Grubb, W. N., Kalman, J., Castellano, M., Brown, C., & Bradby, D. (1991). Readin', writin' and 'rithmetic one more time: The role of remediation in vocational education and job training (MDS-309). Berkeley: National Center for Research in Vocational Education, University of California.

Secondary Vocational Education

Although we did not include secondary vocational programs in this study, other research (Grubb, Davis, Lum, Plihal, & Morgaine, 1991) provides evidence about reforms at the secondary level related to the remedial programs we examined. For a variety of reasons, there has been an upsurge of interest in integrating vocational and academic education. Such integration can serve various ambitious goals, including the reconstruction of many aspects of high school; however, when the purpose of integration becomes the enhancement of basic skills among vocational students, it becomes a form of remediation.

One approach, has been to modify vocational curricula to include more academic or basic skills. These curricula are good examples of the skills and drills approach--providing drills in such conventional subjects as vocabulary and spelling, exercises filling in blanks in sentences, comprehension questions based on short reading passages, and arithmetic problems including word problems--with the vocabulary, reading passages, and word problems drawn from a variety of occupational areas. (The appendix to Grubb et al., 1991, lists a variety of these materials.) But apart from the fact that such materials promote a passive form of learning, they are only weakly connected to vocational skill training because they cover many occupational areas and most examples are trivial. We have never seen such materials used by vocational teachers; several reported that the existing materials are not useful because of inappropriate content, and others commented that teachers need to develop their own materials tied closely to their own vocational subjects.

A different approach has been to give the responsibility for remediation to academic instructors. A few area vocational schools, for example, have hired math and English teachers, who then teach modules to students in vocational classes, collaborate with vocational instructors to provide them ways of reinforcing academic material, work with students in small groups or one-on-one, and teach remedial classes. A more thorough change has been adopted in Ohio's Applied Academics program (Ohio Department of Education, 1990), in which academic instructors are assigned to teach courses in applied math, applied communication, and applied science to vocational students. This allows these classes to be tailored to specific occupational areas; for example, math teachers cover different subjects for electronics students than for drafting and design students; the applied communication class for secretaries covers rules of grammar, punctuation, and usage, while the same course for auto mechanics stresses communicating orally with customers and co-workers, reading instruction manuals, and filling out various forms. Because academic teachers spend some time each week in vocational classes, they become familiar with vocational skills training and can devise curricula that are closely connected to these skills. We saw some remarkable team teaching and some other exemplars of integrating vocational skills training with academic instruction in various Ohio schools. In addition, it was clear that the incorporation of academic instruction into vocational programs provided motivation that would otherwise be missing.

There are, then, some examples in secondary vocational education of remediation linked closely to vocational skills training. When we examine functional context training and its offshoots, in Section Four, these secondary examples provide some insight into the possibilities for integrating remediation with skills training. However, the Ohio approach also contains a serious limitation, one that affects other remedial programs. As long as vocational education or shorter-term job training aim to prepare students for entry-level positions in occupations which require relatively basic academic skills, the level of academic skill instruction will remain low. Although electronics and drafting may require algebra, geometry, and trigonometry, individuals preparing to be secretaries, auto mechanics, and animal care workers need no more than simple arithmetic; and the relatively low reading and writing skills required in most entry level occupations similarly set a ceiling on what it makes sense to teach. Without providing students a vision of a sequence of occupations requiring higher and higher levels of academic competencies, it becomes difficult to justify much more than remedial education in most applied academic courses.

Guessing the Scope of the Remediation System

How large is the current system of remediation? Generating national estimates would be nearly impossible. Some programs (e.g., JTPA) don't collect information which would allow national estimates to be derived; in other cases (e.g., community colleges), estimates are available for individual institutions, but aggregation to the national level would be difficult because of inconsistent data systems among states. The variation in adult education makes it extremely difficult to estimate the magnitude of the largest component of remediation, and the task of converting short-term enrollments to a consistent basis (e.g., full-time equivalents) presents yet another difficulty. We know of no effort to develop national figures.

