Reliability

How reliable are the statistics that comprise the profiles?

Derived from a sample of 174 420 businesses (134 900 unincorporated and 39 520 incorporated), the profiles are generally a robust and reliable information source. Nonetheless, the data in the profiles are estimates derived from a random sample of income tax forms and are, therefore, subject to non-sampling and sampling error. The quality of the estimates depends on the combined effect of these two types of error.

Non-sampling errors

Non-sampling errors are present in data whether a sample or a complete census of the population is taken. These errors may be introduced at various stages of data processing (such as coding, data entry, editing, weighting, tabulation, etc.) and include response errors introduced by tax filers as a result of misclassifications. All efforts are undertaken to minimize non-sampling errors through edits and data analysis, but some of these errors are beyond the control of Statistics Canada. Specifically, Revenue Canada tax forms are designed for the collection of income data for tax purposes and not for statistical purposes.

Sampling error

Sampling error arises because observations are made only on a sample and not on the entire population. Sampling error depends on factors such as sample size, variability in the population, the sample design and the method of estimation. For sample estimation, since inference is made about the entire population based on data obtained from only a part of that population, the results are likely to be different than if a complete census was taken under the same general survey conditions. The most important feature of probability sampling is that the sampling error of an estimate can be measured from the sample itself. Indicators of sampling error have been calculated for individual statistics in the profiles. They can be interpreted as follows:

Data suppressed due to poor quality will appear as blank in the profiles.

Quality and conceptual limitations of using tax data to calculate the profiles (non-sampling error)

• There is no standard for income, expense and balance sheet reporting. For example, some businesses will report rent expenses separately from utility expenses, while others will report one figure for rent and utility combined. This lack of reporting can lead to significant data variability.
• A common variation of non-standardized reporting is the tendency for many businesses to over-report the 'other expenses' category. The data are edited to partially (but not completely) correct this problem.
• Because there is financial incentive to legally minimize their tax bills, businesses may defer income and expenses from one year to the next by utilizing tax rules. As a result, some statistical distortion may be present in the tax reporting of actual business operations in a given year.
• Statistics Canada cannot contact a business for clarifying figures reported on a tax form. This limits the ability to ensure the accuracy and consistency of the data.