Statistics Canada
www.statcan.gc.ca
Common menu bar links
Perceptions of time and stress
Archived Content
Information identified as archived is provided for reference, research or recordkeeping purposes. It is not subject to the Government of Canada Web Standards and has not been altered or updated since it was archived. Please "contact us" to request a format other than those available.
Previous sections discussed differences in how Internet users and non-users allocated their time during the diary day. However, a related question is the extent to which users perceive time and stress in their life in general, and whether their perceptions differ from nonusers.
Table 8 presents several indicators concerning stress levels and certain aspects of time allocation. Although perceptions could not be characterized as vastly different, a slightly higher proportion of non-users’ answers revealed stressful patterns when compared with Internet users.
Non-users were more likely to feel that they didn’t have time for fun, and that they never had extra time on their hands. Heavy users stood out because they were less likely to be stressed, feel rushed or consider themselves a workaholic.
Although fewer Internet users identified with the indicators of stress, it is worth noting that a slightly higher proportion of Internet users asserted that they were willing to cut back on sleep when they felt they needed more time. Referring back to earlier findings (Table 2), both moderate and heavy users (after adjustment for social and demographic factors) typically spent less time sleeping during the diary day, albeit sleep times declined only slightly.
Table 8
General perceptions of time and stress among Internet users and non-users
Again, it is important to consider the extent to which any differences in perceptions of time and stress may have more to do with differences in demographic and socioeconomic characteristics of Internet users and non-users and less to do with Internet use itself. To do this, a multivariate method somewhat similar to the approach used earlier in this study is employed. Logistic regression analysis is used to model the chances that Internet users would indicate that they found most days “quite a bit” or “extremely” stressful compared to non-users, while attempting to account for other factors which likely affect stress levels, such as age, work situation, marital status, income, presence of children and the like (see Text Box 2 for more information).
Text Box 2:
Logistic regression analysis
Logistic regression is used in this section due to the nature of the outcome being modelled. Respondents were grouped into two categories: those stating that most days were “quite a bit” or “extremely” stressful, and those who stated otherwise. In earlier sections, the dependent variable of interest (time in minutes) was measured on a continuous scale, which required different methods of analysis (multiple regression and multiple classification analysis).
The logistic regression models the odds that individuals with different characteristics would report the outcome, in this case, finding most days “quite a bit” or “extremely” stressful, compared to a designated reference group. The influence of each characteristic is modelled while controlling for the influence of other variables in the model. The output of the regression procedure is a series of odds ratios. For each characteristic, values greater than ‘1’ represent increased chances of finding most days stressful, while ratios less than ‘1’ reflect decreased chances of finding most days stressful compared to the designated reference (or comparison group). In this case, non-users are designated the reference group against which comparisons will be made.
For more on logistic regression, odds ratios and their interpretation, see Menard (2001).
Readers will recall from Table 8 that slightly fewer heavy users described their days as stressful compared with non-users. Once accounting for a number of social and demographic characteristics (see the notes below Table 9 for the full list of control variables, and Appendix A for full results of the model), the model revealed that moderate Internet use was associated with reduced chances of being stressed (Table 9). However, it did not find a significant difference in the chances that heavy Internet users would be stressed, compared with non-users. Thus, the earlier differences between heavy users and non-users were largely explained by other social and demographic characteristics included in the model.
Similarly, although moderate Internet use was associated with lower stress levels, the strength of this relationship was dwarfed by other factors much more directly linked with stress (refer again to Appendix A). Most notably, age was an important factor, as middle-aged individuals had the greatest chances of finding days stressful while accounting for other social and demographic factors. Labour force status was also particularly influential, with those who were employed or attending school in the week prior to the survey much more likely to be stressed than retired individuals, household workers, or those seeking employment. In addition, women were more likely to report stress than men, and having children living in the household was also associated with higher stress levels. Interestingly, those with moderate income levels were less likely to report being stressed than individuals coming from low-income households, but those in the highest income group did not differ significantly from those in low-income households in this regard. Put in perspective, each of these factors held a stronger association with the chances that respondents would find most days stressful than their level of Internet use.
A separate model (not shown) examined the chances that Internet users would report “feeling rushed every day”. Similar to the findings above, moderate Internet use was associated with reduced chances of feeling rushed once other social factors commonly related to the chances that an individual would feel rushed were introduced1. Again, the relationship between this measure of stress and Internet use was relatively modest when compared with social and demographic factors more closely connected with the stress levels of respondents.
1. Relative to non-users, moderate users were less likely to report feeling rushed (reflected in the odds ratio .85*), while controlling for other factors in the model. The ratio for heavy Internet users (.83) was also lower than non-users but fell short of being statistically significant (the odds ratio for heavy users could only be described as significantly different from non-users with a 94% level of confidence (p=.06).