However, the California Workforce Literacy Task Force (1990) has developed estimates for California that indicate probable orders of magnitude. These estimates, presented in Table 3, required great time and effort, and they are still subject to many limitations (see some of them noted at the bottom of the table). Still, they indicate patterns for California that we think are true nationwide. Most obviously, the adult education system--provided in California through both adult schools run by school districts and regional occupational centers and programs--accounts for the largest share of remediation, almost two-thirds of total spending. The community college system comprises the second-largest component, spending about fifteen percent of the total. In other states the balance of adult education and community colleges might be different, since some states give responsibility to community colleges for adult education; on the other hand, most other states have relatively smaller community college systems than California. However, the conclusion that remediation in adult education is larger than in community colleges seems correct, and it is consistent with our interview results that most JTPA and welfare programs refer their clients to ABE rather than community colleges. The third largest component, the JTPA system, accounts for roughly seven percent of total spending in the state, much less than either of the other two programs.[32] (In these figures, funds from the state's welfare-to-work programs are spent through other institutions, and, therefore, do not show up as a

Note:

These are estimated funding and numbers served for participants in non-credit or remedial education programs in Fiscal Year 1990-1991, except where noted. CAUTION: Total dollar figure overestimates amounts for the eleven programs with funds listed due to duplicate reporting such as JTPA monies mixed in the Adult Schools' budgets. No funding listing was available for two of the thirteen programs. For these reasons, the total funds given do not accurately state the exact amounts available for adult literacy education. The total numbers served is also misleading because it mixes ADA figures, in which one ADA may involve two or more students, with actual individual participation in some programs. Thus, the numbers served are probably underestimated. Apparently no one knows the exact funding or numbers served in these programs.

Source:

California Workplace Literacy Task Force (1990).

separate amount.) The remaining enrollments and expenditures take place in much smaller programs. In particular, the voluntary programs like California Literacy Inc. and Literacy Volunteers of America are tiny compared to publicly funded efforts. The real action in remedial education takes place in adult education and community colleges; the widespread publicity given to voluntary efforts and to the experimental programs developed by corporations, CBOs, and university researchers misstates the relative importance of such institutions in the existing system.

A second conclusion is that the majority of funds for remediation come from state government, in the form of aid for adult education and community colleges, rather than from federal sources. Table 3 shows that federal support through JTPA is clearly small, roughly $60 million. Support through the Vocational Education Act must be small because only forty-five percent of the state's allocation of roughly $100 million went to community colleges, and much of this funded equipment and other purposes more directly related to skills training. Federal support for remediation through GAIN was probably very small, since GAIN relies on adult education and community colleges for remedial education. In addition, funding through the federal ABE program is similarly small, perhaps $20 to $40 million.[33] The federal share cannot be more than $100 million, therefore, or perhaps ten to fifteen percent of overall expenditures. Indeed, the dominant pattern of cooperation in this system is for federally initiated programs that are badly underfunded relative to what they are asked to do--JTPA and JOBS--to access state-supported ABE and community college programs. Federal funding may be increasing, but it is far from being a major component of the system.

Finally, by almost any account, total funding for the remedial system is large. If California spends $800 to $900 million, then--because California represents roughly ten percent of the country--national spending might be $8 to $9 billion. Even if this estimate is off by fifty percent, the magnitude of remediation is considerable. In bits and pieces, with little planning or discussion, a substantial enterprise has developed.

The Nature of Remediation in the Education and Training System

Despite the enormous variety of remediation, several clear patterns in existing programs emerge. One characteristic--perhaps so obvious that it might be overlooked--is that remedial programs are ubiquitous. In every one of the twenty-three communities we examined, a rich set of institutions provide basic skills instruction and developmental education. This is not to say that the offerings are adequate: Most providers report of being overwhelmed with the demand, and the biggest issue they face will be keeping up with the increasing numbers needing remediation. But there is a rough system in place nearly everywhere.

