The objective of data quality assessment is to evaluate the overall quality of census data. The results are used to inform users of the reliability of the data, to make improvements for the next census and, in the case of two of the data quality studies, to adjust the official population estimates. Quality assurance activities take place throughout the census process. They take place over a span of time commencing prior to data collection (listing of addresses) and ending after dissemination (Reverse Record Check).
Census data are used in the Population Estimates Program at Statistics Canada in calculation of the population estimates, which are used by the Department of Finance in calculating transfer payments between levels of government. In 2006/07, these transfer payments will amount to over $60 billion.
In areas where census questionnaires are delivered by mail, Statistics Canada validates and updates existing address listings during the months prior to Census Day. This listing exercise produces a complete and reliable list of addresses for the mail-out of census questionnaires. In areas where enumerators deliver the census questionnaires, addresses are listed at drop-off time.
This step ensures that all returned questionnaires can be verified so that Statistics Canada knows where questionnaires have not been returned.
All questionnaires are edited for completeness when they are received at the Data Processing Centre. Those that fail edit because they are missing information are sent to be followed up. Households with incomplete forms are then contacted by census staff to resolve missing and incomplete responses.
All households that do not return a census questionnaire are contacted by census staff to obtain a completed questionnaire. This follow-up process starts about 10 days after Census Day. Follow-up contact is done by telephone when numbers are available and by personal visits as required to ensure that a completed questionnaire is obtained from all households.
At the end of every census there are always a few households from which the census has not been able to obtain a completed questionnaire for a variety of reasons. In these cases, census staff is asked to get a count of the number of people who usually live in the household plus any other information that might be available. In no circumstances is staff ever asked to fabricate information. If information is not available, then staff is told to leave the question blank.
Despite best efforts during census data collection, there always remains a number of households that have not returned a questionnaire. On the basis of a sample survey of addresses from which no census questionnaire was received, the Dwelling Classification Survey provides adjustments for the initial population and dwelling release to account for persons in these households. Estimates are derived using data from the sample on whether these addresses are occupied or not and, if so, how many persons usually live there.
This study provides estimates of persons missed by the census. Missed persons either lived at addresses erroneously omitted during address listing or were erroneously omitted from questionnaires for responding addresses. Estimates are provided for each province and territory and for various subgroups of the population (e.g., age-sex groups, marital status).
A sample of persons who "should" be enumerated in the census is traced and interviewed to determine whether they indeed should have been enumerated and, if so, whether they actually were. Persons who have died or emigrated prior to Census Day are identified during the tracing.
The results of this study are the major source of information about persons missed by the census and are used, along with the census population counts and the results of the Census Overcoverage Study, in the Population Estimates Program.
This study provides estimates of persons counted more than once by the census. Estimates are provided for each province and territory and for various subgroups of the population (e.g., age-sex groups, marital status).
For the 2006 Census, double-counting of persons will be detected by matching the census database to itself using a number of different variables and statistical matching techniques. A sample of persons potentially counted more than once is verified to determine the frequency of double-counting and to derive estimates of census overcoverage.
The results of this study are the major source of information about persons counted more than once by the census and are used along with the census population counts and the results of the Reverse Record Check in the Population Estimates Program.