemotional skills development is embedded in programs as diverse as early-childhood nutrition programs for adolescent mothers, K-12 academic curricula, and workforce training programs for vulnerable adults. However, the characteristics that make some programs more successful than others-or even what types of outcomes programs use to measure "success"-are less clear in the literature. This analysis seeks to fill this knowledge gap through a systemic review of socioemotional skills development programs. It uses a new conceptual framework to examine diverse programs that have been rigorously evaluated to answer the following research questions: • What is the existing landscape of socioemotional skills development programs? • How do these programs measure success? • What do we know about programs that work?

conceptual Framework
To categorize the wide range of socioemotional program objectives, this analysis breaks down programs by where they occur, program characteristics, participant characteristics, and outcomes measured. Only programs with randomized or quasi-randomized evaluations are included. The programs in this review are divided into three categories, on the basis of when they are implemented in the life cycle: (a) before-school programs (infants and young children), (b) schoolbased programs (preschool to secondary education), and (c) out-of-school programs (usually targeted, vulnerable populations). Next, the review assesses programs by their specific characteristics: objectives, components and complementary elements, setting, and quality. It also classifies programs by participant profiles, including ages of participants and participant-targeting systems. Importantly, the review further assesses programs by their outcomes. Although all programs emphasize socioemotional skills development, they present a wide variety of outcome indicators used to measure success. Reported outcomes are classified into four broad categories: (a) health related, (b) risk factor related, (c) academic/cognitive related, and (d) economic related.

methodology
Within each of the three life-cycle stages, we identified potential studies and programs from relevant databases, online resources, and experts; screened them in order to include only studies that were written in English after the year 2000, that included randomized or quasi-randomized control trials, that had sufficient sample sizes, and that reported information on standard errors; and determined which papers and follow-ups on the same program were most relevant and were analyzed using the most rigorous approaches.
For programs that targeted children before they entered school, the review includes 21 programs, 42 percent of which were conducted in low-or middleincome countries. They included follow-up periods ranging from three months to 37 years. For school-based programs, the review includes 45 programs, 10 percent of which came from low-and middle-income countries, with follow-up from nine weeks to six years. For out-of-school programs, the review includes 20 programs, 75 percent of which took place in low-and middle-income countries; follow-up periods ranged from less than two weeks to four years (most less than one year).

Key Findings
Before-school programs typically target children younger than age five and tend to focus not only on children but also on their families. Most before-school programs measure outcomes related to academics (such as academic performance and graduation rates) or risk factors (such as criminal activity); smaller percentages measured health and economic outcomes.
School-based programs target a broad range of grades (prekindergarten through 12th grade), often through classroom curricula that usually follow a yearly sequence, that have grade-specific content, and that tend to include all children attending an institution. Of the 45 sample programs, few included curricula for an entire system of prekindergarten through 12th grade, and most focused on lower grades, rather than explicitly targeting adolescents. Most of the school-based programs (59 percent) measured outcomes related to risk factors (such as behavioral variables like aggression toward peers and cooperation); 32 percent measured academic-related outcomes. Few evaluated economic and health outcomes.
Out-of-school programs typically promote skills formation in teenage and adult populations who are not enrolled in, and may not have completed, formal education. Out-of-school programs with socioemotional skills components have a range of objectives, such as helping individuals find jobs or decreasing gender-based violence. Socioemotional skills development is usually only one part of a larger set of outcomes. The programs surveyed targeted ages ranging from 10 to 55 years, but most focused on working-age teenagers and adults, especially those between ages 16 and 30. Labor market outcomes were the goal of about 75 percent of the programs included in this review.
Most successful programs teach socioemotional skills as an embedded component of a broader curriculum that includes active pedagogical, classroom, and training programs: • Among before-school programs in particular, programs with multiple components that target health, cognitive development, and emotional development tend to yield greater and longer-term results. In addition, the type of curriculum followed affects the program's effectiveness; most desirable are curricula where children plan, carry out, and review their own activities while engaging in active learning, or where teachers respond to children's self-initiated play in loosely structured, socially supportive settings. Programs that have more qualified staff and that are implemented with greater intensity and fidelity exhibit greater effects. • Successful school-based programs tend to follow the SAFE approach, that is, they are sequenced, active, focused, and explicit. Programs that take a whole-school approach and prioritize implementation fidelity tend to be more successful. Programs that were integrated into the school day tended to demonstrate more success than after-school programs. • Such findings stress the value of developing socioemotional skills in coordination with other types of skills, as part of a comprehensive, intertwined curriculum-rather than offering separate, stand-alone socioemotional skills "modules." Programs are particularly effective when they target vulnerable populations and, in particular, young children: • Before-school programs appear to have a greater impact than those that take place later in life. However, these early-childhood programs also tend to be more intensive, more targeted toward vulnerable populations, more likely to involve family members, and have longer follow-up periods. • Most school-based programs tend to be universal and offered to all children; whereas evidence on the impact of these programs is more mixed, the greatest effects appear to be on risk factors (such as externalizing or internalizing behavior or aggression toward peers) among vulnerable populations. • A number of out-of-school programs observed small but statistically significant impacts on economic outcomes, such as employment, participation in the formal job sector, number of hours worked weekly, and earnings. However, overall evidence on the programs' effect on employment levels and quality is mixed. Most programs appear to work better for younger participants, for females, and when implemented in urban areas.
School-based and out-of-school programs need longer follow-up periods and more clarity on proxies used to evaluate socioemotional skills as outcomes: • Whereas over one-third of before-school programs included a follow-up period of 10 years or more, only 3 percent of school-based programs and no out-of-school programs did so. Thus, any potential effects of these programs can be observed only over a short time, and the longevity of any positive outcomes is uncertain. • Although programs explicitly target socioemotional skills development, few explicitly state exactly what socioemotional skills they intend to improve. In addition, most impact evaluations do not measure the skills that the program seeks to alter. Future research should include skills assessment in order to measure these outcomes more directly.
indicate that although cognitive skills are set by age 8, noncognitive skills can be modified until the late teenage years; others contend that personality traits are set by age 30 (Costa and McCrae 1990, 2006James 1890;McCrae and Costa 2003). Yet other researchers, including Walsh (2005), indicate that, given the slow development of the prefrontal cortex, socioemotional skills remain malleable even after age 30. Second, investments in socioemotional skills can foster equity and promote social mobility, economic growth, social cohesion, and general well-being. Research indicates that given the positive associations found between socioemotional skills and labor market outcomes, improving these skills could lead to greater equity (Hartas 2011). Since skill formation promotes employability and employment, skills can pave the way for expanding economic growth. Ultimately, skill formation can promote social cohesion, that is, the capacity of societies to manage collective decision making (World Bank 2012). Finally, as the following chapters will show, socioemotional skills increase the well-being of the population, as individuals increase their optimism levels, improve their selfconcepts, and form healthier relationships, among other outcomes.
Despite the increasing importance of the role of socioemotional skills, little systematized and rigorous (experimental) evidence exists documenting the effectiveness of programs that seek to develop these skills throughout the life cycle. The lack of information constrains the policy dialogue with client countries seeking evidence and programs to foster socioemotional skills.
This systematic review 5 aims to fill an important knowledge gap by distilling existing evidence and offering a menu of program approaches to develop key socioemotional skills to influence important life outcomes. This review contributes to the existing literature on socioemotional skills by identifying and organizing programs using a new conceptual framework. The review includes a number of diverse programs that have been rigorously evaluated and that seek to effect different outcomes in multiple contexts, including in developing countries. The inclusion of developing countries is important, as most of the accumulated evidence comes from programs implemented in developed countries where constraints and context may be different from those in developing countries and, thus, could potentially mediate outcomes differently.
In an effort to generate robust policy advice from proven programs, the review casts a narrow net by considering evidence only from programs that had randomized or quasi-randomized evaluations. All of the evaluations analyzed in this review include the information needed to calculate the statistical significance of the effect sizes. The program descriptions (such as components, target populations, objectives, intensity, and costs) were collected from published or publicly available evaluation documents, as an analysis of administrative data was beyond the scope of this review.
The review focuses on programs that are preventive in nature. They include programs that target universal and primarily young populations, as well as programs that aim to prevent negative behaviors largely among atrisk youth. Other important programs, such as those that fight recidivism Heineck and Anger (2010) find that personality traits, and especially locus of control, have an important effect on wages. 2 Employers acknowledge the importance of socioemotional skills in the workplace. For example, using the STEP (Skills Toward Employment and Productivity) Survey in Vietnam, Bodewig and Badiani-Magnusson (2014) find that the most important skills that employers seek in blue-collar workers are job-specific technical skills, followed by behavioral and cognitive skills, such as teamwork and problem solving. With regard to white-collar workers, employers seek individuals with critical thinking, problem-solving, and communication skills. Guerra, Modecki, and Cunningham (2014) identify the following eight skill areas that employers demand (summarized as PRACTICE): (a) problem solving, (b) resilience, (c) achievement motivation, (d) control, (e) teamwork, (f) initiative, (g) confidence, and (h) ethics. According to Job Outlook 2015 (NACE 2015), the five main attributes that employers look for on candidates' résumés are (a) leadership, (b) ability to work in a team, (c) written communication skills, (d) problem-solving skills, and (e) work ethic, all of which can fall under the realm of socioemotional skills. These are followed by analytical and quantitative skills and technical skills, which can be considered cognitive.

evidence of programs that modify socioemotional skills
Given the accumulating evidence on the role of socioemotional skills in shaping key life outcomes, there is growing interest in gathering information on the programs that can modify these skills. For instance, Almlund and others (2011) analyze 15 programs in developed countries. Their results show that some programs can modify personality and behavior and, as such, they may be considered promising mechanisms to address poverty and disadvantage. 3 Guerra, Modecki, and Cunningham (2014) distill findings from 53 successful programs; a key finding from their analysis is that the socioemotional skills valued by employers and commonly targeted by programs that aim to improve labor market outcomes can be taught successfully when aligned with the optimal stage for skills development.
Cost-effectiveness and effects tend to be greatest in programs targeting young children. In a systematic review of 84 research reports, mostly from the United States,  find that socioemotional skills training for individuals under age 18 can help prevent antisocial behavior, especially among at-risk students. 4 In a review of 27 programs, mostly from developed countries and covering primarily programs for young children, Kautz and others (2014) find that although programs aimed at adolescents can work, programs at early stages in life appear to be the most successful. 5 This finding is supported by research from Cunha, Heckman, and Schennach (2010), who find that the optimal stage to invest in noncognitive skills development is during early childhood. 6 To date, most of the evidence available is for comprehensive programs that aim to develop socioemotional skills along with other types of skills and behaviors, making it sometimes difficult to disentangle results. Since most systematic reviews and meta-analyses are not limited to stand-alone programs on socioemotional skills, it is important to interpret results carefully. Such caution is especially important for programs that have not been evaluated using rigorous methodologies.

Impacts of Early Childhood Development Programs
Evaluations of early childhood programs to develop socioemotional skills, which are more readily available than programs later in life, show positive effects. For example, Baker-Henningham and López Bóo (2010) aim to identify the effectiveness of early childhood stimulation in developing countries and report that, regardless of nutritional supplementation, early stimulation benefits cognitive development. Their study is not limited to experimental or quasi-experimental evaluations and does not yield conclusive evidence on the effects of early childhood programs in developing countries on schooling, behavior, or maternal outcomes. Finally, after analyzing experimental evidence from 31 home-visiting programs for pregnant women and parents of young children, Olds and Kitzman (1993) find that not all programs work; the programs more likely to succeed are those that are comprehensive in focus, involve frequent visits and well-trained professionals, and serve at-risk families.

Impacts of School-Based Programs
School-based programs can also prove effective. Heckman, Stixrud, and Urzua (2006) and Kautz and others (2014) find that schooling generates cognitive and noncognitive skills even after controlling for reverse causality (the fact that stronger skills may lead to more schooling). Durlak and others (2011) analyze the results from 213 broad school-based learning programs to develop social and emotional skills, mostly in developed countries, and find that such programs improve social and emotional skills, attitudes, behavior, and academic performance. 7 Lösel and Beelmann (2003) report findings from a meta-analysis of 84 randomized evaluations of skills-training programs to prevent antisocial behaviors in children of all educational levels. Their analysis shows greater effects in studies with smaller samples. They also find that the majority of evaluations reviewed have higher positive effects on social and cognitive outcomes than on antisocial behavior. 8 Finally, for a study published by the Collaborative for Academic, Social, and Emotional Learning, Payton and others (2008) summarize the results from 317 studies of children in kindergarten through eighth grade and find that programs can be effective whether they are implemented during or after school, whether they target students with or without behavioral problems, whether they are implemented in different grades or locations (urban, rural, or suburban), and whether they include diverse racial and ethnic characteristics. 9 These findings are consistent with a more recent study from Domitrovich and others (2013), which focuses on 11 middle school and high school programs in the United States where preliminary findings show promising effects.
After-school programs, primarily from the United States, can also show promising results. Durlak (1997) studied 69 after-school programs implemented from kindergarten through high school in the United States with the objective of fostering socioemotional learning. The review shows that some programs significantly affected behaviors and academic achievement. Those effects were particularly strong for programs that followed the "SAFE approach," that is, they were appropriately sequenced, active, focused, and explicit. 10 A report from the Afterschool Alliance (2014) analyzes more than 70 evaluations of after-school programs in the United States and concludes that those programs generate important gains for children in improving academic performance, safety, discipline, school attendance, and avoidance of risky behavior. 11

Impacts of Training Programs
Out-of-school initiatives with socioemotional skills components are usually training programs that seek to improve job-related outcomes. Based on evidence from developed countries, Heckman, Lalonde, and Smith (1999) and Kluve and others (2007) find inconclusive evidence of the effects of training programs on labor market-related outcomes. In contrast, Martin and Grubb (2001) state that formal classroom training and on-the-job training appear to help female reentrants. Dar and Tzannatos (1999) examine nearly 100 evaluations of active labor market programs and find similar results: training for the long-term unemployed can help when the economy is improving, and youth training generally has no positive effect on employment or earnings.
Programs that foster labor market-related outcomes in the developing world are rarely analyzed, but existing evidence appears promising. Urzúa and Puentes (2010) broaden this literature to include studies of programs around the world. In their analysis of 215 job-training programs, they find that effective programs tend to be intensive and integrated, involve the public sector, and start early. However, only 30 of these programs had experimental evaluations (23 of them in the United States), and those showed positive or neutral results. Similarly, Betcherman and others (2007) gather evidence from 289 studies from 84 countries (42 percent of them in member countries of the Organisation for Economic Co-operation and Development) and observe that evidence in developing countries is lacking and that the level of program evaluation is weak (close to 40 percent of all of the programs included had no evaluation information on outcomes or effects). With the existing information, they report that most programs appear to have positive labor market impacts, but just over half of the programs that had cost-benefit analyses were cost-effective (only 25 had such analyses, and 14 were found to be cost-effective). Further, the impact on youth employment appears to be more favorable in developing and transition countries than in developed countries. Card, Kluve, and Weber (2015) analyze 200 studies, finding that effects are significant two to three years after program completion. They also find greater effects for programs that focus on human capital accumulation and for those that are implemented during a recession.
Both developed and developing countries have many programs for out-ofschool youth focusing on improving employability and employment prospects. However, only the new wave of employability programs tends to include socioemotional skills as part of their comprehensive packages. For example, Honorati and McArdle (2013) identify successful programs in developing countries as those that are demand driven, have a sound governance structure, and take a comprehensive approach (combining different types of skills training with support services). Ibarrarán and Rosas-Shady (2009) find that in contrast to evidence from developed countries, the results from seven comprehensive programs in Latin America and the Caribbean range from modest to meaningful. González-Velosa, Ripani, and Rosas-Shady (2012) analyze programs in Latin America and point out that, because of low levels of investment, only modest (but sustained) effects could be expected. notes 1. "The Big Five taxonomy of personality traits is now widely accepted as the organizational structure of personality traits … consists of conscientiousness, openness to experience, neuroticism, agreeableness, and extraversion" (Pierre and others 2014, 28). Refer to page 17 of their article for further details.
2. The authors measure locus of control, reciprocity, and the Five Factor Personality Inventory.
3. Personality is measured using Rotter's locus of control scale, Rosenberg's self-esteem scale, personal behavior (absences and truancies, lying and cheating, stealing, and swearing or use of obscene words), externalizing behavior, the Big Five scale, effort, initiative, disruptive behavior, and so forth.
4. Social skills were measured through reported antisocial behavior; social skills such as social interaction skills and prosocial behavior; or social cognitive skills like self-control and social problem-solving skills.
5. The authors measure noncognitive skills, including externalizing and internalizing behavior, creativity, verbal intelligence, hostility, beliefs and attitudes, impulse control, and self-esteem.
6. The explanation behind this finding is that higher initial levels of skills increase the productivity of future investments in skill generation (investment in socioemotional skills early in life builds the base for subsequent investment).
7. For a list of the outcomes included in this research, please refer to footnote 7 in their article.
8. For more information, refer to footnote 13 in their article.
9. Payton and others (2008) group social and emotional competencies in self-awareness, self-management, social awareness, relationship skills, and responsible decision making.
10. The author measures social and emotional skills, including "identifying emotions from social cues, goal setting, perspective taking, interpersonal problem solving, conflict resolution, and decision making" (Durlak 1997, 410).
11. These variables were measured through school attendance, class participation, homework completion, attitudes toward school, self-concept, and decision making, among others.

socioemotional skills
The concept of socioemotional skills is typically found in the psychology literature, particularly in publications by psychologists and psychiatrists whose work is related to education or the learning process. Authors such as Payton and others (2008) and Durlak and others (2011) define socioemotional competencies as skills that enable individuals to (a) recognize and manage their emotions, (b) cope successfully with conflict, (c) navigate interpersonal problem solving, (d) understand and show empathy for others, (e) establish and maintain positive relationships, (f) make ethical and safe choices, (g) contribute constructively to their community, and (h) set and achieve positive goals. Accordingly, they define socioemotional learning as the process of acquiring and applying this set of competencies effectively.
Pierre and others state that socioemotional skills "relate to traits covering multiple domains (such as social, emotional, personality, behavioral, and attitudinal)"   2014,8). The NoVo Foundation, a private foundation that supports initiatives for girls and women, defines social and emotional learning (SEL) as "the process of developing fundamental skills for life success ... SEL's skills include recognizing and managing emotions, developing caring and concern for others, establishing positive relationships and making responsible decisions." 2 One advantage of the term skill in the previous definitions is that it denotes malleability, which opens the possibility for programs that can generate change. However, not everyone agrees. Duckworth and Yeager contend that "referring to them as skills may implicitly exclude beliefs (for example, growth mindset), values (for example, prosocial motivation), and other rational attitudes (for example, trust)" (2015).

