Skip over navigation

Journal Issue: America's High Schools Volume 19 Number 1 Spring 2009

Finishing High School: Alternative Pathways and Dropout Recovery
John H. Tyler Magnus Lofstrom

Dropout Prevention

The high costs associated with dropping out make clear the need for programs to help students stay in school. The Dropout Prevention Center/Network lists hundreds of dropout-prevention programs in its online database of "model programs."61 Only relatively few of these programs, however, have been rigorously evaluated for effectiveness. Even fewer have proved effective in achieving this goal. As Mark Dynarski and Philip Gleason write in a report on dropout-prevention programs, "Dropping out is as hard to prevent as it is easy to do."62 Based on the evidence, one might add that it is equally hard to identify confidently the programs that are effective.

In what follows, we group dropout-prevention interventions into two categories. The first is interventions that set dropout prevention as the primary goal and that target specific students or groups of students. The second is interventions that have a broader goal than dropout prevention and a broader target audience than "at-risk" students, but that, nevertheless, aim to lower dropout rates. The first category embraces programs in the regular school or in the community, alternative schools for at-risk students, and smaller learning communities that tend to fit the "school-within-a-school" model and that target at-risk students. The second, broader category includes school restructuring or school reform models. Broadly stated, programs in both categories aim to lower dropout rates through one or more of four mechanisms: increasing school attendance, increasing student school engagement and learning, building student self-esteem, and helping students cope with the challenges and problems that contribute to the likelihood of dropping out.

To date, relatively few evaluations of dropout-prevention interventions could be considered rigorous. One of the largest rigorously conducted evaluations was a late 1990s study of twenty-one different interventions, each funded by the U.S. Department of Education's School Dropout Demonstration Assistance Program (SDDAP). In addition to the SDDAP evaluations, a second source of evidence on the efficacy of dropout-prevention interventions can be found in the Department of Education's What Works Clearinghouse (WWC), which reviews and synthesizes studies of a wide variety of education interventions. The combined findings of the SDDAP evaluation and the WWC synthesis of dropout-prevention programs leave one less than sanguine about the knowledge base about how to lower dropout rates.

The SDDAP evaluation, conducted by Mathematica Policy Research, Inc., included both targeted and broadly defined dropout-prevention efforts. Targeted interventions were usually evaluated through randomized, controlled experiments, while the evaluations of the school-restructuring efforts were quasi-experimental and used observationally similar schools as the comparison group for SDDAP schools. The evaluation looked at sixteen targeted interventions and five school-restructuring projects. Eight of the interventions took place at the middle school level. Two of the targeted interventions at the high school level were community-based programs aimed at helping students who had already left school acquire a GED.

The key finding from the SDDAP evaluations is that "most programs made almost no difference in preventing dropping out in general."63 Some SDDAP programs did make a difference on some outcomes, and we will take a closer look at one of the more successful programs. One of the more consistent positive findings in the SDDAP evaluations, however, involves programs to increase GED acquisition among students who have already left school. Although increasing the GED attainment rate of school dropouts may be a laudable outcome, it seems less clear that it should be considered as successful dropout prevention.

The picture is hardly any brighter when it comes to findings of the What Works Clearinghouse. To date, the first-wave WWC review of dropout-prevention programs has looked at fifty-nine studies of sixteen programs.64 From this group, ten of the programs had undergone evaluations that were rigorous enough to make it possible to reach firm conclusions about program effectiveness.65 These ten programs include a wide range of interventions: counseling and monitoring, school restructuring and curriculum redesign, financial incentives for students and families, and community services designed to mitigate factors that can negatively affect school achievement and success.66

Of the ten programs, five showed promise in reducing dropout rates.67 Two of the five —Achievement for Latinos through Academic Success (ALAS) and High School Redirection— are no longer active. ALAS, a pilot program launched in San Diego during the early 1990s, was designed to address student, school, family, and community factors that affect dropping out. At the end of the ninth grade, 98 percent of the students who were randomly assigned to the ALAS program were still enrolled, compared with 83 percent of the students in the non-ALAS control group.68 Meanwhile, three years after random assignment, 43 percent of the students assigned to the High School Redirection program—an alternative high school program for students considered at risk—had dropped out, compared with 53 percent of the randomly assigned control group.69

The three remaining positive programs represent three distinct approaches to dropout prevention. One, Check & Connect, is a relatively intensive program for (mostly) high school students; a second, Career Academies, fits the school-within-a-school model; a third, Talent Development High Schools, is best described as whole-school reform. We discuss each in turn.

