Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
5 Effectiveness of Occupational Skills Training Programs Skills training programs are generally designed to impart skills relevant to obtaining work in specific occupations. We found that few youth programs exclusively devoted to skills training were undertaken with YEDPA discretionary funds. Moreover, there is a paucity of YEDPA evaluation reports on such programs relative to the number of reports on other types of program. We are left, in effect, with only two programs--the Job Corps (which was developed prior to YEDPA) and New Youth Initiatives in Apprenticeship--that had substantial skills training components and were sufficiently well evaluated and documented to be subjects for our review. Yet each of these programs had special features that limit its applicability to broad segments of the youth population: the Job Corps is a residential program for out-of-school youths that includes much more than skills training; and the apprentice- ship program required a close relationship between employers and school programs dealing with specialized skills. Table 5.1 presents the char- acteristics of these two programs; Table 5.2 summarizes the research design and results of their evaluations. That few skills training programs were developed under YEDPA--which is consistent with a frequently voiced criticism of all CETA programs operating during the 1970s--apparently grew out of several concerns of program administrators. One concern was the belief that below a certain age young people tend to lack the seriousness to make good use of skills training because they have not committed themselves to a particular occupational direction. Another concern is that participants require a sufficiently high level of academic preparation to be able to partici- pate effectively in any but the most general training, and many program applicants lack this level of preparation. In fact, as was noted in Chapter 4 and is discussed in Chapters 7 and 8, some program efforts initially designed to provide skills training were redesigned when it became clear that participants were primarily in need of more basic educational skills training. 108
109 U) .. - U) U] .,, U] · - v ~: o ·e tn o ,' . - E~ n ,' U] o ,, Q4 :' v . ~; E~ o . - .,' ., 4 q~ o s ~ C o U) . ~ ~ U] o .~4 k4 a 4~ U ~ ~s cr ~ L4 ns ~ s E~ C.) a s~ a' q~ q~ o U] Q) .,. V) ~: o ~: o ~, o o ~ h 415 S {d O ~ U) U) . - JJ O 4~ U] ~ ~ ~ ~ O 4~ ~ h C ~ 4 Q C -- U) C Q) Q) O 3 J~ ~ . - U] V h C ·_. ra ~ a C Q U) ~ ~ U] Q5 ~ 1 ~ C ^ U) ~ U] U] U] C 4J ~ O U] . - ~ c 4U a' ~ Q4 J~ 4U C C O O cn - U) y aJ 3 u~ o c ~a) 11 3 o X a: 1 0 C a' U] a' :> ~a~ ~ C ~ 1 '\ ~ a, ~ ~1 - - S ~1 C 3 v ~ns ' - O . - ~ ' - ~ 4) 3 S C ~7 O ~H 1 ~ S JJ S ~ C U] U] ,- ~ '- ~ ~'- ~ S O 3 ~ Q ~C ~ ~ ~S 0 3 ~ ~ Y ~S ~ 0 a) ~l a~ 0 ~ v Ql ~ 0 ~t~l O O ~ a) ~ ~ Ll ~ ~ ~ ~ >1 ~CD S ~ U] ~ ~ ~ ~ ~ ~ O ~ ' - ~ O ~ S ~ ~s c c~ ~ ~ C n =: < - c~ cn ~ U] C ~ :, ~ ~ ~ ~ dP ~O . - m. - 3 =1 L4 O O ~ ~ dP C oP dP 1 dP 1~ ~1~ dP dP d,° 1 n5 0 0 O ~ dP O U)U, ~ ~ tD 1 4~ u;~ C ~ ~ ~ a~ ~ I I I I ~ a~o~ m ~ · - ~4U C O ~O · ~ · · · · · H · · · s ~a) a) ~ JJ ~U] ~ ~ C ~1 ~1 a) ~ ~ (L) a' ~- - 0 ~ ~ ~ ~ ~ s a) ~Y ~ ~ Ll ~ S h ·' U1 3 Q~ ~ ~S U] _ ~a~ ~1 a ·. ~ ~ C Y C ~ Q4 ~ - - C O ~ ·'= U] ~ O ·' ~ 3 ~ V ~ C ~ 5 0 ~ ~ C 0 u~ C a~ c c .Q ~ ~ ~ a' a' 0 0 ~ 0 O ~ ~ ~ ~ S ~ ~ ~-- s ·- ~ Q4 c a) ~l 0 0 ~ Q' ~ ~ ~ ~ 3 - - ~ ~ ~ . - ~ . - Ll ~ :D t,) ~ ~1 V u~ ~q v ~ ~ c a' 0 c 0 O ~ ~ ' - ~ ~-- O ~ C 1 ._1 ~ .,.C CJ1 C ~ ~ O c u' O c - -Q4 ~ 1 C -l ~ 4 ~ JJ ~1 C C O J~ ·~1 54 ' - ~ .-1 C ~ J~ ~ (V Q4 0 ~ ~ ~ ~a . - JJ QJ u, JJ a) c L4 ns 0 u~ ~ (.) -_~ al J~ nS U] Q] C U~ ~ ~ C C ~ ~ O ~ ~ U] U] a~ ::) ~c a~ c~ ~ Q] J~ - ~ a~ ~ 0 ~a) C) c Q5 c CJ) ~ J~J .,4 ~ ~ - - O ~ O U] O U] ~ ~ C a\ J~\ U) C 15 01 ~ ~ H ~a) JU ._d ~ C ' s ·r4 a' U] (0 O S 4~ ~ U) Q] L4 a) ~ ~ ~ Q~ 0 s ~ a' c ~O u, a) ~ ' - ' - ~ 0 a) ~C c s O ~3 ~ · - u] Z
u] 4U o c~ a' - ~ 1 u, 3 C: O Q. U] 0 a ~ ~r; U] U] U] S S~ U. a' ~; ·a ,1 - E~ U] y U] o .~1 8 . U~ U] v q~ h o >1 O - O .o O a== a O h Q 4J ~ o o O V O C O 4U N ~ a' v ~ 4U U) ~ o P4 S o 110 a >1 L~ a U] ~3 · - O S u~ ~ dP ~ a' ~ :) 3 a' ~ ~ ~ 3 ~ a' 1 a' Ll S ~ O O ~ 3 ~ a~ ·~ ~' LJ O ~ 3 ~ ~ - )- O ~1 3 (U a' ~ ~ 0 ~ V O ~ ~ 4U 0 ~ O q ~ ~ ~ U] ~ ~ O ~ ~ ~ ~ =, ~t(5 ~ ~ S a~ ~q~ U] u a' O s :^ ~ ~ ~ O 1 ~ O ~ O ~ ~ C ~ O O 4J ~ ~P ~ S ~ ~ ~ ' - O 4~ ° V 1 ~S ~5 0 u~ ~r tD U] 0 4 ~: s5 1 ~1 1 3 1 3 0 3 a' ~ O ~O O JJ O dP dP O dg dP ~ dP dP O ~ ~ ~ o ~ 0 ~ ~u, q~ co ~ oo q~ ~ ~s u: ~ ~ ~ ~ ~5 JJ11 11 C: 11 11 ~11 11 {: n u,0 ~0 P~ c: v . - ~s cn E ~LD ~_ u, u, U? ~r 4 Y ~u~ ~s:: 1 a) 1 . ~ O 0 3 ' - q ~ ~ ~I dP ~ O tU 05 a ~tI5 CO + Q) + ~S JU O V V q ~- - O ~ ~ ~ ~ ~ ~ O 0 a) a n ~ n ~ n ~ u' ' O ~O 0 ~ ~ -- n -- 0 ~ a' 0 rs ~ - ·~ u~ 0 ~ u~ v ~ a' ~ :: ~ ~Q) u, ~ ~ ~ ~ O ~ ~a) 4 0 ~ r~ ~5 ~ u~ ~ ~ ~ O ~ ~ ~V o ~ a' ~ ~ ~ u~ ~C · q~ ~ ~ ~ ~ 0 - ~ V ~ v ~ tl~ + a~ - u~ ~ 0 ~ a u~ u~ ~ a, 0 ~5 C: ~ a~ u ~ ~ ~ 4J ~ m~ u~n~ a' ~ ~ ~ ~ ~ 3 a' c ~ ~ ~ ~ 1 ~ V ~ V ~ a' Q) a' ~ c~ v ~ ~ v ~ ~ m-~ ~ ~ v v O ~ ~O ~O H H (,) P:; ~H Z U] ' 4~ V 5 ~ S U) O ~ Q4 ~S v ~ 0 3 ~ c: O s ~ 4U 0 O ~ on O Ul ~ ~ O U) ·,1 Q, H) ns C: ~ 3 ~ U, ~ O Q] C: ~ ~ ~ ~ ~ '- ~ ~ O O ~ ~ S ~O ~ v ~ ~ 3 ~v ~ ~n n ol ~ uc ~ P. O ~ ~ ~ ~ O ~ ~ 0 s ~ O ~ ~s 0 s v ~ ~ ~ c v ~ v a) q~ ~ a' ~ ~ O? X: ~ 0 0 0 0 h - a' ~ 3 ~r O 0 4~ ~0 0 ~ 0 0 a) ~ q~ r~ ~Q a, 1 . - a' ~, u, s ~ ~ ~ 4~ ~ ~ s ~ - - ~ u, 0 0 ~ ~ a c' ~ ~ v n 3 H.- 0 a Z U] o r~ a, H
