Interpretation was based on descriptive, univariate analysis of the data.
For the most part, we focussed on summarizing key patterns and significant differences between the main demographic variables.2 Demographic groups significantly above and below the national average in the various measures were systematically highlighted.
When differences by target groups or a particular socio-demographic variable are examined in isolation, it should always be borne in mind that this variable may be correlated with one or even several other variables and that after having taken into consideration their interaction, the importance of the variable may be greatly reduced, to the extent that it is no longer significant.
Therefore, the demographic differences outlined in this report should be interpreted judiciously. Nevertheless, knowledge of the overall demographic differences will help facilitate understanding of the mind-set of consumers.
While overall results may be considered accurate to within +/-2.4%, nineteen times out of twenty, results of sub-groups contain larger margins of sampling error. Note that sampling error is the only potential error that can be measured. In addition, results of any survey may contain non-sampling errors including coverage error, measurement error, non- response error and, in some cases, other types of error. The quality of a survey rests on the effective management of these sources of potential error to achieve the lowest total survey error.
Also note that public opinion surveys cannot be exact measures of perceptions, but only approximations at a certain point in time.
All figures are based on the weighted sample. Percentages over .5 are rounded up.
Footnote 1 Region, age, gender, language, education, area of living.