soft skills
The Oxford Dictionaries website defines soft skills as all "personal attributes that enable someone to interact effectively and harmoniously with other people." 3 Note that the term soft skills is typically found in business and management literature. In fact, more than 80 percent of the papers found in the Education Resources Information Center-or ERIC-database that include the specific term soft skills present findings from analyses of issues in postsecondary education. For example, those papers often analyze programs that target managers and executives or programs that promote leadership skills in business.
Bancino and Zevalkink define them as "the cluster of personality traits, social graces, facility with language, personal habits, friendliness, and optimism that mark people to varying degrees" (2007,22). Kaplan, Sorensen, and Klebanov (2007) and Kaplan, Klebanov, and Sorensen (2012) mention similar characteristics as those embraced by Bancino and Zevalkink, including creativity, listening skills, and team skills. Similarly, Heckman and Kautz use the concept of soft skills to describe "personality traits, goals, motivations, and preferences that are valued in the labor market, in school, and in many other domains" (2012,451).
Although the previous definitions are not entirely overlapping, the concept of soft skills is similar to that of socioemotional skills.

noncognitive skills
Another commonly used term is noncognitive skills, an expression used primarily by economists. They typically use noncognitive skills to refer to the bundle of skills that are not covered in standard definitions of cognitive skills (numeracy, literacy, and so forth). According to Messick, "once the term cognitive is appropriated to refer to intellective abilities and subject-matter achievement in conventional school areas ... the term non-cognitive comes to the fore by default to describe everything else" (1978,2).
However, leading scholars outside of economics find the term unfitting. As pointed out by Duckworth and Yeager (2015), using the term noncognitive has a number of shortcomings. First, because the term is broad, it is "difficult to define with precision, often misinterpreted because of lack of consensual definitions, hard to measure ... and representative of heterogeneous rather than homogenous categories" (Duckworth and Yeager 2015). Heckman and Rubinstein suggest that the term is unhelpful and elusive, given that "no single factor has yet emerged to date in the literature on non-cognitive skills, and it is unlikely that one will ever be found, given the diversity of traits subsumed under this category" (2001,145).
Another often cited limitation of using the term noncognitive skills is that it indirectly implies that all that is contained under the noncognitive skills category (for example, personality, self-control, grit, perseverance, dependability, persistence, self-esteem, optimism, and time preference) is unrelated to cognition. This implication is inaccurate in psychology, as facets of psychological functioning involve processing information and thus are inherently cognitive (Duckworth and Yeager 2015). 4

character skills
In the psychology literature, Colman defines character skills as "the aggregate or combination of psychological traits that distinguish a person from others" (2009,125). In the economics literature, Heckman and Kautz refer to the term character skills "to describe the personal attributes not thought to be measured by IQ tests or achievement tests" (2013, 10). They suggest character skills include "perseverance ("grit"), self-control, trust, attentiveness, self-esteem and self-efficacy, resilience to adversity, openness to experience, empathy, humility, tolerance of diverse opinions, and the ability to engage productively in society" (2013, 6). Detractors of this terminology point out that character and virtue are admirable qualities with a positive connotation and that psychological traits or attributes do not necessarily need to be positive or negative.

personality traits and temperament
A number of studies use personality and temperament to define socioemotional skills. Psychologists have a long-established tradition of studying the determinants, characteristics, and instruments to measure with a level of precision an individual's personality traits and temperament. According to the American Psychological Association, "personality refers to individual differences in characteristic patterns of thinking, feeling, and behaving." 5 Sanson, Prior, and Kyrios (1990, 179) state that temperament "refers to early-appearing, intrinsic individual differences in the style of behavior, as distinct from its content or motivation." Psychologists have made significant advances in personality measurement, and today there is a wide array of taxonomies to measure personality. Some of the most commonly used include the Big Five, the Big Three, the MPQ, and the Big Nine. 6 This diversity of instruments should not be interpreted as a lack of agreement on the measurement of personality and temperament. Rather, the diversity should be seen as an ability to measure finer grades embedded in the measurement of personality and temperament (Rothbart, Ahadi, and Evans 2000).
Of these measurement instruments, the Big Five is a well-accepted and widely used taxonomy. It measures the following five facets of personality: (a) openness to experience, (b) conscientiousness, (c) extraversion, (d) agreeableness, and (e) neuroticism. The five are usually referred to as OCEAN.
Whereas there is generally agreement on the measurement of personality, there is a lack of consensus on the extent to which personality can be used to denote the presence or absence of socioemotional skills. A key disagreement in the literature relates to the practice of equating "personality traits" to competencies, skills, attitudes, beliefs, or values, given that personality and temperament tend to remain stable over time. However, research has found that personality traits and temperament are mutable and can be altered by experiences (Almlund and others 2011).

twenty-First century skills or competencies
The term 21st century skills describes "a comprehensive set of skills that, along with content mastery, are what all sectors can agree are essential for success." 7 The 21st century skills encompass a wide range of competencies, including learning and innovation skills (creativity and innovation, critical thinking and problem solving, and communication and collaboration); information, media, and technology skills (information and media literacy and information and communication technology); and life and career skills (flexibility and adaptability, initiative and self-direction, social and cross-cultural skills, productivity and accountability, and leadership and responsibility). Whereas the 21st century skills are not labeled as socioemotional skills, it could be argued that some of the competencies can fall in the realm of socioemotional skills. Although Duckworth and Yeager (2015) embrace the skills, they contend that referring to the skills as 21st century skills might be misleading. The name implies that such skills have only come to light recently and are a requirement for the needs of this century when, in fact, some of them (for example, self-control and gratitude) have been around and deemed important well before the turn of the 21st century.

life skills
A term that has been used for some time to capture socioemotional abilities is life skills. The United Nations Educational, Scientific, and Cultural Organization (UNESCO), the United Nations Children's Fund (UNICEF), and the World Health Organization (WHO) have used the term extensively, yet it is interesting that these organizations acknowledge that there is no universally accepted definition of life skills (Singh 2003;WHO 1999). According to UNESCO's working definition, life skills encompass "a mix of knowledge, behaviour, attitudes, and values and designate the possession of some skill and know-how to do something or reach an aim. They include competencies such as critical thinking, creativity, ability to organise, social and communication skills, adaptability, problem solving, and ability to co-operate on a democratic basis to shaping a peaceful future" (Singh 2003, 4;italics in original). UNICEF and WHO define life skills as "psychosocial abilities for adaptive and positive behavior that enable individuals to deal effectively with the demands and challenges of everyday life. They are loosely grouped in three broad categories of skills: cognitive skills for analyzing and using information, personal skills for developing personal agency and managing oneself, and interpersonal skills for communicating effectively with others." 8 These skills are linked to health, peace education, human rights, citizenship education, and other social issues.
This review will subsume the various terminologies described in table 3.1 under the term socioemotional skills. This examination is done with the understanding that these categories include certain levels of overlap and important differences. In fact, 21st century skills include many job-specific skills such as problem solving and critical thinking that are not contained under the socioemotional terminology. c h a p t e r 4

Conceptual Framework
A number of programs foster socioemotional skills to achieve diverse objectives, such as child development or labor market-related outcomes. Such programs offer a range of different services and target diverse population groups. For example, some programs offer monthly home visits to provide early stimulation activities for children who live in marginalized households and communities. Other programs focus on providing classroom instruction to develop specific skills for youths or young adults. Those differences, along with where and how a given program is implemented, must be taken into account when determining the program's overall success. Drawing lessons from diverse program offerings and target groups is complex and calls for a way to organize a program's characteristics, results, and outcomes. The first and most important task in categorizing programs is to identify whether they occur inside or outside the formal education system. Programs that take place in a school tend to have a different set of characteristics than those that happen outside of school. For example, school-based programs typically target universal populations. Out-of-school programs focus on specific groups, with a vulnerability criterion attached to participation.
Another important distinction is the age of participants. Programs built around the school system usually target students ranging from ages 4 to 18-those in preschool through grade 12. Out-of-school programs are more likely to target (a) infants and younger children for early childhood development programs or (b) adolescents and young adults ages 11 to 25.
The programs included in this review are organized into three categories, according to when they are implemented in the life cycle: (a) before school (infants and young children), (b) school based (preschool to secondary education), and (c) out of school (usually targeted, vulnerable populations).
As a general rule, programs for infants and young children tend to have longer follow-ups on participants that generate more and richer information on outcomes and effects. School-based programs target enrolled students with a separate curriculum that focuses solely on socioemotional skills. Programs outside the school usually focus on objectives (including labor market-related outcomes) that are above and beyond the development of socioemotional skills. Figure 4.1 presents the conceptual framework that will be used throughout this book. This stylized image organizes the programs in this review according to age of beneficiaries and relationship with the formal education system.
To launch the review of programs, we identified what matters most in program features and participant characteristics for achieving program objectives and outcomes (figure 4.2). Every program is immersed in an environment that inevitably mediates its outcomes. An analysis of the broad social, cultural, economic, and political context in which each program takes place is beyond the scope of this review.

program characteristics
The review analyzes evidence from programs with randomized or quasirandomized evaluations that included the information needed to calculate the statistical significance of the effect sizes.
We will carefully assess the main features of each program, including (a) objectives, (b) components and complementary elements, (c) program setting, and (d) program quality levels. Although all programs selected for this review aim to foster socioemotional skills, they have a wide range of objectives, including (a) promoting young children's learning potential, (b) improving school readiness, (c) reducing children's developmental risks associated with poverty, (d) reducing high school dropout rates, and (e) improving the economic prospects (labor participation, earnings, socioeconomic status) of jobless youths. This wide range of objectives translates into different program designs and, ultimately, outcomes and effects. Few of the programs reviewed focus exclusively on the development of  The second factor that we will analyze is the program components and complementary elements. Although programs have a commonly shared goal (fostering socioemotional skills), no two programs are alike. Some programs are narrowly constructed with only one component, whereas other programs are broad based and include multiple components. In some cases, components of the broad-based programs have elements that are intertwined. Examples of component activities include family or community training, classroom instruction, and experiential learning. Complementary elements can include, for example, health access stemming from participating in early childhood services to mentoring and networking opportunities for eligible beneficiaries of training.
We also study the setting in which program services are provided. The setting is important given that socioemotional skills are usually moderated by diverse environmental characteristics. Bearing this fact in mind, we analyze the different effects of programs that occur in diverse environments. For example, programs that are implemented in institutions (nurseries, schools, or the workplace) occur in controlled, structured, and organized environments compared with programs implemented at home or in multipurpose community or training centers.
Another important feature that we analyze relates to the quality of programs. The analysis will show that (a) program content (for example, curricula),  (b) delivery method (for example, pedagogy, duration, intensity, and class or group size), and (c) characteristics of individuals delivering the program (for example, their qualifications, professional training, and experience) are all critical elements of any program and determine the achievement of desired outcomes.

participants' characteristics
An essential distinction among programs is related to the characteristics of eligible participants (for example, age, socioeconomic background, educational attainment, vulnerability, personality, and previous work experience). Our analysis will focus on two main aspects: age group and targeting mechanisms. As stated previously, most of the available research on the process of skills formation is structured around specific ages. The consensus is that the skills-formation process is dynamic, and compelling evidence indicates that the life cycle contains critical or sensitive periods for optimal learning (Cunha and Heckman 2007). This evidence lends support to using age group as an area of interest, since it has important effects on the dynamics of a program and the program's objectives. Both the literature and our review of programs indicate that program participation early in life usually involves small groups or even one-on -one interactions with parents or caregivers and their children. Such programs tend to have a strong health component and aim to foster both cognitive and socioemotional development in ways that differ greatly from school-based programs that focus on children between the ages of 4 and 18. Similarly, our review of programs shows that those aimed at adults most often seek to foster socioemotional skills as a way to shape participants' immediate labor market outcomes.
Our analysis finds that even though programs target specific age groups, most programs use intense screening processes to select their beneficiaries, targeting those with low socioeconomic status, high risk, vulnerability, or unemployment. A program's targeting mechanism is important insofar as it moderates outcomes and objectives. For example, early childhood programs targeting infants with few environmental resources attempt to diminish inequity in cognitive and socioemotional development; programs that target the unemployed usually develop traits that ease labor market-related outcomes. 1 outcomes Although programs (or their components) share the common goal of fostering socioeconomic skills, each is unique in the outcome indicators that are selected to determine success. Also, the monitoring and evaluation strategies attached to each program and the data collection mechanisms vary considerably. For example, some programs rely on self-reported beneficiary data, others use information collected by blind observers or specialists, and many use administrative data.

Search Methodology
In this chapter, we describe the search strategy applied to identify relevant literature and programs seeking to foster socioemotional skills. Given our program classification, a methodologically comprehensive review had to be conducted in three stages according to life cycle. In the first stage, we identified literature and program evaluations of early childhood development programs, that is, those undertaken with children before they enter the formal and compulsory education system. In the second stage, we focused on school-based programs (from preschool to secondary education). And in the third stage, we reviewed literature on programs that take place outside the formal and compulsory school environment. Segmenting our search strategy was deemed necessary as key definitions, relevant programs, and associated literature vary according to the age of the population and program settings. Within each stage, the search contained three phases: • Identification: During this phase, we identified potential studies by examining specialized and multidisciplinary databases, analyzing the references section of pertinent papers, searching the websites of universities and international development institutions, and consulting with experts, among others. Papers and programs selected in this phase aimed to affect socioemotional skills. • Screening: The papers selected in the first phase were further analyzed to determine if the complete program file and component details were readily accessible, if they were written in English after the year 2000, if they kept a strict methodological approach (randomized or quasi-randomized control trials), if they had sufficient sample sizes, and if they reported information on standard errors. • Inclusion: The screening phase usually resulted in the identification of more than one paper evaluating the effects of a particular program. When that was the case, we kept the relevant follow-ups and papers that had a wider variety of outcomes analyzed with a rigorous approach. For this segment of the literature, we reviewed four different sources: (a) papers identified using the search engine of the American Economic Association under the sole keyword early childhood, (b) papers and policy documents referenced by international development agencies, (c) programs identified by University of Chicago professor James Heckman, 1 and (d) papers referenced in revised metaanalyses or in other documents analyzed. During each of these phases, special effort was made to include credible research that took place in developing countries.
That effort resulted in the identification of 118 potential studies that aimed to affect socioemotional skills. Of those studies, 111 were accessible, written in English, and methodically strict. After eliminating multiple followups for the same program, 82 studies were included in the early childhood stage.

search phase 2: school-Based programs
For the school-based portion of the literature, a set of relevant keywords and search terms were used to scan a short list of online newspapers and websites to identify relevant programs, researchers, and institutions working on socioemotional skills at a school level. 2 Three sources were useful (Gelles 2015; Kahn 2013; Kimball and Smith 2013). All were written for a general audience, referenced academic institutions, and cited researchers working on the topic. From those articles, we identified nine relevant institutions, six researchers who were repeatedly mentioned, and four programs directly related to our topic of interest.
Using this information, we searched for relevant academic literature in peer-reviewed journals and electronic databases. In particular, we extended our initial list of keywords with the names of the programs previously identified and conducted systematic searches on electronic databases and policy discussion websites. The following electronic databases and institutions were reviewed: We analyzed 247 reports and found 98 academic articles evaluating the effects of different programs aimed at fostering socioemotional skills. We kept only those that were written after the year 2000 and targeted universal population groups.
Excluded from our search strategy were programs that targeted specific conditions or groups of individuals. We acknowledge the limitation of excluding programs that target specific conditions or groups of individuals, but the available alternatives were not time efficient and cost-effective. However, we decided to include some insights from targeted programs that were mentioned in other compilations or meta-analyses.
In cases where more than one paper was written about a specific program, we decided to keep only those that were relevant follow-ups and those that used a wider variety of outcomes.
We ended up managing a sample of 48 different studies and 45 different programs. Most of those 45 programs were conducted in a school environment and targeted populations between the ages of 3 and 18.

search phase 3: out-of-school programs
Our search strategy for out-of-school programs was broader than in the first two phases. The reason was that no unified pattern existed to organize the literature by a small set of words, concepts, or ideas, 3 and therefore there was no unique place to compile all programs.
Our approach was as follows: • We identified papers and policy documents that most international development agencies highlighted in their reports on youth employability. We were particularly careful to include evidence regarding developing countries. • We reviewed the programs examined by Heckman and Kautz (2013). • We created a database of effective out-of-school violence-prevention programs as suggested by the Office of Juvenile Justice and Delinquency Prevention 4 and the National Registry of Evidence-Based Programs and Practices of the U.S. Substance Abuse and Mental Health Services Administration. 5 Of the 130 documents initially identified, 42 records were screened. All others were neither up-to-date nor high-quality research (randomized control trial or quasi-experiment) written in English. Once we eliminated the multiple evaluations of the same program, the resulting list of out-of-school, labor market-related programs included 20 different programs and eight compilations. Figure 5.1 illustrates the flow of the literature review process that we conducted to identify studies and literature of socioemotional skills. As a result, the analysis includes more programs than previous systematic reviews. Further, it includes as much evidence from developing countries as possible. Finally, it does not concentrate on one particular socioemotional skill or one particular life stage.

Before-school programs
The early stages of life are critical for an individual's healthy development. However, the unequal opportunities faced by the most vulnerable at-risk infants lead to an achievement gap that emerges even before preschool (Melmed 2008;Rubio-Codina and others 2014;Shady and Paxson 2007). Early childhood programs that improve nutrition, cognitive skills, and socioemotional skills have gained the attention of scholars and policy makers, as these programs have the potential to prevent this achievement gap from forming early in children's lives and, therefore, to promote equality. Additionally, such programs have a high rate of return and yield greater results for vulnerable children than do equal investments made during a later stage in life (Heckman 2008;Heckman, Stixrud, and Urzua 2006;Heckman and others 2009;Rolnick and Grunewald 2007). Because initial levels of skills affect skill acquisition (a phenomenon described as "skills beget skills"), investments early in life provide the foundation to increase the efficiency of further investments (Cunha and others 2005). Seeking to capture the existing evidence regarding the effectiveness of early childhood programs on socioemotional skills, we analyze the effects of 21 before-school programs, 10 of which take place outside the United States. 1 All of the programs that were included were rigorously evaluated through randomized control trials or quasi-experimental techniques, with follow-up periods that range from two months to 37 years. Almost half of the analyzed programs measure results more than five years after program implementation.
Programs that begin before the formal schooling begins typically target children younger than 5 years old, and some begin even before birth (see tables 6.1, 6.2, and 6.3). These programs focus not only on young children or newborns but also on their families. Outside the United States, all these programs target at-risk families (in the United States, seven programs have universal coverage).  Total 7 8 7 7 7 7 10 9 9 10 10 10 11 11 10 9 6 3 Note: Shading indicates that the program includes that particular age group. ECD = early childhood development; HIPPY = Home Instruction Program for Preschool Youngsters; PIDI = Proyecto Integral de Desarrollo Infantil. Overall, before-school programs can be categorized into two broad groups: (a) home visiting programs and (b) center-based programs.

