The design of LTC RAPs is decentralized with individual employers given great discretion over what goals they attempt to achieve and the critical elements of the training program, including the curriculum, length of the apprenticeship, and types of training provided. LTC RAPs include four different occupations, with different training requirements both across occupations and within occupations. A key issue is whether the sites program goals and the interventions they administer are uniform enough that the program can be evaluated as a whole. Successful evaluations of multi-site programs require that the goals and the activities of the different sites be at least roughly the same; if they are not, then it is not clear what intervention is being evaluated.
In the RTI International/Urban Institute analysis of the administrative data and in our site visits, we found that while there was variation across sites, there was enough uniformity across sites in our judgment to evaluate the program as a whole (Anderson et al., 2010; Kuehn et al., 2011 ). For example, in all of the sites visited, the goals were to improve the long-term care workforce in order to improve quality of care and to create more attractive jobs for apprentices. These goals help sponsors to meet state certification requirements, reduce errors in caregiving, reduce turnover, and create career opportunities for apprentices.
Moreover, while the length and content of the components of the LTC RAP programs varied, they all included at least the basic structure of the apprenticeship programs -- OJT, related instruction, peer-mentorship, and a wage increase upon the successful completion of the program. While the distinction between routine supervision and OJT was sometimes unclear, all of the training programs visited involved substantially greater levels of training than is typically required by federal and state regulation. Although one site used its LTC RAP for entry-level training of all new employees, most sites used the LTC RAP for advanced training and mentoring of employees who had already received basic training and had leadership or other personal qualities that management wished to develop.
There are, however, substantial differences in the content of the training provided and modalities of delivery across occupations in terms of the skills learned to meet the needs of different service settings and populations. Different settings also have an impact on how some of the training can be administered. For example, while CNAs work in a nursing home where supervisors are readily available, HHAs provide services in the homes of individuals, usually without direct supervision. Moreover, there may be a substantial difference between the program described in the agreement with DOL and the program that is actually implemented. As a result, other observers may judge the programs too heterogeneous to analyze as a whole; with that judgment, separate evaluations would have to be done of each occupation, which would have a major impact on the sample sizes potentially available for analysis. One partial solution to this problem is to use the four occupations as control variables in the multivariate analyses.