A second characteristic of this system is that--in theory--it is structured to provide a hierarchy of programs from the lowest levels of literacy (and, to a lesser extent, math competency) to the highest. A tripartite structure of programs exists in most communities. Individuals who test at the lowest levels--for example, under a fourth grade level of equivalency--are typically referred to volunteer literacy programs using one-on-one tutoring, sometimes associated with libraries. The next highest stage includes ABE (or pre-GED) programs, often described as covering the equivalent of fourth to seventh or eighth grade instruction. In turn, they prepare individuals for GED programs that are designed to help individuals to pass the GED. Because the GED is widely interpreted as the equivalent of a high school diploma, individuals who have passed the GED are considered out of the remedial system and ready for college.[34] This tripartite structure is sometimes a matter of state policy: In Tennessee, for example, individuals below a 4.9 grade level are sent to literacy programs; those between grades 5 and 8.9 go to basic skills courses; and those between grades 9 and 12.9 enroll in GED courses. More often, such a division has developed informally, as programs assess what levels of students they can handle.

Within community colleges a slightly different structure exists, but there is still a tendency to have a three-part set of offerings. The goal is usually entry into the first college-level English course rather than completion of the GED; from that standard, community colleges offer courses that are one and two levels down from the college level, with many, though not all, offering a third level for individuals without any reading skills. Therefore, a well-developed remedial program will have three levels of reading, three levels of writing, and three levels of math courses, and it will accommodate a range of individuals that includes JTPA and welfare clients. It will also differentiate reading courses into those for native speakers and those for non-native speakers. These courses then lead to the college-level English and math courses that prepare individuals for transfer to four-year colleges.

In theory, then, the system of remediation in many communities allows individuals to start at any level, move through increasingly difficult material, and then receive a GED or move into college-level courses. In practice, however, the mechanisms of tracking students are poorly developed. Welfare-to-work programs give caseworkers the responsibility for making sure that welfare clients make progress, but this tracking mechanism doesn't always work well. Some community colleges have developed student tracking systems which provide information on the progress of students (e.g., see Palmer, 1990); these can inform students if they lag behind in a sequence of courses and alert guidance counselors who can then investigate why students are not making adequate progress (as in the Miami-Dade system described in Roueche, Baker, & Roueche, 1985). However, these tracking systems are not by any means uniformly in place, and the resources that community colleges have for follow up if students fall behind in their programs are limited. In practice, then, a smooth continuum of courses--with mechanisms helping students make the links among pieces of the continuum and providing guidance or tutoring if they falter--exists in very few areas, though a few community colleges come close.

Yet another restriction on the continuum of remediation is that most programs have relatively modest ambitions. Most JTPA and welfare-to-work programs hope to advance their clients one or two grade levels, and provide so little time--as little as four weeks in many cases--that even this much progress seems unreasonable. The time in ABE for most students is also relatively short, as well as quite erratic, so that gains in most cases are limited to a grade level or two; at the most, adult education programs hope that their students can pass the GED, but at the same time, many adult instructors recognize that the GED is not very helpful in obtaining employment. Community college programs are less subject to limitations in their ambitions, since the stated goal in most of them is to enable students to enter college-level courses and then to progress to a vocational or academic degree; but here, too, rates of noncompletion are high. Limited funding, particularly in JTPA, welfare, and adult education, is partly responsible for limited ambitions, and, of course, there are high dropout rates in adult programs. As a result, what appears to be a continuum of remedial education in many communities in practice is difficult for individuals to negotiate.

The curriculum in remedial programs appears to have changed substantially over the past fifteen years. Virtually all remedial programs report extensive use of materials that are individualized, self-paced, and often open-entry/open-exit, rather than operating with the rigid starting times associated with conventional schooling. (Of course, when institutions such as community colleges provide both lab settings and classroom-based discussion sessions, the classroom portions must follow a conventional schedule.) Many curriculum materials are also competency-based, so individuals progress to new units or subjects when they pass a competency test; conversely, those who fail to pass such tests are given additional lessons and practice in the specific skill until they can master it. These characteristics are generally true of both print-based curricula and computer-based methods; indeed, many remedial instructors reported their preference for computer curricula. The curricula include a battery of individual tests that make it easy to identify skill levels, and the computer presents lessons in sequence without any intervention from a teacher.