Home Visiting Programs
Home visiting programs occur in the home, where children's most relevant interactions take place. Most of these programs aim to foster child development by improving parenting practices, increasing sensitive responses, and raising the level of stimulation (language and cognitive opportunities) at home. Basic health services and the provision of micronutrients complement some programs.
Program services are delivered by a wide variety of individuals, ranging from nurses to trained parent educators. They also vary in their intensity, with some lasting three months and others up to two years. The associated cost also varies considerably: some programs cost US$500 per child per year, whereas others cost US$13,600 per participant over a three-year period.
Home visiting programs have the advantage of allowing for ample participation, since providers can reach difficult-to-access locations and can adapt the services to families' schedules. In addition, when entering a home the provider gains a comprehensive view of the child and his or her family, which facilitates tailoring the service to meet the child's specific needs. (For examples of home visiting programs, see boxes 6.1 and 6.2.)

Center-Based Programs
Center-based programs typically provide a combination of education and stimulation services with nutrition at a child care center. As these programs take place before formal schooling, they can be strongly oriented toward promoting school readiness or more heavily oriented toward providing care and guaranteeing child well-being, depending on the curriculum. They sometimes include health checkups and a parental education component. Although they promote early childhood care and stimulation, these programs also facilitate the ability of the child's main caregiver to get a job. (For examples of home visiting programs, see boxes 6.3 and 6.4.) Box 6.1 home visiting program: the Jamaican study The Jamaican Study was a home visiting program that targeted growth-stunted children ages nine months to 24 months living in poor neighborhoods in Jamaica. Participants were randomly assigned into one of four groups: (a) psychosocial stimulation, consisting of two years of weekly home visits with trained community health aides (paraprofessionals) with a curriculum that promoted language development (for example, mother-child conversations, labeling things and actions), parenting skills, educational games, the use of praise, and improvement in the self-esteem of both the child and the mother; (b) nutritional supplementation, consisting of one kilogram of milk formula provided weekly for two years; (c) both psychosocial stimulation and nutritional supplementation; and (d) a control group.
Short-term outcomes suggested that all three groups improved early cognitive development, but in general, psychosocial stimulation outperformed nutritional supplementation. In the long term, only the psychosocial stimulation group improved cognitive and character skills, increasing the average earnings of participants at age 22 by 42 percent relative to the control group. Findings suggest psychosocial stimulation can have substantial effects on labor market outcomes and might reduce inequality in later life. The characteristics of center-based programs vary immensely. For example, the service providers of Proyecto Integrado de Desarrollo Infantil in Bolivia are mothers in the community, whereas the service providers of the Parenting Practices program in Canada are pediatric nurses with master's degrees. Additionally, some of the programs involve four to seven counseling sessions for parents, whereas others involve 10 hours of daily care. These differences are reflected in the costs, which range from US$500 for 18 months (Integrated Early Childhood Development Intervention in Colombia) to US$67,000 for five years of the program (Abecedarian).
Outcomes in the literature usually include child abuse and neglect, child health and safety, home environment, parental responsibility and sensitivity, parental harshness, depression, parental stress, child cognition, and child internalizing and externalizing behavior problems. Programs usually exhibit short-term positive results, and outcomes vary from modest to multiple, positive, and persistent among treated children (see boxes 6.3 and 6.4).

Box 6.2 home visiting program: nurse-Family partnership, United states
The Nurse-Family Partnership is a home visiting program that targets first-time, unmarried, low-income mothers (usually adolescents) with the objective of improving prenatal conditions and early parenting. Trained nurses do the following: (a) encourage mothers to improve their health-related behaviors, (b) improve children's health and development by helping parents provide more competent care, and (c) improve families' economic self-sufficiency by helping parents make appropriate choices regarding family planning, finishing their education, and finding work.
The program has been replicated in several locations and is one of the few programs with long-term follow-ups. Treated families receive between 0 and 16 home visits during pregnancy and between 0 and 59 home visits from birth through the child's second birthday. a During the first month, prenatal visits are weekly, after which they taper off to biweekly visits until the child is born. After birth, weekly visits resume for the first six weeks, and then biweekly visits continue until the child is approximately 20 months old. The final four visits, leading up to the child's second birthday, occur monthly.
Treated children improved their vocabulary skills, their grade point averages, and their results in achievement tests relative to the control group. Furthermore, they had a lower rate of behavioral problems in the borderline or clinical range 9 years later, and they had a lower probability of arrest and conviction 15 years later. Girls in the program had fewer children and used less Medicaid than the comparison group. Effects were stronger for girls.
Sources: Eckenrode and others 2010; Kitzman and others 2010;Olds 2006;Olds, Henderson, Cole, and others 1998;Olds, Henderson, Tatelbaum, and others 1986;Olds and others 1986;Olds and others 1997;Olds and others 2007;Olds and others 2010;Olds and others 2014;Zielinski, Eckenrode, and Olds 2009; Nurse-Family Partnership website, http://www.nursefamilypartnership.org. a. The evaluations of the program measure intention to treat. Therefore, if families who were assigned to the treatment group were not visited, their outcomes are still captured in the treatment group.

Box 6.3 center-Based program: save the children's early childhood Development programme, mozambique
Save the Children's Early Childhood Development Programme, carried out in rural areas of Gaza Province in Mozambique, seeks to improve children's cognitive, social, emotional, and physical development. It has two main components: (a) high-quality early stimulation, psychosocial support, and literacy and numeracy instruction in community-based preschool centers that take care of children for 3 hours and 15 minutes every day; and (b) monthly parenting meetings oriented toward strengthening positive parenting practices.
Communities are ultimately responsible for managing and sustaining the centers, which are staffed with volunteer teachers who are mentored, trained, and supervised. The program costs about US$2.50 per month per child.
Among other effects, an evaluation of the program found significant increases in primary school enrollment, improvements in cognitive and problem-solving abilities, and progress in motor skills and socioemotional and behavioral outcomes. However, language, stunting, and wasting did not show significant improvements.

Box 6.4 center-Based program: highscope perry preschool program, United states
The HighScope Perry Preschool Program had two components: a 2.5-hour preschool curriculum on weekdays during the school year and weekly home visits by teachers. The latter component was designed to involve parents with the classroom curriculum and to improve the household environment. The program targeted low-income African American children who were at high risk of school failure, and it is well-known for the outcomes it was found to have during long-term follow-up (participants were followed until they were age 40).
Impact evaluations found that participants faced less grade repetition than nonparticipants; scored better on intellectual and language tests from ages 3 to 7; scored better on school achievement tests at ages 9, 10, and 14; and scored better on literacy tests at ages 19 and 27. Furthermore, they achieved higher levels of schooling and greater high school graduation rates. At ages 15 and 19, participants had significantly better attitudes toward school than did nonparticipants.
With regard to economic performance, at age 40 participants had higher employment rates and greater median annual earnings. They were also significantly more likely to own a house or a car and reported having more in savings accounts than did nonparticipants. Importantly, participants had significantly fewer lifetime arrests and arrests for violent crimes than did nonparticipants, as well as fewer arrests for property and drug crimes and fewer subsequent prison or jail sentences.

outcomes of Before-school programs
The discussion that follows presents the main findings of programs that occur before formal schooling and that have among their components improvement in participants' socioemotional skills. The task is particularly challenging because, as Heckman and Kautz (2013) suggest, there are more differences among evaluations than there are evaluations, making it difficult to understand exactly why some programs are more successful than others.
In particular, (a) programs target different populations; (b) specific program components are not usually tested or evaluated; (c) even within a particular population, samples are heterogeneous in many aspects; (d) identification problems might be present in some evaluations; (e) small samples decrease the statistical power and make it hard to detect significant impacts even when they exist; (f) not all programs measure the same outcomes; and (g) most programs have only short-term follow-ups. In other words, because of external validity considerations, caution was necessary when giving general conclusions.
We will face this challenge using two basic inputs: our own compilation of program characteristics, outcomes, and results (see appendixes); and (b) a group of selected reports and compilations available in the literature. Findings and conclusions will be presented using the same framework that we used in the previous section, that is, organizing programs according to their relation with the formal education system and the ages of the participants involved.
This section starts by presenting the primary outcomes documented in the programs analyzed and the available meta-analysis. It will then provide an overview of the program characteristics that mediate the achievement of those particular results, and, finally, it will describe consistent findings regarding the effect of the participant's profile over the results of the programs.
To get an approximate idea of the kinds of results that were captured by program evaluations at each stage of the life cycle, we first grouped the statistically significant and unique outcomes of each program by whether they relate to health, risk factors, academics, or economic aspects. For example, for Proyecto Integral de Desarrollo Infantil in Bolivia, we counted one result for health (increments in weight and height for children), two results related to academics (motor skills and language skills), one result for risk factors (statistically significant effect in psychosocial skills), and another result related to economics (future earnings). Then, we added the number of outcomes in each category across programs. Finally, we calculated the percentages by dividing the outcomes measured in each category by the total number of results (measured in the sum of all of the categories). 2 Most of the outcomes documented by our sample of studies of before-school programs are related to academics or risk factors. As mentioned before, programs that target individuals before formal schooling tend to have longer follow-up periods than do other types of programs. In particular, (a) 38 percent of the effects were related to academic activities, such as schooling, literacy, math skills, intelligence quotient (IQ), school performance, and graduation rates; (b) 31 percent of the effects were related to risk factors, such as criminal activity, arrests, and externalizing and internalizing behavior; (c) 18 percent were related to health variables, such as drug use, fertility, and physical health; and (d) 13 percent were related to economic variables, such as welfare use, employment status, earnings, and marital status (see figure 6.1).
In the short term, before-school programs foster cognitive skills. In the long term, they improve behaviors and reduce welfare use, which indicates that they affect socioemotional skills. A careful analysis of the studies shows that child care and preschool education programs implemented in different contexts and with different methodologies appear to have consistent short-term effects on cognitive and academic performance. Long-term studies find that the cognitive gains diminish over time, but the programs yield increased rates of high school graduation and lower rates of teen pregnancy, arrests, and substance use and abuse. As suggested by Heckman and Kautz (2013), these findings are indicative of the benefits that early prevention can bring by promoting socioemotional skills during early childhood. 3 For example, for the Project Student/Teacher Achievement Ratio (STAR) program in the United States, the effects of class quality on test scores fade by eighth grade, but gains in noncognitive measures persist. A relationship exists between cognitive and noncognitive skills, and, for example, being more persistent leads to individuals' studying for longer periods, which in turn yields higher test scores. Both the HighScope Perry Preschool Program and the Abecedarian program (also in the United States) find lasting effects on adult outcomes as well, including criminal activity and graduation, despite the fact that effects on test scores fade.  Finally, combined data from Abecedarian and Project CARE (character actualization requires education) in the United States show that the effect of early childhood programs on the IQs of eight-year-old children is mediated by improvements in infants' responsiveness to people and objects in their surroundings. Reported rates of return for early childhood programs are extremely high. For example, Heckman and others (2010) and Heckman and Kautz (2013) estimate that the rate of return of the HighScope Perry Preschool Program is about 7 percent to 10 percent a year. Behrman, Cheng, and Todd (2004) estimate that Proyecto Integrado de Desarrollo Infantil in Bolivia has a ratio of earning benefits to cost of 1.7 to 3.7. Additionally, the Abecedarian program was found to have benefits of US$2.50 for every US$1.00 invested. Finally, evaluations of the Nurse-Family Partnership find that the cost of the program was recovered from low-socioeconomic families before the children reached age four (Eckenrode and others 2010;Olds, Henderson, Cole, and others 1998;Olds, Henderson, Tatelbaum, and others 1986;Olds and others 1986;Olds and others 1997;Zielinski, Eckenrode, and Olds 2009. However, some early childhood programs do not work as envisioned and even have negative results. For example, Loeb and others (2007) find that although exposure to a center-based program before kindergarten had positive results in prereading and math skills for children, it had similar-sized but negative effects on a teacher-reported behavioral measure that captured self-control and a variety of interpersonal skills. Similarly, Magnuson, Ruhm, and Waldfogel (2007) find that prekindergarten intervention increases reading and mathematics skills at school entry but also increases behavioral problems and reduces self-control. Finally, Bouguen and others (2014) find that a preschool program in Cambodia had negative or insignificant effects on some indicators of child development.
The extensive differences in results might derive from the heterogeneity of the program characteristics (the dissimilar design elements and implementation realities behind each intervention). As illustrated in figure 6.2, six elements that the literature finds alter the effects of before-school programs include (a) a program's components, (b) its curriculum, (c) its intensity, (d) the qualifications of its personnel, (e) its teacher-to-child ratio, and (f) the fidelity of its implementation (including context-specific responses to the intervention). The effects of each of these six elements are reviewed next.
First, programs with multiple components that target health, cognitive development, and emotional development tend to produce greater and longer-lasting effects. These conclusions may suffer from confounding factors, since multiplecomponent programs tend to be more intense. Additionally, the Jamaican Study and the Integrated Early Childhood Development Intervention show that nutritional programs tend to be less effective than stimulation programs. Furthermore, as mentioned by Kautz and others (2014), programs that involve parents are usually more effective. That finding can be explained by the fact that if the parent changes his or her practices, the home environment is transformed. After the program ends, the child still benefits from a more supportive or stimulating environment, thereby increasing the effect and the persistence of the programs.
Second, as expected, the type of curriculum a program follows affects the program's effectiveness. The HighScope Preschool Curriculum Comparison Study (Schweinhart and Weikart 1997) in the United States measured the effect of implementing three different curricula: (a) the HighScope model, where children planned, carried out, and reviewed their own activities while engaging in active learning; (b) a traditional nursery school model in which teachers responded to children's self-initiated play in a loosely structured, socially supportive setting; and (c) a direct instruction model that involved teachers following a script to directly teach children academic skills. This study found that people born in poverty experienced fewer emotional problems and felony arrests if they attended a preschool program that used either the HighScope model or a traditional nursery school model than if they attended a preschool with the direct instruction model. No significant differences were found between the effects of the HighScope and the nursery school models when participants were 23 years old. When compared with the direct instruction model, beneficiaries of the HighScope and the nursery school models required less treatment for emotional impairment or disturbance, had engaged in fewer acts of misconduct, had experienced fewer arrests for felonies, had committed fewer property crimes, were more tolerant, were more involved in volunteer work, and were more likely to plan to graduate from college. The three models had similar effects on IQ, school achievement, and high school graduation rates.  find that for home visiting programs, using a clearly defined curriculum with a formal structure is key to inducing and promoting optimal program results. Programs where the content of the visits varied dramatically from family to family had inconsistent outcomes.
Third, intense programs, which have greater coverage with regard to hours per day, days per week, or even years of coverage, result in greater effects. For example, in the Nurse-Family Partnership trial in New York, the group that received home visits from public health nurses during both pregnancy and infancy experienced greater benefits in most child-related outcomes than the groups that received the intervention in either pregnancy or infancy but not both (Eckenrode and others 2010;Olds, Henderson, Cole, and others 1998;Olds and others 1986;Olds and others 1997;Zielinski, Eckenrode, and Olds 2009. Fourth, programs with more qualified personnel tend to be more effective. In the Nurse-Family Partnership in the United States, having a paraprofessional in charge of the visits instead of a nurse led to the program losing its effect for the general population (effects were found only for high-risk children) (Kitzman and others 2010; Olds and others 2007; Olds and others 2010). In the STAR program in the United States, students with more experienced teachers in kindergarten had higher earnings (Chetty and others 2010). Howard and Brooks-Gunn (2009) state that home visitors' credentials matter, but their importance depends on the objectives or goals. In general, professional home visitors such as nurses are preferred, but for those programs that seek primarily to improve mother-infant attachments (as opposed to promoting healthy behaviors), social workers or trained paraprofessionals can also be beneficial.
Fifth, programs with greater teacher-to-child ratios are more effective. In the STAR program in the United States, students in small classes (13 to 17 students) performed better on standardized tests, such as the SAT, and were more likely to attend college than were those in larger classes (22 to 25 students) (Chetty and others 2010; Krueger and Whitmore 2001). These results were stronger when the small class size occurred during the child's earliest school years and diminished when small class size occurred after first grade. Being in a small class did not affect student self-concept and motivation (Word and others 1990). According to , the HighScope Perry Preschool Program in the United States showed this positive effect of small class size with up to two adults teaching in a class of up to 20 students.
Sixth, critically, having a well-designed program is insufficient: interventions that do not follow the designed standards and practices have fewer or even negative effects. The fidelity of implementation (understood as faithfulness to a program's standards and practices) can be low because of administrative or operational limitations. Countries should ensure that they have the capacity to implement the programs that they design successfully. In Cambodia, Bouguen and others (2014) attribute the negative effects of the program they evaluated to implementation constraints, low take-up rates, and context-specific behavioral responses to the intervention. 4 In many early childhood programs, implementation fidelity is low. Araujo, López-Bóo, and Puyana (2013) find that the quality standards stipulated by early childhood programs in Latin America do not translate into the daily practices of the operators of these services. In particular, they identify that (a) center-based programs usually involve more children and fewer adults than those reported by program directors, forcing a child-to-caregiver ratio that might hinder beneficial outcomes; (b) although programs are de jure completely funded by local authorities, parents are usually "encouraged" to pay a fee for the services they receive from the centers; and (c) great heterogeneity exists in the daily schedule of activities. Further, they suggest that ensuring positive effects requires ensuring fidelity or compliance with implementation standards.
Another important finding is that the effects of before-school programs tend to yield superior benefits among the most vulnerable populations (see figure 6.3). This finding strongly supports the use of early childhood programs as mechanisms to decrease countries' inequality levels.
Even when effects are widespread, they tend to be greater for children born in high-risk environments (with younger, single, or uneducated parents; in families with few socioeconomic resources; or with low birth weight). For High-risk environment Age, marital status, and educational level Context Outcomes example, for the Nurse-Family Partnership in Tennessee (United States), significant effects were found for children at risk in the areas of grade point average at age 12, achievement tests, arithmetic achievement, and aggression, among others (Kitzman and others 2010;Olds and others 2007;Olds and others 2010). The Parents as Teachers (PAT) program in the United States also achieved greater benefits for very low-income parents and their children, as well as for Latino groups who were less educated (Wagner and Clayton 1999;Wagner, Spiker, and Linn 2002). For the Parent-Child Home Program (United States), infants that started with the lowest IQs also obtained the most benefits (Levenstein and others 1998).
Further, effects tend to be greater for low-income, first-time adolescent mothers and for those who are psychologically vulnerable.  find that programs targeting adolescent mothers tended to differ in effectiveness from programs that enrolled mothers with other backgrounds. The greatest effects were usually found among low-income, first-time adolescent mothers. Another important group that benefited most from this type of intervention was psychologically vulnerable mothers (that is, mothers who were rated high in depressive symptoms and low in self-mastery).
The greater effect of programs on vulnerable beneficiaries could be attributable to early childhood programs leveling the playing field. That is, it could be that these programs level out the preexisting differences in stimulation and even in nutrition among children and parents of varying socioeconomic strata. Before-school programs offer a stimulating environment for infants and toddlers of all backgrounds. In their absence, children from low-resource households would face a harsher environment, so the programs provide a setting that is more productive than their homes. In contrast, infants with plentiful resources would have been highly stimulated at home even if the programs did not exist, so they benefit less from them.

