(Page 9 of 16) - Faith-based Providers of Urban Social Services: A Story of Niche Effectiveness authored by Monsma, Stephen.

impact each of these factors, or variables, had on the employment status of the respondents, if all
of the other variables are held constant. Thus, for example, it answers the question of what was
the independent impact on employment of having taken part in a faith-based welfare-to-work
program if other variables, such as education, race, and gender are the same.

The results of this analysis are reported in Table 4. We considered 30 hours or more of

work a week as full-time employment. At six months there were four statistically significant
relationships. Being female and being unemployed at the baseline significantly decreased one’s
chances of finding full-time employment, and being unmarried significantly increased the
chances of one doing so. Most interestingly for our purposes here, being in a for-profit program
significantly increased one’s chances of finding full-time employment (as compared to being in a
government program), even when all the other variables listed in Table 5 are held constant. Both
types of faith-based programs had positive coefficient scores when compared to the government
programs, but they were too small to say with confidence whether or not these scores resulted
merely by chance. One cannot say with confidence that their clients were doing better at full-
time employment than were the government programs. The same can be said for the
nonprofit/secular programs, although their coefficient scores were ever lower than those of the
faith-based programs. It is also clear that it would have been better to have had a larger number
of respondents than we had in this study. Achieving statistical significance is a function of both
the observed patterns and the number of respondents included in one’s study. Even when the
observed patterns do not change, they are less and less likely to be have been caused by chance
as the numbers of respondents included in the study increase.

TABLE 4 GOES HERE

Table 4 shows there were some changes at twelve months. Being female or unemployed

at the baseline no longer decreased one’s chances of finding employment, while being unmarried
continued to increase one’s chances of finding a job. Meanwhile, being in a for-profit program
continued significantly to increase one’s chances of finding full-time employment, as it did at six
months. Both types of faith-based programs continued to have positive coefficient scores; both
continued to fail to reach statistical significance.

These findings demonstrate the effectiveness of the for-profit programs in enabling their

clients to find full-time employment. At both the six month and twelve month interviews, even
when one holds constant gender, race, baseline employment status, work-related skills (including
education), marital status, and the number of dependent children, the for-profit program clients
did significantly better than the government program clients in finding full-time employment.
This was true of no other type of program, including the faith-based programs.

From our knowledge of these 17 programs three possible reasons suggest themselves for

this for-profit program success. One is the focused, job-training emphasis of the for-profit
programs included in this study. The for-profit programs were vocational education programs,
and although they had some supportive services in addition to the job skills classes, the emphasis
was clearly on the hard skills training. Also, both had internship programs, which in themselves
may have been a factor in their clients’ success in finding full-time employment. Second, the
for-profit programs clearly felt themselves under the gun and in need of producing positive
results. The use of for-profit firms to provide welfare-to-work services in Los Angeles County
had been highly controversial and accompanied by high-profile newspaper stories.

As a result