Findings in Brief

• Implementation of the CET approach is difficult, and the fidelity of program services to the original CET model varied greatly across the sites, affecting both implementation and program impacts.



• In the four sites implementing the CET model with high fidelity, access to the program increased youths' participation in training activities substantially above the level for the control group and increased the percentage of youth completing a training certificate. In the medium- and low-fidelity sites, the impacts on service receipt and completion were smaller.



• In the high-fidelity sites, access to the program produced substantial positive impacts across a range of employment-related outcomes for young women. These included impacts on the percentage ever working, on employment rates at the follow-up survey, and (quite probably) on total earnings -- though the small sample prevents a statistically significant finding.



• For young men, the results in high-fidelity sites were either negative or negligible. As in the past, it has proved difficult to improve the employment prospects of low-income young men.



• In the medium- and low-fidelity sites, impacts were either negative or negligible across a range of outcomes for the full sample and for all key subgroups.



• These findings are similar to prior results from CET-San Jose, where implementation was very strong and the research findings were largely driven by positive impacts for young women. The employment impacts reported here occurred during a period of strong demand for labor that provided unusual job opportunities for youth with low skills, so the longer-term follow-up that is now under way is important because it will extend into the recent economic slowdown.

Implementation of the CET Model in the Study Sites

• The CET approach is difficult to implement - only four of the twelve study sites put all the key aspects of the model in place - and sites with greater fidelity of implementation produced stronger impact findings.

Implementation of the model was strongest among four of the established western sites that were part of the network of programs that CET developed and operated as it gradually expanded its operations. These high-fidelity sites were able to put in place all the key aspects of the program. Other sites that were newly established, that were operated by organizations other than CET, or that shared both characteristics had much more difficulty implementing the full model. Six sites implemented it with medium fidelity, and two did so with low fidelity. Because of their location in the western states, the four high-fidelity sites served many Hispanic clients. An analysis of site impacts that controlled for the characteristics of each site's sample revealed that differences in findings across sites were largely driven by fidelity of program implementation rather than by differences in sample characteristics. Thus, much of the analysis focuses on grouping the sites by the strength of their implementation.

The weaker-implementing sites tended to lack intensive participation in training and strong organizational stability. CET offers full-time services and seeks to involve participants intensely in education and training to quickly prepare them for work. Only the high-fidelity sites were able to generate the intense participation called for in the CET approach. Successful implementation also requires the sustained attention of leaders committed to the approach. This was present in the high-fidelity sites, but the medium- and low-fidelity sites tended to have turnover in leadership, funding changes that led them to depart from the CET approach, or even closure of the programs because of administrative problems or changing priorities.

• In the high-fidelity sites, access to CET significantly increased participation in skills training during the early months of follow-up.

In the high-fidelity sites, survey respondents in the program group reported an average of 218 hours of skills training in the first six months of follow-up (including zeros for those who reported no participation in skills training), compared with 36 hours for control group members -- an impact of 183 hours. In the medium- and low-fidelity sites, the comparable average increases were only 62 hours and 20 hours, respectively.

• Access to CET significantly increased receipt of training credentials, with the largest increase occurring in the high-fidelity sites.

By the end of the follow-up period, 39 percent of program group members reported receiving a training credential, compared with 25 percent of control group members, for a difference of 15 percentage points. In the high-fidelity sites, the difference in credential receipt was 24.9 percentage points; smaller impacts were found in the medium- and low-fidelity sites.

Impacts on Employment and Other Outcomes

• The period covered in this report was one of strong demand for low-skilled labor, which led to unusually favorable employment outcomes for members of the control group.

Youth with low skills had better job prospects within the strong economy of the mid to late 1990s than in other recent periods. Thus, to produce positive employment impacts, the program faced an unusually high benchmark of employment among control group members. Although youth typically apply for a training program like CET when they are unemployed, nearly 20 percent of the control group reported employment in the month following their entry into the research sample. Over the follow-up period, this employment rate gradually rose; in the final month of follow-up, more than 60 percent of the control group -- and about 55 percent of those without a high school education -- reported employment. As a further illustration, in the high-fidelity sites, 83 percent of women in the control group and 100 percent of men reported working at some point during the follow-up period.

• In the high-fidelity sites, the CET model led to a substantial increase in women's employment and earnings and to an increase in marriage, childbearing, and arrests for younger sample members, but it had negative impacts on young men's employment and earnings.

Given the strong labor demand reported above, the program's employment impacts are especially interesting. For women in high-fidelity sites, the program increased the percentage who ever worked during the follow-up period (from 83 percent to 92 percent) and who were working at the time of the follow-up survey (from 47 percent to 61 percent). In addition, it appeared to increase earnings in the last six months of follow-up, from $3,610 to $4,228, and earnings over the entire period, from $12,325 to $14,094 -- but neither of these results was statistically significant, given the relatively small sample size. For men, access to CET led to a decline in the percentage who ever worked in the follow-up period (from 100 percent to 93 percent) and to a decrease in earnings in the last six months (from $6,391 to $4,954).

These different results by gender are related to the higher levels of control group employment for men and to shifts in participants' industry and occupation that the program produced for each gender. For women, the CET model produced a shift away from retail trade and toward other industries (especially transportation) and a shift away from service occupations to clerical, with an increase in the percentage of women earning $7 to $9 per hour. For men, access to CET led to shifts into the construction and manufacturing industries, without an accompanying move into the occupations for which they were trained. In combination, these shifts produced a slight (statistically insignificant) increase in hourly earnings but a decline in average hours worked (largely because of a decline in the percentage of men working more than 40 hours per week). It appears that although the training may have induced shifts in industries, the new skills did not position the men to maintain their work hours over time in these more skill-intensive settings.

• In the medium- and low-fidelity sites, impacts were either negligible or negative for both employment-related and other outcomes.

Most impacts in the lower-fidelity sites were not statistically significant. But access to CET reduced the employment rates and the third-year earnings of youth who lacked a high school diploma at program entry. Further, the programs in these sites also reduced the employment rates of women and program entrants older than 19 and reduced the third-year earnings of teenage entrants. These negative impacts highlight the importance of implementing the program strongly.

Implications

The findings in this report clearly highlight the importance of implementing the CET approach strongly; the sites showing high fidelity to the CET model produced more substantial impacts. The results at this stage for young women appear similar to those in earlier studies: Where CET is implemented well, there are positive impacts on employment-related outcomes. Unfortunately, the current findings also continue a pattern of disappointing results for young men within employment and training evaluations. In part, this pattern appears to be linked to young men's relatively high employment rates in the absence of special training and to the inability of training to shift them into more stable and higher-paying jobs. The findings could reflect the short duration of the training and the reluctance of employers to hire young men (especially those of color) for more demanding, more rewarding jobs.

It is also important to remember the context in which these findings rest: The very strong demand for low-skilled workers boosted the employment and earnings of the control group to higher levels than in past studies, making it more difficult for the program to produce positive impacts. The longer-term follow-up that is now under way will extend into the current economic slowdown, providing an opportunity to see whether the enhanced skills that are produced by the CET model will help young people to maintain their employment and earnings better than their counterparts who do not have access to this program.

Notes