Loprest, 2001, for a review of leaver study surveys). There has been very little exploration into the quality of data from these studies: for example, whether the surveys are representative of their universe and how different measures of key variables, such as employment and wages, compare with measures from administrative data sources. For studies of formally and informally diverted populations (see Chapter 2), data issues for both survey data and administrative data are in some ways likely to be more severe. Sometimes very little information is gathered for cases that have been diverted, so there is very little information with which to track individuals in order to survey them. Some areas are using past participants in other social welfare programs who are not or have not participated in TANF to identify informally diverted cases. For these studies, there is the additional burden of finding these cases with what may be out-of-date information.
Surveys at the city, county, or local level are also being conducted as part of the Three-City Study, The Project on Devolution and Urban Change, and the Los Angeles Families and Neighborhood Survey. In a unique example of a national survey conducted at a state level, Iowa State University, in conjunction with the Census Bureau, conducted a modified SPD survey in Iowa (with modifications to questions that were of particular relevance to state policy makers) to explore the feasibility of conducting state-level surveys that could be integrated with the national-level surveys (Nusser et al., 2000). The state of California has also recently launched a large telephone survey to collect information about health and health care access. This survey will also collect information relevant to studying welfare policy.
Devolution has contributed to the growing demand for more localized data on low-income populations. States do not have a great deal of experience in sponsoring or conducting surveys, and thus far, data quality for some state-level surveys has been less than adequate. DHHS-ASPE has recognized the lack of experience for such surveys and has taken steps to help develop state-level capacity to conduct surveys or to manage surveys conducted through contractors. So far these efforts have been geared mostly towards those states and local areas that have grants to study those who leave or are diverted from cash assistance. For this group of states, ASPE has held conferences that provide technical assistance for conducting surveys and has hired a contractor with survey research expertise to provide technical assistance for these states. ASPE staff have also compiled information relevant to developing better surveys, such as survey instruments that include welfare-relevant questions and references to key survey methodology literature. Funding for further enhancements to surveys of welfare leavers and divertees for three states and two county groups with previous grants to track