A Dropout-Prevention Program: Check & Connect
The Check & Connect70 model, developed through a partnership between the University of Minnesota, local public schools, and local community service organizations, was originally funded by the Department of Education. The Check & Connect model "was initially developed for urban middle school students with learning and behavioral challenges and was designed to promote students' engagement with school and learning, and to reduce and prevent dropping out. The model is currently being replicated and field-tested for youth with and without disabilities in grades K–12 in urban and suburban communities."71 Broadly speaking, Check & Connect works with and coordinates services among the student, family, school, and community to help the student succeed and stay in school.

The signature feature of Check & Connect is the assignment of a "monitor" to each student in the program to be the student's mentor and case worker. In the Check component, the monitor continually assesses the student's school performance, including attendance, behavior, and academics. Monitors are trained to follow up quickly at the first sign that a student is struggling in any of these areas. The Connect component combines individualized attention to the student with the coordination of services and information about the student across school personnel, family, and community service providers. The program carries a minimum two-year commitment to students and families, including the promise and ability to follow highly mobile youth from school to school so that students do not lose services when they move from their original program site.

In two separate experimental evaluations, Check & Connect showed positive effects on staying in school and progressing through school. One study showed that ninth-grade students enrolled in Check & Connect were substantially less likely than control group members to have dropped out of school by the end of the year—9 percent compared with 30 percent. Another study showed that by the expected graduation year, 39 percent of students in the Check & Connect treatment group had dropped out of school compared with 58 percent of the control group. The high dropout rate associated with both groups indicates the level of dropout risk present in the population targeted by Check & Connect. The cost of implementing the Check & Connect model was about $1,400 per student during the 2001–02 school year.72

The School-within-a-School Model: Career Academies
Career academies are another intervention that rigorous evidence shows effective in lowering dropout rates, at least for students most at risk of dropping out.73 The career academy model has three key features. First, it is organized as a school-within-a-school: students in a smaller and more personal learning atmosphere stay with the same teachers over the three or four years of high school. Second, it includes both academic and vocational coursework, with the two integrated in the curriculum and in pedagogy. And, third, it uses partnerships between the academy and local employers to build links between school and work and to provide students with career and work-based learning opportunities.

Begun in the 1970s, the career academy model has both evolved in concept and grown in numbers over time. Today some 1,500 career academies nationwide serve a much wider set of students than the "vocational ed" students who were seen as the original constituents of the academies.

The most important study of career academies is an experimental evaluation of more than 1,700 students who applied for admission to one of nine career academies across the nation. The study found that among high-risk youth, the career academies reduced the baseline dropout rate of 32 percent by 11 percentage points and that in the students' projected twelfth-grade year, 40 percent of the high-risk academy students had earned enough credits to graduate compared with only 26 percent of the high-risk students in the control group.74 The best cost estimates are that in 2004 the per-pupil cost of educating a student in a career academy was $600 more than the average per-pupil cost of non-academy students.75

High School Reform Models: Talent Development High Schools
High school reform models do not usually state "dropout prevention" as the sole objective for school restructuring. Nevertheless, these reform models often have goals related to dropout prevention, in particular increasing students' school engagement and academic achievement. Common components of many reform models include: reorganizing schools into smaller "learning communities"; focusing instruction and curricula on careers or on intensive or high-level English and math instruction, or both; increasing family involvement; and sometimes focusing on a college preparatory curriculum for everyone.

