As preparation for a national probability sample survey focused on internal migration in the People's Republic of China using either a one-time design or a panel design, a pilot project is proposed to test the feasibility of alternative sampling, study design, and questionnaire construction strategies. A sample of 1,550 households, drawn from 31 sites selected to represent villages, towns, and cities with high and low rates of in-migration and out-migration and with high and low average income will be interviewed, using a questionnaire that solicits complete migration histories from respondents and other household members, including place moved to, date of move, and reasons for move. Also included will be standard sociodemographic questions, health measurements, and measures of personal characteristics such as risk-taking propensities and resourcefulness, that are likely to affect both the propensity to migrate and migration outcomes. The main difficulty in studying migration in China is to secure complete coverage of the population. Most Chinese surveys rely on population registers, but these tend to under-count migrants, especially short term migrants. Two main strategies to locate migrants will be tested: supplementing registers with information from local officials, and doing complete household listings. Encounter sampling will be used to check the adequacy of these strategies. The feasibility of locating and interviewing immigrants using information provided from household members at the place where they are permanently registered will also be assessed, both to secure a check on the adequacy of proxy reports regarding the characteristics of migrants, and to assess the likelihood of being able to locate individuals who have moved in the event that a panel design proves possible. In this same vein, proxy reports by parents on migrants' premigration characteristics will be evaluated. Finally, alternative data collection formats will be assessed, to determine the most effective ways of soliciting complex migration histories and other information, particularly health measurements, in the course of a personal interview.