In contrast, when Cross (1976) reviewed adult education in the early 1970s, most programs provided a relatively uniform curriculum, with progress based on seat-time--the amount of time spent in the program--rather than acquired competencies. She recommended individualizing instruction, mastery learning methods, and self-paced methods as ways of allowing individuals to progress through a series of skills at their own pace; she argued, as did other proponents of mastery learning, for substituting an educational process in which the amount of time remained constant for all students and the amount of learning varied, with one in which the amount of learning was constant while the time to master particular skills could vary. Since then, evidently, these recommendations have been widely embodied in curriculum materials, with a "new orthodoxy" widely practiced.

As part of the new orthodoxy, the majority of remedial programs in our sample of communities and the majority of those we visited, follow the pedagogy we label skills and drills. In this approach, complex competencies--the ability to read, for example, or the ability to use mathematics in various forms--are broken into smaller discrete skills such as the ability to decode words, or to recognize the point of a three-sentence paragraph, or to add two-digit numbers with carrying. Students drill on each of these subskills until they have mastered them (i.e., until they can pass a small exit exam), and then they move on to the next most difficult skill. While we will examine the assumptions underlying skills and drills more carefully in Section Three, "The Nature of Effective Programs: The Conventions and the Structure of Skills and Drills," it is important to recognize that most of remedial education follows this approach (a few exceptions are described in Section Four, "Alternatives to Skills and Drills"). Remedial education is provided in a bewildering variety of institutions, with many different funding sources and with individuals attending for many different purposes. In addition, there is no national curriculum, no textbook approval process like the one that standardizes K-12 texts in many states, and no mechanisms like college entrance requirements and the SAT examination to standardize the curriculum.[35] In spite of these differences, there is still a stunning sameness to the instructional methods and curriculum materials in remedial programs.

Finally, almost no remedial program in our sample of communities linked its curriculum in any way to the vocational skills training that would normally follow or, in the case of concurrent programs, that students are taking simultaneously. There is increasing recognition that many individuals learn best when competencies are taught in some concrete application (or "contextualized"), and "functional context literacy training" has become a popular notion in some circles, but these principles have not yet been embodied in curriculum materials, teaching methods, or program philosophies. Several administrators commented that the lack of connection generates motivational problems when individuals fail to see the relevance of abstract skills and drills to their occupational futures, and these administrators expressed the desire for some integration; however, almost none of them had found the time, resources, or curriculum materials to do so.

Coordination in Remedial Programs: Its Status and Value

Because so many programs provide remedial education, almost every community we surveyed has many providers. The offerings in the Motlow State Community College SDA in Tennessee--a six-county rural area--provide a good example. The SDA contracts with one area vocational school and four non-profit CBOs to provide remediation to JTPA clients, and, in a sixth county, the SDA operates a remedial program itself. The area vocational-technical school provides a GED program as well as basic education for JTPA students in its vocational programs. The community college has a developmental studies program for entering students who score low on a mandatory assessment, and the local school district provides ABE programs as well as a JTPA 8-percent program. Nearly every education and training institution participates in remediation then. The only exception is that there is still no welfare-to-work program, though JTPA recruits at local welfare offices. There are, too, some exceptions to the general pattern of multiple remedial programs: In southwest Wisconsin, Southwestern Wisconsin Technical Institute provides virtually all remediation, at its main campus or in off-campus programs, as do the Heart of Georgia Technical Institute in Oconee County, Georgia, and the adult education system, widely described as "the only game in town," operated by the county school board in Broward County, Florida. However, these are clearly exceptions; in most communities, several types of remediation co-exist.