school-Based programs
Programs that foster socioemotional skills during the school year are both abundant and strategic. On the one hand, Heckman and others (2010) and Heckman and Kautz (2012) have found that important periods of brain malleability occur during the school years. Additionally, since vital personality traits-such as conscientiousness 5 -increase monotonically from childhood into late adulthood, school environments allow for continuous skills development (Borghans and others 2008). For instance, school-based programs analyzed in this book, such as Montessori and Family Check-Up, target a broad range of school grades (kindergarten through 12) as potential beneficiaries with classroom curricula that follow a yearly sequence, have grade-specific content, and adjust program length and intensity to children's school progression. Finally, these programs multiply the impacts of investments made in disadvantaged children during early childhood. 6 Operational or practical advantages result from working with groups of beneficiaries who are already present in an organized school atmosphere. According to Rutter (1987), the structured and supportive environment developed in classroom settings eases the implementation of socioemotional programs. Furthermore, since elementary school participation is widespread in most countries, school-based socioemotional programs have higher enrollment and retention rates. Besides, since schools are already established and working, programs that are built on top of school curricula can be implemented and scaled up more quickly. Finally, as  state, school-based programs may have relatively low costs as they can involve group training delivered by regular teachers.
Even though myriad programs aim to develop socioemotional skills in schools, in this book we analyze 45 programs that are widely mentioned in the literature and that have been rigorously evaluated with follow-up periods that range from nine weeks to 23 years. 7 Ten of these programs are located outside the United States.
Unlike most early childhood development programs, school-based programs tend to be universal at a school or classroom level, benefiting all children and adolescents attending the institution. When targeting exists, it occurs mostly at a geographical level. For example, the Linking the Interests of Families and Teachers program covers all first-and fifth-grade elementary school children and their families living in neighborhoods characterized by high rates of juvenile delinquency.
These programs use different nomenclatures to refer to their objective population: some define specific age groups, others particular school grades or taxonomies, and others explicit life periods. To facilitate the analysis, we refer mostly to the grade taxonomy (from prekindergarten to 12th grade, for children ages 3 to 18). When analyzing the school grades covered by the programs, two basic conclusions emerge. First, curricula for the entire system from prekindergarten to 12th grade are scant. Second, the availability of socioemotional curricula is greater for early childhood and childhood age cohorts. In our sample of 45 programs, only 2 covered the whole schooling range. Additionally, whereas 7 programs exclusively target prekindergarten and kindergarten students, only 2 programs focus explicitly on adolescents. 8 This outcome can be easily observed in table 6.4, which illustrates the distribution of the programs according to the school grades offered. It is organized so that programs that include higher grades are located at the top of the list. Horizontally, it shows the grade range covered by each program in our sample. Vertically, it shows the supply of curricula by grade. Table 6.5 presents the school grades that were included in the impact evaluations. The pattern is similar to that shown in table 6.4 in the sense that most of the empirical evidence is available for early childhood and childhood ages, precisely where the supply of programs is greater. Information is limited regarding the impact on adolescent populations. Analysis of the impact evaluations shows that knowledge of these programs is limited by the fact that the majority of the evaluations focus on a subgroup of the academic grades covered by each intervention (see boxes 6.5 and 6.6). Since the evaluations cover a smaller grade range than the programs themselves, they leave important gaps in the longitudinal understanding of these types of programs. That is, questions remain unanswered regarding cumulative impacts, variations of intensity along different years, and the effects of a particular sequence of curriculum content over the process of skills development, among other matters.
The short follow-up periods in the evaluations do little to increase the understanding of the real impact of these programs. Among the 45 programs that were studied, only 8 analyzed the situation of the beneficiaries five or more years later, and only 2 of those (Family Check-Up and Positive Action) targeted children older than age five. Furthermore, close to half of them followed participants for a year or less after the beginning of implementation. 9 Consequently, questions remain regarding whether or how the results will persist over time.

Box 6.6 school-Based program: i can problem solve, United states
Boyle and Hassett-Walker (2008) present the results of a two-year evaluation of the I Can Problem Solve (ICPS) program. The goal of ICPS is to teach children to think about solutions, anticipate consequences, and solve problems. It comprises a specific curriculum of activities, a two-day teacher training component, and instructional sheets for parents. The intervention was implemented in kindergarten and first-grade classrooms in a racially and ethnically diverse urban school district (city and state not disclosed). The district's schools were matched according to (a) the number of students and kindergarten classrooms, (b) racial and ethnic student body composition, (c) socioeconomic status, and (d) language proficiency of students. Treatment was randomly assigned at the school level (226 students) to one of three groups: a two-year ICPS group (n = 96), a one-year ICPS group (n = 106), and a control group (n = 24). (None of the control group participants were African American.) Teachers were required to assess baseline and postevaluation student behaviors using two different behavior rating scales. The findings suggest that the results were positive and monotonically increasing with years of implementation. That is, the authors claimed the effectiveness of ICPS in increasing prosocial behaviors and in reducing aggressive behaviors. (Those outcomes are not directly related to the program's objective.) Box 6.5 school-Based program: al's pals  evaluated the effectiveness of the Al's Pals early childhood program's outcomes associated with socioemotional learning. The program combines a twoday teacher training component, in-class curriculum, and parent awareness. Thirty-three classrooms were preselected (unclear criteria), 17 were randomly assigned to receive the Al's Pals curriculum (218 children), and 16 remained as control groups (181 children).
Classrooms that received the Al's Pals intervention showed significant positive changes in socioemotional competence and prosocial skills as measured by the Preschool and Kindergarten Behavior Scales and Child Behavior Rating Scale-30. They also showed improvement in positive measures of coping. Control classrooms showed no significant changes in any of these measures.
Sources Lynch, Geller, and Hunt 1998; The objective of most programs that tackle socioemotional skills in a school environment is to enhance prosocial characteristics in participants, such as the ability to solve conflicts and be assertive. 10 Some also have the intention of decreasing risk factors related to early aggressive, violent, or antisocial behaviors or related to substance abuse. Others seek to improve the classroom climate and conduct as well. The importance of improving the classroom climate is explained  Brown and others (2010), who suggest that a positive classroom climate has been associated with greater self-esteem, perceived cognitive competence, internal locus of control, mastery motivation, school satisfaction, academic performance, and less acting-out behavior. Moreover, positive teacher-child relationships have also been associated with classroom quality and the process of learning (NICHHD 2003).
To teach all students in a classroom or school to identify and effectively handle emotions or social situations that could generate problems or conflicts in the absence of proper training or awareness, the programs that were studied have three main components. First and foremost, they are often based on a socioemotional syllabus that is integrated into the educational curriculum. Second, since teachers usually deliver the curriculum, the programs usually include teacher training on implementing the socioemotional component (incorporating specific dynamics or exercises into their existing school classes), as well as on class and group management. Third, the programs usually have a set of household activities that reinforce concepts learned at the school and that involve parents in their children's education.
For example, the 4Rs (reading, writing, respect, and resolution) Program is a literacy-based curriculum in conflict resolution and socioemotional learning, composed of 21 to 35 grade-specific lessons. To teach this curriculum and achieve positive rules and norms and safe and secure classroom environments, teachers receive 25 hours of training and ongoing coaching. Finally, a component called 4Rs Family Connections consists of activities for children and parents at home (see box 6.7).

Box 6.7 school-Based program: 4rs program, new York city
Brown and others (2010) used a novel measurement tool and a clustered, randomized control trial evaluation strategy to assess the short-term effectiveness of the 4Rs (reading, writing, respect, and resolution) Program on classroom quality. The 4Rs Program includes a comprehensive class curriculum, teacher training, and take-home activities to develop social and emotional skills. Participants were 82 third-grade teachers in 18 urban public elementary schools in the New York City area.
Classroom quality was measured using the CLASS assessment tool, which combines three different dimensions of the class environment: (a) classroom organization, (b) classroom emotional support, and (c) classroom instructional support. Results showed positive improvements on their measure of classroom quality, and effects were robust to differences in teacher's socioemotional factors. Among the study's limitations pointed out by the authors were (a) the small sample, which limited the ability to control observable factors in the estimation; and (b) the nature of the intervention and the design of the evaluation, which made it impossible to map specific components of the intervention to classroom quality.
By implementing each of these components, programs alter various environments where people learn and interact. Curricular modifications and teacher trainings alter classroom norms or school-wide practices (or both) to induce specific behaviors indirectly. Household activities alter parenting practices and modify the family environment. Other programs, not included in this review, encourage extracurricular activities (usually sports or music) as a way to create an environment that requires sharing responsibilities with others and to boost the learning of socioemotional skills (see box 6.8). 11 Most programs foster skills directly, through a school curriculum, and indirectly, through parenting and the household environment.
An analysis of the frequency of each of these components shows that class curriculum and teacher training are the most recurrent features of socioemotional programs in school environments. Figure 6.4 illustrates how many programs in our sample include each of these four components. Since programs usually target one or more components, the number of programs for some components is more than 30.

Box 6.8 after-school program: Big Brothers Big sisters of america, United states
Some programs, like Big Brothers Big Sisters of America, are extracurricular and might involve only a portion of all students. In this intervention, children attending school who were referred by school staff were matched to a trained volunteer mentor, who spent 45-60 minutes with the child about once a week.
Matches often chose how they spent their time together, although all of the programs had some degree of structure (that is, the activities from which matches could choose are at least partly outlined by the program). In a few cases, the activities in which matches engaged were predetermined. More often, programs offered suggestions for the meetings. Most of the contents did not focus on academics but instead emphasized creative activities (such as drawing and arts and crafts), playing games, and talking about various topics (such as friends, family, academic issues, and the importance of staying in school).
A randomized control trial involving 1,139 youths in grades four through nine in 71 schools nationwide found positive and significant effects of the program on (a) academic performance (overall, in science and in written and oral language), (b) the quality of classwork and the number of assignments turned in by students, (c) scholastic efficacy (feeling more competent academically), (d) students' college expectations, (e) the level of serious school misconduct, (f) the number of absences without an excuse, and (g) children's willingness to start skipping school. The program found no benefits in the out-of-school areas that were examined, including substance use, misconduct outside of school, relationships with parents and peers, social acceptance, selfesteem, and assertiveness. The program cost is an estimated US$1,000 per student per year.  Unfortunately, little evidence illustrates the relative effectiveness of each of these components. The reason for the lack of evidence is that school-based programs that seek to promote socioemotional skills do not normally test the effectiveness or intensity of their specific components. Out of the 45 programs included in the analysis, only the evaluations of Project SAFE, school-based programs in Colombia, and Becoming a Man evaluate the relative effect of the different components. Furthermore, only the evaluation of I Can Problem Solve incorporates the level of exposure in the evaluation design, allowing for an analysis of the persistence of the results (in particular, one can compare a group that has been in a program for two years with a group that was treated only during the first year, along with a control group for the two years).
The remaining programs capture the effect of exposure time by analyzing the impacts of the program after the first year and comparing them with the impacts in the second year. However, this method does not allow one to see how long the effects persist once the program is no longer delivered.

outcomes of school-Based programs
This section focuses on the main findings from evaluations of school-based programs that foster socioemotional skills. The primary outcomes of these types of programs will be discussed, followed by important program and participant characteristics.
Most of the impacts of school-based programs are related to risk factors (figure 6.5). In fact, our list of statistically significant outcomes for school-based Outcomes programs was distributed as follows: 59 percent related to behavioral variables, such as internalizing and externalizing behavior, aggression to peers, and cooperation; 32 percent were academic-related outcomes, such as grades, standardized test results, and classroom climate; and 7 percent related to health variables, such as drug use and high-risk sexual behavior. The minimal evidence related to important economic variables indicates the lack of longer follow-up periods in the evaluations of this part of the literature. Durlak and others (2011) conducted a meta-analysis with 213 studies and conclude that those programs significantly improved students' socioemotional skills and attitudes toward self, school, and others. Both internalizing behaviors (such as self-esteem, depression, and emotional distress) and externalizing behavior problems (such as aggression toward others and conduct problems) were also effectively addressed with these types of programs. In a systematic review of 87 rigorous research reports,  find that a large number of programs revealed positive and statistically significant outcomes with regard to antisocial behavior, social skills, and social-cognitive skills. In a meta-analysis of only programs that sought to reduce aggressive and disruptive behavior, Wilson and Lipsey (2007) find that all outcomes were positive and statistically significant. 12 The impact of these programs on academic performance (standardized test scores and grades) is also significant, and as documented by Payton and others (2008, 6), there is "an average gain on achievement test scores of 11 to 17 percentile points." 13 Finally, the programs have effects on the health of participants.
The positive effects of these programs are generalizable for both in-school and after-school settings and in diverse geographical contexts (urban, suburban, and rural areas). However, after-school programs tend to have smaller outcomes than school-based programs (Durlak and others 2011; Payton and others 2008).
When measured, these programs also seem to have positive rates of return. Belfield and others (2015) conducted a benefit-cost analysis of six prominent socioemotional learning programs: (a) 4Rs, (b) Positive Action, (c) Life Skills Training, (d) Student Success through Prevention (Second Step), (e) Responsive Classroom, and (f) Social and Emotional Training (the first five programs are in the United States; the last one was implemented in Sweden). They find that all of the programs have a positive return, as their benefits exceed their costs. In particular, the average benefit-cost ratio is almost 11 to 1, which means that for every dollar invested in the programs, there is a return of US$11.
However, even though school-based programs appear to have positive impacts on risk factors, they have certain limitations on their effectiveness. For example, • Programs might be effective in inducing changes only in evident and not complex risk factors. The evaluation of Second Step (United States) shows that programs might be more effective at reducing evident forms of disruptive and aggressive behaviors, such as physical aggression, rather than more complex forms of aggression, such as homophobic teasing and sexual violence (Espelage and others 2013). • The effectiveness of some programs might depend on the incorporation of schoolwide or community-wide components. After analyzing programs aimed at fighting bullying in school, Vreeman and Carroll (2007) find that programs that aim to change behavior based strictly on social cognitive principles are not effective. They suggest that effective programs need to address systemic issues and social environments. Thus, effective programs tend to incorporate a wholeschool approach that involves the entire school community.
• The Cambridge-Somerville Program (United States) had negative effects on its participants. This program provided medical assistance, tutoring, summer camps, and parent-teacher meetings for 13-year-old children who had behavioral problems (Kautz and others 2014). A randomized control trial that measured the impact of the program found increases in the drinking habits, serious mental diseases, heart problems, blood pressure, crime rates, and mortality rates of its participants. According to McCord (1978), the program did not create a sense of autonomy among participants. Once removed, it led to the original unfavorable behaviors while generating rejection and resentment for the lack of support.
Further research is needed to evaluate whether the documented effects are permanent or transitory. Many studies have not included postprogram follow-up assessments, and existing evidence related to the subject is mixed. Using 24 studies with a median follow-up period of 52 weeks, Payton and others (2008) find that the effects of the programs on socioemotional skills, positive attitudes and social behaviors, conduct problems, and academic performance persist over time. However, they also find that those long-term effects are less strong than the effects immediately after the intervention. By contrast, upon evaluating Across Ages, Aseltine, Dupre, and Lamlein (2000) find that the impacts of the program (designed to reduce substance use in high-risk youth) disappeared in a six-month follow-up.
Additionally, the question of what works better and under which circumstances has not been properly answered in the literature. As mentioned before, evaluations vary profoundly, which increases the difficulty of pinpointing the characteristics that drive program success (figure 6.6). Hahn and others (2007) argue that although some school programs are likely to have greater impacts than others, the characteristics (or the setting) that make programs more effective are not clearly understood. Jones and Bouffard (2012) note that research linking specific program components to outcomes is rarely found in the literature. Wilson and Lipsey (2007) find that multicomponent programs do not have a greater impact on the reduction of problematic behaviors than single-component school-based programs. Payton and others (2008) and Durlak and others (2011) find that this observation can derive from the challenges associated with the implementation of multicomponent programs.
Furthermore, the programs in our sample that analyze the effect of complementing the school-based component with a parental-home component have mixed results. As box 6.9 describes, the Project SAFE program found greater results when targeting both the school and the home environments. However, the

Box 6.9 impact evaluations that include an analysis of an intervention's components: project saFe and a school-Based intervention in colombia
To evaluate the impact of Project SAFE (Strategies Aimed at Family Empowerment), Kumpfer and others (2002) compare four groups of students: (a) children receiving the I Can Problem Solve (ICPS) intervention, a school-based universal primary prevention curriculum delivered by trained teachers in classes three to five times a week in 20-minute lessons; (b) students receiving the ICPS intervention combined with the Strengthening Families (SF) Program (a 14-session family skills training program that consists of skills training for the parents, skills training for their children, and skills training for the family); (c) students receiving the ICPS intervention and only the parenting component of the SF Program; and (d) control students not participating in any of the programs. The authors find that the most effective program delivery was the simultaneous administration of ICPS and SF. The ICPS intervention alone affected school bonding and selfregulation. The ICPS with the parenting component of the SF had significant effects on social competence and self-regulation. Finally, the implementation of ICPS and the full SF Program affected school bonding, parenting skills, family relationships, social competency, and behavioral self-regulation, all of which are associated with the risk of substance abuse.
For the impact evaluation of a school-based intervention in Colombia (Klevens and others 2009), students were divided into three groups: (a) some received the teacher-only intervention, where teachers received 10 four-hour weekly workshops on standard classroom management techniques and strategies for shaping children's behaviors; (b) some received a box continues next page results of the school-based programs in Colombia found that involving schools alone had greater effects than combining parent and teacher intervention. Several authors (Durlak and others 2011; Payton and others 2008) agree that the most successful socioemotional programs in school settings are sequenced, active, focused, and explicit (grouped together under the acronym SAFE). Sequenced programs have a planned set of activities that develop skills in a step-by-step approach. Active programs use active forms of learning like role playing, which provide the opportunity to use the recently acquired skills. Focused programs devote a sufficient amount of time and attention to socioemotional skills training. Finally, explicit programs clearly identify and specify the socioemotional skills that they aim to develop as a result of the intervention (Arthur and others 1998; Bond and Hauf 2004;Durlak 1997;Durlak, Weissberg, and Pachan 2010;Dusenbury and Falco 1995;Gresham 1995). These authors find that those four characteristics reinforce one another; an intervention that is focused but not sequenced or active will not be as effective as one that is sequenced, active, focused, and explicit. Durlak, Weissberg, and Pachan (2010) explain that the presence or absence of those characteristics in different programs moderates the programs' effects.
Complementing the finding regarding SAFE programs, Weissberg and Greenberg (1998) report that programs will most likely promote and sustain socioemotional skills formation when they are relevant to the challenges faced by children; when they have longer implementation periods (teaching skills over several years); and when they accompany socioemotional skills training with the opportunity to use the skills learned, a characteristic captured under the SAFE approach as "active." The length and intensity of the programs also appear to have important effects. For example, the evaluation of the STEPS to Respect program finds that children who are part of the program for two years experience greater improvements in problem behavior on the playground than children who participate for only one year (Frey and others 2009). Something similar is found with the Unique Minds program, where the level of program exposure directly influenced the results for cognitive, social, and emotional skills and grade gains; self-efficacy; and problem solving. Finally, Boyle and Hassett-Walker (2008) find that students who participated in the I Can Problem Solve program for two years showed less teacher/parent combined intervention that included the same teacher's training as well as a one-hour weekly parenting intervention that covered almost the same topics as the teacher's training; and (c) a third group served as a control.
The results of the intervention were surprising, as the teacher/parent combined intervention had a smaller effect on prosocial behavior.