111 PROGRAM FOR OUT-OF-SCHOOL YOUTHS Job Corps The Job Corps is in many respects unique. It is distinguished by the population it serves, the comprehensive nature of the services it offers, its stability as a program, and the quality of the evaluation that is available on it. We note that these last two points are probably not unrelated. The Job Corps is a largely residential program for out-of-school, economically disadvantaged youths between 14 and 21 years old. It provides a range of services including remedial (basic) education, vocational skills training, and health care to enrollees for about 30 weeks (the average stay during the subject evaluation). The Job Corps serves a severely disadvantaged population: about 90 percent of Job Corps enrollees were either from households below the poverty line or receiving welfare benefits; more than 75 percent were minorities, and 30 percent were female. Moreover, despite the fact that the median age of Job Corps enrollees was about 18, median reading levels were at or below the sixth-grade level. The Job Corps has existed for 20 years; few programs have had such stability. The program served about 102,000 youths in fiscal 1985 in 41,000 slots; enrollees averaged just under 5 months participation. At the time of the evaluation we reviewed, about 70,000 participants were being served per year. Although the Job Corps has been substantially modified since it was first established in 1964, most evaluations of the program prior to the 1982 study we reviewed were based on the experience of those who par- ticipated in the Job Corps during the mid-1960s. A series of surveys by Louis Harris and Associates served as the primary data source for researchers attempting to estimate the impact of Job Corps. These early studies had conflicting findings. For example, one study (Cain, 1968) found that participants earned $188 to $260 per year more than "no-shows" (those who enrolled but never participated) 6 months post- program. Another study (Woltman and Walton, 1968) found no significant difference between the earnings of Job Corps enrollees and early terminees (those who remained in the program less than 3 months) 18 months after participation. Taken together, these early findings suggested that Job Corps had a short-term impact that decayed (faded) fairly quickly (Goldstein, 1973~. Our assessment of the Job Corps is based on an evaluation conducted by Mathematica Policy Research, Inc. (Mallar et al., 1982~. This evaluation was the most extensive and sophisticated of the studies of the Job Corps undertaken over the years. And, unlike most evaluations of other youth employment programs that we reviewed, this study: · was based on a large sample of program participants (2,800) and nonparticipants (1,000) who were similar in most respects to Job Corps participants. The nonparticipants were youths eligible for Job Corps who were residing in geographic areas where Job Corps enrollment was low.
