In large part, the expansion of job training programs has reflected a concern with particular economic problems, especially those of unemployment, underemployment, and poverty. Job training programs were first discussed as a response to the unemployment created by the 1960-61 recession. Since then, periods of recession and unemployment, and specific kinds of unemployment (like the increases in dislocated workers unemployed as a result of substantial economic changes) have generated interest in job training programs as potential solutions. And the expansion of welfare programs supporting low-income families during the 1960s generated the realization that poverty was not likely to wither away of its own accord, and one solution has been to propose job training. Of course, the problems of poverty, unemployment, and underemployment are closely related, and so -- despite a proliferation of programs for specific purposes -- it is not surprising to see an overall pattern to the job training programs in the United States.
In the process, an important distinction has emerged between education and job training. The difference is not always clear, since some short-term, job-specific education programs look quite similar to job training. However, there are at least six differences between the two. First, job training programs are generally much shorter. Many of them last 10 - 15 weeks, with part-day attendance, so that the number of contact hours may be as low as 40 hours; the average program length in the recent JTPA evaluation was 3.5 months (Orr et al., 1994, Exhibit 3.18). In contrast, the shortest common postsecondary education programs -- for occupational certificates -- generally last two semesters (or about 30 weeks) of full-time enrollment, involving 360 to 1,000 contact hours; and two-year Associate programs dominate the programs of community colleges.[1]
Second, education programs -- particularly in community colleges and other two-year colleges, and area vocational schools -- are generally open to all members of the population; but job training programs are open only to those who are eligible -- for example, the long-term unemployed or dislocated workers in JTPA, or welfare recipients in welfare-to-work programs. (The issue of eligibility reflects the origin of job training as a solution to particular economic problems: only those who have suffered from these problems are eligible, not the population as a whole.) By construction, then, job training programs enroll individuals who have had particular problems in employment; while some problems may be due to overall employment conditions, others may be due to deficient skills, behavioral problems, and other personal traits.
Third, most education programs take place in educational institutions that are well-institutionalized and standardized -- in high schools, community colleges, and four-year colleges. In contrast, job training services are offered in a bewildering variety of educational institutions, community-based organizations (CBOs),[2] firms, unions, and proprietary schools, making it difficult to determine how services are organized and provided.
Fourth, the kinds of services provided in education programs are relatively standard: for the most part they offer classroom instruction, with both academic and vocational courses, often including labs, workshops and other "hands on" activities. Job training programs offer classroom instruction too, in both basic (or remedial) academic subjects like reading, writing, and math, as well as in vocational skills; but they also offer on-the-job training[3], where individuals are placed in work sites, presumably to learn on the job; work experience, where individuals work for short periods of time; job search assistance, in which clients are given some training in how to look for work, write resumes, file job applications, interview for jobs, and the like; and job clubs, in which clients are required to spend a certain amount of time looking and applying for jobs. Some programs provide counseling as well, to give clients both information about labor market opportunities and "life skills" like the abilities to plan. Job training programs also support placement efforts somewhat more often than educational institutions do, reflecting another division between education and job training: those in educational institutions are likely to declare that they are responsible for "education, not employment", while those in job training are more likely to accept that they have a responsibility for placing individuals as well as training them appropriately. Unfortunately, the variety of such services is so great, and the forms they take is so varied, that it is difficult to know precisely what takes place. As a result the evaluations of specific program components (reviewed in Section III.5) is not particularly comprehensive and is generally inconclusive.
Fifth, the goals of education programs are typically quite broad, and generally encompass political, moral, and intellectual purposes as well as occupational ends; but job training programs are focused exclusively on preparing individuals to become employed. In the case of welfare-to-work programs, the single goal is to get welfare recipients employed as quickly as possible so they can move off the welfare rolls. Because the goal of job training programs is so unambiguous, and because there are no intrinsic benefits to job training -- no one would declare that being in a job training program is fun, or a social activity, or a normal part of growing up, as Americans might say about schools and colleges -- there has been a long history of evaluating them to ascertain their effectiveness. These evaluations have also become increasingly sophisticated over time -- certainly much more sophisticated than those of education programs. The results have also influenced public policy in a way that has not been true in the education system because the political pressures in education -- the support of parents for programs that benefit their own children, including such diverse offerings as those aimed at low-income students, or limited-English-speaking (LEP) students, or gifted students, for example -- are generally lacking in job training programs, where the only justification for public support is their reduce of unemployment, poverty, and the receipt of welfare.
