« Swing Voters or the Base? | Main | The Value of Turning Research Questions Upside Down »

11 October 2005

McDermott Paper

The Political Psychology and Behavior Workshop hosted Monika McDermott of the University of Connecticut who presented a paper (pdf) about the effects of AFL-CIO endorsements on Congressional candidates (see Barry's post for more details).

This paper uses data from an experiment as well as data from the American National Election Survey (ANES) and reaches similar findings in both analyses. The most notable findings are that an AFL-CIO endorsement activates liberal cues and increases the likelihood that a liberal will vote for the Democrat. Conservatives, on the other hand, don't appear to be influenced by a labor endorsement of the Republican candidate.

One question raised by an audience member was about the role of information. Aren't those who classify themselves as "liberal" or "conservative" likely to be high interest / information voters? If so, it only seems natural that they would receive and correctly process an AFL-CIO endorsement.

I was personally curious about whether or not the AFL endorsement is an ideological cue or whether it's really just another partisan cue. Two things suggest it is in fact an ideological cue: 1) in the experiment, the party of the candidate is given; and, 2) McDermott says ideology (interacted with the endorsement) does a better job of predicting vote choice than party id (interacted with the endorsement). In any event, I think the paper would benefit by addressing this issue more directly.

Others raised the question of how endorsements were distributed geographically. Are the endorsements focused in certain regions (for instance, Michigan or Illinois, and not in the South). Are races where there are endorsements somehow different from races where the AFL has not endorsed a candidate?

Another interesting point was made about whether or not respondents know what the AFL-CIO is or stands for. One possibility is that respondents view any endorsement as positive. What if, instead of the AFL-CIO, the candidate was endorsed by the ABC-DEF? Could high information liberal Democrats be driving the positive relationship McDermott finds between ideology and the endorsement?

Again, the nice thing about this paper is that there in an experiment and some external validation of that experiment by the ANES so McDermott's findings are not so easy to dismiss.

Posted by Andrew Reeves at October 11, 2005 9:01 AM

Comments

Kenny, McBurnett, and Bordua (BJPS April 2004) produce a similar finding about the asymmetry of group endorsements: NRA endorsements have bigger effects on Republican candidates than on Democrats.

Posted by: Barry Burden at October 11, 2005 10:14 PM