Box 6.9 impact evaluations that include an analysis of an intervention's components: project saFe and a school-Based intervention in colombia (continued)
Taking Stock of Programs to Develop Socioemotional Skills http://dx.doi.org/10.1596/978-1-4648-0872-2 relational aggression than both students who had only one year of the program and students in the control group.
As with programs focused on early childhood development, implementation fidelity appears to affect a program's outcomes.  argue that the lack of significant effects of the Social and Emotional Aspects of Learning program results from implementation problems. Chang and Muñoz (2006) find that varying degrees of implementation fidelity of the Child Development Project implemented by Project CARE led to different results: when including all schools, there were no differences between all treated schools and controls. However, when comparing high-quality treatment and control schools, statistically significant effects were found in variables that included reading and number of student referrals. Finally, Payton and others (2008) and Durlak and others (2011) find that the absence of a careful record of implementation problems might hinder the promotion of better results in school-based programs. Further, they state that programs with reported implementation problems yielded fewer positive outcomes than those with no reported problems.
Multiple evaluations find that school-based programs that target socioemotional skills development are effective both for the general population and for particular ethnic groups. For example, the evaluation of Fast Track PATHS (Promoting Alternative Thinking Strategies) documents effects both for students of multiple ethnicities in elementary schools located in neighborhoods where delinquency and juvenile arrests are high and for students in more typical American public schools (CPPRG 2010;. Similarly, significant outcomes are achieved by Family Check-Up, a program implemented both among children in urban schools and among ethnic-minority adolescents (Connell and Dishion 2008;Connell, Klostermann, and Dishion 2012;Fosco and others 2013;Stormshak, Connell, and Dishion 2009;Stormshak and others 2011;Van Ryzin, Stormshak, and Dishion 2012;Fosco and Dishion 2012). Further, the evaluation of Positive Action also documents impacts for racially diverse students in both Hawaii and Chicago (Lewis and others 2013; Snyder and others 2013).
However, the effect of school-based programs on socioemotional skills is greater for students who exhibit higher initial levels of problem behavior or more risk factors (see figure 6.7). Some programs-such as 4Rs, Linking the Interests of Families and Teachers, and Mindful Awareness Practices-find effects on the entire population but report effects of greater magnitude on highly vulnerable students (those with the greatest levels of initial aggression or with the lowest level of executive functions). The greater effects of these programs on groups with multiple risk factors are also documented by . They explain that universal programs have lesser impacts on behavior because they affect the learning process of children who would not develop serious behavioral problems even in the absence of the program. Thus, in the long run, this group exhibits no major differences in behavior from those exhibited by a control group. Lösel (2001) warns against generalizing this finding to extremely high-risk youngsters, stating that an inverted U-shaped relationship between risk level and program effect appears to be most plausible. Following this theory, the most effective school-based programs would target those youngsters who are at risk, but not at extreme risk (see figure 6.8).
For their part, Hahn and others (2007) suggest that program effectiveness might decrease with the age of participants. Payton and others (2008) and Durlak and others (2011) also report that a student's mean age is usually significantly and negatively related to skill outcomes. However, they do not provide tentative channels or explanations for this result.

out-of-school programs
This section characterizes a set of programs that aim to promote socioemotional skills outside the school environment. As Kuckulenz (2007, 1) states, "Postschool learning is an important source of skill formation that accounts for as much as one-third to one-half of all skill formation in a modern economy." That conclusion is driven not only by the fact that most of the world's population is out of school, 14 but also by the documented productivity of skill investment in teenage and adult populations (Kautz and others 2014). Socioemotional skills are particularly relevant in this life cycle, since as Kautz and others (2014, 78) mention, "If the early years have been compromised, it is more effective in the adolescent years to focus on developing non-cognitive skills rather than cognitive skills." Having said this, out-of-school programs with socioemotional skill components have an ample range of objectives. For example, some seek to promote job placement for the unemployed, others intend to reduce gender violence or to improve stress-related coping strategies, whereas still others aim to decrease recidivism among individuals with substance dependence or conduct disorders. Although all of these programs play important roles in social protection systems around the world, this book will not cover programs specifically designed for people who have expressed addictions or conduct disorders. Additionally, of the out-of-school programs analyzed in this section, 75 percent are in countries other than the United States.
Most (75 percent) of the programs analyzed in this section seek to improve labor market outcomes. These programs usually include the promotion of socioemotional skills as part of a larger set of components, such as work-related skills training, mentoring, and on-the-job training. Of the programs studied, 10 percent aim to prevent gender-based violence by fostering socioemotional skills and providing educational activities and communication campaigns. Another 5 percent are oriented toward the reduction of risky sexual behaviors among Latino immigrants in the United States, and, finally, 5 percent provide intensive remedial education, social skills training, and personal development.
The group of individuals targeted by out-of-school programs range from age 10 to over age 55. However, since most of the programs included in the sample have job placement as an objective, they focus on potentially economically active populations (adolescents and young adults). Furthermore, as can be seen in table 6.7, which illustrates the distribution of programs in our database organized by targeted age, most programs focus on younger cohorts. Although all of the programs in our sample include youths ages 16 to 30 as part of their target audiences, only 30 percent include individuals over age 30.
A careful analysis of the participant profiles of these programs shows that most focus on unemployed or underemployed youths with a low socioeconomic background. Some programs, such as National Guard Youth ChalleNGe (United States), Apprenticeship Training for Vulnerable Youth (Malawi), Juventud y Empleo (Dominican Republic), and Kingston Young Men's Christian Association (YMCA) Youth Development Programme (Jamaica), explicitly benefit adolescents who have dropped out of school. These programs have the intention of reintegrating youths into the education system or equipping them with the necessary skills to enter the labor market. A wide variety of programs benefit women; 20 percent of them were exclusively aimed at women, contributing to the amelioration of gender inequality in the labor market. 15 The following components are included in most programs that seek to improve labor market outcomes for vulnerable groups: • Classroom instruction: Instruction normally includes technical and theoretical training in a particular field, as well as academic education in math, language, and writing. Professional training services are usually provided only for low-skill occupations like bakery work, bartending, hair styling, tailoring, and carpentry. • Counseling or mentoring: Programs often provide their beneficiaries with the opportunity to contact a professional, role model, or more experienced individual to advise them on how to interview, where to seek jobs, and how to pursue additional training. • Internships or other workplace-based training activities: Many programs offer hands-on training opportunities in the workplace, including internships or apprenticeships. Internship programs are more common in developing countries and consist of less-intensive interactions that usually last from three to six months and that do not involve classroom training. Apprenticeship programs, which are very common in Europe, last for three to four years and involve highly structured training that combines in-class knowledge (that relates to the particular occupation) with workplace practice (Kautz and others 2014 explicit classroom curriculum. Other times, it is incorporated in the classroom dynamic and taught simultaneously with the particular field or the academic subject. Some programs also transmit socioemotional skills during the apprenticeship, internship, or job-training periods. Seeking to offset some of the opportunity costs associated with leaving the labor force while training, programs like Jóvenes en Acción in Colombia and Juventud y Empleo in the Dominican Republic also provide stipends to their participants. Additionally, programs like Jordan New Opportunities for Women (NOW) provide subsidies to employers that hire the programs' graduates. Previous evaluations of such wage subsidies have found limited results, which are normally attributed to low usage rates of subsidies and to possible stigma effects (Groh and others 2012). Figure 6.9 illustrates the components found in the set of programs gathered in our literature review and their frequency. Classroom training is by far the most commonly recurring feature of the programs analyzed. It is followed by counseling and mentoring, and by workplace-based job training. Few programs include a subsidy component.
The average length of the programs in our database is 10 months (median of 6 months). However, programs vary substantially: whereas Team Awareness lasts only five hours, the Kingston YMCA Youth Development Programme can take up to four years to complete. Sixty percent of the programs are full time and last more than a week. When present, the on-the-job training component lasts from one to six months. Only two programs (Job Corps and the National Guard Youth ChalleNGe Program) involve high-intensity arrangements where beneficiaries live on the program's premises. As Honorati and McArdle (2013)   Information on costs was available for 45 percent of the programs. Cost information normally includes the entire program, not just the socioemotional skills formation module or component. For these programs, investments ranged from US$85 to US$16,500 per participant. Taking into account that the higher costs are associated with the two residency-based programs in our sample, the costs of out-of-school programs are low when compared with programs in the other stages of the life cycle. On the basis of an analysis of job-oriented programs in Latin America, González-Velosa, Ripani, and Rosas-Shady (2012) suggest that the small per capita investment should be associated with modest expected results. The programs they analyzed cost between US$330 and US$750.

outcomes of out-of-school programs
This section synthesizes the central findings regarding out-of-school programs with components that strengthen socioemotional skills. Following a description of the main outcomes of these types of programs is a discussion on the aspects of program design and implementation that alter a program's impacts. Finally, the ways that various participant characteristics can hinder or facilitate the attainment of positive results are summarized.
The majority of the out-of-school programs analyzed in this book are oriented toward improving labor market outcomes (see box 6.10 and table 6.9). Consequently, most of the documented program outcomes are related to job status and earnings and to behavior. In fact, 38 percent of the outcomes fall under the economic-related category and a similar percentage of the outcomes (35 percent) are related to behavioral aspects, such as leadership skills, self-esteem, conflict resolution abilities, confidence, and arrests and convictions. Furthermore, 17 percent of the outcomes are related to health (such as obesity, and drug and alcohol consumption), and 10 percent are related to academic aspects (such as enrollment, hours in education, and certification). See figure 6.10.

Box 6.10 out-of-school program: Juventud y empleo, Dominican republic
Juventud y Empleo is a workplace-based intervention that targets Dominican adolescents between ages 16 and 29 who live in poor neighborhoods; are not attending school; lack a high school diploma; and are unemployed, underemployed, or inactive. The program provides them with 150 hours of vocational training, 75 hours of socioemotional (life skills) training, and a three-month internship to increase the likelihood of their finding a job, to reintegrate them productively into society, and to prevent them from participating in delinquent activities. While participating in the training, participants receive a stipend of close to US$3 per day. At 18 to 24 months after the intervention, overall employment showed few effects, but job quality for males was affected significantly (17 percent increase in formal employment, 7 percent increase in monthly earnings for those employed), especially in Santo Domingo. The program also had a positive effect on teenage pregnancy, and in women's expectations and perceptions about the future. With regard to socioemotional skills, the program improved leadership skills, persistency of effort, and conflict resolution.
Six years after the intervention started, employment and earnings continued to show no significant effects. However, formal employment for men had sustained positive effects (25 percent more men in the treatment group have a job with health insurance) and for both genders in Santo Domingo (when comparing the participants and the control group, 30 percent more individuals participating in the intervention have jobs with health insurance). In Santo Domingo, the program also increases earnings by 25 percent for women participants over the control group. Box 6.10 out-of-school program: Juventud y empleo, Dominican republic (continued) The programs' effects on employment, formality levels, number of hours worked per week, and wages or earnings tend to be small but statistically significant. However, evidence regarding program effects on employment levels and quality of employment is mixed. For example, the impact evaluations of Juventud y Empleo (Dominican Republic), Jóvenes en Acción (Colombia), and entra21 (Argentina) found no significant effects on employment levels, but they found significant improvements in formality (employment with health insurance, a written contract, formal wages) and earnings for those employed. Similarly, the Year Up program (United States) significantly increased the annual earnings of beneficiaries, primarily through higher hourly wages and better job placements, but with no significant changes in employment rates. In contrast, programs like Job Corps (United States) and Galpão Aplauso (Brazil) increased employment rates but had no effect on the quality of the jobs offered to participants. 16 When measured, job-oriented programs had mixed effects on risk factors. For example, after 75 hours of basic or life skills training that aimed to strengthen trainees' self-esteem and work habits, Juventud y Empleo had an effect on participants' aspirations, conflict resolution skills, leadership, self-esteem, and order and self-organization. 17 It also affected persistence of effort among women and leadership skills among men. In Jordan, Questscope Non-Formal Education was found to have had significant effects on conduct problems (as measured by the Strengths and Difficulties Questionnaire, which screens for behavioral issues in five domains of psychological attributes), 18 men's prosocial behaviors, connectedness, and hyperactivity, although the effects on self-efficacy and social skills  Interestingly, the two programs that offer residential services in our sample (Job Corps and National Youth Guard ChalleNGe) have short-term effects on arrests and conviction rates. However, the effects of both programs on criminal behavior fade over time. Kautz and others (2014) attribute this phenomenon to the fact that since youths are housed in the program, they are "incapacitated" from committing any crime in the short term. Once they return to their original environment, they revert to their usual behaviors and criminality. That, however, does not mean that modifying aggression should not be considered a viable program outcome for programs that target extremely disadvantaged youth: the Kingston YMCA Youth Development Programme in Jamaica finds statistically significant effects on the aggressive behavior and aggressive propensity of its participants.
Finally, some of the programs have statistically significant effects on academics-related variables, such as knowledge of information and communication technology; knowledge of life skills; enrollment in educational institutions; number of hours spent on education; and attainment of degrees, certificates, or diplomas. Those effects were expected, as most of the programs have a component that trains participants on particular job-related skills.
Cost-benefit analyses find that out-of-school programs that target labor market outcomes and are not residential appear to be cost-effective (see figure 6.11). As Card, Kluve, and Weber (2010) mention, few studies include the information required to conduct a cost-benefit analysis. However, the impact evaluations that do capture cost-effectiveness have found positive results. For example, the evaluation of Procajoven (Panama) found that the overall costs of the program were recovered in 12.6 months. Additionally, Attanasio, Kugler, and Meghir (2011) estimate that Jóvenes en Acción (Colombia) yielded internal rates of return of 35 percent overall and 21.6 percent for women. Further, Alzúa, Cruces, and Lopez Erazo (2013) calculate that entra21 (Argentina) had an internal rate of return of 19 percent, which, although positive, was lower than the rate found for similar programs in the region. Finally, several studies of apprenticeship training in European countries find high rates of return for participants, in the range of 9 percent to 15 percent (Adda and others 2013; Bougheas and Georgellis 2004;Clark and Fahr 2002;Winkelmann 1996). Consistent with those findings, Hollenbeck (2008) estimates that in the state of Washington (United States), social and governmental returns (for example, taxes received on earnings and reductions in spending on unemployment insurance, Temporary Assistance for Needy Families, food stamps, and Medicaid benefits) to apprenticeship programs administered by the state's Department of Labor and Industries are greater than 20 percent within the first 2.5 years after apprentices leave the program.
By contrast, residential programs that foster job placement do not appear to be cost-effective. Programs such as Job Corps and National Guard Youth ChalleNGe have very high costs for trainee housing and meals and for teaching costs. Even with important benefits, those costs lead to net earnings effects of close to zero (Bloom, Gardenhire-Crooks, and Mandsager 2009;  These results are limited by the fact that out-of-school programs have short follow-up periods. In fact, only 25 percent of the programs analyzed in this book have a follow-up period that exceeds two years. Shorter follow-up periods may lead to an underestimation of the effects of active labor market programs 21 and may limit the number of outcomes that can be measured. For example, evaluations that cover only two months after an intervention cannot reasonably expect to find effects on fertility outcomes. This limitation restricts the accuracy of the rates of return calculated for these programs. In addition, it also increases the difficulty of making claims regarding the longevity of the results. For example, the evaluation of Jordan NOW found extremely large immediate effects from providing a subsidy to employers as an incentive to hire graduates of the program. Nevertheless, the effects disappeared 14 months after the start of the intervention. This finding is consistent with reports from Calero and others (2014), who report that the impact of wage subsidies peaks at 12 months and then fades.  Kautz and others (2014) argue that this lack of long-term evaluation might be incorrectly fueling the perception that programs targeting adolescents are not as effective as programs that target earlier life cycles. For instance, if labor marketoriented programs have low impacts over the short run but significant impacts over the long run, calculated estimations might underestimate their effects. This problem is observable in the evaluation of the United States' Workforce Investment Act, which appeared to have negative impacts in the short run but actually resulted in higher earnings in the long run (Heinrich and others 2013). Caliendo, Schmidl, and Uhlendorff (2011) observe a similar phenomenon in Germany, where short-term negative impacts are followed by positive significant impacts after 10 months. Card, Kluve, and Weber (2015) analyze 200 econometric evaluations of active labor market programs and find that long-term evaluations seem more favorable: even though average impacts are close to zero in the short-term, impacts become positive and significant two or three years after the intervention.
Because of the diversity of out-of-school programs, identifying their effective elements is difficult. However, evidence suggests that program components matter, and that "workplace-based programs that teach non-cognitive skills appear to be the most effective remediation programs for adolescents" (Kautz and others 2014, 82). In particular, • There appears to be a greater benefit from including socioemotional skills compared with having mainly cognition and academic learning (Kautz and others 2014). In the evaluation of Jordan NOW, Groh and others (2012) find that although life skills training had only slight effects on employment (no effects in the short term and a slight effect one year after program completion but only for those outside of Amman), it improved life outlook and reduced depression. Additionally, in the evaluation of the Ninaweza program (Kenya), Alvarez de Azevedo, Davis, and Charles (2013) find that the inclusion of a socioemotional skills component had a significant effect on knowledge acquisition, life skills, and the probability of obtaining a job. Oddly, the group in this program that went without life skills training experienced greater increases in their weekly income. • Furthermore, on-the-job training appears to be a constant component of effective out-of-school programs (Calero and others 2014; Urzua and Puentes 2010). One reason could be that this type of program involves the transmission of field-specific job-related skills, as well as noncognitive skills, such as punctuality, personal presentation, and teamwork. Furthermore, workplace-oriented training eases matching of workers and firms. The prospect of future employment may also serve as an important incentive for participating youth, increasing their effort to meet the requirements of the particular profession. However, because these programs integrate cognitive and socioemotional education with particular incentives for satisfactory performance, they do not allow one to discern the relative effectiveness of each of these elements.
Our limited evidence on the use of residential services to boost labor market outcomes is not promising. Only two programs in our sample of preventive programs offer residential services: Job Corps and National Guard Youth ChalleNGe. These programs take place in the United States, target at-risk youth and highschool dropouts, usually take more than a year to complete, and include training and remedial education as well as mentoring. Over the short term, they have substantial impacts on employment, educational attainment, and crime. However, the effects of Job Corps on earnings and the effects of National Guard Youth ChalleNGe on high school graduation and criminality fade over time.
Further analysis of the programs included in this book shows that, as expected, the quality of the programs mediates the magnitude of their effects. For example, the impact evaluation of Questscope Non-Formal Education (Jordan) found both significant negative effects in low-quality centers and important positive effects in high-quality centers. Also, the low-intensity socioemotional intervention called Jordan NOW was found to have a limited effect on women's graduation from public community colleges. 22 Most programs appear to work better for younger individuals and females, and when implemented in cities (see figure 6.12). The impact of both Procajoven (Panama) and Jóvenes en Acción (Colombia) on the probability of employment, number of hours worked weekly, and labor earnings is limited to women. Procajoven also has greater effects in Panama City. Ibarrarán and others (2012) also find that Juventud y Empleo (Dominican Republic) has greater benefits for younger individuals and for those living in Santo Domingo, but more in men's formal employment. entra21 (Argentina), Questscope Non-Formal Education  (2010) report that youth programs lead to fewer and less positive results in labor outcomes. In an analysis of more than 200 econometric evaluations, Card, Kluve, and Weber (2015) find larger gains for females and participants that enter the program after long-term unemployment. With regard to location, multiple authors argue that programs aimed at improving labor market outcomes appear to have greater effects in developing countries. For example, Urzua and Puentes (2010) and Ñopo and Saavedra (2003) find more positive evidence in Latin America and the Caribbean than in more industrialized countries. Betcherman, Olivas, and Dar (2004) document similar results after reviewing 69 impact evaluations of unemployed and youth training; the impacts they found in the 19 programs implemented in developing countries were greater than those found in the United States and Europe. notes 1. Appendix A describes the before-school programs that were analyzed.
2. The calculation focused on statistically significant outcomes only. Including both significant and nonsignificant outcomes would have been ideal, as that would have provided information on the objectives of the programs and their effectiveness. However, most evaluations do not report nonsignificant outcomes, resulting in no knowledge regarding further program objectives.
3. Only the Nurse-Family Partnership and the Jamaican Study find effects on cognitive abilities that persist over time.
4. Wealthier children who first attended primary school switched to preschool, but children from poorer families who were attending primary school left the formal education system when the official age of primary school was enforced.
5. Conscientiousness is defined as the degree to which a person is willing to comply with conventional rules, norms, and standards.
6. Currie and Thomas (1995) document that Head Start's impact on vulnerable minorities decreases significantly when these children stop attending the program and return to their initial environment.
7. Appendix B describes the school-based programs that were analyzed.
8. To avoid redundancy, prekindergarten and kindergarten programs were included in the preceding section concerning early childhood programs in an institutional setting.
9. See appendix B for further details.
10. According to Eisenberg and Mussen (1989, 3), prosocial behavior refers to "voluntary actions that are intended to help or benefit another individual or group of individuals." 11. Because of space and time limitations, this book does not analyze programs that take place during after-school hours or in the summer.
12. Standardized mean difference is defined as the difference between the treatment and control group means on an outcome variable divided by their pooled standard deviation.  (2015) find that average impacts may be close to zero in the short run, but become more positive two to three years after program completion.
22. Jordan NOW provides 45 hours of socioemotional training over a nine-day period, with a maximum of 30 participants in each training group. The course covers effective communication and business writing skills (for example, making presentations and writing business reports and different types of correspondence), team building and teamwork skills (for example, characteristics of a successful team, how to work in different roles within a team), time management, positive thinking and how to use it in business situations, excellence in providing customer service, and résumé and interviewing skills.