Finally, job training programs differ from education programs in constituting a separate kind of system, a "second chance" system in some ways parallel to but disconnected from the "first chance" educational system. Over the course of 150 years, the education system in the United States has developed a well-articulated series of offerings from kindergarten (now often extended to pre-school programs) to the university level. But for those who have left this system without adequate skills, the job training system can be interpreted as a second chance to get back into the mainstream of the labor force. In general, the establishment of this second chance system is one manifestation of a generous American impulse: to provide opportunities to individuals through various forms of learning, and to be inclusive of all who might benefit from such activities. However, this second-chance system is much younger than the education system; it spends less, is more disorganized, has lower status, and is poorly institutionalized so that it cannot resist purely political pressures. As a result it has been subject to revision by nearly every President so that it lacks the stability of the education system. Given these differences, it is not surprising to find that the job training system is not especially effective -- partly, as I will argue in sections IV and V, because of the separation of education and job training in separate arenas, a division that has been detrimental to both of them.
In this monograph I review the effectiveness of job training programs in the United States, concentrating on the most recent and most sophisticated evaluations.[4] Section I describes in greater detail the variety of job training programs described, explaining why it is appropriate to consider a wide variety of programs. Section II outlines the preferred methodology of recent evaluations using random assignment methods, and clarifies both the strengths of this approach and its inevitable weaknesses. Section III then presents a series of results, first for job training programs (III.1), then for welfare-to-work programs (III.2), and finally for special experimental programs (III.3). These results are followed by the outcomes for different population groups (III.4), for different types of services (III.5), for the effects of programs over time (III.6), and for different programs (III.7). Finally, in Section III.8, I present some recent benefit cost-analyses. A series of tables drawn from the evaluations accompanies this section, so that the reader can see concretely the results of the major evaluations.
The major question these evaluation address, of course, is whether job training programs been successes or failures. A conventional reading of the evaluations is that many (though not all ) job training programs lead to small but statistically significant increases in employment and earnings, and (for welfare recipients) small decreases in welfare payments; where cost-benefit analyses have been done, the social benefits usually outweigh the costs (but not always). One might conclude that these programs have been successful and should be continued. However, the gains in employment and earnings are, from a practical standpoint, quite small: they are insufficient to move individuals out of poverty, or off of welfare; their effects very often decay over time, so that their benefits are short-lived; and as they are currently constructed they certainly do not give individuals a chance at a middle-class occupation or income. In my interpretation, therefore, the successes of job training programs have been quite modest, even trivial -- and that dismal conclusion requires some understanding of why that might be true. In Section IV, therefore, I present a series of possible explanations for the weak results of job training programs. The reasons for failure are necessarily more speculative than are the outcome results in Section III, which are based on harder data and (in many cases) random assignment methods; but some understanding of why job training programs have had such modest results is necessary to develop recommendations that could remedy such programs through public policy, or create more effective programs from the start.
Although the benefits of current job training programs have been small, they problems they address -- unemployment, underemployment, and welfare dependency -- are too serious to ignore. Therefore, rather than abandoning job training, the appropriate response is to determine how to reform them. The conclusion therefore presents a vision of how job training programs could be structured in ways that avoid the reasons for failure outlined in Section IV. This vision has in fact been embodied in current federal legislation -- the School-to-Work Opportunities Act, passed in May 1994 -- that applies to high schools and community colleges. But its implications for job training programs have not yet been developed, and so the purpose of the conclusion is to clarify how reforms now proposed for the education system might benefit job training programs as well. In the end such reforms could eliminate the unproductive division between education and job training that has developed over the past thirty years.
[1] Of course, student attendance patterns can reduce these longer education programs to individual courses. Area vocational schools often offer short programs of part-time attendance over one semester of about 15 weeks, and are therefore no more intensive than many job training programs.
[2] CBOs are private organizations, incorporated under state laws, that provide many different kinds of social and educational services. Some of them are identified with groups of individuals -- e.g., they represent the black community, or Hispanic migrant workers, or the disabled, or older women returning to the labor force; some of them represent particular services, like child care or homemakers services for the elderly. In general, there is minimal government regulation of CBOs, and their quality varies greatly.
[3] One of the issues here is whether there is any training involved in on-the job training, or whether it is the same as short-term work experience; see the discussion in Section IV.3 below. For evidence that much on-the-job "training" does not provide training, see Kogan et al. (1989).
[4] This monograph is not a formal meta-analysis, which would require statistical analysis of outcome results, suitably standardized, from a large number of studies. A formal meta-analysis has been carried out by Fischer (1995), and will be cited as appropriate; however, like most meta-analyses, the summarization of evidence makes it difficult to understand what might have caused the outcomes. Instead, I take the approach of presenting results from a number of specific evaluations in a series of tables illustrating the kinds of findings that have been typical.
