Sampling and non-sampling errors

View the most recent version.

The underlying spsd/m data can be affected by both sampling and non-sampling errors derived from SLID and CIS as inferences about the entire population are based on information obtained from a sample. For the 2022 reference year, for example, the final CIS response rate was 70.1%. Sampling errors associated with survey estimates are measured using coefficient of variation for CIS estimates. Additionally, imputation is carried out for income variables as well as variables related to labour, school attendance, food security, housing, and utility costs. Non-sampling errors, on the other hand, are those related to coverage errors, non-response errors, response errors, interviewer errors and other types of processing errors Surveys and statistical programs - Canadian Income Survey (CIS) (statcan.gc.ca).

The table below provides a comparison of income estimates for 2009 from the SPSD/M and the MEA concept by aggregated income category. While the SPSD/M provides estimates on a cash basis of accounting to derive household after-tax income, the MEA applies accrual adjustments to a number of income items to derive the MEA concept of household disposable income. As described in the SNA 2008:

"Accrual accounting records flows at the time economic value is created, transformed, exchanged, transferred or extinguished. This means that flows that imply a change of ownership are entered when the change occurs, services are recorded when provided, output at the time products are created and intermediate consumption when materials and supplies are being used."

As shown in the table, SPSD/M income items are generally lower than the MEA, due mainly to differences in income concepts between the two sources. For example, SPSD/M market income is lower than the MEA concept of primary income, due mainly to the inclusion of employers’ social contributions. As well, transfers to corporations are lower in the SPSD/M as the MEA concept includes transfers from households to corporations for pension plan entitlements.

As mentioned above, there are a number of other reasons why income estimates in the SPSD/M do not match the MEA income concept. In each case, as the table suggests, elements are included in the MEA macroeconomic estimates of household disposable income that are not included in the SPSD/M microeconomic estimates. Macroeconomic statistics are compiled within an integrated system which considers income to accrue from productive activity and tracks income flows consistently among all institutional sectors. So it follows that logical proxies need to be identified to distribute directly observable and missing elements.