Factors associated with GOL use: a multivariate approach

View the most recent version.

In order to test the effects of various socio-demographic and Internet use characteristics on the likelihood of using the Internet to connect with government (see Methodology), a logistic regression model was used. Variables were selected based on a review of available literature, as well as the results of the descriptive analysis.

GOL users online longer, more frequently, and for a wider variety of activities

The results from the model show that, all things considered, the primary predictor of whether an Internet user will search for government information online, or communicate with government using the Internet, was the number of activities in which the individual was engaged. As the number of activities increased (from nine or fewer, to ten or more), so did the odds of an Internet user connecting with government via the Internet (odds increased by 3 times) (Chart 5). The activity variable used in this model is considered a proxy indicator for breadth of use—that is, while frequency and hours capture intensity of use, the activity variable appears to suggest a level of sophistication with Internet use, based on the number of activities for which the Internet is being used. A respondent who is capable of conducting a variety of activities via the Internet, such as buying goods or services, conducting electronic banking, paying bills, making travel arrangements, or researching investments, for example, is considerably different from the novice user who is just managing the use of email from home. As pointed out by Eastin and LaRose (2000), Internet use requires something more than simply computer skills—a level of interactivity is required that goes above and beyond basic computer literacy.

In addition to the number of activities variable, a number of other demographic factors played a significant role in determining whether or not an Internet user accessed government information online. These included marital status and gender. The relative influence of these factors varied. The odds of using government online services were slightly higher for men and for married individuals (Chart 5). Not surprisingly, education was an important predictor—with post-secondary education, the odds of GOL use increased by 1.6, compared with having a minimum of a high school education alone. Similarly, for an individual with a university degree, the odds of GOL use more than doubled compared with having only a high school diploma or less.

The presence of children under the age of 18 was not found to be a significant predictor of using the Internet to access government-related information. In addition, geographic location—whether the GOL user resides in an urban or rural location—did not have a significant influence on the use of government online services. This is of particular interest, since geographic location has been found to have a significant influence on general Internet use; urban dwellers are more likely to use the Internet than their rural counterparts (McKeown, Noce and Czerny 2007). The analysis here suggests that urban-rural location may be an important factor influencing Internet use, but it does not influence specific GOL use. In other words, rural Internet users were just as likely as urban Internet users to access GOL.

Other important predictors in the model included the duration of online use, as measured by the number of years an individual had been online. The odds of GOL use increased by 1.5 for individuals who had been online for five years, compared with those who had been connected for less than five years. Frequency and intensity of use also significantly increased the odds of Internet use to access government online information (odds ratios of 1.5, and 1.2, respectively).