Program Findings: What Works (or Doesn't Work) in Fostering Socioemotional Skills?
This chapter presents our best effort to synthesize the most important findings of the extensive literature on the rigorous evaluations of programs that seek to promote socioemotional skills, usually alongside fostering several other skills or behaviors. The findings below summarize results along the following dimensions: targeting, focus, impacts, replicability, and evaluation.

targeting
Most programs reviewed are oriented toward individuals with high vulnerability. Before-school programs tend to benefit (a) children in low socioeconomic strata or who belong to minority groups or (b) children whose parents are poorly educated, have low occupational status, have low cognitive skills, or are in their teenage years. School-based programs are often carried out in institutions located in high-risk areas, including regions with low income, considerable school absenteeism, high mobility, or elevated crime rates. Finally, out-of-school programs focus primarily on low-income unemployed individuals, those who have dropped out of the education system, or women. With regard to age, gender, and location, programs tend to target all age ranges, are mostly gender neutral (although a few out-of-school programs focus on women), and focus on urban areas. Because of the differences in the age groups covered, the process for entry into the different programs varies greatly throughout the life cycle: whereas beforeschool programs require parental consent and interest, school-based programs normally require consent only from school officials, though some programs do seek parental consent. Finally, out-of-school programs are sought out by youths and adults who make their own decisions about whether to join.

Focus
Another important finding is that few programs focus exclusively on the development of socioemotional skills. For example, most before-school programs also seek to provide care or to support cognitive development or school readiness; school-based programs seek to improve classroom climate and to decrease aggressive, violent, or antisocial behaviors and substance abuse; and outof-school programs tend to have labor market-related components. This finding is expected, because socioemotional skills development is normally part of a broader curriculum.

impacts
Most of the before-school programs focus on highly vulnerable children and their families. They have important effects on cognitive skills and academic outcomes in the short run, whereas in the long run they improve behavior and risk factors, which indicates that they can modify socioemotional skills. These programs appear to be particularly effective when targeting vulnerable populations, as well as when they have multiple components, involve parents, and are intense with regard to instruction time and the teacher-to-child ratio. Unlike most early childhood development programs, school-based programs tend to be universal at a school or classroom level, benefiting all children and adolescents in the classroom or attending the institution. The evidence regarding these programs is less widespread, but available results show a particularly strong effect on such risk factors as internalizing and externalizing behavior, aggression toward peers, and cooperation. Long and intense programs tend to be more effective, as do those that target vulnerable or at-risk populations.
Programs for out-of-school children and youths are usually designed to achieve immediate labor market outcomes (for example, job placement, formal employment, and higher wages). Although some of these programs show positive and statistically significant effects on socioemotional skills, the impacts tend to be small. These programs seem to be cost-effective, but long-term evaluation is lacking.
Early childhood programs appear to have a greater impact than those that take place later in life. Kautz and others (2014) state that only very early childhood programs (before age three) improve IQ (intelligence quotient) in a lasting way. Further, they suggest that the most promising adolescent programs integrate aspects of work into traditional education, breaking down the rigid separation between school and employment. Kautz and others (2014) explain that early childhood is a sensitive period for cognitive and noncognitive skills development. Furthermore, through dynamic complementarity, investments in the early years make later investments more productive. They also state that the most successful programs target preschoolers (after age three) and primary school children, improving later-life outcomes by developing socioemotional skills. However, as will be explained below, these results could partially reflect the fact that early childhood programs have longer follow-up periods.
Throughout the life cycle, socioemotional skills acquisition appears to explain important life outcomes. Several programs throughout the life cycle find longterm effects on earnings despite no indication of cognitive gain. This finding suggests that the long-term effects result from a change in beneficiaries' socioemotional skills. Among these programs are the HighScope Perry Preschool Program, the Abecedarian Program, and the Career Academies Program (before-and duringschool programs).
replicability Finally, there are questions regarding replicability, as small-scale programs appear to be more effective than massive ones, and external research on scaling-up programs frequently shows that positive effects diminish with dissemination (Durlak and DuPre 2008). Kautz and others (2014) find that the reason behind this outcome in early childhood development programs is that large-scale programs are less intensive and involve parents less frequently. Others argue that this outcome is the result of higher levels of program fidelity in impact evaluations of small-scale programs: normally, the researcher or highly trained university students implement the intervention, whereas the program itself is normally run by individuals with lower education attainment levels, less training, or less commitment to the project.
The diminishing positive effects with dissemination might be explained by, among other causes, the following: the programs are seldom designed for universal coverage; and when they are scaled up, they experience lower participation rates and lower retention, have lower levels of quality and fidelity, and are unable to meet families' needs (Dodge and others 2014; Howard and Brooks-Gunn 2009). Lösel and Beelmann find greater effects in rigorous impact evaluations of school-based programs that have smaller sample sizes and state, "in large studies, difficulties in maintaining program integrity and homogeneity of samples or treatments may reduce design sensitivity" (2003,99). They also find that when small evaluations are taking place, programs tend to be delivered by the study authors, research staff, or other highly qualified individuals-rather than by more typical staff-leading to greater effects than one finds when programs are implemented widely. Hahn and others (2007) also point to the difficulty of providing accurate conclusions because of reporting implementation problems.
However, evidence also shows that some large-scale, massive programs can be effective. For example, the school-based Chicago Child-Parent Center program in the United States had significant effects on high school completion, years of education, juvenile and violent arrests, school dropout rates, grade retention, and use of special-education services. The replicability of the results relies on maintaining such features as expenditure per child, training for providers, and parent involvement. Furthermore, the external validity of the results can be compromised when a program is replicated in a different context or country with dissimilar conditions. In some developing countries, for example, where malnutrition levels are higher, programs that effectively foster socioemotional skills in more developed countries or settings cannot be expected to achieve the same results.

evaluation
Programs that target early stages in life have longer follow-up periods. Participant outcomes were recorded for 10 or more years after the intervention for 36 percent of the before-school programs, 3 percent of the school-based programs, and none of the out-of-school programs. The length of the follow-up directly affects the results captured by each program, for the following reasons.
First, the outcomes that can be measured in programs with short follow-up periods are more constrained. Before-school programs often capture effects that can be observed only with the evolution and growth of participants. For example, they measure behavioral effects with outcomes that evolve over time, from playground practices and school suspension to arrests, substance abuse and use, early sexual activity, and even reproductive and childbearing practices. This measure is profoundly richer than what is measured in out-of-school programs, which focus on immediate job-related outcomes.
Second, the persistence of any intervention's effects can be measured only through observations made over an extended period. All evaluations that are restricted to a few months might leave the reader questioning whether the findings might vanish after a certain time frame. Furthermore, this lack of information could mislead those responsible for policy design and implementation. For example, without the 40-year follow-up applied in the Jamaican Study, policy makers could not have learned that providing stimulation was more effective than providing nutrition with regard to cognitive and character achievements for stunted children living in poor, disadvantaged neighborhoods.
Third, programs with extensive follow-up periods provide rich information that allows researchers to understand the process of skills acquisition behind the observed impacts (Kautz and others 2014). For example, one could hypothesize that the HighScope Perry Preschool Program's long-term positive effects on criminality are the result of socioemotional skills acquisition rather than cognitive gains, because the follow-up 37 years later found that the effects on test scores faded, whereas behavioral results did not.
An additional finding is that although the programs explicitly target socioemotional skills development, few explicitly indicate what socioemotional skills they intend to improve. Further, most impact evaluations do not directly measure the skills that the program intends to alter. As discussed throughout this book, the effects of the different programs are captured in the behavior of the beneficiaries, yet rarely do researchers look to see if an intervention modified personality traits, risk behaviors, or mind-sets. Future research should include skills assessment in order to measure whether or not such changes occur as a result of the intervention. The evaluation of Juventud y Empleo and Galpão Aplauso are exceptions, because they use Rosenberg's self-esteem scale, Duckworth's Grit scale, and the Social and Personal Competencies Scale. The Juventud y Empleo intervention had a statistically significant effect on socioemotional skills, whereas Galpão Aplauso had no effects on this area. Furthermore, the literature might contain blind spots on socioemotional skills formation, as the effect of environmental investments to improve socioemotional learning appears underexplored. Only one intervention in our sample (Program H in Brazil) evaluated the impact of a community-wide "lifestyle" social marketing campaign to promote condom use and change gender norms and behavior.

Direction for Future research
Although evidence regarding socioemotional skills is promising, important questions still remain unanswered. Future studies should define and measure socioemotional skills directly so as to identify the triggers that lead to behavioral alterations. As mentioned before, a key area for future research is related to the identification of longer-term effects, particularly for school-age and out-of-school populations. It would also be important to measure the relative effectiveness of program intensity versus program duration as well as the way in which the overall intervention is structured and delivered.
Further, impact evaluations of multicomponent programs should capture the cost-effectiveness of the different components. An important area of research lies in the identification of the ideal variations within programs that achieve maximum effects within dissimilar target populations. Other areas of future research include the documentation of the externalities of the programs, which could be achieved by assessing the effects of school-wide versus classroom-wide programs. Much could be gained by documenting the costs of the programs, by including longer follow-up periods, and by collecting direct measures of socioemotional skills. references Dodge, Kenneth A., W. Benjamin Goodman, Robert A. Murphy, Karen O'Donnell, Jeannine Sato, and Susan Guptill. 2014. "Implementation and Randomized Controlled Trial Evaluation of Universal Postnatal Nurse Home Visiting." American Journal of Public Health 104 (S1): S136-S143.  Beneficiaries: Young children (from pregnancy until the child enters the school system) from poor families who need intensive and comprehensive support services Objective: "Enhance the physical, social, emotional, and intellectual development of children in low-income families; provide support to their parents and other family members; and assist families in becoming economically selfsufficient" (Pierre and others 1997, 1 The program had no significant effects. The variables that were measured were mother's employment status, level of employment (number of hours worked) of parents, income, level of dependence on public assistance, mother's educational status, parental attitudes linked to abusive or neglectful behavior, quality of the home environment, parent-child interaction, and health behaviors of mothers during subsequent pregnancies. Note: The program was not implemented according to design. Only 33% of the families were enrolled for five or more years.

Child-Parent Psychotherapy
Beneficiaries: Children ages 3 to 5 years who witnessed marital violence Objective: Alleviate symptoms and restore child's development toward healthy functioning Description: Child-parent psychotherapy (CPP) is a relationship-based treatment that integrates modalities from psychodynamic, attachment, trauma, cognitive-behavioral, and social learning theories. The treatment consisted of 55 weekly sessions that lasted approximately 60 minutes each, distributed over a 1-year period. Cost: n/a Authors: Lieberman, Ippen, and Van Horn (2006) Type: RCT Location: United States Follow-up: 6 months later Sample: 75 preschool-age child-mother dyads were randomly assigned to CPP (42) or case management plus community referral (33).

Play and Learning Strategies (PALS) program
Beneficiaries: PALS I-mothers with infants ages 6 to 12 months; PALS II-mothers with toddlers ages 24 to 36 months Objective: Teach parents responsive parenting skills to support their child's socioemotional cognitive, and language development. Description: Each week for three months, families are visited by coaches who implement 1.5-hour sessions that include the discussion of the mother's practice during the preceding week, a description of the behavior targeted in the visit, an observation and discussion of educational videos demonstrating the skill, videotaping of mothers practicing the skill with their child, reviewing the videotaped practice, and planning ways to integrate responsive behavior during the upcoming week.
Authors: Landry, Smith, and Swank (2006) Type   [maternal use of rich language for children who received PALS I and PALS II (+)**]; verbal encouragement for children who received PALS II (+)**. Smaller effects were found with PALS I, even though they were greater than those found with DASs.

Nurse-Family Partnership
Beneficiaries: Nulliparous women who were pregnant less than 29 weeks. Most were (a) low income, (b) unmarried, and (c) teenagers. Objective: Provide maternal and early childhood health programs to families most in need. Description: Families received sensory and developmental screening, as well as transportation services for prenatal and well-child care. Additionally, families received monthly prenatal and infancy home visits by public health nurses (during pregnancy and the first two years of their children's life). The nurses tried to (a) improve pregnancy outcomes by helping women improve their health-related behaviors, (b) improve children's health and development by helping parents provide more competent care, and (c) improve families' economic self-sufficiency by helping parents make appropriate choices regarding family planning, finishing their education, and finding work, and by linking them to appropriate services. (2010)    high-risk women (approximately 90% of the women were African American, 85% were low income, and almost all were unmarried; their average age was 18), who were randomly assigned to receive (a) 255 free developmental screenings and referrals for their child before age two, (b) 245 screenings offered in treatment 2 plus paraprofessional home visiting during pregnancy and the child's first two years of life, (c) 235 screenings offered in treatment 1 plus nurse home visits during pregnancy and infancy. Term used: Socioemotional development Health related: Infant and childhood mortality from preventable causes 6 years later(−)*; children's injuries or ingestions before age 2 (−)**; substance use by mothers-cigarettes, alcohol, and marijuana (−)**; mother's subsequent live births (−)**; mother's subsequent low-birth-weight newborns (−)* Behavior related: Behavior problems in the borderline or clinical range (−)**; aggression for high-risk children (−)** Academic/cognitive related: GPA for children in risk group at age 12 (+)*; results on achievement tests for children in high-risk group at age 11 (+)**; enrollment in formal out-of-school care by ages 2 to 4.5 (+)**; intellectual functioning and receptive vocabulary scores at age 8 (+)**; for high-risk children, arithmetic achievement (+)** were Mexican American, 36% were white, 15% were African American, and 84% were unmarried; average age was 20), who were randomly assigned to receive nurse home visits or developmental screening and referral to treatment for their child Term used: Socioemotional development Health related: Subsequent pregnancies for mothers (−)**; interval between first and second birth for mothers (−)**; health care encounters for children's injuries or ingestions at age 2 (−)**; says hospitalized for injuries or ingestions at age 2 (−)** Behavior related: Emotional/behavioral problems in children under age 6 for children visited by nurses (−)*; internalizing problems at age 9 for children visited by nurses (−)*; dysfunctional attention at age 9 for children visited by nurses (−)**; for high-risk children visited by paraprofessionals: deregulated aggression (−)**; behavioral regulation (+)** Academic/cognitive related: For children in high-risk environment who were visited by nurses: receptive language at ages 2, 4, and 6 (+)**; sustained attention at ages 4, 6, and 9 (+)**. Those receiving visits by paraprofessionals presented differences only in visual attention/task switching at age 9 (+)*. Economic related: Months using welfare (−)**   (47 in HIPPY and 66 as control), which received services in 1991. Children were randomly assigned to HIPPY or to high-quality preschool. In Arkansas, children who attended HIPPY were matched over several characteristics with those who did not. student's adjustment to classroom (+); enjoyment of books; listening and paying attention; task orientation, self-direction in learning, initiative, and interest in school (+) Academic/cognitive related: Test on reading and language arts (+); classroom adaptation (+) Note: No p-values reported in the paper.

