The activists participating in the study were randomly selected from the Democracy for America (formerly Dean for America) volunteer database. Only active members of the campaign¹ were eligible for participation. Democracy for America contacted these activists by e-mail on behalf of the Pew Research Center for the People & the Press. Each potential participant received an initial e-mail request; one follow-up message was sent if they had not completed the survey within approximately two weeks. Respondents were directed to a website managed by the Pew Research Center, where they could complete the survey. Interviews were completed from Sept. 13, 2004 through Dec. 14, 2004.

A primary objective of the study was to assess the opinions of Dean activists before and after the Nov. 2 presidential election. Accordingly, two samples were drawn: a sample was interviewed in September and re-interviewed after the election (Wave I); a separate sample was interviewed only after the election (Wave II). The September survey drew 3,925 respondents, for a response rate of 13%. Slightly more than half (51%) of those responding to the September survey also completed the post-election re-interview.

In the second sample, one-in-five (19%) of those invited to participate completed the survey. The lower response rate in September is believed to be due to a server error that prevented some potential respondents from logging on to the survey website during one of the days on which invitations were being sent out.

Auxiliary information² in the Democracy for America database enabled the Pew Research Center to assess how those who responded in the study compare with all the activists the Dean campaign had on file. The most important difference is that the study participants were more likely than the average activist to have contributed money to the campaign. Two-thirds of those participating in the study (66%) contributed money to the campaign compared with only about half (49%) of all activists in the database. Aside from contributions, however, the study participants and the entire pool of activists had similar levels of campaign involvement.

This segment of the most engaged activists, while somewhat overrepresented in the study, tended to mirror the political values and demographic characteristics (gender, race, etc.) of typical members of the Dean campaign. Voting for Dean to opt out of public financing (74% among very active, 45% among less active) was one notable exception, as was the tendency for the most engaged activists to be somewhat older. Half (50%) of Dean’s very engaged activists were age 50 or older, compared to 36% of the less engaged activists.

On major political issues, however, Dean activists tended to hold the same views regardless of how involved they were in the campaign. Very engaged and less engaged Dean activists voted for Kerry in equal proportion (roughly 97%) and were equally likely to believe that using military force in Iraq was the wrong decision (99%). The two groups of activists also expressed similar opinions on less one-sided questions such as trade agreements, keeping troops in Iraq, and the role of churches in politics. This overall lack of difference between very engaged Dean activists and less engaged ones suggests that the study’s results are generally representative of the intended population: active members of the Dean campaign.

To ease the burden on individual respondents, most questions in the survey were divided across different forms of the questionnaire so that no respondent would have to answer all items. The table below shows the number of interviews in each of the waves and forms of the questionnaire, along with the margin of error for estimates based on each.