4.0
Detailed Sub-Group Analysis

The previous chapter concentrated on the principal reasons why Canadians did not have access to the Internet at home. In this section, we use key demographic characteristics of Internet users and non-users to highlight where the differences and similarities between each group lie.

Multivariate Regression Analysis

Because there are both an overlap in demographic characteristics and strong currents that cut across a number of lines (for example, youth, regardless of income, are more apt to be Internet users), a regression analysis to examine the likelihood of being an Internet user and/or having access at home is the next reasonable step in the analysis of the digital divide.

A logistic regression analysis examines the extent to which changes in independent variables increase or decrease the likelihood of an event. The “event” in this case is having access to the Internet in the past three months or having access to the Internet at home.

Regression Model with Key Demographic Variables

Testing the likelihood of recent and home Internet access by key demographic variables, we find that education, age, household income, gender and, to some extent, geography are all significant factors in determining the level of access. For data points in 1999 and 2000, an increase in education and household income levels are both positively linked to increased likelihood of Internet use. Higher Internet use and access levels are positively correlated with younger Canadians and men, as well as Canadians living in urban areas. In the 1999 survey, however, the urban-rural variable proved to be a non-significant factor in determining the level of home Internet access assuming all other demographic variables were held at a constant value.