Statistics Canada
Symbol of the Government of Canada
Warning View the most recent version.

Archived Content

Information identified as archived is provided for reference, research or recordkeeping purposes. It is not subject to the Government of Canada Web Standards and has not been altered or updated since it was archived. Please contact us to request a format other than those available.

6. Dwelling Classification Survey

6.1 Introduction

6.2 Methodology

6.2.1 Stratification and sample selection

6.2.2 Field interviews

6.2.3 Processing, coding, and editing

6.2.4 Census whole household imputation

6.3 Estimates

6.3.1 Unoccupied dwellings

6.3.1.1 Occupied dwellings misclassified as unoccupied

6.3.1.2 Housing stock overcoverage

6.3.2 Non-response dwellings

6.3.2.1 Persons added in non-response dwellings

6.3.2.2 Dwellings not in the housing stock misclassified as non-response

6.1 Introduction

As described in Section 5.7, census data are adjusted for occupied non-respondent dwellings. The number of people living in these dwellings is estimated by the Dwelling Classification Survey (DCS). These estimates are used in census processing to specify how many people should be imputed during whole household imputation (WHI). The second objective of the DCS is to measure three types of dwelling classification error.

One of the potential sources of error in a census is the misclassification of dwellings. When a questionnaire is not returned from a household, the enumerator has to determine if the dwelling is occupied or not. Two types of errors can occur. First, an occupied dwelling can be incorrectly classified as unoccupied. Census dwelling and population undercoverage are the result of this classification error because the dwelling is excluded from the census database. Second, an unoccupied dwelling can be incorrectly classified as occupied. When this error occurs, no questionnaire will be received for this dwelling and it will be subject to non-response follow-up (NRFU) as described in Section 4.3. The dwelling will be considered as a non‑respondent dwelling and therefore subject to imputation. This would add persons to the census database when, in fact, no one is living at that dwelling. That is, this classification error results in population overcoverage. Estimates from the DCS are used to adjust census data for both of these coverage errors.

The third component of dwelling classification error measured by the DCS is the error incurred when marginal dwellings or dwellings under construction are classified in error as dwellings. Since the dwelling would be classified as unoccupied, no population overcoverage results as only occupied dwellings can be classified as non-respondent dwellings and therefore be subject to imputation. However, there is dwelling overcoverage. Census data is not adjusted for these dwellings so census estimates of the housing stock include some degree of overcoverage.

6.2 Methodology

6.2.1 Stratification and sample selection

The DCS target population was all non-response dwellings and all unoccupied dwellings excluding dwellings in collective collection units (CU), canvasser CUs and Indian reserves CUs. These areas were excluded because of cost and operational considerations.

The sample size for the DCS was set at 1,405 CUs. The sampling frame consisted of all self‑enumeration CUs with the exception of Indian reserves. Consequently, the sampling frame for the territories included only Whitehorse, Yellowknife, Hay River and Fort Smith. The sample design was as follows. First, CUs in Whitehorse, Yellowknife, and Hay River and Fort Smith in the Northwest Territories formed one stratum. All of these CUs were selected for the DCS sample. All of the CUs in Prince Edward Island formed a second stratum from which a simple random sample of 44 CUs was selected.

The remaining CUs were grouped into urban and rural strata. A CU was considered urban if it initially had been part of a census metropolitan area (CMA) or a census agglomeration (CA) that had 40,000 or more occupied dwellings. Further, all of the CUs within a crew leader district (CLD) were considered urban if more than 50% of the CUs in the CLD were urban. All of the remaining CUs formed the rural strata. Urban CUs were stratified by CMA and CA. A simple random sample of at least five CUs was selected within each stratum. From past census data, it was determined that five CUs was an appropriate workload for an interviewer. There were 812 urban CUs in the sample. In order to control field costs, the rural sample was chosen to be geographically close. This was done via a two-stage stratified simple random sampling design. In the first stage, CLDs were selected within each province. In the second stage, five CUs were selected from each of the selected CLDs. There were 593 rural CUs in the sample.

