Three simple tools to evaluate the quality of fertility estimates from birth histories. Application to African Demographic and Health Surveys

Bruno D. Schoumaker, Université Catholique de Louvain

In this paper, I present three simple tools that can be implemented easily to identify potential data quality problems in fertility estimates. The three tools rely on a simple idea: good quality data provide consistent estimates across surveys or across methods; in consequence, inconsistencies across surveys or across methods reflecting data quality issues. The three tools are based on this same idea, but use different methods, and as a result are sensitive to different types of data quality problems. Used together, these tools enable researchers to obtain a quick diagnostic of the quality of fertility estimates. Stata commands are provided and allow users to apply these tools in a user-friendly way. Simulations and Demographic and Health Surveys from a large number of countries are used to illustrate the tools in a variety of situations.