View the most recent version.

Modified on February 2, 2009

Housing and Dwelling Characteristics
Reference Guide, 2006 Census

Catalogue no. 97-554-GWE2006003

• Definitions and explanations of variable concepts
• Questions
• Classification
• Data
• Data quality
• Data comparability - Structural type of dwelling
• Collective dwelling
• Shelter Cost Universe - Exclusion of dwellings on a farm that are occupied by the farm operator

Definitions and explanations of variable concepts

Introduction to the dwelling universe

The dwelling universe consists of Collective dwellings and Private dwellings.

Housing and dwelling characteristics variables collected by the census are: Structural type of dwelling, Rooms, Bedrooms, Period of construction, Condition of dwelling, and Value of dwelling. Data on structural type of dwelling were collected for all Private dwellings while data on the remaining variables were collected only for Private dwellings occupied by usual residents in Canada.

Questions

Housing and dwelling characteristics data for private households were obtained from Questions H1 to H8 on the 2B and 2D questionnaires. The 2B questionnaires were used to enumerate a 20% sample of all private households in Canada, while the 2D questionnaires were used to enumerate private households on Indian reserves and in remote areas.

Data for structural type of dwelling were coded by trained enumerators.

2006 Census questionnaires

Classification

Dwellings can be classified into many subuniverses based on whether they are collective or private dwellings, whether they are suitable for year-round occupancy, whether they are unoccupied or occupied, and whether they are occupied by usual residents or foreign and/or temporary residents. Complete definitions for each subuniverse are available in the Dwelling universe index of the 2006 Census Dictionary.

Figure 19 2006 Dwelling universe

Data

2006 Census

• Dwelling counts highlight tables for 2006 Census
• Housing and dwelling characteristics cross-tabulations for 2006 Census
• Profiles for 2006 Census at each level of geography

2001 Census

• Dwelling counts highlight tables for 2001 Census
• Housing and dwelling characteristics cross-tabulations for 2001 Census
• Profiles for 2001 Census at each level of geography

1996 Census

Note: Subprovincial data comparisons over several censuses should be made with caution since geographical boundaries may have changed.

Data quality

Data quality verification in place for the 2006 Census

The overall quality of the dwelling variables from the 2006 Census is acceptable. However, users of the 2006 Census data are cautioned that imputation rates, in general, have increased. Considerable effort is made throughout the entire process to ensure high standards of data quality; but, the resulting data are subject to a certain degree of inaccuracy. The evaluation of housing and dwelling variables consisted of the following at the provincial level:

• Examination of total imputation rates;
• Comparison of the distribution of unedited and edited data to determine if any data bias is introduced by imputation;
• Historical comparison with data from the previous census(es);
• Comparison with other sources of data as applicable.

To assess the usefulness of census data for their purposes and to understand the risk involved in drawing conclusions or making decisions on the basis of these data, users should be aware of the following data quality indicators for the housing and dwelling variables.

Structural type of dwelling

• Data for this variable were coded by trained enumerators in the field.
• One point six percent (1.6%) of the records were blank or invalid and required imputation (1.7% for 2001 Census).
• The comparison of edited and unedited data indicated that imputation did not alter the overall percentage distribution of structural type of dwelling.
• Changes in 2006 affect historical comparability. See 'Data comparability - Structural type of dwelling' below for more details.

Rooms

• The total imputation rate for rooms at the Canada level was 8.4% (5.8% for 2001 Census).
• The total imputation rate was a combination of non-responses and invalid data resulting from inter-variable (rooms and bedrooms) edits.
• The number of rooms in a dwelling may be lower in the 2006 Census than in the 2001 Census due to a change in data capturing procedures. Only dwellings where the respondent provided the number of rooms using a fraction, such as 4.5 rooms, are affected. The value was rounded down in the 2006 Census. In the 2001 Census and before, the value was rounded up. This change in procedure was done to increase data quality by better reflecting the number of rooms in the dwelling. The fraction part of the response typically represents a bathroom, which according to the census definition should not be counted. An unexpected trend was observed in Quebec's data that could likely be attributed to this change in processing. In Quebec, there was a decrease in five and six room dwellings and an increase in three and four room dwellings in apartment buildings. This observation is inconsistent with historical trends and Quebec apartments are commonly advertised with a fraction in the number of rooms; thus, the change in data capturing procedures may be the cause.

Bedrooms

• The total imputation rate for bedrooms at the Canada level was 6.7% (5.7% for 2001 Census).
• Data processing did not result in any large distortion of data for bedrooms. Comparison of unedited and edited 2006 Census data showed that data processing affected the distribution of number of bedrooms by increasing the proportion of occupied private dwellings with two or fewer bedrooms by 1.1 percentage points at the expense of dwellings with three or more bedrooms.