All of the unoccupied dwellings and non-response dwellings in the sampled CUs formed the DCS sample of dwellings, a total of 32,345 unoccupied and 6,788 non-response dwellings. Table 6.2.1 shows the distribution of the sample by province and territory.

6.2.2 Field interviews

All dwellings in the sampled CUs that were classified as unoccupied on Census Day or classified as occupied but for which no census form had been returned, were to be checked again in late June or early July 2006 to determine the true occupancy status of the dwellings on Census Day. A DCS questionnaire was used for this purpose.

The timing of this operation was left to the discretion of each regional office (RO). In order to determine occupancy status and collect other information, enumerators were instructed to contact current occupants, neighbours, landlords, or any other person with some knowledge about the dwelling. Up to three contact attempts were made for each dwelling. If the dwelling was found to have been occupied on Census Day, the number of occupants on Census Day was obtained.

6.2.3 Processing, coding, and editing

All completed questionnaires were sent to Ottawa for processing.

Some preliminary edits and general grooming were then performed before the questionnaires were sent for data capture (key entry). Once data capture was completed, the questionnaires were subjected to an extensive set of consistency edits. The questionnaires failing edits were examined manually in order to resolve the inconsistencies.

For each dwelling in the DCS sample found to have been occupied on Census Day, the DCS questionnaire was consulted to determine whether another address was listed where the household members may have been enumerated. If an alternate address was given, then the Visitation Record (VR) and the census questionnaires for the alternate addresses were checked to see if the household members were indeed enumerated elsewhere. If they were found to have been enumerated elsewhere, they were considered as already having been enumerated and therefore they need not add to overcoverage by being included again. The dwelling itself, however was added to the occupied dwelling count.

At this point in processing, the unoccupied dwellings and the non-response dwellings in the sample were separated and the classification of these dwellings was confirmed against final census listing. The questionnaires completed for each sampled CU were matched to the final census listing of unoccupied dwellings. If a match could not be found, the sampled dwelling was discarded and no further processing was required. Dwellings listed as unoccupied on the census list for which no DCS questionnaire was received were considered as total non-response and went onto the next step of processing. Similarly, the final census listing of all dwellings for which a census questionnaire was not received was used to establish which of the DCS dwellings for which a DCS questionnaire was not received would be considered as total non-response.

Total non-response was addressed by a weighting adjustment while item imputation was used for item non-response. The procedure was the same for the unoccupied dwellings and non-response dwellings. When there was no information for a dwelling, the design weights of the respondents were adjusted to account for the design weight of the non-respondents. The adjustment was done separately for each of the Montréal, Toronto and Vancouver CMAs, for the remaining urban areas in each province and territory, and for the rural areas for each province and territory. Item non‑response for occupancy status, number of usual residents, and dwelling type was addressed by imputation. Occupancy status was imputed first and then used in the imputation of the other variables. Design weights were then adjusted so that the sum of the adjusted weights for each subprovincial area equaled the number of unoccupied/non-response dwellings.

6.2.4 Census whole household imputation

Once the DCS estimates were produced, census data were adjusted for non-respondent dwellings and for occupied dwellings classified in error as unoccupied. This was done in the whole household imputation (WHI) step of census processing as follows for the non-respondent dwellings; unoccupied were handled in a similar, but simpler, fashion. First, within a DCS post‑stratum all the non-respondent dwellings were identified. Second, any non-respondent dwelling for which field collection had obtained the number of usual residents was deemed to be occupied and assigned the recorded household size. Finally, an additional number of non-respondent dwellings were randomly selected and declared to be occupied. The selection was done such that the final number of non-respondent dwellings converted to occupied dwellings in the post-strata equaled the DCS estimate of occupied dwellings in the non-respondent dwelling universe. This process resulted in all private dwellings on the database being classified as either occupied or unoccupied.

A second procedure was used to impute the household dwelling size and other variables for the selected non-respondent dwelling. Household size was determined by randomly selecting a dwelling from all dwellings that had completed a census questionnaire in the same CU. The complete record from this donor household was then assigned to the non-respondent dwelling. If no donor was found, then only a household size was assigned.

