Methodology
The target population for the main sample of Cycle 18 of the GSS was all persons aged 15 years and over residing in Canada, excluding:
- residents of the Yukon, Nunavut, and Northwest Territories; and,
- full-time residents of institutions.
In the survey, all respondents were contacted by telephone. Households without telephones were therefore excluded. These persons, however, represent less than 2% of the target population. Interviews are not conducted by cellular telephone so persons with only cellular telephone service are also excluded; again, this group makes up a very small proportion of the population, less than 3%.
Data for Cycle 18 were collected monthly from January to December 2004 inclusively. The sample was evenly distributed over the 12 months to represent seasonal variation in the information. The main sample was selected using Random Digit Dialing (RDD).
From the 31,895 households in the GSS Cycle 18 sample, 23,766 usable responses were obtained, producing a response rate of 75%.
Data limitations
The figures which appear in this report are estimates based on data collected from a small fraction of the population and subject to error. The error can be divided into two components: sampling and non-sampling error.
Sampling error is the difference between an estimate derived from the sample and the one that would have been obtained from a census that used the same procedures to collect data from every person in the population. The size of the sampling error can be estimated from the survey results and an indication of the magnitude of this error is given for the estimates in this report. If the estimated sampling error is greater than 33% (relative to the estimate itself), it is considered too unreliable to publish and the symbol 'F' is printed in table cells where this occurs. Although not considered too unreliable to publish, estimates with an estimated error between 16.6% and 33.3% should be "qualified" and used with caution. These are identified with the symbol 'E'.
All other types of errors, such as coverage, response, processing, and non-response are non-sampling errors. Many of these errors are difficult to identify and quantify.
Coverage errors arise when there are differences between the target population and the surveyed population. Households without telephones represent a part of the target population that was excluded from the surveyed population. To the extent that this excluded population differs from the rest of the target population, the estimates will be biased. In general, since these exclusions are small, one would expect the biases introduced to be small.
To the extent that the non-responding households and persons differ from the rest of the sample, the estimates will be biased. The overall response rate in terms of questionnaires that were complete enough to be useful for analysis was 75%. Non-response could occur at several stages in this survey. There were two stages of information collection: at the household level and at the individual level. As such, some non-response occurred at the household level, some at the individual level. Non-response also occurs at the level of individual questions so the overall response rate for some questions will be below 75%.
|