112 · gathered data on the participant and comparison groups for a reasonably long time after the program so that it was possible to establish the degree to which postprogram effects exist and persist or decay. The third follow-up interview was conducted 42-54 months after the program period. · had low rates of attrition in the follow-up samples of participant and comparison group members. The third follow-up survey was completed by 70 percent of those who completed the original baseline questionnaire (65 percent of participants and 75 percent of comparison group members). · took measurements on a wide variety of factors that could be affected by, or affect, the Job Corps experience, including educational attainment, the value of economic production by Job Corps participants, receipt of welfare and other transfers, the extent of criminal activity, unemployment rates, employment rates, hours worked, and wage rates. · used a comparison group methodology in a way that was as careful and technically sound as the state of the art permitted. The essential finding of the Mathematica evaluation is that the Job Corps "works." In particular: · On average, participants in the Job Corps were employed about 3 weeks per year (13 percent) more than nonparticipants up to 3-1/2 years postprogram, and their earnings gains after leaving the Job Corps were estimated to be $567 per year higher in 1977 dollars (28 percent) for enrollees than they would have been in the absence of the Job Corps experience. The amount of time that Job Corps enrollees received cash welfare or unemployment compensation benefits was lower by 2 weeks per year and 1 week per year, respectively, compared with nonparticipants. Table 5.3 presents evaluation results up to 4 years after Job Corps participation. · Participants' educational attainment increased substantially while they were in the Job Corps: the probability that enrollees would receive a high school diploma or its equivalent (General Equivalency Diploma) within the first six months after leaving the Job Corps was .24 for enrollees compared with .05 for comparison group members. · Overall, the health of Job Corps participants was better than that of nonparticipants after the program; participants reported about 1 week less per year of serious health problems. · Criminal activity, as indicated by rates of arrest, was significantly lower for participants during the period of the program, and after leaving the program participants had fewer arrests for serious crimes than nonparticipants. · After an initial 6-month postprogram period, when enrollees fared worse than the comparison group in terms of employment and earnings, the aggregate positive effects of Job Corps emerged and persisted at a relatively stable rate throughout the 4-year follow-up period. This outcome suggests that the main effects of Job Corps do not stem from job placement.
113 TABLE 5.3 Estimated Job Corps Effects on Employment and Earnings, Including Military Sector: First Through Fourth Postprogram Years Months After Employed Weeks Worked Hours Worked Weekly Earnings Termination (fraction of time) per Six Months per Week (1977 dollars) 0-6 6-12 12-18 18-24 24-30 30-36 36-42 42-48 -0.018 0.070* 0.113* 0.081* 0.081* 0.075* 0.068* 0.040 -0.47 1.82* 2.94* 2.11* 2.11* 1.95* 1.77* 1.04 0.79 3.19* 5.19* 3.76* 3.44* 3.35* 3.20* 1.62 -0.21 10.27* 15.64* 9.42* 10.20* 11.59* 10.12* 5.47 *Statistically significant at the .05 level or below in a two-tailed test. SOURCE: Data from Mallar et al. (1982:Tables IV.10 and IV.ll). These overall effectiveness estimates included all participants, early leavers as well as those who followed the Job Corps course to completion. The study also provides estimates of differences in effects according to categories of program completion. Program completers composed 40 percent of the sample, while partial completers and early dropouts each accounted for 30 percent. Program completers benefited most. Partial program completers, those who stayed at least 90 days and completed at least one specific segment of a vocational or educa- tional program but not an entire program, benefited about one-third as much as completers. Early dropouts were found to benefit little or not at all. The authors note, however, that econometric methods for con- trolling for bias in selection into the three completion categories would not prove effective. Program effects were estimated separately for males (representing 70 percent of corps members and 70 percent of the follow-up sample), females without children (21 percent of the follow-up sample), and females with children present (9 percent of the follow-up sample). The estimated effects on employment and earnings are similar in magnitude for males and females without children, though somewhat more erratic over the postprogram period for females. For females with children present, the employment and earnings effects are both lower and more erratic than for the other two groups. The authors hypothesize that the lower effect for women with children may be due to the higher proportion of very young children among the participants than among the comparisons. This difference is apparently due to delays in childbearing by participants during the in-program period, followed by a higher rate of postprogram births than among comparison group members. The result is the presence of a greater number of young children among participants in the postprogram period.