Period of construction

• The total imputation rate for this variable was 7.5% (5.2% for 2001 Census).
• Edit and imputation did not result in any large distortion of data for Period of construction.
• Data for 2006 Census is comparable to the 2005 Survey of Household Spending (SHS) although census counts for dwellings constructed between 1961 to 1970 are lower than SHS estimates by 8.8%.

Condition of dwelling

• The total imputation rate for this variable at the Canada level was 4.4% (2.2% for 2001 Census).
• The comparison of edited and unedited data indicated that imputation did not alter the overall percentage distribution of condition of dwelling.
• As in previous censuses, the data distribution in 2006 differed somewhat from the distribution from the 2005 Survey of Household Spending (SHS). Census reported fewer dwellings requiring regular maintenance and more dwellings requiring minor repairs than SHS. A high degree of comparability could be seen for the major repairs response category. This may be due to the reverse order of listing for the three response categories as well as different illustrative examples in the census and SHS.

Value of dwelling

• The total imputation rate for this variable at the Canada level was 15% (13% for the 2001 Census).
• The comparison of 2001 and 2006 data for value of dwelling corroborated the general upward trend in housing prices seen in the multiple-listing sales from the Canadian Real Estate Association and Statistics Canada's New Housing Price Index. As these two sources are not directly comparable to the Census data, variations in the size of the increase were expected and observed; nevertheless, the general upward trend was corroborated.
• Post-censual evaluation of data for Value of Dwelling has revealed that for some smaller communities there are a few high values of dwelling that cause substantial differences between the average and median values of dwelling. In most cases, the few high values of dwelling reflect the range of different housing characteristics in the community. However, in some instances, the high values of dwelling may reflect a response error where the value was over-reported. Furthermore, in some smaller communities, high non-response rate for the Value of Dwelling question resulted in some high values of dwelling being estimated during data processing. Data users should consider both average and median values of dwelling as well as the community housing characteristics when examining data for Value of Dwelling in small communities.

Census subdivisions most affected by high values of dwellings are:

Capital H (Part 2) (British Columbia)
Hart Butte No. 11 (Saskatchewan)
Barkmere (Quebec)
Northern Rockies B (British Columbia)
Silver Beach (Alberta)
Sainte-Praxède (Quebec)
Terra Nova (Newfoundland and Labrador)
Rushoon (Newfoundland and Labrador)
Eldon No. 471 (Saskatchewan)
Improvement District No. 4 (Alberta)
Estérel (Quebec)
Senneville (Quebec)
Greater Vancouver A (British Columbia)
La Morandière (Quebec)
Stanbridge Station (Quebec)
Lestock (Saskatchewan)
Poplar Bay (Alberta)
Bonnyville Beach (Alberta)
Spalding No. 368 (Saskatchewan)
Jarvis Bay (Alberta).

Data comparability – Structural type of dwelling

Concepts and definitions of dwelling variables have not changed from the 2001 Census with the exception of the Structural type of dwelling variable. Changes in instructions provided to enumerators and changes to the enumeration process affect the historical comparability of the Structural type of dwelling variable. Users should take this into consideration when making historical comparisons.

The Structural type of dwelling variable is collected by trained enumerators. Improvements to the enumeration process have resulted in a better identification of hard-to-find dwellings such as basement apartments. As a result, structures that may have been classified in previous censuses as single‑detached houses because there was no outside sign of an apartment are more likely be classified as apartments – either in a duplex or a building that has fewer than five storeys, as appropriate.

The additional classification instructions to enumerators clarified how certain types of dwelling should be classified – mainly those attached to other dwellings. In particular, the definition of a duplex (building with two apartments, one above the other) is broadened to include duplexes that are attached to other structures. In the 2001 Census and earlier, an apartment in a duplex attached to other dwellings or buildings was classified as an 'apartment in a building that has fewer than five storeys'. For 2006, these attached duplexes are included, along with detached duplexes, in a new category, 'apartment or flat in a duplex'. This new category replaces 'apartment or flat in a detached duplex'.

Both changes affect areas that received a Census questionnaire by mail (roughly two-thirds of Canada). The additional classification instructions affect only those areas where the particular types of attached dwellings are found. Montreal is particularly affected.

Comparisons of structural type of dwelling data for Canada between the 2001 and 2006 censuses show a decrease in share for 'single-detached house' (-2.1%), an increase in share for 'apartment or flat in a duplex' (+1.8%) and an increase in share for 'apartment in a building that has fewer than five storeys' (+0.4%). These changes are a combined result of the additional classification instructions, improvements to the enumeration process, and real changes that have occurred since the 2001 Census. The contribution of each of these three factors cannot be determined.

