Classification-Error Models and Labor-Market Dynamics
比较了两种修正劳动力市场流动数据中测量误差的方法，发现它们调整方向相反，并指出其中一种方法依赖的理论假设不现实且被数据拒绝。
Measurement error is often a severe problem in longitudinal data such as labor-market flows. Two correction methods have been used for labor-market flows, but the methods tend to adjust the data in opposite directions. One method shows a rigid labor market with long unemployment spells, whereas the other shows the opposite. Because the methods appear theoretically similar, one might interpret the differing results as an indictment of both procedures. This article documents and explains the differences between the two methods. More important, it argues that the differences result from the reliance of one method on assumptions that are theoretically implausible and rejected by the data.