Constructing Cohort Data From Inter-census Data Based On Population Track Matching Method

Census data is typically a cross-sectional data. We try to find a method to track the records of same individuals in the two censuses, which can extract the information of the same cohort of people from two censuses. Compared by directly using cross-sectional data, analyzing and comparing the cohort data reflects more detailed variations of characteristics between two census years, not only the variation of characteristics,but the variations caused by which group of people.By exploring the’Population Track Matching Method’, this article extracts cohort data from inter-census data. The method is based on the principle of "Family Structure Stability" and the principle of "Space Related". By putting individuals back into families they belonged to and using the mutual constraints between family members, in order to match "same person" from "one family" in two censuses, realizes the idea of constructing cohort data from inter-census data.The main contributions of this article are as follows, classifies area into different units, first tracking in the small units, then expands to big units; holds household size as logic line, relies on principles of "Family Structure Stability" and "Space Related" sets two match acquirements according to the change of "education degree"This article can be divided into four chapters. The first chapter is introduction, expounds the important points of constructing cohort data from cross-sectional data. The second chapter is about the specific technical route of population track matching method. The basic rules of matching records from two census data are the principle of "Family Structure Stability" and the principle of "Space Related". Household size change as the logical line and match acquirements which provided by census table fields:"gender","birth" and "nationality" must remain the same,"education degree" can change in a reasonable range. When there are one-to-many, and many-to-one and many-to-many match results, using address as auxiliary information to fix the "same person". The third chapter applies’Population Track Matching Method’to fifth population census data and sixth census data of Huangpu District, Shanghai to find cohort people. First uses each community as small unit, then expands to the whole district as big unit to track "same person" of "one family" from two censuses. The fourth chapter summarizes the pros and cons of’Population Track Matching Method’. The method finally reaches to27.85%household matching rate of fifth census, and44.95%household matching rate of sixth census.