Despite the number of remedial programs and the proliferation of funding mechanisms, we heard little complaint about duplication and overlap.[36] One reason is simply that the need for remediation and ESL is much greater than the resources available; most providers would welcome additional programs or additional funding, rather than seeing others as competitors. A second reason is that coordination--in the form of referring individuals to other programs--seems relatively good. Referrals from JTPA and welfare-to-work programs, predominantly to adult education but also to community colleges, are especially common. While there are complaints about paperwork (especially for welfare-to-work programs with their complex reporting requirements), there were no complaints about the unwillingness of other programs to refer their clients, nor were there claims that political allegiances and turf issues prevent cooperation--as there frequently are for job skills training.

Cooperation in the form of referral is partly caused by the desire not to duplicate services and--particularly for JTPA and welfare programs that do not see themselves as educational institutions--by the desire not to expand into another area. In addition, referral is also driven by a motive we have referred to as cost-shifting (Grubb & McDonnell, 1991). That is, programs like JTPA, with a limit on funding, and welfare-to-work programs, without adequate resources, are constantly looking for ways to expand services by shifting costs to other programs--particularly to institutions (e.g., adult education and community colleges) which have open-ended, enrollment-driven funding. This is a fiscal motive for cooperation, not one driven by a concern for the quality of services; with only a few exceptions, the administrators in our sample communities refer their clients to ABE and community college programs because they don't want to reinvent the wheel, not because they have any evidence about the effectiveness of these programs. Indeed, few of the JTPA and welfare-to-work programs we interviewed had established any policies about the content of remediation; and few knew much about the content of basic education in their area.

However, in another sense there seems to be little coordination. As we pointed out earlier, there are few mechanisms of tracking individuals through remediation. In addition, while a few communities have established central councils which provide information to individuals seeking basic education, most have not. As a result, individuals approaching the education and training system are likely to feel bewildered and to find a way into a program almost accidentally (Hull, 1991). In this sense, then, coordination in most communities is poor, even though cooperation in the form of referrals is common.

In this context, a crucial question is whether cooperation in providing remediation is a good thing. One troubling aspect of the referral process--given a firm convention within adult education (reviewed in the next section) that policies and goals should be carefully established--is that few programs develop policies of any kind before they refer clients to remediation. Referral seems expedient, rather than principled or planned; some administrators who admit or even boast that they are not educators, especially in JTPA and welfare-related programs, seem relieved to find another institution providing remediation so that they need not have to think about it. While the resulting division of labor may seem rational, it does not necessarily result in individuals receiving the education they need.

Since there are few mechanisms to track individuals, it is generally impossible to tell whether a person referred from JTPA to an ABE program ever enrolls, completes, or manages to enter job skills training. While we know of no evidence, we suspect that many individuals referred to other programs become "lost." The rates of completion in most remedial programs are relatively low to start with for a variety of reasons, ranging from personal difficulties in individuals' lives to the uninteresting curriculum in many programs. Moreover, adding a change among institutions introduces an additional barrier.

Third, a process of referring individuals to another program for basic skills instruction makes it impossible to link remediation to job skills instruction. Even though the effectiveness of functional context literacy training or of programs integrating remediation with job skills training remains unclear, there are still motivational advantages to linking these two components.[37] However, the referral process generally requires individuals to complete remediation or to achieve a specific level on a test before entering job skills training, a sequence that seems likely to eliminate the lowest-performing individuals most in need of both remediation and job training.

Finally, we remain concerned about the breadth and effectiveness of programs to which individuals are referred. Individuals are most commonly referred to ABE programs; however, these are also relatively limited, at least compared to community colleges' developmental education, in the subjects they offer, in the levels of difficulty they provide, and in the instructional methods they use. The dropout rates from ABE are generally thought to be very high--even though it is impossible to find substantial evidence--and there is virtually no outcome evidence. Community college programs are more likely to offer remediation in reading, writing, math, and a variety of employability and life skills, as well as courses ranging from the elementary to the college level; the best of community college programs are much more likely than ABE programs to offer both classroom-based and individual or lab formats and to experiment with alternatives to skills and drills. Nevertheless, here too dropout rates seem to be high and effectiveness unclear; given the variety of community college offerings, it is disconcerting to find other programs referring clients to them without clear guidelines, policies, or evidence of effectiveness.

