ABSTRACT
In this chapter, we look at sources of demographic data, and common errors and gaps. We then discuss the modelling choices that need to be made when faced with imperfect data.
We review traditional data sources, including vital registration systems, household registration systems, population censuses, and household surveys. We then turn to new sources of data that are generated as a side product of administrative and commercial processes. We look at the strengths and weaknesses of different data sources. We discuss attractive features of the aggregate-level datasets used in this book, compared to individual-level datasets. Aggregate data series raise fewer ethical and privacy issues. They can also generally extend much further into the past than individual-level series, and are found in poor countries as well as rich ones.
Most of the data that demographers work with in practice have measurement or coverage errors of varying levels of seriousness. Depending on whether one or multiple demographic series are modelled, and whether data are treated as having measurement errors or not, researchers and practitioners wishing to apply the methods from this book can choose from three frameworks.
