Abstract
{\it Objective:} Determine whether the number of job coaches provided by local nonprofit disability service providers (``boards'') significantly affected the employment outcome for individuals with mental retardation (MR) in South Carolina in 1997.
{\it Study Design:} Health, demographic and IQ information for individuals with MR (IQ 20--74) were matched with employment information derived from written surveys supplied by the boards. The study group comprised 431 employed and 6659 unemployed individuals. Logistic regression was used to assess the relationship between employment outcome and the number of job coaches per individual served.
{\it Results:} Job coach numbers per board varied from 0.00 to 2.48 job coaches per 100 individuals, with mean and median values of 0.97 and 0.87 coaches per 100 individuals. The effect of coaches on employment likelihood was approximately two times greater for individuals with low (IQ 20--39) than for high IQ (IQ 40--74), approximately three times greater for individuals in counties with low (3--6%) or intermediate (6--9%) unemployment and approximately ten times greater for individuals located in highly (>75%) urbanized counties. The likelihood of employment given the addition of one job coach per 100 individuals increased by factors of 0.37 to 2.49 in rural or moderately urban areas and by 3.79 to 25.70 in highly urban areas.
{\it Conclusions:} These results suggest that expenditures on job coaches are effective, and that such expenditures may be most effective if applied in urban counties with low or intermediate unemployment rates.