In sum, the provision of remediation may be an area of education and training where cooperation in its common form of referral may not be desirable. Unless there are clear guidelines for remedial programs, methods of tracking individuals among programs, and better evidence of effectiveness, referral seems like a mechanism for claiming to address the skill deficiencies of the adult population--but without any of the conditions necessary to ensure success.

Information and Evaluation

A striking characteristic of the existing remedial system is the lack of information. Many providers--especially JTPA programs that delegate remediation to subcontractors, and welfare-to-work programs with incomplete management information systems and individuals lost in the system--could not even tell us how many individuals were enrolled in remedial or basic education. Fewer still could provide any systematic information about completion rates or other measures of intensity--that is, how many individuals had completed various proportions of a program. Administrators usually provided estimates of enrollments and completion rates, but it was clear that their estimates were often very rough. While most programs claimed to carry out evaluations of their students, almost none of them sent the evaluation materials they promised to us. Most curricula require pre- and posttests of individual students to monitor their progress through a sequence of skills, and these tests are used to evaluate individual students; even so, the information they provide is not used to evaluate the program as a whole and is, in fact, not suited to such an evaluation. The almost complete lack of systematic data means that these programs do not have the information necessary to show evidence of their success to others (except success measured by continued enrollments), to be self-conscious about their performance, or to improve their offerings.

Where evaluations have been carried out, they are often seriously flawed. The most common method of evaluation is to compare the pre- and posttests of groups of students enrolled in remedial courses without any comparison or control group.[38] Conventionally, increases in scores are presented as evidence that the program works. However, there are many other plausible explanations for such an increase. One is that only students making substantial progress stay in remedial programs until the posttest. Since many students in developmental education (from which most of these evaluations come) are concurrently enrolled in other courses, these other courses rather than the remedial offerings may be responsible for test score changes. Regression to the mean--many students selected on the basis of low test scores improving over time merely because their initial low score was the result of chance rather than low achievement--and practice effects--students improving because of gaining practice with test-taking methods--may also be responsible. The upshot is that it is impossible to conclude anything about the effectiveness of remedial programs from this approach to evaluation.

In other cases, gathering data for an appropriate control group provides somewhat better evidence of effectiveness. For example, the New Jersey results cited above (Morante et al., 1984) compare attrition rates for those students who have completed remediation, those who have started but dropped out of remedial courses, those in need of remediation who have not enrolled, and those not needing any developmental education; the results from Miami-Dade Community College in Tables 1 and 2 in effect contain ten comparison groups. There are still problems with these evaluations, since different groups of individuals needing remediation may not be comparable and are certainly different from those not needing remediation; but the findings with comparison groups are still more persuasive than simple pre- and posttest comparisons. In addition, the outcome measures used--retention within the community college and, in Florida, passing the CLAST--are more meaningful than test scores because scores may simply reflect test performance rather than any progress toward educational goals.

However, even in evaluations with comparison groups, the analyses typically ask an inappropriate question. A comparison of outcomes between a group that has completed some education and another that has failed to complete any education can answer only whether more of that education is effective compared to less of it. These evaluations can, therefore, establish whether remediation is worth doing at all; if the results are negative, remediation should presumably be abolished. But almost no one proposes eliminating remedial education: The competencies of many students entering community colleges, JTPA programs, and welfare-to-work programs are too limited to ignore. The appropriate question is to ask what kinds of remedial programs are most effective for which students. This requires an evaluation in which the outcomes of different approaches to remediation--varying, for example, in the intensity, the relative balance of classroom-based and lab-based instruction, the use of teacher-based methods versus computer-based methods, the use of skills and drills versus the alternatives, or the reliance on materials drawn from occupations (as in functional context training) versus context-free materials--are compared. Such evaluations would help teachers and administrators improve existing programs, rather than providing evidence only for decisions about eliminating programs. However, we have uncovered no evaluations of this form, nor is there much pressure within remedial education for such evaluations.

