« Plamegate is over (I hope) | Main | the other side »

September 12, 2006

campaigning: what works

I just voted in the District of Columbia primary election, having been personally approached by at least five candidates (mostly at my Metro stop), called at home by three live volunteers, sent countless letters, and barraged with a dozen or more "robocalls" (automated phone messages). Which form of outreach works best?

With help from CIRCLE, Young Voter Strategies has released a publication called Young Voter Mobilization Tactics (pdf). The heart of the report is a list of findings about the cost-effectiveness of various forms of campaigning, when directed at people under 30. We are able to present estimates of the extra votes cast per dollar spent in various ways. These estimates are based on genuine field experiments, in which organizations contact a random group of potential voters and leave the rest carefully untouched. They then compare actual voter records to see the difference in turnout rates. They also keep track of expenses. By dividing the expenses by the difference in turnout rates, we can estimate the efficiency of each method.

Top findings:

1. Youth can be efficiently mobilized. Costs can be as low as $8/additional vote, which is very attractive: campaigns generally expect a much lower yield from broadcast advertisements.

2. The more personalized and interactive the contact, the more cost-effective it will be. Door-to-door campaigning is highly efficient. Robocalls are a complete waste of money. They certainly make me mad. I would almost vote against the candidates who purchased them, except that I fear dirty tricks. (Pretending to send a robocall on behalf of your opponent might work quite well.)

3. Despite repeated efforts to find more effective messages, it appears that the medium matters, not the message. For example, if you organize a phone bank, it doesn't matter whether your callers use positive or negative scripts, simply provide information, or invoke civic duty. I find this a strange result, because calling someone is a communicative act, and I would think that what is communicated would matter. But perhaps the very fact that people are contacted makes them feel valued and encourages them to vote.

We do this kind of research, by the way, as part of a larger strategy. We hope to convince parties and candidates that it pays to mobilize young people. When young people are contacted about elections, they are more likely to discuss issues, understand the system, and feel efficacious. Therefore, giving them more attention in elections should make them more active participants.

September 12, 2006 8:32 AM | category: none


I have a question regarding #3 above. You say "Despite repeated efforts to find more effective messages, it appears that the medium matters, not the message". Did these studies track how the contacted voters voted?

It would seem to me that the measure of message effectiveness would not be whether or not the person goes to vote, at least not for most groups doing outreach. If the medium by which the message gets delivered is the important factor in getting people to the polls, an idea which I've long believed to be true (which is why I personally do outreach through concerts, canvasing on campuses, etc), than the effectiveness of the message, and the messenger, to get people to vote in a certain way is even more important.

Anyway, I am definitely curious to know if any of these studies looked at this. Otherwise I believe that the Agenda-Setting research of Iyengar and others would hold true here too (i.e. whatever comes to mind easiest will be what individuals will use when they make their decision for how to vote). But maybe this is simply outside of the scope of what these researchers are interested in.

September 14, 2006 2:43 PM | Comments (2) | posted by Alex Urevick-Ackelsberg

It's a good question. The outcome in all these studies is voting (i.e., whether the contacted person voted). It's possible that some messages are more effective at persuading people to vote for X instead of Y.

September 14, 2006 9:07 PM | Comments (2) | posted by Peter Levine

Site Meter