Cyber Campaigning
19
** statistically significant at p<.01, * p<.05, two-tailed.
Logistic regression results showing parameter estimates and standard errors predicting the probability of
web use and having a web page. All variables are scored zero or one except age (years) and length of
party membership (years). For party membership, the excluded category are Liberal-National
candidates.
Source
2001 AES, candidates file.
What influence did these very diverse aspects of campaign activity have on web
utilization, measured together? The results of the two logistic regression models in Table
5 suggest that they and some influence on whether or not the candidate had a web page,
but not on web use. In terms of having a web page, as we saw in the previous table there
is a strong and significant association between spending time on media interviews and
having a web page. But having more party workers is also an important resource, as well
as preparing for the election campaign as early as possible. By contrast, receiving
numerous and diverse kinds of party support, through leaflets and visits, does not
influence the presence of absence of a web page. There are, then, important links between
various facets of the election campaign and web utilization, and most particularly with
whether or not the candidate had a web page. The next section whether or not these
campaign activities had any effect on the outcome of the election.
Evaluating the Electoral Consequences
So far we have examined the extent of web utilization among party candidates in the
2001 Australian election, the links to party and other resources, and the inter-relationship
with other facets of the election campaign. This section brings these components
together by examining the effects of campaign activities and web utilization on the vote.
25
The sample included upper house (Senate) as well as lower house(House of
Representatives) candidates and since the electoral system (and to some degree the
character of the campaign) is different for each house, the results in Figure 1 are
restricted to lower house candidates.
26
Most importantly, the results in Figure 1 show that web utilization by candidates has
a direct effect on the vote. Web use by candidates increases the first perfect vote by 3.1
percent, and having a web page has a slightly larger effect, at 3.5 percent, both net of
other aspects of campaigning. The standardized coefficients (in parentheses) suggest that
the overall importance of web utilization is about as important as all the conventional
methods of election campaigning combined, with the exception of party workers. Indeed.
In the case of the latter, each additional party worker increases the vote directly by 2
percent, and indirectly via a substantial influence on whether or not the candidate has a
Convention