Measuring Preferences in Social Surveys

点击翻译按钮获取中文摘要

DESCRIPTION (provided by applicant): Innovative methods for the measurement of preferences of individuals in population-based socio-economic surveys will be developed, tested and implemented in a longitudinal household-based survey. The focus will be on inter-personal preferences, attitudes towards risk and inter-temporal preferences. These domains of preferences play a key role in many models of individual and family behavior, yet they are seldom explicitly measured in socio-economic surveys. This project will make progress towards filling that gap. Drawing on methods that have been successfully applied in the experimental social science literature, study subjects will participate in carefully-controlled, well-designed behavioral experiments. Subjects are paid based on their decisions, making it costly for them to misrepresent their preferences. These measures of preferences will be carefully validated and will serve as a "gold standard". The same methodology will be adapted to a household-based setting, extensively tested and fielded in the second round of a household survey to measure the same 3 classes of preferences. The measures from the survey will be compared with the "gold standard". The baseline household survey will collect information on the same preference domains from the same respondents using hypothetical questions and well-developed survey instruments. These preferences measures will be compared with the two sets of experimental-based measures. All the measures of preferences will be validated against predictions about behaviors of individuals and families drawing on the richness of information collected in two rounds of the household survey. Experimentally induced variation in individual resources within households will shed additional light on the contributions of these measures of preferences to understanding individual and family behavior. The project will, thereby, provide a thorough evaluation of the feasibility, costs and benefits of including experimental based measures of preferences in future population-based household surveys.