Authors: Eckenrode and others
However, the aspects mentioned above are reported as statistically significant. The following variables were not statistically different between groups: math scores, school attendance, special education, classroom grades, teacher ratings of classroom behavior, curiosity, and using assistance.

Integrated Early Childhood Development Intervention in Colombia
Beneficiaries: Children ages 12 to 24 months located in 96 small municipalities in Colombia Objective: Reduce some of the developmental risks associated with poverty. Description: During 18 months, female community leaders delivered psychosocial stimulation through weekly home visits with play demonstrations or micronutrient sprinkles that were given daily or both combined.

Parents as Teachers (PAT) program
Beneficiaries: Low-income women with children no older than 6 months. The program starts in pregnancy or infancy and continues through kindergarten entry. Objectives: (a) Increase parental knowledge of early childhood development, (b) improve parenting practices, (c) detect developmental delays and health issues early, (d) prevent child abuse and neglect, and (e) increase children's school readiness and success. In addition to home visits, the program provides health and developmental screenings, group meetings, and referrals to resource networks. Description: These monthly (or even more frequently if at-risk family) home visitation services are provided by trained parent educators (with at least some college education) who provide information about children's development, model and involve parents in developmentally appropriate activities, and respond to parents' questions and concerns. Additionally, the program involves parent group meetings to discuss child development and to build informal support networks with other parents, periodic developmental screenings, and referrals to community services as needed.
Authors: Wagner and Clayton (1999) Type: RCT Location: United States: RCT 1, Salinas Valley, California; RCT 2, Los Angeles, San Bernardino, San Diego, and Santa Barbara Counties in California Follow-up: When children were 2 years old Sample: Two RCTs: RCT 1 was composed of 497 Latino families where parents had limited English proficiency; 298 were assigned to treatment (receiving an average of 20 visits of 28-50 minutes) and 199 to control (which received age-appropriate toys and referrals if their children were found with delays); and RCT 2 was composed of 704 families with teenage parents who received (a) 177 PAT services alone, (b) 174 who received case management, (c) 175 who received PAT and case management, and (d) 178 untreated. Term used: Social development Health related: Immunization for children receiving the case management intervention (−)**; child abuse or neglect in Teen PAT, for the group receiving PAT and case management (−)** Behavior related: Self-help development for Latino families (+)**; social help for Latino families when using multivariate analysis (+)* Academic/cognitive related: Cognitive development in Teen PAT, for the group receiving PAT and case management (+)**; when using multivariate analysis, cognitive development measured by DPII for the Latino PAT, and the groups receiving case management and the combined intervention in Teen PAT (+)** Note: No benefit with regard to parenting knowledge or practices.

Parent-Child Home Program
Beneficiaries: Low-income families with children ages 2 to 4 Objective: Improve parent-child interactions to strengthen children's cognitive development and early literacy Description: The program consisted of 46 twice-weekly paraprofessional or volunteer home visitation services spread over seven months in each of two years. In this play-filled and nondidactic home-based intervention, the toy demonstrator models verbal interaction with the child centered on toys and books that are permanently assigned to the family and encourages the mother to assume responsibility for the interaction. The program has a cognitive curriculum, a socioemotional behavior curriculum, and a "parenting" curriculum. Cost: US$4,500 per child for two-year program Authors: Levenstein and others (1998) Type: RCT Location: United States: Pittsfield, Massachusetts Follow-up: 16-20 years later Sample: 123 students who at age 2 (1976-80) had been recruited for the Parent-Child Home Program. The control group is very small (15).

Healthy Families Beneficiaries:
New or expectant parents deemed to be at risk of abusing or neglecting their children Objectives: (a) promote positive parenting skills and parent-child interaction; (b) prevent child abuse and neglect; (c) support optimal prenatal care, and child health and development; and (d) improve parents' self-sufficiency Description: Paraprofessionals make home visits starting in pregnancy and continuing until the child reaches age 5. Cost: n/a Authors: Dumont and others (2006)

Parenting Practices Canada
Beneficiaries: Children at risk for mental health problems attributable to poverty and/or their parents' inexperience, lack of educational attainment, and young age Objectives: Improve the parent-child relationship and indirectly enhance the resilience capacity among at-risk children. Description: Two pilot studies were conducted to compare the effects of the following two types of social support for families on parent-child relationships: (a) support with extensive and intense parent training targeted toward adolescent parents in the newborn period Authors: Letourneau and others (2001) Type: RCT Location: Canada Follow-up: 11-13 weeks later Sample: 52 families divided into two groups: (a) 18 families received support with extensive and intense parent training, and (b) 34 families received support without extensive and intense parent training.

Term used: Social skills
Behavior related: Improved psychophysiological arousal and orienting at age 11. This is seen through skin conductance amplitudes (+)***; skin conductance rise times (+)***; skin conductance recovery times (+)**; slow-wave electroencephalogram measure during rest (−)***; electroencephalogram measure during continuous performance task conditions (−)*** (mothers received six weekly home visits delivered by a master's-prepared pediatric nurse); and (b) support without extensive and intense parent training targeted toward other at-risk parents (for example, low-income, low educational attainment, and single-parent status) of 3-to 4-year-old children identified as having developmental delays. Cost: n/a

Durham Connects
Beneficiaries: All resident births in Durham, North Carolina, between July 1, 2009, and December 31, 2010 Objective: Improve infant health and well-being. Description: A four-to seven-session program assesses family needs (parenting and child care, family violence and safety, parental mental health and well-being, and health care) and connects parents with community resources.  Beneficiaries: Children ages 0-6 living in vulnerable villages in Indonesia objective: Improve poor children's overall development and readiness for further education within a sustainable quality early childhood education and development system. Description: The project had three main components: 1. Provision on integrated early childhood education and development services. These services were delivered through community-driven mechanisms in targeted poor communities. The program helped communities identify early childhood education and development service needs, prepare a proposal, and then implement this proposal. 2. Ideation of a sustainable system to ensure quality in early childhood education and development services. This includes the generation and implementation of standards and quality-assurance systems and the institutionalization of early childhood education and development at the district and provincial levels.  Description: Program provides full day (10 hours/day) child care, nutrition, and transportation year-round, from infancy to kindergarten (approximately 250 days per year). Teacher-child ratios ranged from 1:3 for infants and toddlers to 1:6 for older children. Cost: US$67,000 for the five years of attendance ($11,000 in year 1; US$16,000 in each of years 2 and 3; US$12,000 in each of years 4 and 5) Sample: 112 children, mostly African American, born between 1972 and 1977, who were believed to be at risk of retarded intellectual and social development (51 in control group and 53 in treatment); 104 children in follow-up for 20 years Terms used: Social development; social adjustment; noncognitive skills Academic/cognitive related: IQ until age 1 (+)** then faded; grade repetition (−)**; assigned to special education (−)**; high school graduation by age 19 (+)**; college attendance (+) **; educational attainment at age 40 (+)*** Economic related: Earnings (+)***; homeownership (+)** Note: Annual social rates of return: 7%-16% (US$9-US$13 for every US$1 invested).

Montessori
Beneficiaries: 3-to 5-year-old children Objective: Foster a child's natural inclination to learn. Description: Montessori schooling is a 100-year-old system that naturally incorporates practices that align with mindfulness and are suited to very young children. Montessori education is characterized by multiage classrooms, a special set of educational materials, long periods spent in student-chosen activities, collaboration, the absence of grades and tests, and individual and small-group instruction in both academic and social skills.

Early Child Development Program
Beneficiaries: Children born in three regions (13 provinces) of the Philippines Objective: Improve survival and developmental potential of children, particularly the most vulnerable and disadvantaged. Description: Children participated in interdepartmental programs on nutrition, health, early education, and social services programs with the following components: (a) minimizing the health risks to very young children, (b) contributing to the knowledge of parents and the community about child development and encouraging their active involvement, (c) advocating for child-friendly policies and legislation, (d) improving the ability and attitude of child-related service providers, and (e) mobilizing resources and establishing viable financing mechanisms for ECD projects. Cost: n/a Authors: Armecin and others (2006) Type: Intent to treat difference in propensity score matching Location: The Philippines Follow-up: 3 years Sample: 6,693 children ages 0 to 4 years Term used: Social-emotional skills Health related: Weight for height (+)**; proportion wasted (−)**; anemia (+)**; hemoglobin (+)** Behavior related: Social-emotional skills (+)**; self-help (+)**; social development (+)** Academic/cognitive related: Cognitive development (+)**; expressive language (+)**; gross motor skills (+)**; fine motor skills (+)**; receptive language (+)**

Proyecto Integral de Desarrollo Infantil (PIDI)
Beneficiaries: Poor children ages 6 to 72 months in urban areas Objectives: Improve child nutrition and provide environments that are conducive to learning. Description: Day care, nutrition, and educational services are provided to children who live in poor, predominantly urban areas. Under the program, children are cared for in groups of 15 in homes in their own neighborhood. The community selects local women to become home day care mothers. These nonformal, home-based day care centers, with two to three caregivers, provide integrated child development services (play, nutrition, growth screening, and health referrals). The women receive child development training before becoming educators but are usually not highly trained. Cost: US$516 yearly per beneficiary Authors: Behrman, Cheng, and Todd (2004) Type: Propensity score matching Location: Bolivia Follow-up: n/a. They compare children in the program for short (<2 months) and longer durations Sample: 1,227 children for round 1; 2,420 children for round 2; and 364 children for both rounds as controls.

Term used: Psychosocial development
Health related: Height for children up to 36 months (+); weight for children younger than 36 months (−) Behavior related: Psychosocial skills for those ages 37 to 54 months (+)* Academic/cognitive related: Bulk motor skills, fine motor skills, language skills for those ages 37 to 54 months (+)* Economic related: Earnings (+); earning benefits/cost: 1.7-3.7 Beneficiaries: 4-to 6-year-old children from deprived backgrounds. These children were at risk for low educational achievement and failure to develop to their full potential because of poverty and their families' low educational level. Objective: Improve overall child development. Description: Under TEEP, children were sorted into one of three environments: (a) an educational day care center (educational group), (b) a custodial day care center (custodial group), or (c) a home (home group). Randomization of children within each environment happened only if there was excess demand for that particular service. Mothers of approximately half of the children in each care environment were randomly assigned to receive maternal training emphasizing educational activities with the child plus support for the mother through group meetings and guided discussions. Cost: n/a Authors: Kagitcibasi, Sunar, and Bekman (2001); Kagitcibasi and others (2009) Location: Istanbul, Turkey Follow-up: 22 years later. Intervention carried out in 1983-85. Sample: 255 participants divided into the following groups: (a) 27 received mother training and educational day care, (b) 40 received mother training and custodial day care, (c) 23 received mother training in a home care environment, (d) 37 did not receive mother training while in educational day care, (e) 65 did not receive mother training while attending a custodial day care center, and (f) 63 did not receive mother training while in a home care environment. The 22-year follow-up gathered information on 131 of the 255 participants (39% attrition).

Term used: Personality and social development
Behavior related: For children with trained mothers, immediate postprogram positive self-concept (+)*; aggression (−)*; school adjustment for children with mother training (+)**; emotional problems for children in the home environment immediately postintervention (+)***; child self-confidence six years later for mother-trained children (+)**; child self-confidence six years later for children in the custodial group (−)** Academic/cognitive related: IQ scores for children with mother training postintervention (+)***; IQ scores for children who were in educational or home environment postintervention (+)*; achievement test scores postintervention (+)**; school grades for the children in the educational environment at age 3 (+)**; school grades for the children with mother training at age 5 (+)* [no effect of mother training in grades after primary school]; school enrollment seven years after the program for those with mother training (+)***; failed years in middle school for the custodial group (+)**; college attendance for those with mother training (+)*

Economic related:
Ownership of a credit card for those with mother training (+)**; occupational status of those who attended an educational day care center (+)**; ownership of a computer for those who attended an educational day care center (+)* Other: Indicators of integration into modern urban life, such as owning a computer. For children with trained mothers, immediately postprogram: mother's attentiveness to and direct interaction with the child (+); involvement with children in cognitive-oriented activities (+); educational aspirations and expectations for the child (+); positive disciplinary strategies and praise (+)

Early Childhood Parenting Programme in Rural Bangladesh
Beneficiaries: Poor children under age 3 and their mothers Objective: Promote physical and mental development of children Description: In this program, 90-minute weekly education sessions are offered by trained women, known as facilitators, to groups of about 20 mothers. The topics include common diseases and oral rehydration solutions, hygiene, sanitation, breastfeeding, weaning foods, micronutrient deficiencies, stages of cognitive and language development, how parents can help children learn, how to encourage language development, positive discipline, gender equality, and child rights. The facilitators had some secondary education; to deliver the program, they received 17 additional days of basic training with a manual of 40 topics, four days a month of supervision, and monthly refresher courses. preventive health (+)*** Behavior related: Knowledge of mother (+)***; amount and quality of stimulation and support provided to a child in the family setting-home inventory (+)***; stimulation (+)*** Note: The parenting mothers did not communicate differently with their children while doing a picture-talking task, and children did not show benefits in nutritional status or language comprehension. Behavior related: Socioemotional indicator (ASQ) (+)**; caregivers' belief that physical punishment is appropriate (−); daily routines between parents and children (+)***; self-sufficiency activities between parents and children (+)** Academic/cognitive related: Primary school enrollment rates (+)***; enrollment at appropriate age in primary school (+)**; hours spent on schooling and homeworkrelated activities (+)***; cognitive and problem-solving abilities (+)**; fine motor skills (+)*; school attendance of older siblings of beneficiary children (+)*** Economic related: Caregivers working in the 30 days before the interview (+)* Other: Caregivers report a significant increase in satisfaction with their children's preparation for future school.
Note: Some of the principal measures of communication and language development are not significantly different between groups and remain alarmingly low. No differences are noted in rates of stunting and wasting, which was expected since a child's growth potential is largely determined by age 3. The study finds mixed impacts in children's health (less diarrhea and fewer skin problems, more illnesses). teacher-taught social-emotional curriculum. Thirty lessons that were delivered once a week included the topics of compliments, basic and advanced feelings, a self-control strategy, and problem solving. In addition to the lessons, teachers generalized the concepts of the curriculum through detailed extension activities (such as group games, art projects, and books) that were integrated into the existing typical preschool programs (such as language arts, music, and art).  school-based preventive intervention in literacy development, conflict resolution, and intergroup understanding that trains and supports teachers from kindergarten through grade 5 on how to integrate the teaching of social and emotional skills into the language arts curriculum and to achieve positive rules and norms, as well as safe and secure environments. The program has three primary components: (a) a comprehensive seven-unit, 21-35 lesson, literacy-based curriculum in conflict resolution and socialemotional learning (provided to teachers in a standardized, grade-specific teaching guide); (b) 25 hours of training followed by ongoing coaching of teachers to support them in teaching the 4Rs curriculum with a minimum of 12 contacts in one school year; and (c) a parent component (4Rs Family Connections) that consists of activities for children to do with their parents at home.       victimization at school (−)*; misconduct at school (−)*; involvement in positive youth activities (+)***; gets along well with others, is sought out by his/her fellow students, and has many friends (+)*  In high implementation subsample: sense of efficacy (+)**; delinquent behaviors (−)**; friends' drug use (+)*; friends' delinquent behaviors (+)**; friends' positive involvement in school (+)**; gets along well with others, is sought out by his/her fellow students, and has many friends (+)**; does not hesitate to state opinions, even when others disagree with his/her views (+)*; is engaged in class-participates in discussions, stays on the topic, and generally takes an active part in whatever the class is doing (+)**; appears to be socially awkward and inept-tends to say the "wrong thing" and to be rebuffed or ridiculed by other students (+)**; considers others' feelings, treats them with respect, and offers and gives help to those who need it (+)* Academic/cognitive related: Sense of efficacy (+)***.
In high-implementation subsample: trust in and respect for teachers (+)**; positive teacher-student relations (+)***; educational aspirations (+)***; educational expectations (+)*; GPA (+)***; achievement test scores (+)**; comes to class and completes assignments on time, tries to learn the material, and does the best work he/she can (+)*; task orientation toward learning (+)* Other: Sense of school (+)*; positive teacher-student relations (+)**; liking for school (+)**; attendance at religious services (+)* Note: No significant differences between control and treatment in academic outcomes, attendance, or suspension.   Boyle and Hassett-Walker (2008) Type: RCT Location: United States Follow-up: 2 years Sample: 226 students were randomly assigned in three groups: (a) 96 students received instruction for two consecutive years; (b) 106 students received only one year of program instruction, either in kindergarten or first grade; and (c) 24 students were assigned to the control group for two consecutive years.

Term used: Prosocial behaviors
Behavior related: Preschool Social Behavior Scale for those with two years of instruction (+)***; Hahnemann behavior rating scale for those with one or two years of instruction (+)***; rational aggression for those with two years of instruction (−)**; overt aggression of those with two years of instruction (−)****; prosocial behavior for those with two years of instruction (+)***; passivity for those receiving two years of instruction (+)**; prosocial behavior for those who received either one or two years of instruction (+)*** Note: There is a modest additive effect of the program, with students who received two years of instruction becoming more prosocial and less aggressive than students who received one year of instruction.