114 Effects of the program on crime are estimated from self-reported arrest data gathered in interviews with participants and comparison group members. The major effects are estimated to occur during the period in which participants are in the program, when total arrests and incarcerations were significantly lower for Job Corps members than for the comparison group. In the postprogram period, the estimates show an overall reduction in arrests (statistically significant for males), a reduction in theft arrests, an increase in auto-related arrests, and no effect on time in jail. When the benefits and costs of the program were estimated--in a quite detailed and sophisticated benefit-cost analysis--the study found that from the view of society as a whole, the net present value of benefits exceeded costs by $2,300 per enrollee (in 1977 dollars). From the view of the participants, benefits exceeded costs by $2,400 on average. For nonparticipants (i.e., private benefits and costs), a net cost of $115 per enrollee was incurred, representing a net redistribu- tion of resources from nonparticipants to Job Corps participants. The estimated magnitude of the benefit-cost difference is particularly sensitive to the assumptions regarding the magnitude of the effect of the program in reducing crime. The evaluation assumes that actual arrests were underreported by 70 percent among members of the Job Corps sample; this assumption is based on a study done for the evaluation of the Supported Work program. Considerable attention has been devoted to the issue of the correlation between self-reports and official reports of criminality and arrests in the criminal justice field. There is no generally accepted differential between self-reported and official data on criminality that supports the use of any given numerical factor to increase the self-reported incidence of arrests.) However, even when it is assumed that no postprogram crime-reduction benefits are associated with Job Corps, the net present value of the program to society is still positive, about $500 per enrollee. The Job Corps evaluation was extensively reviewed by outside experts in 1982 at the request of the Office of Management and Budget The reviewers did not find any major problems, though one had some doubts about the adequacy of the selection bias corrections. We have some remaining reservations about the Job Corps evaluation that are largely technical in nature. Random assignment to the Job Corps and to a control group was ruled out by the Department of Labor at the outset. Given that constraint, the comparison group strategy seems to have been well conceived and, HA recent summary of research on the use of self-reported measures of delinquency indicates widely varying estimates of underreporting depending on method of administration (questionnaire or interview) and sample characteristics (sex, race, socioeconomic background, school status, and previous contact with police). Underreporting appears most serious among black males with previous delinquent offenses (Hindelang et al., 1981~.
115 for the most part, well executed. Comparison group members were drawn from geographical areas that were similar to the home areas of Job Corps members but that had low rates of previous enrollment in the Job Corps. Within these areas, sample members were drawn with selection probabilities in proportion to their similarity to Job Corps partici- pants in terms of age, poverty status, race, and education. Beyond controlling for measured characteristics when estimating effects, the evaluation attempted to control for selection bias by modeling the selection process using the methods suggested by Heckman (1979~. While the selection bias correction appears to have worked well, more detailed information than that available in the report is required to remove some residual doubts as to whether this correction dealt completely with self-selection problems. The efforts at correc- tion go considerably beyond those usually applied when dealing with comparison groups (rather than randomly assigned controls groupsJ, but the nature of the comparison group renders the evidence of program effects less convincing than it would be had a randomly assigned control group methodology been used. The Job Corps provides a comprehensive set of services and whether the comprehensiveness is central to the effectiveness of the program has not been directly evaluated. Apparently, earlier reports did provide some estimates of the difference in impact according to the members' exposure to particular components, but those results were not reported in the Mathematica study and mention is made of selection bias problems in making assessments. Since participants are not randomly assigned to the various components, self-selection factors seem likely to be confounded with the actual effects of the component in which a participant is enrolled. We do not have sufficiently detailed evidence that allows us to isolate the elements of the Job Corps program and determine whether (or for whom) the residential element of the program is critical; whether the health component is essential; or whether the skills training offered adds to any effects that the basic education elements may have created--or vice versa. Since women represent about 30 percent of Job Corps enrollees, the desire to obtain reliable estimates by sex led to the selection of a comparison group that was 50 percent female. Unfortunately, it appears that the enrollee and comparison groups for women were not adequately matched on child-responsibility status. According to the final follow-up report, almost none of the female corps member sample initially had children present, but by the time of the final follow-up interview approximately 50 percent of the women had children present. Thus, female Job Corps members with children represented about 2 percent of the sample in the first 6 months of the postprogram period and 15 percent of the sample during the last 6 months of the study period (42-48 months after leaving the program). In fact, baseline data from an earlier report indicate significant differences between participant and comparison group members, particu- larly among females (Kerachsky and Mallar, 1977~. Approximately 3 percent of the female participants were pregnant at the baseline com- pared with 12 percent of comparison group females. Female participants were significantly more likely to be black, Hispanic, or members of
