How Campaigns Facilitate Connections

March 4, 2003

Kirsten A. Foot, Steven M. Schneider, Elena Larsen, Meghan Dougherty, and Michael Xenos

Campaigns Co-Producing the Electoral Web Sphere

The electoral Web sphere is increasingly a realm where citizens look to become informed and perhaps persuaded about a candidate or an issue, to involve themselves in the political process by connecting with other individuals, groups and organizations, and to mobilize others to become politically active. A Web sphere is a dynamically defined, interlinked set of Web materials oriented around a common event, within a particular timeframe. By creating links (“outlinks”) from candidate Web sites to Web sites produced by other political actors, campaigns participate in the co-production of an electoral Web sphere within which citizens and others who use the Web can engage in political action.

To evaluate how campaign sites facilitated connections in the 2002 electoral Web sphere, we analyzed a sample of 515 Web sites produced by candidates running for a U.S. House, Senate or gubernatorial office. We examined campaign sites for the presence of at least one outlink to other Web sites produced by each of the following types of political actors: civic/advocacy groups, government bodies, press organizations, and political parties. We did not count the total number of outlinks on a site.

Prevalence of Outlinking

Eighty-four percent of all campaign sites had at least one outlink to a site produced by a civic/advocacy group, government body, press organization or political party. Third party candidate sites were most likely to outlink, and had the highest average number of outlink types. Ninety percent of third party candidates sites outlinked, in contrast to 82% of major party campaign sites. The mean number of outlink types by third party campaign sites was 1.97, versus 1.56 by major party campaign sites. Campaign sites for House offices were only slightly more likely than Senate or gubernatorial candidates to outlink at all, however, House sites had a modest yet significantly higher average number of linked-to site types (1.73 average for House versus 1.56 for Senate and 1.30 for gubernatorial). Incumbency status, race intensity and candidate gender were not significant factors. The average number of site types linked to by all candidates (1.66) indicates that nearly all campaign sites tend to link to only one or two of the four other types of political actor sites examined in this analysis. (See Table 1)

Table 1: Frequencies of Outlinks in Candidate Web Sites

Candidate Characteristics	% sites with outlinks	Mean # of outlink types	Sites Examined
All Candidates	84%	1.66	629
Party Type*
Major Party	82%	1.56	478
Third Party	90%	1.97	151
Office Sought*
Governor	79%	1.56	94
House	85%	1.73	458
Senate	82%	1.30	77
Incumbency
Incumbent	84%	1.65	158
Challenger	84%	1.65	471
Campaign Intensity
Low	82%	1.71	347
Medium	87%	1.64	159
High	83%	1.50	123
Candidate Gender
Male	83%	1.64	536
Female	86%	1.74	93
* indicates statistical significance at .05 or greater for independent variable in multi-variate regression

Campaign Linking Practices

Campaign linking practices varied not only in the average numbers of site types linked to, but also by the specific types of sites linked to. Sites produced by government bodies were most likely to be linked to by campaign sites (51%), followed by political party (43%), civic/advocacy group (38%) and press (32%) sites.

However, campaign sites for third party candidates tended to link to different types of sites than campaigns affiliated with major parties. Third party candidates were more likely to link to sites produced by civic or advocacy groups (59% third versus 32% major party) and political parties. Candidates for House offices were twice as likely as Senate candidates to link to civic or advocacy groups (42% versus 21%). Major party campaign sites were more likely to link to government sites (56% major versus 35% third party). Incumbent candidates were more likely to link to government sites than were challengers (62% versus 47%), a finding which may be explained by incumbents’ existing relationship with a government body. Candidates in low intensity races were more likely than candidates in medium or high intensity races to link to civic or advocacy groups. (See Table 2)

Table 2: Percentages of Campaign Sites with Outlinks to Specific Site Types

	Characteristics of Candidates	% of candidate sites with outlink to type of site
		Civic/ Advocacy Group	Political Party	Press	Govern- ment
	All Candidates	38%	43%	32%	51%
	Party Type	*	*
	Major Party	32%	35%	32%	56%
	Third Party	59%	69%	31%	35%
	Office Sought	*	*
	Governor	35%	39%	36%	46%
	House	42%	45%	32%	53%
	Senate	21%	39%	26%	44%
	Incumbency
	Incumbent	34%	40%	27%	62%
	Challenger	40%	44%	33%	47%
	Campaign Intensity
	Low	43%	48%	29%	51%
	Medium	36%	38%	35%	54%
	High	28%	38%	36%	47%
	Candidate Gender
	Male	37%	43%	32%	51%
	Female	46%	44%	31%	51%
* indicates statistical significance at .05 or greater for independent variable in multi-variate regression

Summary

In brief, party affiliation, office sought, incumbency status and the intensity of a race all contribute independently to our analysis assessing the impact of candidate characteristics on particular practices of linking (or not linking as the case may be). Since each outlink on a site represents a choice on the part of the site producer, patterns of linking on campaign sites may reveal strategic assumptions about the strategies or different views of the electoral Web sphere by different kinds of campaigns. Sites with more outlinks to a greater range of other site types may reflect an integrationist perspective on the electoral Web sphere. Linking helps citizens make connections within the electoral Web sphere more easily. Fewer outlinks on campaign sites may reflect a more isolationist perspective on the part of the campaign, or simply a desire to increase the “stickiness” of the campaign site.

Methods & Sources:

This analysis was conducted by researchers associated with WebArchivist.org, a research group directed by Professors Kirsten Foot and Steven Schneider at the University of Washington and the SUNY Institute of Technology. This analysis is part of a broader project investigating the role of the Internet in the 2002 elections. The project is sponsored by a grant from The Pew Charitable Trusts.

Campaign Web sites were identified through searches on publicly available and subscription-only Web sites, including politics1.com, www.greenpapers.com, www.washingtonpost.com, www.google.com and the National Journal. For this report, approximately 515 sites were analyzed during each of the final 12 weeks of the campaign. The sites for these analysis sets were chosen by randomly sampling sites from races at a variety of competitiveness levels, with an emphasis on competitive races. Campaign Intensity was measured using competitiveness rankings available on publicly available and subscription-only Web sites, including www.washingtonpost.com and the National Journal.

Citations:

We encourage you to use the reports published on PoliticalWeb.Info in your own publications. When referencing this report, please consider using the following citation: “How Campaigns Facilitate Connections,” by Kirsten A. Foot, Steven M. Schneider, Elena Larsen, Meghan Dougherty, and Michael Xenos. PoliticalWeb.Info, Kirsten A. Foot & Steven M. Schneider (Editors), March 4, 2003, http://politicalweb.info/reports/connections.html.

© 2002-2003 University of Washington