More information on WHI can be found in Dick (2007).

6.3 Estimates

Census data are adjusted for non-respondent dwellings and for occupied dwellings that are classified in error as unoccupied using DCS estimates. The estimates are given in Section 6.3.1.1 and 6.3.2. Census data are not adjusted for marginal dwellings or dwellings under construction that are classified in error as dwellings. Section 6.3.1.2 presents estimates of the number of marginal dwellings and dwellings under construction that are classified in error as dwellings and therefore erroneously included in the housing stock.

6.3.1 Unoccupied dwellings

6.3.1.1 Occupied dwellings misclassified as unoccupied

Table 6.3.1.1.1 gives the estimated number of dwellings classified as unoccupied that should have been classified as occupied and the corresponding error rate for unoccupied dwellings by urban and rural1, by province and territory, for the three largest CMAs, and by type of dwelling. For comparison, Table 6.3.1.1.2 gives the same estimates for the 2001 Census. Table 6.3.1.1.3 gives the estimated number of persons living in occupied dwellings misclassified as unoccupied. Table 6.3.1.1.4 shows the number of households and persons added to the initial 2006 Census counts to adjust for these misclassifications.

Table 6.3.1.1.1 shows that 17.4% of all dwellings classified as unoccupied were actually occupied. This is an increases from 12.7% found in 2001. The misclassification of dwellings was much more prevalent in urban areas (25.7%) than in rural areas (8.1%). Both areas show increases from 2001. Among the three largest CMAs, there was a large decrease in the rate of misclassification in both Toronto and Vancouver while a large increase occurred for Montréal. Increases in the misclassification rates occurred for all provinces except Prince Edward Island where it remained stable. The misclassification rate decreased for apartment buildings of five or more stories between 2001 and 2006 but the rates increased for all other types of dwellings.

Among the provinces and territories, British Columbia had the highest misclassification rate, 25.5%, followed by the Yukon Territory, 23.5%, Alberta, 21.4%, Quebec, 21.1% and Ontario, 16.0%. The rates for the other provinces and territories ranged from 14.9% for Nova Scotia to 7.3% for Newfoundland and Labrador. Among the three largest CMAs, the 2006 rate of misclassification is very high in all three areas with the rate in Montréal (34.0%) being higher than the rates in Vancouver (25.5%) or Toronto (23.3%). Among the types of private dwellings classified in the census, the rate of misclassification is lowest in single-detached houses (17.4%) and highest in apartments in buildings of five or more storeys (39.6%). The rate of misclassification in the 'Other' category, which includes semi-detached houses, row houses, duplexes, apartments in buildings with fewer than five storeys, mobile homes and other movable dwellings, is also high at 38.2%.

Because of error in the initial classification of dwellings, approximately 162,897 households were not enumerated in the 2006 Census. This is the number of households added to the census during WHI. Table 6.3.1.1.4 shows the number of households and persons added to adjust for occupied dwellings misclassified as unoccupied.

6.3.1.2 Housing stock overcoverage

Table 6.3.1.2 shows the estimated number of unoccupied dwellings not in the housing stock and the corresponding error rate for unoccupied dwellings for various geographic areas. No adjustments are made to the census database to account for dwellings not in the housing stock that were erroneously classified as unoccupied.

The enumeration of unoccupied dwellings that fall outside the housing universe results in overcoverage of dwellings. Dwellings are considered to be outside the housing universe if they are used for commercial purposes, if they are not habitable year round, or if they are double counted in the census. This last situation can happen when the dwelling appears to have two addresses associated with it, or when two questionnaires are mistakenly returned for a dwelling which no longer contains a separate apartment within it.