From our survey of remediation, then, a picture emerges of a large and expanding system with enormously varied institutional sponsorship and different funding sources. A variety of individuals have gained access to the system by enrolling in community colleges and adult schools, by applying to JTPA, and by being forced or by volunteering to participate in welfare-to-work programs. With its great variety and its lackadaisical data collection, it becomes difficult to describe the particulars of this system. However, some general characteristics stand out: the availability of several distinct levels of the system, usually poorly articulated; and the dominance of pedagogical methods which rely on individualized, self-paced, competency-based materials, driven by a skills and drills approach and unconnected to either the academic education or vocational training for which remediation presumably prepares people. While there is general recognition of high rates of noncompletion, there is little evidence about completion rates, or even--in some cases--a consensus about how to define completion. While there are high hopes for these programs, there is almost no evidence to indicate which of these programs are effective at all, and still less that would enable teachers and administrators to improve them.

In the next section, we address the issue of effectiveness in greater detail. Despite the lack of evaluation, there is extensive literature on the characteristics of good programs--literature based on experience and convention. This accumulated wisdom forces us to examine the pedagogy of existing programs more carefully, as a way of coming to terms with the deficiencies of existing programs.

[32] A footnote acknowledges that the methodology for the JTPA estimate is "very rough." We suspect that the JTPA figure is an overestimate, since it implies that twenty-two percent of total JTPA spending in the state went for basic education, which strikes us as too high. However, the basic conclusion that JTPA is not a major contributor to the overall system of remediation is surely true, simply because of the overall scope of JTPA and its emphasis on job skills training.

[33] The federal program provides about $200 million annually. California usually receives about ten percent of federal funds, but early results from the National Evaluation of Adult Education programs indicate that California had twenty-two percent of clients in October 1990, suggesting the state might have received as much as $40 million.

[34] However, the GED is widely described as requiring only an eighth or ninth grade reading level, and those who have examined it closely claim that individuals can pass it with only a fifth or sixth grade reading level (Quinn & Haberman, 1986).

[35] There is a partial exception: The goal of most adult education and of many JTPA and welfare programs that refer their clients to adult education is to enable individuals to pass the GED. Because the GED is a conventional multiple-choice test of reading comprehension and arithmetic computation, preparation for the GED leads naturally to skills and drills.

[36] There is one possible area in which duplication may be a problem: The assessment of skill levels may take place several times for one individual. For example, a JTPA client is typically assessed upon entering the program, particularly to determine which programs he or she can enter; another assessment may be carried out upon referral to a particular skill training program; and if that training provider also incorporates some basic skills instruction, there may be a third assessment to ascertain exactly where in a sequence of reading and math skills to begin. Since these assessments--typically, conventional multiple choice tests--are especially trying for individuals with low skill levels (Lytle, Marmor, & Penner, 1986), it is possible that multiple assessments contribute to high dropout rates. For other evidence that the assessment process is threatening and can contribute to dropping out, see Hershey (1988).

[37] Several administrators in our sample of communities remarked that individuals find remedial programs boring, and fail to see the relevance of curriculum materials devoid of any occupational content to the jobs they hope to enter. Conversely, CET--one of the few to integrate remediation into skills training--developed their approach because of the dropout problems they experienced when remediation and skills training were sequential.

[38] On the dominance of pre- and posttest comparisons, specifically in remedial mathematics programs, see Akst (1986). Only twenty-four percent of evaluations used a comparison group, defined as students needing remediation but not taking the remedial program.

Grubb, W. N., Kalman, J., Castellano, M., Brown, C., & Bradby, D. (1991). Readin', writin' and 'rithmetic one more time: The role of remediation in vocational education and job training (MDS-309). Berkeley: National Center for Research in Vocational Education, University of California.