Quick links (pdf): Survey Questions | Survey Results

YEAR ONE (2005-06): Discourse Analysis of Blogs, Viral Videos, Bush in 30 Seconds, and The Daily Show

The project examined four sites, beginning with discourse analysis of the content of the four kinds of sites and productions:

• 150 independently produced 30 second Quicktime movies for MoveOn.org Bushin30Seconds campaign
• 110 web-logs that engage political discussion of media representation of U.S. foreign policy, particularly with respect to the invasion of Iraq
• 50 online discussions (threads, blogs, comments posted to blogs) that address Jon Stewart and The Daily Show, with particular focus on Stewart’s 2004 appearance on the CNN talk show Crossfire
• 40 independently-produced multimedia “memes” that address U.S. political issues

YEAR TWO (2006-07) Interviews with Producers/Writers of the Digital Media

From the sample surveyed (see below), we selected 35 persons to be interviewed through semi-structured interviews generally 45 minutes to 1.5 hours in length. These questions were developed through collaborative small research clusters; piloted; and revised to invite open-ended inquiry.

Interviews were then transcribed and coded, using a variety of inquiry methods. The primary themes that emerged are currently being compared across the four groups; and analyzed to understand interviewee diverse motivations, relationship to media as consumer and producer, and online practices and behaviours.

How we Approached the Four Sites

This process of identifying discourses is similar to typical qualitative analyses that identify themes or codes (which may or may not be discursive). It is most similar to that of Grounded Theory (Glaser & Strauss, 1967; Glaser, 1994b). From these codes, categories are established. The coding scheme is constantly revised using the method of Constant Comparison (Glaser, 1994a), new categories are constantly created and others collapsed to ensure the most meaningful categories are used. Comparisons are also made across texts and movies to establish commonalties and develop larger more meaningful categories.

The Daily Show

In this early phase of the project, it was our intent to gather online data related to The Daily Show (TDS) that could be used in later discourse analysis. After a preliminary scan of blogs referring to The Daily Show, we decided to concentrate our efforts on Jon Stewart’s appearance on CNN’s Crossfire in October 2004. This proved to be a rich source of data. Not only did a Google search using the terms “Jon Stewart” and “Crossfire” yield 260 000 results, but discussion threads and blog posts related to Stewart’s appearance revealed in-depth discussions of mainstream media failings, government corruption and the role of journalists in democracy.

In order to generate as wide a range of opinions and ideas as possible, we collected data from a variety of sources, including discussion threads, fan websites, blogs and comments posted to blogs. We located these sources using a variety of search engines, most particularly Google Groups, Google Blog, Blogger and Technorati. In all, data was collected from more than 60 sites.

Both during and after our data collection, we employed critical discourse analysis in order to identify common themes and discourses within the postings. Frequent discussions with those researching the Bush in 30 Seconds Campaign and the political blogs allowed us to compare discursive strategies and to identify both similar and unique categories of analysis.

Bush in 30 Seconds Movies and Animations

The research team spent months conducting a close visual and textual analysis of 150 finalists of Bush in 30 Seconds (many of whom are featured in our links section). The research team analyzed each frame, visual choice, text, and narration of all 150 Bushin30 movies.

Political Blogs

Our intent in the preliminary phase of the project was to gather data that could be used to structure later data collection and analysis. In the political blog section of the project, this meant identifying and reading a wide range of political blogs.

The first step in this process was locating the blogs we intended to examine. Since our primary interest was on the intersection of politics, the media, and the Iraq War, we searched for blogs that addressed all three of these issues. To gain a more thorough understanding of how the issues were seen by the blogosphere, we sought a wide variety of blogs and bloggers. Our final selection encompassed a range of ages, locations, racial and ethnic backgrounds, political views, and audience sizes. In the end, we identified and read 110 political blogs.

In order to maximize the variety of the blogs we selected, we combined several different techniques for locating the blogs, including:

1. Looking up well-known blogs of which we had pre-existing knowledge.
2. Searching on both Google and Google Blog Search using keywords related to our topics of interest.
3. Searching Technorati and The Truth Laid Bear.
4. Following links from other blogs.

Once we had generated a list of blogs that addressed the topics we were interested in, we identified posts that addressed project areas, and analyzed them for common themes and discourses. These findings were then used when constructing the survey and interview questions.

Survey and Quantitative Methodology

The quantitative component of the project involves the administration of an online survey to individuals whose online contributions (blogs, memes, video productions, etc.) were tracked and analyzed using CDA. Four versions of the survey were developed for each of four groups: The Daily Show (TDS) viewers who participated online discussion about TDS; meme producers, political bloggers, and creators of Bush in 30 Seconds media. All four groups responded to 42 core questions that were the same for each group (see Appendix A), as well as a core group of demographic questions. Unique questions for each group were also added (see Appendix B), which were aimed at gathering information about their motivations for their online contributions.

Survey Instrument Development

The survey was developed by the research team. It incorporates several questions from a Pew Survey on media habits, as well as World Values Survey questions regarding political views. These reasons for inclusion questions are two-fold: first, they provide a point of comparison for analysis; second, they provide an added layer of validation.

Once developed, the survey was validated by this research team in two ways. First, four [LEP1] experts reviewed the survey and provided input. These experts included individuals from various North American universities with expertise in survey design, as well as one survey researcher from a prominent research firm in the United States. The survey was revised based on expert feedback. Next, the survey was piloted with eight individuals representative of the survey sample. Pilot respondents were asked to provide feedback, which was incorporated. As well, the data gathered from the pilot were tested. Appropriate revisions were made, and the survey was live.

Sampling

Non-probabilistic convenience sampling was used to obtain respondents. This sampling method was selected because little is known about the actual population given the vast nature of Internet blog and digital media producers and contributors.

In order to identify a sampling frame, the researchers logged and recorded the contact information of creators of blogs and other digital media being catalogued and studied as part of this research. A total of 600 individuals were identified. Personalized emails were sent to all individuals except The Daily Show group. This latter group received a bulk email. In one week, follow-up emails were sent to remind individuals to participate. All invitational correspondence included a covering letter outlining the purpose of the research, and providing the URL to access the survey.

Data Analysis

Data has been analyzed in several ways. First, summary statistics are presented to provide an overview of media habits of the groups sampled. Second, analyses of variance will be explored among groups, and among various demographic characteristics (e.g., gender, country or region of residence, etc.). Related to this, analyses of variance between these respondents and general populations of the US and possibly other countries will be explored for questions duplicated from the Pew survey and the World Values Survey. Finally, factor analysis will be performed to determine what clusters of media habits, perceptions, and political views may exist.