Aboriginal Peoples Survey (APS)
Summary of changes
Activity on this program started: 1991
Target population - The target population of the 2017 APS is 15 or older while the 2012 APS included people aged 6 to 14.
The 2012 APS collects unique and detailed data on education, employment and health, data which are not available from any other source. For example, although the 2011 National Household Survey collected data on level of education and on major field of study, the 2012 APS addresses additional topics such as number of schools attended, exposure to Aboriginal languages, school climate and support, frequency of reading, participation in extra-curricular activities, peer influences and plans for further schooling.
Target population - Contrary to 2001 and 2006, the 2012 target population no longer covers the Aboriginal ancestry-only population and no longer covers Indian reserves and First Nations communities in the territories. In the provinces, certain Indian reserves and settlements were covered in 2001 and most of them were covered in 1991.
Instrument design - Prior to 2012, the APS used a paper questionnaire format. The questions in the 2012 APS have been designed for use in a computer assisted interviewing (CAI) environment.
Sampling - The most important difference in design between the 2012 APS and previous cycles is that the 2012 sample was selected from the 2011 NHS, a voluntary survey with a complex design, rather than from the mandatory long census questionnaire used in previous censuses and distributed to about one in five households in Canada.
Error detection - Previous cycles of the APS had used a paper-and-pencil method of data collection, so error detection and editing at the time of collection was minimal, leaving most editing to be done during data processing. With the use of Computer-Assisted Interview (CAI) methods in 2012, more editing could be done during data collection.
Imputation - For the 2006 APS, Aboriginal respondents could report Aboriginal ancestry but not self-report as being Aboriginal. In 2012, the ancestry question was not included in the APS, and so Aboriginal respondents could not have Aboriginal ancestry, without Aboriginal identity. As well, following imputation for the 2012 APS, all respondents had both Aboriginal identity and an identity including at least one of the three specific Aboriginal groups.
Estimation - The use of the Sigma-gap method to reduce excessively large weights was a change from previous cycles of the APS. Its use was made necessary by the high variability of the NHS weights compared with the weights associated with the long form questionnaire from previous censuses. Another change was the bootstrap method developed for the complex sampling design combining NHS and APS sampling.
Quality evaluation - The questions on Aboriginal identity underwent major changes between 2006 and 2012. The methodological differences are due to the change in the target population between the two cycles and the fact that the sample in the 2012 APS was drawn from respondents to the NHS rather than from respondents to the 2006 Census long form. Therefore, making comparisons between population estimates from the 2006 APS and the 2012 APS is not recommended.
Statistics Canada was mandated to coordinate a third APS after the 2006 Census, thus the frequency of the survey changed from "occasional" to "Quinquennial".
The Royal Commission on Aboriginal Peoples (RCAP), an important user of the 1991 data, recommended that APS be conducted regularly to monitor the demographic and social conditions of Aboriginal people. The federal government responded to the RCAP recommendations through its Aboriginal action plan, Gathering Strength. In this plan, the need for relevant and current data was recognized. Statistics Canada was mandated to coordinate a second APS after the 2001 Census, thus the frequency of the survey changed from "one time" to "occasional".
The Aboriginal Peoples Survey is conducted for the first time.