Many different reform models have been tried over the years, most without rigorous evidence of success. One exception is Talent Development High Schools (TDHS), a reform model for large high schools that face persistent problems with student attendance, behavior, performance, and dropout rates. The model, developed at Johns Hopkins University, calls for schools to reorganize into small learning communities that feature a curriculum designed to prepare all students for high-level English and math courses, along with measures to increase parent and community involvement in the school. Begun as a partnership between Johns Hopkins and a high school in Baltimore, the TDHS program now includes schools in forty-three districts in fifteen states across the nation.76 The added cost is about $350 per student per year.77

A research design that followed twenty cohorts of ninth graders for up to four years in high school in Philadelphia found that 68 percent of the students in TDHS schools were promoted to tenth grade compared with 60 percent of the comparison group.78 These positive TDHS findings are notable as it has been hard for high school restructuring efforts to document positive results on outcomes of interest, including keeping students in school. At the same time, the findings should probably be viewed with some caution because they are based on a quasi-experimental research design.

Other Programs
As noted, there are many, many dropout-prevention programs, most of which are "stand alone" programs and many of which are much larger than either ALAS or Check & Connect. As examples, the Valued Youth Program served 108 schools in twenty-four cities in the United States and Brazil during 2002–03, along with an unknown number of schools in Great Britain; the Teen Outreach Program served more than 13,000 students across sixteen states during the 2001–02 school year.79 These and other larger-scale programs, however, have not been rigorously evaluated, and thus in spite of their apparent popularity, their effectiveness in reducing dropout rates remains unknown.

One program that has been rigorously evaluated through random assignment is the Quantum Opportunities Program (QOP). An intensive and relatively expensive program that offers comprehensive services that begin in the ninth grade, QOP can last for up to five years, providing services even after a student drops out. In six of seven QOP demonstration sites, the cost of the program ranged from $22,000 to $28,000 per enrollee (in 2006 dollars) over the full five years of the demonstration, and labor costs in another QOP demonstration site made the program there even more expensive. In spite of the high costs and intensive nature of the QOP model, experimental evaluations do not offer evidence that QOP participants were more likely to advance in or complete school than were the control group non-participants.80 These examples suggest that one cannot use a program's popularity or size, cost, or even intensity as evidence of effectiveness.81

An examination of the dropout-prevention interventions that show measurable results shines some light on what it likely takes to reduce a student's chance of dropping out. Successful programs have some or most of five elements in common. The first element is close mentoring and monitoring of students. With restructuring models, this mentoring occurs as part of the movement to smaller schools or to school-within-a-school models. The normally high adult-student ratio in a smaller learning environment would have to be higher still to reach the level of monitoring found, for example, in Check & Connect. In the High School Redirection model, teachers are encouraged to serve as mentors as well as instructors, and classes are kept small to foster high levels of individual attention. The second element is case management of individual students. Again, case management is most likely to happen in a restructuring model with a movement to a smaller learning community. The remaining three elements are family outreach; curricular reforms that focus either on a career-oriented or experiential approach or an emphasis on gaining proficiency in English and math, or both; and attention to a student's out-of-school problems that can affect attendance, behavior, and performance.

In closing, we note one complication in designing and implementing dropout- prevention programs. Namely, although common risk factors are important in helping to identify potential dropouts, they are relatively inefficient predictors of who will in fact drop out.82 For example, the risk factors that best predict dropout for high school students are high absenteeism, being over-age by two years, having low grades, and having a child. Using these factors should help identify a group of students with the highest probability of dropping out. Mark Dynarski and Philip Gleason found that these factors would in fact identify a group where one in three students would actually drop out. Although this rate is higher than the baseline 15 percent dropout rate that Dynarski and Gleason find based on the full sample of high school students, one could still question the use of these predictors to assign students to dropout-prevention programs. After all, a program serving students based on these predictors would serve many students who would not need the services and would fail to serve many students who would need them.83 Because most programs use a common set of risk factors to target students for intervention, Dynarski and Gleason's work helps to explain why so few programs show positive results, and it challenges program designers and practitioners to develop better ways to identify potential dropouts.