Furthermore, in the 2001 Census, apartments in a building that has fewer than five storeys were further distinguished by whether or not there was direct ground access; the 2006 Census did not make this distinction. Postcensal data evaluation for the 2001 Census revealed a data quality issue for these dwellings. As a result, the two categories were aggregated into 'apartment in a building that has fewer than five storeys' and this category is directly comparable with the same category from previous censuses and the 2006 Census.

Comparability between the 2001 and 1996 censuses is described in Dwellings, Households and Shelter Costs, 2001 Census Technical Report.

Collective dwelling

The classification of collective dwellings during census enumeration has evolved over time and has become quite complex. New to the 2006 Census, enumerators completed a Collective Dwelling Profile Form for each collective dwelling prior to enumeration. The profiling form collected information on services provided, the capacity, contact information, and the existence of buildings associated with the facility. This profiling process improved the classification of dwellings for the 2006 Census.

Despite the new profiling form for the 2006 Census, classification of collective dwelling types remains a challenge. To illustrate, it is becoming increasingly difficult to distinguish between nursing homes and residences for senior citizens as new establishments that contain both facilities become more common.

New categories of collective dwellings were added in the 2006 Census. The 'Special care' category includes nursing homes, residences for senior citizens, and chronic care and long-term care hospitals and related institutions. The 2001 category 'Establishments for children and minors' is separated into 'Group homes for children and youth' and 'Homes and treatment centres for children with psychiatric disorders or developmental disabilities'. 'Police lock-up facilities' is a new category. The 2001 category 'Other shelters and lodging and rooming with assistance services' is separated into 'Shelters for abused women and children' and 'Other shelters and lodging and rooming with assistance services'. Finally, in 2006, the 'Merchant and government vessels' category includes Canadian Armed Forces vessels at sea or in port, Coast Guard vessels, and merchant vessels.

In view of the classification issues related to collective dwellings, care must be taken in interpreting the data on individual collective dwelling types and making comparisons with previous censuses.

Users should also be cautioned that the Statistics Canada 2006 classifications are not expected to agree with classifications used in administrative data or other sources, since the facilities are classified by their functioning, rather than by their names or official status. Furthermore, census data are not collected for foreign and/or temporary residents while these residents may be included in administrative data sources.

Shelter Cost Universe – Exclusion of dwellings on a farm that are occupied by the farm operator

Shelter costs (consisting of utilities and cash rent payments for renters; and utilities, mortgage, property tax, and condominium fee payments for owners) are not collected nor calculated for dwellings on a farm that are occupied by the farm operator. The reason is that farm operators living on the farm they operate may have difficultly separating the shelter costs associated with the operation of the farm and the personal shelter costs. Since at least the 1971 Census, farm dwellings occupied by the farm operator have been excluded when calculating shelter costs.

In the Census, dwellings on a farm that are occupied by a farm operator are self-identified by the respondent. Respondents are instructed to check a box immediately preceding the shelter cost question (Question H6 to H8 on the 2006 Census long form) if they are a farm operator living on the farm they operate. The wording of the instruction has remained almost the same since 1981; however, the layout has changed slightly over the different censuses. In the 2006 Census, the check box was placed closer to the instructions and it was not aligned with other question check boxes which possibly made it less visible for the respondent.

Dwellings on a farm that are occupied by a farm operator make one to two percent of all private occupied dwellings in the 2001 Census (about 171,925 out of 11,562,975 private dwellings). The number of these farm dwellings has been decreasing from census to census. However, from the 2001 to 2006 Census, a larger than expected decrease was observed. In the 2006 Census, the number of these farm dwellings decreased by 73,555 dwellings (from 171,925 in 2001 to 98,370 in 2006); whereas in the 2001 and 1996 Census, the decrease was about 15,000 for each of these censuses. A possible explanation for the change in response pattern could be the different layout of the farm dwelling check off box since neither the wording nor the processing rules of this indicator variable changed between the censuses.

As a result, shelter cost tables and statistics may include some dwellings that should be excluded because they are farm dwellings occupied by the farm operator but the farm operator did not identify his dwelling as such on the questionnaire. Given the small number of farm dwellings occupied by the farm operator relative to the number of private dwellings, the impact at the Canada level of not excluding some of these dwellings is negligible; however, in small areas where these dwellings are common, caution should be used in making historical comparisons of shelter costs as past censuses likely excluded more of these farm dwellings.