View the most recent version.

The Labour Force Survey produces estimates based on information drawn from a sample survey of households. Somewhat different estimates might have been obtained if a complete census had been taken using the same questionnaire, interviewers, supervisors, processing methods, and so forth. The difference between the estimates obtained from the sample and a complete count taken under similar conditions is called the sampling error of the estimate.

While the sampling error is not known, it can be estimated from the sample data. One measure used is the coefficient of variation (CV), which is the standard deviation expressed as a percentage of the estimate. Since it can be very time-consuming and expensive to compute CVs for a large number of estimates from a complex survey such as the LFS, an indirect measure of reliability may be used. Generally speaking, the larger the estimate, the smaller its CV. Analysis has shown that LFS estimates of less than 1,500 typically have high CVs, making them unreliable.

In this publication, absence rates at the national level are considered reliable enough if they are derived from estimates of at least 1,500. For example, in 1997 the estimated number of male full-time employees aged 65 and over was 32,700. Since the estimated number of these men with absences was below the reliability threshold of 1,500, no rates are shown. Estimates not reliable enough to be published are shown as 'F'.

For provinces and regions, reliability thresholds are as follows:

Errors that are not related to sampling may occur at almost any phase of a survey operation. Interviewers may misunderstand instructions, respondents may make errors in answering questions, answers may be incorrectly entered on the questionnaire, or errors may be introduced in the processing and tabulation of the data. These are all examples of non-sampling errors.

Over a large number of observations, randomly occurring errors will have little effect on estimates derived from the survey. However, errors occurring systematically will contribute to biases in the survey estimates. Considerable time and effort was taken to reduce non-sampling errors in the survey. Quality-assurance measures, implemented at each stage of the data collection and processing cycle, included the use of well-trained and highly skilled interviewers, the observation of interviewers to detect problems of questionnaire design or misunderstanding of instructions, the use of procedures to ensure that data-capture errors were minimized, and the provision of coding and edit quality checks to verify the processing logic.