Who Said What?

Shanto Iyengar
Special to washingtonpost.com
Tuesday, September 19, 2006; 8:30 AM

With the midterm elections rapidly approaching, TVs across the country will be swamped with "issue ads" championing one party over the other. How effective are these general campaigns, which usually don't focus on individual candidates?

washingtonpost.com teamed up with the Political Communication Lab at Stanford University to figure out which issues work best for the two parties. More than 4,000 readers participated in the washingtonpost.com - Stanford survey. Because these volunteers overwhelmingly self-identified as Democrats, the study was also issued to a more diverse group of registered voters. The results:

  • Overall, many people have already made up their minds about the election. Advertising will still make a difference, but mostly in getting supporters to the polls.
  • There was a significant decrease in party defection when partisans watched ads in which their party addressed an "owned" issue.
  • Ads featuring terrorism and immigration boost Republican votes, and people who list these issues as their chief concerns will generally vote Republican. Ads highlighting a tough immigration stance and stating that "we need a fence" are particularly effective for the GOP.
  • Ads featuring Iraq and the economy help the Democrats. Most people who cite these issues as most important will vote Democratic.
  • The Study

    As the 2006 campaign kicks into gear, Americans will inevitably be subjected to a barrage of "issue advocacy" in the form of television ads sponsored by the political parties and their surrogates. These ads typically urge the audience to vote Democratic or Republican, but without reference to individual candidates.

    Our findings suggest that Republicans face an uphill task to retain control of Congress. At present, voters disenchanted with President Bush and critical of the decision to go to war outnumber those on the other side. These disgruntled voters overwhelmingly intend to vote Democratic. Moreover, the traditional Republican "money advantage" may be to little avail -- our results suggest that issue ads will do little to attract or persuade voters; instead, their chief effect is likely to be to push Democrats and Republicans into their respective camps.

    Between August 22 and September 6, invitations to participate in an online research study about campaign advertising appeared in the politics section of washingtonpost.com. In response, more than 4,000 users of washingtonpost.com completed the study. Based on past experience, we anticipated that this "volunteer" sample would be tilted heavily in favor of Democrats. (It was.) For purposes of comparison, therefore, we also administered the study to a more diverse group of 860 self-identified registered voters drawn by the opinion research firm of Polimetrix.

    Participants in each experimental condition first watched a pair of campaign ads -- always one from each party (or affiliated group) -- and then completed a survey of their attitudes. The ads used were actual ads previously aired either by a party committee or groups such as MoveOn.org. The ads focused on the domestic economy (jobs and gas prices), illegal immigration, the Iraq war, or terrorism. Our objective was to find out how much, if at all, Democrats or Republicans stand to gain or lose by emphasizing these particular issues.

    Given where the parties stand on these issues, the conventional wisdom suggests that the economy and Iraq should cut in favor of Democrats, while immigration and terrorism should be the favored issues for Republicans. We tested this notion by randomly assigning study participants to watch different pairings of issue ads. For instance, some watched a MoveOn.org ad insinuating that Republicans were in cahoots with the oil companies and an ad from a pro-Republican group urging the construction of a border fence. Others watched  Republican National Committee and Democratic Senatorial Campaign Committee spots on national security. This paired arrangement allowed us to observe variations in voting preference under several different combinations of issues.

    View the eight ads used in the study

    CONTINUED     1        >

    © 2006 The Washington Post Company