9709635 Philipson Much of positive economic analysis relies on data that are obtained through surveys which record the behavior or experiences of included sample members or through social experiments in which treatments are assigned by investigators. However, although the production of data, and the functioning of the market for observations in which it takes place is a concern universal to all fields of positive economics, greater emphasis and sophistication in systematic analysis has been paid by economists to the consumption of data, rather than its production. This project will seek to remedy this unbalanced focus by studying incentives used in data production to facilitate identification and estimation of population parameters of economic interest. In particular, this project will utilize a labor economic approach to data production. In the labor market for observations, the supply of observations by sample members determines both the quantity and quality of data analyzed. The project will develop wage discrimination schemes that may be used on the sample for identification of population parameters to be estimated. Such schemes will allow for correcting estimates for production biases due to poor quantity or quality of data. The project relies on the idea that if wage discrimination is random within a survey, then it alters the supply behavior of sample members across the same types of populations. This in turn allows for separating out poor supply behavior from the population parameters of ultimate interest. The project seeks funds for both theoretical work and empirical analysis of a data set produced by the principal investigator. These data include the implementation of the random wage discrimination discussed in the theoretical analysis as an experimental module on a national survey of US physicians produced by The National Opinions Research center (NORC) at the University of Chicago.
