Replication data for: Heckman selection model to correct HIV prevalence estimates for survey non participation

We recently published revised estimates for national HIV prevalence among adult men and women from the Zambia 2007 Demographic and Health Surveys (DHS) using Heckman-type selection models to correct for potential selection bias due to non-participation in HIV testing during the DHS survey. Our goal here is to briefly describe the rationale for the use of Heckman-type selection models in this context and provide annotated Stata code that can be used for conducting this type of analysis in the Zambia 2007 DHS and, after adaptation, in other datasets.

我们近期基于2007年赞比亚人口与健康调查（Demographic and Health Surveys, DHS）数据，采用Heckman型选择模型（Heckman-type selection models）对该调查中成年男女的全国HIV流行率估算值进行了修正，以校正因调查期间未参与HIV检测所引发的潜在选择偏倚。本文旨在简要说明在此研究场景下使用Heckman型选择模型的理论依据，并提供带注释的Stata代码：该代码既可用于开展2007年赞比亚DHS的此类分析，经适配调整后也可推广至其他数据集。