Suchergebnisse

IntroductionMortality inequalities by income and education levels have historically been estimated using an area-based approach in Canada. Although useful in measuring socioeconomic inequalities overtime, this method underestimates the level of inequality and only allows the examination of a single dimension at a time.
Objectives and ApproachTo create a series of census linked datasets that allowed for the examination of health inequalities across different socioeconomic dimensions. Specifically, five census cycles (beginning with the 1991 Census) were probabilistically and deterministically linked to different health outcomes (mortality, cancer, hospitalization) to create the Canadian Census Health and Environment Cohort (CanCHEC). Each dataset was created using a similar methodological approach which allowed for the measurement of these health inequalities over time. Mortality inequalities by both income and education level (including multidimensional) for all causes and cause-specific groups were examined.
ResultsFive census linked datasets were constructed that followed mortality for a period of up to 20 years. The 1991 CanCHEC includes 2.6 million adults, the 1996 and 2001 CanCHECs include 3.5 million adults respectively, and the 2006 and 2011 CanCHECs include 5.9 and 6.5 million people respectively. Findings revealed a stair-stepped gradient in all-cause and cause-specific mortality by educational attainment and income quintile across each time period. The lowest mortality rates were among the university educated and richest income quintile and highest mortality rates among those with less than high school graduation and the poorest income quintile. The gradient differed by cause of death groupings. Over the 25-year time period, the mortality gradient trend varied by socioeconomic dimension and cause of death.
Conclusion/ImplicationsThese data show clear mortality inequalities by socioeconomic position across the different time periods. These linked datasets can help advance knowledge in understanding health inequalities in Canada as well as provide a tool for on-going surveillance of health inequalities by different socioeconomic dimensions.

IntroductionLinking the many decades of census data collected during Canada's settlement allows researchers to investigate the movement patterns of early settlers, changes in regional demographics, and intergenerational mobility. This research leverages new methodologies and previously untranscribed individual attributes to link for the first time Canadian censuses from 1852 to 1921.
Objectives and ApproachThis work aims to build upon prior efforts (1871-1901 linking) to link the decennial Canadian censuses spanning from 1852 to 1921. We use a more complete transcription of the censuses, that has recently become available to researchers through The Canadian People's project.
Our approach to this task begins with reproducing the results of previous work using data from this new transcription. From there, we add additional time-invariant individual characteristics as features to our classification model. We also explore newer methodologies designed to leverage household information during the linkage process, with the goal of increasing the linkage rate.
ResultsWe describe the differences between the different methodologies we use, and the steps we took to clean and standardize the data. We compare the links produced by the different methodologies in terms of the number of links yielded, their quality (false positive rate), and certain aspects of the bias present in the resulting collections of links. We discuss the challenges and potential approaches to dealing with sections of the population who remain difficult to link.
We expect the advancements in record linkage methodologies for historical populations coupled with this more complete transcription of the censuses to offer advantages over prior approaches when linking these censuses. We expect the resulting linked data to offer new insight into Canada during this time period.
Conclusions/ImplicationsThe resulting collection of linked data over this time period should characterize approximately three generations of early Canadians. This linked data will be passed on to other researchers and will allow us to better understand the changing experiences of the Canadian population during these early stages of Canada's development.

This article explores the erratic history of counting by race on the Canadian census. It argues that the political development of racial classifications on Canadian censuses has been shaped by the interactions among evolving global ideas about race, the programmatic beliefs of international epistemic communities of statisticians and census designers, and domestic institutions involved in the administration of the census. First, Canadian census designers drew from shifting global conceptions about the nature of race and racial difference, which normatively defined the legitimate ends of race policies. Second, Canadian census designers often paid heed to the programmatic beliefs of the international statistical community about the appropriateness of collecting racial data. Finally, evolving political institutions involved in the administration of the census mediated these transnational ideas, molding them to fit the Canadian national context through institutional and cultural translative processes. Theoretically, this research makes the case that focusing on interactive political development can augment the theoretical toolbox of American political development, enabling a more comprehensive picture of the emergence, dynamism, and persistence of the Canadian racial order.

IntroductionFor a number of years, Statistics Canada has been evaluating the potential of increasing its use of administrative data into its Census Program. The research conducted so far has revealed that it could be possible to produce the census population counts by using existing administrative data.
Objectives and ApproachThe building of the Statistical Population Register (SPR) is one step towards achieving the use of administrative data into the Canadian Census Program. In addition, the SPR, in combination with the Business Register and the Statistical Building Register, would support a more efficient production of statistics via a multi-register-based system in the future. The SPR is created by linking numerous administrative data sources (federal, provincial, municipal and private). The in-scope Canadian population is then identified and extracted from the SPR.
ResultsThe presentation will focus on the reasons as well as the goals that had to be met in the initial research project in order to demonstrate the potential of using administrative data within the Census Program. The current state of the project will be highlighted by presenting high-level results at the Canadian, provincial and territorial levels. This is accomplished by comparing the Statistical Population Register's in-scope population to its reference, Statistics Canada's official population counts.
Conclusion/ImplicationsDespite promising results, areas of improvements have already been identified and work is under way in order to improve the quality of the upcoming Statistical Population Register. The final section of the presentation will be devoted to the future research agenda of the Census Program.