Statistics Canada
Symbol of the Government of Canada

4.0 Survey methodology

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.

4.1 The survey universe

The 2005 Survey of Financial Security was carried out in all ten provinces, the territories were not included. Those living on Indian reserves and crown lands and official representatives of foreign countries living in Canada and their families were also excluded from the survey. Members of religious and other communal colonies, members of the Canadian Forces living in military camps and people living in residences for senior citizens were excluded, as were people living full time in institutions, for example, inmates of penal institutions and chronic care patients living in hospitals and nursing homes. The survey covers about 98% of the population in the ten provinces.

Information was not gathered from persons temporarily living away from their families (for example, students at university) because it would be gathered from their families if selected. In this way, double counting of such individuals was avoided.

4.2 Survey content and reference period

With a few exceptions, the reference period for the information was the time of data collection (May to July 2005). For the asset and debt information respondents were asked to provide an estimate of the value or amount as close to the survey date as possible, recognizing that their most recent statement may have been as of the end of the previous calendar year, or for the last financial quarter. Some of the information was collected for each person in the family 15 years of age and over.

The assets and debts, however, were collected for the family as a whole, because they often cannot easily be assigned to one person in the family. Specifically, the following information was collected:

From each family member 15 years of age and over:

  • demographics (age, sex, marital status);
  • ethno-cultural characteristics;
  • education;
  • current employment;
  • income, for the calendar year 2004.

From each family member 25 years of age and over:

  • previous employer pension plans
  • pension plan benefits

From each family member 45 years of age and over:

  • retirement information

For the family unit as a whole:

  • financial and non-financial assets;
  • equity in business;
  • debt in the form of mortgages, vehicle loans, credit card and line of credit debt, student loans and other debt.
  • distribution of registered plans investments
  • distribution of mutual funds investments

4.3 The sample

The total sample for the 2005 Survey of Financial Security was 9,000 dwellings; it was drawn from two sources.

The main sample, drawn from an area frame, consisted of 7,500 dwellings. This area sample was a stratified, multi-stage sample selected from the Labour Force Survey (LFS) sampling frame. Dwellings selected for this survey had not previously participated in a labour force or financial survey conducted by Statistics Canada. Sample selection comprised three steps: the selection of clusters (small geographic areas) from the LFS frame, field listing of all addresses within each selected cluster, and the selection of dwellings within these selected clusters. At the time that the SFS sample was selected the LFS frame was using 2001 Census geography.

The second portion of the sample, 1,500 dwellings, was drawn from geographic areas in which a large proportion of family units had what was defined as "high-income". This sample was included to improve the representation in the sample of high income families, as a disproportionate share of net worth is held by such higher-income family units. For purposes of this sample the income cutoff was total family income of at least $200,000 or investment income of at least $50,000. The latter was used to take into account those family units that may not have high income from employment but have substantial assets that generate investment income.

4.4 Data collection

Data were collected during a personal interview using a paper questionnaire.

For families, the interview was held with the family member with most knowledge of the family's financial situation. If necessary, follow-up was done with other family members. Proxy response was accepted. This allowed one family member to answer questions on behalf of any or all other members of the family, provided he or she was willing and able to do so.

To reduce response burden, for the questions on 2004 income, respondents could give Statistics Canada permission to use the income information from their T1 tax return. Close to 80% of survey respondents gave their consent to use these administrative records.

4.5 Data processing and quality control

In-house scanning software was used to capture survey data from the questionnaire. A quality control operation was applied to ensure that pre-specified quality standards were achieved. Data then passed through an automated edit system to identify inconsistencies and potential errors in the data.

4.6 Imputation of missing data

Missing responses were imputed for all key fields in the questionnaire. Where possible, information was imputed deterministically, using other information reported by the respondent. For example, when the respondent could not estimate the value of their vehicle, the reported make, model and year was used to impute a value. This value was determined by consulting a reference book. When deterministic imputation was not possible, hotdeck imputation methods were used in most cases, and for all components of income and net worth, nearest neighbour techniques were employed. These methods involve identifying another individual or family with similar characteristics to become the "donor" and provide the imputed value. Income data obtained from tax returns are considered complete and thus do not require imputation.

Table 4-1 shows the percentage of asset and debt values that was determined through imputation.

4.7 Weighting

The estimation of population characteristics from a survey is based on the premise that each sampled unit represents, in addition to itself, a certain number of unsampled units in the population. A basic survey weight is attached to each sample record to indicate the number of units in the population that it represents. Adjustments are then applied to the basic survey weights in order to improve the reliability of the estimates.

The basic weights are first inflated to compensate for non-response. This adjustment was applied within groups of sample units that are geographically close and the area and high income samples were adjusted separately.

A frame allocation factor was then applied. Since the high income frame overlaps completely with the area frame, units on the high income frame were eligible to be selected from either the high income or the area frame. The frame allocation factor was applied to non-response adjusted weights to account for this increased probability of selection.

The weights are then further adjusted to ensure that estimates of relevant population characteristics would respect known population totals from sources external to the survey. The population totals used for the SFS were based on Statistics Canada's Demography Division population counts for different province - age - sex groups. The weights were also adjusted to ensure that the number of 1-person and 2-person families, and the number of 1-person and 2-person households agreed with known totals by region.

Additionally in 2005, two new sources of weight adjustments were introduced. The first adjustment was based on administrative data from the T4 file. Weight adjustments were made to ensure that the survey distribution of earnings reflected approximately the same distribution as the T4 population. The second new adjustment made use of Survey of Labour and Income Dynamics (SLID) data to improve estimation. SFS as the smaller sample survey borrowed strength from SLID, the larger sample survey to not only improve SFS estimates but also to bring estimates for the 2 surveys more in line with each other.