The Dwelling Classification Survey estimates of the number of unoccupied misclassified as dwellings are not used to adjust the census database because of the degree of subjectivity associated with classifying a dwelling as suitable for year-round occupancy. A dwelling must have a source of heat or power and provide complete shelter from the elements to be considered as suitable. It is sometimes difficult to tell whether or not a dwelling is habitable such as when the dwelling is a cottage, when the dwelling is under construction and almost complete, or when the dwelling has deteriorated.

Dwellings outside the housing stock account for 35.5% of all dwellings classified as unoccupied. Among the provinces and territories, the incidence of dwellings outside the housing stock having been classified as unoccupied ranges from 8.3% in Prince Edward Island to 46.1% in Manitoba. The problem is evenly found in urban areas (35.0%) and the rural areas (36.0%). For the three largest CMAs, the rate ranges from 29.9% in Montréal to 54.4% in Toronto.

Dwellings actually outside the housing stock represent 2.5% of all private dwellings in the 2006 Census. This is an increase from the 2001 error rate of 1.7%. Among the provinces and territories, the error ranges from a low of 0.9% in the Yukon Territory to a high of 3.8% in Newfoundland and Labrador. For the three largest CMAs, the error ranges from 1.0% in Montréal to 2.3% in Toronto.

6.3.2 Non-response dwellings

6.3.2.1 Persons added in non-response dwellings

Table 6.3.2.1.1 gives the estimated number and rate of occupied non-response dwellings in the census by urban (> 50,000) and rural, by province and territory, for the three largest CMAs, and by type of dwelling. Table 6.3.2.1.2 gives the number of persons estimated by the DCS to be living in these non-response dwellings while Table 6.3.2.1.3 gives the same information for the 2001 DCS.

Table 6.3.2.1.1 shows that 70.9% of all dwellings classified as non-response were actually occupied. The census did a slightly better job of classifying non-response dwellings in urban areas (72.1%) than it did in rural areas (64.1%). At the province and territory level in 2006, the Northwest Territories had the highest rate of correctly classified non-response dwellings at 91.0%, while New Brunswick had the lowest rate at 59.5%.

Among the three largest CMAs, the 2006 rates for occupied non-response dwellings ranged from 68.3% in Vancouver to 75.6% in Montréal. Finally, when examining types of private dwellings, the rates for occupied non-response dwellings ranged from 74.3% in apartments in buildings of five or more storeys to 78.9% in the 'Other' category, which includes semi-detached houses, row houses, duplexes, apartments in buildings with fewer than five storeys, mobile homes and other movable dwellings.

Table 6.3.2.1.2 shows the number of non-response dwellings in the 2006 Census, and gives the number of persons added in those dwellings through the DCS. Table 6.3.2.1.3 shows the same data from the 2001 DCS. In 2006, a total of 571,521 persons were added to the census in 259,894 dwellings. The comparable 2001 numbers are 317,587 persons in 143,681 dwellings.

6.3.2.2 Dwellings not in the housing stock misclassified as non-response

Table 6.3.2.2 shows the 2006 Census dwelling classification error from dwellings erroneously classified as non-response because they should not have been included in the housing stock. Section 6.3.1.2 provides the definition of dwellings outside of the housing universe and comments on the difficulty in determining whether a dwelling should be included in the housing stock.

At the national, dwellings outside the housing stock account for 9.0% of all non-response dwellings. The error rate is slightly higher in the rural areas (10.5%) than in the urban areas (8.7%). For provinces and territories, the incidence of dwellings outside the housing stock having been classified as non-response ranges from 2.4% in Prince Edward Island to 11.2% in both Alberta and British Columbia. For the three largest CMAs, the rate ranges from 8.4% in Montréal to 15.6% in Vancouver. At the national level, non-response dwellings outside the housing stock account for 0.3% of all private dwellings. For provinces and territories, this error ranges from a low of 0% rounded in Prince Edward Island to 0.4% in both Alberta and British Columbia. For the three largest CMAs, the error ranges from 0.2% in Toronto to 0.7% in Vancouver.

Note:

  1. Urban refers to urban areas with a population of over 50,000 persons. The remaining geographies constitute the rural areas.

   Previous page | Table of contents | Next page