This page is part of an undergraduate assignment at Davidson College.
The Effects of Endorsements on Media Coverage in the 2004 Democratic Primary Presidential
Grant Cooper, John Elliott, and Matt Garfield
Home |
Class Web Page | Combined Tables | Printer Friendly Intro | Davidson College |
In conjunction with Political Science 467, “Political Manipulation of the News,” our group looked to see if newspapers or politicians’ endorsements had an effect on print and broadcast media coverage. Our individual papers and results are accessible through the links above.
The main focus of our study is to understand what politicians do to influence media coverage. Specifically, we want to know if endorsements raise candidate coverage. Other factors we believe will influence coverage are the progression of time, the weekly news cycle, the number of candidates, and the approach and departure of primaries.
We evaluated the candidates we believed to have the best chance in the race, which includes Wes Clark, Howard Dean, John Edwards, Dick Gephardt, John Kerry, and Joseph Lieberman. In order to capture the main portion of the primaries we began our data collection with September 1, 2003 and continued it until March 3, 2004.
After compiling a list of endorsements from the press releases on each candidate’s website we used regression to determine their influence on media coverage. (To look at the Data please see our Tables).
We found that endorsements did not have a large impact on media coverage. Politicians’ endorsements only raised coverage in isolated instances. For example party officials’ endorsements raised the level of Dean’s broadcast coverage. Newspaper endorsements had more of an impact on newspaper coverage than broadcast coverage, as newspapers were not significant at all in raising the level of broadcast coverage.
In the regression model with both the newspaper and politician endorsement variables, the Party officials variable is only significant once in the newspaper coverage model and in once in the broadcast coverage model. In broadcast media, it has a significant negative effect on Dean, and with respect to newspaper coverage, it has as positive effect on Kerry.
Looking at newspaper endorsements is interesting because although past research shows that newspaper endorsements have a positive effect on media coverage, our results only found that Kerry’s newspaper endorsements tremendously affected his coverage negatively bringing down his national paper coverage significantly. None of the broadcast variables were significant.
Newspaper circulation was also a surprising variable because it was not significant related to any of the candidates in either broadcast or newspaper coverage.
We also found several other factors that explain media coverage. The number of progressed days in the primary was significant in influencing media coverage in print and broadcast outlets. As the days pass, candidates receive more coverage. A possible explanation for this is as primaries approach and the horse – race increases the media begins to cover the candidates more. All six candidates’ media coverage increase as time progressed.
The number of candidates also has a significant role on coverage in both mediums. As the number of candidates decreases, the amount of coverage the remaining candidates receive will increase. For example, John Kerry receives 3.276 stories a days when there are six candidates, but when there are only 4 candidates he receives 4.99 stories per day on broadcast outlets. When a candidate drops out the media has more time to focus on less people, which results in a rise in coverage.
The news cycle played an important role in the amount of coverage a candidate received. Our study showed a difference, however, in how that cycle materialized in the two mediums, print and broadcast. In print outlets Mondays were significant, but in broadcast outlets Mondays, Tuesdays, Wednesdays, and Fridays were significant. The difference can be accounted by the nature of the medium. Modern technology enables television to report more quickly than newspapers, so it is possible that they can run more breaking stories all week long. Newspapers, however, have a chance to compete on Mondays after the generally slow news period of the weekend passes.
As the Iowa primary approached the number of stories each candidate received increased in both outlets. After the primary the candidates that did not do very well, Lieberman and Clark received less coverage. The number of stories for Gephardt did increase greatly, largely because he didn’t do as well as expected and dropped out of the race one day after the Iowa primary, so the increase only reflects one day of data. The same pattern is true as the South Carolina and New Hampshire primaries approach. The candidates’ coverage increases as the primaries approach. John Edwards receives 6.32 stories in broadcast outlets when the South Carolina primary is 116 days away. When it is only 15 days, Edward’s stories jump to 9.32. After the South Carolina and New Hampshire primaries, candidates’ coverage dropped, with exception of Clark whose coverage rose after South Carolina.
There are many factors that influence the media coverage. Our anticipation that endorsements would raise the level of coverage for candidates did not prove true in this study, except for in isolated instances. Time, the number of candidates, and the approach and departure of primaries did however influence coverage as we predicted.
© Davidson College, 2004, Department of Political
Science, Davidson College, Davidson, NC 28035
Send comments, questions, and suggestions to Dr.
Patrick Sellers, Professor of Political Science.
Created: 4/29/2004. Last updated: 5/03/2004.