Project SAFE (Strategies Aimed at Family Empowerment)
Beneficiaries: Students from kindergarten to grade 5 Objectives: Prevent substance abuse. In particular, the I Can Problem Solve program (ICPS) is directed at enhancing problem-solving and critical thinking skills. The Strengthening Families (SF) program was originally designed for children of drug abusers.
Authors: Kumpfer and others (2002) Type: RCT Location: United States Follow-up: 9 months Sample: 655 first-grade students from 12 rural schools randomly divided into the following groups: (a) 256 received the Behavior related: Parenting skills for the group receiving ICPS+SF1 (+)***; family relationships for ICPS+SF1 (+)**; social competency for ICPS+SF1 (+)** and ICPS+SF2 (+)***; behavioral self-regulation for all groups (+)*** Academic/cognitive related: School bonding for all programs except for ICPS+SF2 (+)*** Head Starts and 14 elementary schools were involved in the project. These schools were matched on variables such as size, geographic location, and demographics of the children. The sample consisted of 1,746 kindergarten or first-grade students who were nested in 160 classrooms, which in turn were nested under 119 teachers (40 teachers had 2 or more classrooms). Classrooms were in culturally diverse schools with high rates of poverty. Term used: Social emotional and self-control skills Behavior related: School readiness and conduct problems (−)**; social competence; child disengagement (−)**; number of positive feelings identified by children (+)***; emotional selfregulation; conduct problems (−)** Academic/cognitive related: Effective discipline by Head Start teachers (+)**; positive classroom management strategies Other: Involvement with parents (+)** Note: Overall, children who were initially most at risk benefited most from the intervention. Furthermore, all of the student behavioral outcomes showed strong teacher-level effects (the teacher played an important role in the effects).  and nutrition and physical activity. In fifth grade, 28 lessons were taught across the same health topics. The intervention was implemented in classrooms over 12 weeks in grade 4 and 14 weeks in grade 5 during a normal class period of 40-50 minutes by the classroom or health teacher, who received a 12-hour curriculum training course with follow-up support provided as needed.    Note: Most intervention effects were moderated by school environment, with effects stronger in less disadvantaged schools. Effects on aggression were greater in students who showed higher baseline levels of aggression.      The family intervention strategies offered during the elementary school years consisted of universal and selected components. Multiple-session parenting workshops (for example, "Raising Healthy Children, " "How to Help Your Child Succeed in School, " and "Preparing for the Drug Free Years") were offered to all parents at intervention schools. For families of students identified as high risk because of academic or behavioral problems, in-home services tailored to the specific risk factors of these children were provided to reinforce curricula covered in the parenting workshops. Cost: n/a

Resolving Conflict Creatively
Beneficiaries: Children in grades 1 through 6 Objective: Change the mental processes and interpersonal behavioral strategies that lead children to engage in aggression and violence by teaching them constructive conflict resolution strategies and promoting positive intergroup relations Description: The universal, school-based intervention involves violence prevention and intergroup understanding. The intervention has two major components: (a) training and coaching of teachers to support them in implementing a curriculum in conflict resolution and intergroup understanding (teacher training and coaching), and (b) the delivery of that curriculum through classroom instruction for children provided by the trained teachers (classroom instruction). Additional features of the program include peer mediation, principals' training, and parent training.      Behavior related: Student self-efficacy (+)***; socioemotional competencies (+)**; attention and concentration (+)***; social and emotional competence (+)***; authority and compliance problems (−)***; aggression (+)*** Academic/cognitive related: Problem-solving skills (+)**; math grades (+)** Note: No significant effect for classroom climate, standardized reading, or math test.

School-based programs in Colombia
Beneficiaries: First-and second-grade students in public schools Objective: Reduce aggressive and antisocial behavior by first-and second-grade students in a resource-poor setting. Description: The study tested two school-based programs. The first was a teacher-only intervention that consisted of teacher training that focused on standard classroom management techniques (arranging the physical environment to reduce opportunities for conflict, establishing and consistently enforcing clear rules, and instituting routines and procedures) and strategies for shaping children's behaviors (modeling and reinforcing appropriate behavior consistently in the classroom's daily activities and interactions). Teachers received 10 four-hour weekly workshops, a manual, and weekly written and verbal feedback. The second program was a teacher/parent combined intervention that included the same teacher training as well as parenting intervention that covered almost the same topics as the teacher training (the session on redesigning the classroom was replaced by one of prosocial behavior). This session lasted one hour, once a week for 10 weeks. Cost: n/a Authors: Klevens and others (2009) Type: RCT with pre-and postassessments Location: Pereira, Colombia Follow-up: 6 months after the program starts (just after it finishes) Sample: 2,491 children in 12 public schools: (a) 5 schools with 40 first-and secondgrade classrooms and 881 students participated in the teacher-only intervention group, (b) 4 schools with 39 first-and second-grade classrooms and 787 students were assigned to the teacher/parent combined intervention group, and (c) 3 schools with 41 classrooms and 823 students were assigned to a wait-list control group. Term used: Antisocial/prosocial behavior Behavior related: Aggressive behavior (−)***; antisocial behavior for the teacher-only intervention (−)***  Behavior related: Attitude toward forgiving a bully (+)**; attitude toward forgiving a friend who returns a book he/she had stolen (+)**; not fighting a friend who lies about one (+)**; self-reporting verbally abusing others (−)** Note: Statistically significant behavioral effects were not detected, although a downward trend was seen in the intervention group.

Linking the Interests of Families and Teachers (LIFT)
Beneficiaries: All first-and fifth-grade elementary school boys and girls and their families living in at-risk neighborhoods characterized by high rates of juvenile delinquency Objective: Affect a set of interrelated antecedents of the conduct disorders of children in high-risk neighborhoods. Description: Universal preventive intervention consists of (a) parent training aimed at teaching parents how to create a home environment that is most conducive to the ongoing practice of good discipline and supervision (parents met in groups of 10 to 15 families once each week for six weeks, free child care was provided, and a prize was raffled at the end of every session); (b) a classroom-based social skills program consisting of 20 one-hour sessions that included classroom instruction and discussion on specific social and problem-solving skills, skill practice in small and large groups, free play in the context of a group cooperation game, and review and presentation of daily rewards; (c) a playground behavioral program; and (d) systematic communication between teachers and parents. Cost: n/a Authors: Reid and others (1999)

Becoming a Man
Beneficiaries: Children attending grades 9 and 10, referred by school staff Objective: Provide both academic and nonacademic remediation for disadvantaged youths who are falling behind and at great risk of dropping out.

Description: The socioemotional intervention called
Becoming a Man includes school programming that exposes youths to prosocial adults and provides youths with social-cognitive skills training using principles of cognitive behavioral therapy. Youths have the chance to participate in up to 27 one-hour, weekly group sessions during the school day over the school year (the program lasts for three-quarters of an academic year). The intervention is delivered in groups of fewer than 15 youths. Students skip an academic class in order to participate in the program. The program can be delivered by college-educated people without specialized training in psychology or social work. The curriculum also includes efforts to develop specific social or social cognitive skills, such as generating new solutions to problems, learning new ways of behaving, considering another's perspective, thinking ahead, and evaluating consequences ahead of time.

Authors: Cook and others (2014)
Type: RCT Location: United States: Chicago, Illinois Follow-up: 1 year Sample: 106 disadvantaged males in grades 9 and 10 in public high schools (95% are black and 99% are eligible for free or reduced-price lunch) divided into three groups: (a) 24 children received just the nonacademic intervention, (b) 48 received both the nonacademic intervention and the academic intervention, (c) 34 received status quo services. Term used: Social-cognitive skills Academic/cognitive related: Math test cores (+)**; math grades (+)*; nonmath courses failed (−)**; expected graduation rates (+)**; absence without an excuse (−)* Note: The confidence levels are too wide to say anything about which of the two intervention arms is more effective. Additionally, there was contamination among the different groups and spillover.  The academic intervention is delivered by intensive, individualized two-on-one math tutoring provided for one hour per day every day by well-educated, committed people without formal teacher training. Cost: Approximately US$4,400 per participant (ranging from US$3,000 to US$6,000)

Social and Emotional Training (SET)
Beneficiaries: Children in grades 1 through 9 in Swedish schools Objectives: Train students to improve self-control, social competence, empathy, motivation, and self-awareness. Description: SET is a classroom-based intervention that focuses on developing the following five functions in students: self-awareness, emotional management, empathy, motivation, and social competence. The program was delivered by regular classroom teachers during scheduled hours: to grades 1 through 5 twice a week in two 45-minute sessions, and to grades 6 through 9 in one 45-minute session per week over the school year. The program is guided by detailed manuals for the teacher, one volume for each grade. It also includes a student workbook for each grade. Altogether, the program consists of 399 concrete exercises, some of which are inspired by similar programs in the United States. Teachers are instructed and supervised monthly.
Interaction between school and parents is emphasized.

Learn and Serve America
Beneficiaries: Students in grades 6 through 12 Objectives: Help young people (a) develop as responsible citizens, improve their academic skills, and develop as individuals, while helping to meet the unmet needs in the public safety nets of the United States; (b) promote the integration of service learning in schools and academic curriculum; and (c) promote the delivery of needed services in the community. Description: The program helps young people develop as responsible citizens, improve their academic skills, and develop as individuals through involvement in meaningful service linked to structured learning activities. Costs: Approximately US$149.12 per participant (1995-96 dollars) Author: Melchior (1998) Type: Quasi-experimental Location: United States Follow-up: 1 year Sample: 1,052 participants in grades 6 through 12 attending 17 middle school and high school sites across the country Term used: Life skills Health related: Ever pregnant or made someone pregnant (−)* Behavior related: Cultural diversity (+)**; service leadership (degree to which students feel they are aware of the needs in a community, are able to develop and implement a service project, and are committed to service now and later in life) (+)***; total civic attitudes (+)***; volunteered for a community organization or got involved in other community service in the last six months (+)***; hours or support to community service initiatives in the last six months (+)*** Academic/cognitive related: Engagement in school (+)***; math grades (+)**; science grades (+)*; core GPA (+)* Note: Results are greater for specific subgroups. For example, the impact was statistically significant for the educationally disadvantaged students on English, math, and science grades; overall and core GPA; and a number of at-risk behaviors. For the economically disadvantaged students, impacts were significant for school engagement, days absent, and aggressive behaviors (fighting, hurting someone, or using weapons). For females, the program had statistically significant effects on the use of illegal drugs in the past 30 days, aggressive behavior, and the number of at-risk behaviors. Further, the program had effects on English, math, and social studies grades; GPA; and course failures of minority students.  Beneficiaries: Girls from grades 6, 7, and 8 in the poorest districts of Rajasthan, India Objective: Create soft skills in girls and their communities. Description: The program consists of democratic elections (by both boys and girls) of 13 girls from grades 6, 7, and 8 to meet after school several times per month to participate in life skills games that stimulate growth and development and build confidence, leadership, and self-esteem. They practice setting goals for their school or community and are told to pass on the life skills games to other students. Cost: n/a Authors: Delavallade, Griffith, and Thornton (2015) Type: RCT Location: Rajasthan, India Follow-up: 1 year Sample: 30 schools randomly assigned as follows: (a) 10 received the standard program, (b) 10 received a modified program in which girls were randomly selected rather than elected, and (c) 10 were controls and did not receive the program Terms used: Soft skills, life skills Behavior related: Expectations of democratically elected girls about age at marriage (+); expectations about career prospects (−); educational expectations and aspirations for nonelected girls in democratically elected program (−); partition of friendship groups between those who are randomly selected and those who are not (+) Note: Main effects on the standard program where girls are elected.

Strengthening Families Program (ISFP) and Preparing for the Drug Free Years (PDFY) program
Beneficiaries: Sixth graders Objective: Strengthen parent and child skills that have been found to delay the onset or reduce the use of substances. Description: The ISFP consists of seven sets of sessions that are conducted once per week for seven consecutive weeks and are held on weekday evenings in participating schools. The ISFP includes separate, concurrent one-hour training sessions for parents and children, followed by a joint one-hour family session. The PDFY program is delivered in five two-hour training sessions held on weekday evenings once per week for five consecutive weeks (four sessions are for parents only, and another session on peer-resistance skills is for parents and children). Cost: n/a Authors: Spoth, Randall, and Shin (2004) Type: RCT Location: United States: Iowa Follow-up: 6 years Sample: 667 families of sixth-grade students enrolled in 33 rural public schools were randomly assigned to the ISFP (11 schools and 238 families), the PDFY program (11 schools and 221 families), or a minimalcontact control condition (11 schools or 208 families).

Chicago Child-Parent Center Program
Beneficiaries: Disadvantaged, African American, inner-city, 3-and 4-year-old children Objectives: Promote academic success by providing a school-stable learning environment during the preschool and primary grade years and actively involve parents in their children's education. Description: Comprehensive education, family, and health services, including half-day preschool at ages 3 to 4, half-or full-day kindergarten, and school-age services in linked elementary schools at ages 6 to 9. Child-to-teacher ratios of 17:2 in preschool and 25:2 in kindergarten. Parental participation includes interaction with other parents in workshops, reading groups, crafts projects, classroom volunteering, school events, and field trips, and completing high school. The program also includes health and nutrition services, such as screening, speech therapy, and nursing Cost: Mean per-child expenditures in 1996 for one year of preschool and one year of school-age participation are US$4,350 and US$1,500, respectively.

Tools of the Mind
Beneficiaries: 3-and 4-year-old children attending a state-financed, full-day "Abbott" preschool education program located in high-poverty school districts Objective: Build strong foundations for school success by promoting intentional and selfregulated learning in preschool-and kindergarten-age children. Description: A child care program with a curriculum is focused on self-regulation, while teaching literacy and math. During pretend play, children must inhibit acting out of character, remember their own and others' roles, and flexibly adjust as their friends improvise. Cost: According to What Works Clearinghouse, the first year costs about US$3,000 per classroom. Curriculum guides cost an additional US$100.
The second year costs about US$1,500 per classroom.
Authors: Barnett and others (2008) Type: RCT Location: United States: New Jersey, One school with a high proportion of children from low-income and non-Englishspeaking families

Peace Works
Beneficiaries: Preschool children ages 4 to 5 years Objectives: Promote preschoolers' social skills and reduce behavior problems Description: The program instructs teachers and parents in how to use activities and ''I-Care Rules and Language'' to encourage empathy and fair play, express feelings, avoid conflict, manage anger, and interact more positively with others. The curriculum models conflict resolution, anger management, and communication skills to promote children's positive interactions, cooperation, and emotional self-regulation.

Work related
Procajoven Beneficiaries: Low-income unemployed 18-to 29-year-olds, and first-time job seekers with complete secondary education Objective: Improve prospects for jobless youths and disadvantaged groups. Description: The program has two modalities. The first, called the insertion modality, provides short-term training for low-income unemployed 18-to 29-year-olds. Classroom training has two parts, job readiness skills and technical training (120 and 150 hours, respectively), followed by 172 hours of internship in a firm. The second modality, called the transition modality, focuses on the transition of the first-time job seekers with complete secondary education, providing job readiness and a longer internship (344 hours). Cost: The insertion modality costs US$611. The transition modality costs US$375. Both cost calculations include a transfer to participants of US$255.

Galpão Aplauso
Beneficiaries: Disadvantaged favela youths (a) younger than age 29, (b) living in households with total monthly income below the minimum wage, and (c) not involved in drugs or gang activity Objectives: Improve the socioeconomic situation and employability of youths by teaching them basic skills, life skills, and vocational skills and provide placement services. Description: Galpão Aplauso (or the Sociocultural and Productive Integration of At-Risk Youth Project) is an innovative labor training program that uses expressive arts and theater as a pedagogical tool. The program consists of a combination of vocational, academic, and life skills training, delivered through a pedagogic method that uses arts and dance. Program duration is approximately six months, five hours a day, five days a week, delivered in three shifts-morning, afternoon, and evening. It comprises 300 hours of vocational training (mainly construction related, soldering, and woodshop); 180 hours of training in academic and basic skills, including remedial courses in both mathematics and Portuguese; and 120 hours in life skills (social harmony and socioemotional development). Cost: The average cost per youth is R$810 (US$385) a month, or R$4,680 (US$2,225) for the entire curriculum (2014 dollars).

Authors: Calero and others (2014)
Type: RCT Location: Rio de Janeiro, Brazil Follow-up: 1-2 months for the first cohort; 4-5 months for the second cohort; and 2-3 months for the third cohort. Sample: 380 individuals: 162 in the treatment group, 195 in the control group, and 23 nonrandomly assigned treated youths (preselected because of high vulnerability and high achievement, high need, or because participating sibling enrolled). Attrition was 77%. Term used: Life skills/emotional skills Economic related: Employment during final week after four to five months (+)**; employment in salaried job during final week (+)***; monthly labor income (+)*; saving as main expenditure (+)** Note: No impact on personality-related traits (measured through Grit and the social and personal competencies scales, which capture leadership, behavior in situations of conflict, self-esteem, abilities to relate to others, order, and empathy and communication skills). No impacts on social activities or risky behavior.  training in life skills, work experience through internships, and job placement support. Cost: n/a treatment 1, which consisted of ICT and life skills, along with on-the-job experience through internships and job placement support; (b) 350 received treatment 2, which included everything in treatment 1 except for the life skills training; and (c) 810 acted as a control group and received none of the program. Term used: Life skills training Note: Although both treatment groups faced statistically significant impacts on knowledge of ICT and weekly income, treatment 1 had greater impacts with regard to ICT knowledge and smaller impacts on weekly income when compared with treatment 2.

Jordan New Opportunities for Women (Jordan NOW)
Beneficiaries: Women graduating from eight public community colleges Objective: Assist female community college graduates in finding employment. Description: The project tested two programs: (a) a subsidy paid to employers equivalent to the minimum wage for up to six months if they hired the graduate and (b) 45 hours of employability skills training. The training course was 45 hours over nine days (5 hours per day), with a maximum of 30 participants in each training group. The course covered effective communication and business writing skills (such as making presentations and writing business reports and different types of correspondence), team-building and teamwork skills (such as characteristics of a successful team, how to work in different roles within a team), time management, positive thinking and how to use it in business situations, excellence in providing customer service, and résumé and interviewing skills. Cost: The cost of the voucher was JD 150 each month, for six months. The cost of the soft skills training was US$400 per assigned graduate. Most participants reside at a center while training and receive meals and health and dental care and can participate in student government and recreation activities. Job Corps provides placement services to help participants find jobs or pursue additional training. Cost: About US$16,500 per participant (1995 dollars)

Empowerment and Livelihood for Adolescents (ELA) program
Beneficiaries: Adolescent girls ages 14 to 20 Objective: Empower Ugandan women along economic and reproductive dimensions. Description: The program provides (a) vocational skills to enable adolescent girls to start small-scale, income-generating activities, including hairdressing, tailoring, computing, agriculture, poultry rearing, and small trade operations; and (b) life skills, including sexual and reproductive health, menstruation and menstrual disorders, pregnancy, sexually transmitted infections, HIV/AIDS awareness, family planning, rape, management skills, negotiation and conflict resolution, leadership among adolescents, bride price, child marriage, and violence against women. In addition, the clubs also host popular recreational activities, such as reading, staging dramas, singing, dancing, and playing games. The intervention is delivered from designated "adolescent development clubs" (a fixed meeting place in each community) rather than in schools.