Modified on April 17, 2008
Journey to work data are generally comparable from 1971, 1981, 1991, 1996, 2001 and 2006. However, there are some changes between censuses.
The place of work question was included in almost every census since 1971 in almost the same format. Nonetheless, some changes were made over the years.
In 1991, a write‑in box for postal code responses was added. Three changes were implemented in 1996. The 'No fixed workplace address' response category was made explicit. The 'County' write‑in box was removed after investigation indicated that most persons did not respond and many of those who did respond confused 'county' with 'country' and wrote 'Canada' as a response. Clearer instructions, with an example of a complete civic address response, were included to assist respondents in providing a complete workplace address.
The 2001 place of work question was changed slightly to include the phrase 'most of the time'. The question reads 'At what address did this person usually work most of the time?'
The 'No fixed workplace address' category did not exist in 1971. Persons who had no usual place of work address, and who did not report to a headquarters or depot at the beginning of each shift, were instructed to write the address where they most often worked in the job described. Persons having no fixed workplace address were included with persons who worked at a usual place or in the 'Not stated' category. In 1971, the 'Not stated' category accounted for 9.2% of the respondents, many of whom worked in the construction or primary industries – industries which typically account for much of the 'No fixed workplace' category.
In 1981 and 1991, persons who had no fixed workplace address, and who did not report to a headquarters or depot at the beginning of each shift, were instructed to write 'No usual place of work' in the address write-in boxes. These instructions did not appear on the questionnaire, but were included in the Census Guide. In 1996, a 'No fixed workplace address' response category was explicitly added to the place of work question, thereby reducing response burden.
The changes over censuses associated with each of the stages in collection and processing have not significantly affected the comparability of journey to work data. However, there are some processing changes of which data users should be aware.
In 2006 and 2001, the workplace location of persons working in census metropolitan areas or census agglomerations was coded to a block-face, dissemination block or dissemination area representative point. The workplace location of persons working outside census metropolitan areas and census agglomerations are coded to census subdivision representative points.
For the 2006 Census, workplace locations are linked to representative points that are generated from the Spatial Data Infrastructure (SDI). For the 2001 Census, the representative points are generated from the National Geographic Base (NGB). Block-face representative points are used when the street address ranges are available in the NGB (normally in large urban centres and some small urban centres). When address ranges are not available, dissemination block, dissemination area or census subdivision representative points are used.
In 1996 the situation was similar to that of 2001. The workplace location of persons working in most urban areas was coded to a block-face, street intersection or enumeration area representative point. The workplace location of persons working in rural areas and some urban areas was coded to census subdivision representative points. These points may change from census to census.
Prior to 1996, workplace location data were first coded to the census subdivision and then coded to the census tract in separate coding operations. Census tract data were coded for all census-tracted areas in 1971 and 1981. In 1971 and 1991, census tract workplace location data were coded from a subsample of respondents residing in prescribed areas. From the 1991 Census, census tract workplace location data were produced only within the province of Ontario. For 1991 and earlier years there are differences in the counts from the census subdivision and the census tract coding due to differences in the coding methods and the sampling.
A significant change occurred in the edit and imputation of 1981 data. In 1971, non‑responses to the place of work question were reported as 'Not stated'. However, in 1981, the 'Not stated' category was dropped and non‑responses to the place of work question were changed to a specific response through imputation. Imputation was performed on both the 'place of work status' and 'workplace location' (census subdivision level) variables. However, census tract data were not imputed. Since the 1991 Census, location data have been imputed for all missing workplace geographies.
In 1971, commuting distance was calculated to the nearest half mile, between the residential enumeration area representative point and the workplace census tract representative point. Values of 251 miles or more were all stored as 251 miles. In 1996, the distance was calculated in kilometres, to the nearest 0.1 km, between the residential enumeration area representative point and the workplace location representative point. In 2001, the distances are calculated in kilometres, to the nearest 0.1 km, between the residential block representative point and the workplace location representative point. Values of 200.1 kilometres or more are all stored as 201.0 kilometres.
The comparability of workplace location data between censuses is affected by the conceptual changes in geography (such as the definitions of census division [CD], dissemination area [DA] and designated place [DPL]) and the changes in the census subdivision, census metropolitan area, census agglomeration and census tract boundaries. Because of the large number of geographic areas and possible boundary changes between censuses, data users are encouraged to exercise caution when comparing workplace location data between censuses.
Municipal restructuring in many provinces has resulted in a significant number of municipal dissolutions and incorporations affecting the number of census subdivisions and their boundaries.