The ipums approach to harmonizing the world’s population census data

IPUMS integrates census microdata from around the world into one consistent database, harmonizing variable codes across countries over a 60 year period. The variables in the files provided by National Statistical Offices utilize a wide variety of classification schemes. To enable comparative research requires coping with this empirical reality of pre-classified microdata. The aim is to impose order while minimizing loss of detail. This paper describes the general principles IPUMS follows in the harmonization process. It discusses the challenge of applying international standards to demographic data produced over a long period of time by dozens of statistical offices. The preferred strategy involves the use of composite codes in which leading digits for a variable apply generally across countries and trailing digits retain details not universally available. That approach is not always viable, however, and the paper describes a number of other strategies to cope with the variety of challenges posed by international population data harmonization.