Gannon Brenda
Department of Economics, Irish Centre for Social Gerontology, National University of Ireland, Galway, Ireland.
Health Econ. 2009 Jul;18(7):743-59. doi: 10.1002/hec.1399.
Self-reported disability status is often relied upon in labour force participation models, but this may be reported with error for economic or psychological reasons and can lead to a bias in the effect of disability on participation. In this paper, we explore the possibility that reported limitations in daily activities are mis-reported, in particular for those who define their labour force status as disabled/ill, and assess if economic incentives influence this group to mis-report. The main questions we wish to address therefore are: (1) was there state-dependent reporting error and did economic incentives play a role, and (2) did this change over the years 1995-2001? Using a generalised ordered response model, we compute cleansed measures of disability that correspond to predicted responses individuals would have made if employed. Unobserved differences between the employed and non-employed may exist; therefore, we control for this via correlated random effects. The results indicate that the disabled/ill group did over-report and the difference between actual and predicted probabilities only marginally changed between 1995 and 2001. The extent of this measurement error is lower once we control for unobserved heterogeneity.
劳动力参与模型中常常依赖自我报告的残疾状况，但由于经济或心理原因，这种报告可能存在误差，进而可能导致残疾对参与度影响的偏差。在本文中，我们探讨了日常活动报告限制被误报的可能性，特别是对于那些将自己的劳动力状况定义为残疾/患病的人，并评估经济激励措施是否会影响这一群体误报。因此，我们希望解决的主要问题是：（1）是否存在与状态相关的报告误差，经济激励措施是否起了作用；（2）在1995年至2001年期间这种情况是否发生了变化？我们使用广义有序响应模型，计算出与个体若受雇会给出的预测响应相对应的经过修正的残疾衡量指标。就业者和非就业者之间可能存在未观察到的差异；因此，我们通过相关随机效应来对此加以控制。结果表明，残疾/患病群体确实存在过度报告的情况，实际概率与预测概率之间的差异在1995年至2001年间仅略有变化。一旦我们控制了未观察到的异质性，这种测量误差的程度就会降低。