116 other minority groups than females in the comparison group (84 percent compared with 62 percent) and also more likely to never have been married (94 percent compared with 73 percent). The Job Corps studies are among the few we reviewed that did not present data on the characteristics of participants and comparison group members at various stages of the evaluation. Therefore, we are not able to determine the levels of childbearing among comparison group women over the study period. This makes the fertility and family- formation outcomes of Job Corps particularly difficult to interpret: since the participant and comparison groups were apparently not initially comparable, later differences may indicate the presence of self-selection bias. Other studies show that more highly motivated women tend to postpone childbearing and marriage and that the presence of children inhibits program participation. The evaluation design used for the Job Corps does not allow one to determine whether Job Corps participation actually induced delays in childbearing and family formation (see Margaret Simms, in this volume). The differences in measured characteristics between Job Corps participants and comparison groups members suggest that there may also be differences in unmeasured characteristics (e.g., motivation and aspirations). The possibility of self-selection into the program is a strong argument for the use of a random assignment experimental design, since statistical techniques may not adequately control for the factors that determine program entry and postprogram outcomes. PROGRAM FOR IN-SCHOOL YOUTHS New Youth Initiatives in Apprenticeship The New Youth Initiatives in Apprenticeship program sought to pro- mote the use of registered apprenticeship positions, outside the construction trades, by developing linkages between the schools and employers with registered apprenticeship positions. Employers were provided subsidies for one-half of apprentice wages, up to a maximum of $1,700 per student apprentice year. The program began operation in four sites in 1977 under the sponsor- ship of the Department of Labor's Bureau of Apprenticeship Training; one of those sites operated for only 1 year. Four additional sites were funded by the Office of Youth Programs in 1978, bringing the total number of sites operating in 1978 to seven. According to the evaluation (Williams et al., 1981:6~: "The new [OYP] projects included targeting economically disadvantaged students to participate as student appren- tices, an activity not specifically mandated in the original demonstra- tion effort. n Despite this, the data indicate that the newer projects may have been less successful in enrolling minorities than the more established projects. The New Youth Initiatives in Apprenticeship program was reviewed in two reports. The report by Williams et al. was more comprehensive and competent, and we based our assessment on that report alone. The follow-up analysis was based on samples of about 600 student
117 apprentices and 500 matched comparison group members. The data on the student apprentices indicate that they were generally nonminority (80 percent) and male (89 percent) and had a grade average of B- in high school and a high rate of graduation from high school (96 percent). This program does not appear, therefore, to have reached the heavily disadvantaged segment of the youth population. The results of the participant-comparison group contrasts show, on average, small and not statistically significant differences in annual earnings ($290 above the $10,000 annual average) or wage rates. Apprentices tended to be more concentrated in machine trades occupations and comparison group members in clerical and sales occupations. The evaluation study also surveyed several hundred employers of the student apprentices in the eight sites where the program operated. The employer survey indicated that three-fourths of cooperating employers were small businesses, employing fewer than 50 workers. Employers appeared to be more attracted by the screening and training functions performed by the program than by the wage subsidies provided. Employers were predominantly in manufacturing (44 percent) and services (38 percent); only 10 percent had unionized work forces. A multivariate analysis indicated that the number of apprentices employed in the postprogram period was positively and significantly related to four factors: the total number of employees; being a manufacturing firm; being a union firm; and the number of years the program had been in operation in the site. Though the evidence was ambiguous, it suggested that those employers who rated stipends as important retained fewer apprentices. Two aspects of the research design seem worthy of note. First, all of the postprogram interviews were conducted in the last six months of 1980, which provided only 6 months of post-high school data for the 40 percent of the sample from the new sites and, at most, 1-1/2 years of post-high school information for 14 percent of the group. With such a heterogeneous group and relatively short follow-up period, one cannot be sure whether sufficient time had elapsed for program effects to emerge. Second, the comparison group sample was drawn (after the program began) from the same high schools as the participants. It is natural to question whether there is some selection bias--despite matching on characteristics--since the members of the comparison group presumably either had an opportunity to join the apprenticeship program and did not do so or were specifically not chosen to participate in the program. Since so few significant findings emerged from the participant-comparison contrasts, we did not pursue this issue further. Because of the characteristics of the participants we cannot determine whether this type of program might be more effective among more disadvantaged youths. In addition, it is difficult to tell whether the failure to enroll significant numbers of more disadvantaged youths is inherent in the nature of the program or simply due to the characteristics of the sites where it was implemented. Given the character of the participant population and the nature of the program, no generalizations can be drawn from the evaluation of the
118 New Youth Initiatives in Apprenticeship program to youth employment and training efforts generally. CONCLUSIONS While it would be misleading to attribute the Job Corps results to skills training efforts in general, the measured effects of Job Corps indicate that effective skills training can be provided for economically disadvantaged youths. At the same time it is clear on the basis of both Job Corps and the less effective New Youth Initiatives in Apprenticeship program that the staff capacity and other resources needed to mount skills training efforts are not acquired quickly or Inexpensively.