Sample Attrition and Labor Supply Response in Experimental Panel Data
利用西雅图和丹佛收入维持实验的数据，评估样本流失对劳动力供给反应估计的影响，发现流失偏差可能不严重，方法可推广至其他面板数据。
Sample attrition is a potentially serious problem for analysis of panel data, particularly experimental panel data. In this article, a variety of estimation procedures are used to assess the importance of attrition bias in labor supply response to the Seattle and Denver Income Maintenance Experiments (SIME/DIME). Data from Social Security Administration earnings records and the SIME/DIME public use file are used to test various hypotheses concerning attrition bias. The study differs from previous research in that data on both attriters and nonattriters are used to estimate the experimental labor supply response. Although not conclusive, the analysis suggests that attrition bias is probably not a serious enough problem in the SIME/DIME data to warrant extensive correction procedures. The methodology used in this study could be applied to other panel data sets.