Political Scientists Love Puzzles

At the recent New England Political Science Association annual meeting some of the most interesting work was presented by Matthew MacWilliams. Matthew was a Benjamin Franklin Scholar at The University of Pennsylvania where he graduated Phi Beta Kappa. He is a PhD candidate (ABD) at the University of Massachusetts. His current academic research interests include election forecasting using social media metrics, campaign communications, social media, political behavior, the Supreme Court, the politics of climate change, the politics of health care reform, interest group lobbying of the judiciary, campaign finance, and political campaigns. He was nominated in the fall of 2013 for a distinguished teaching award at the University of Massachusetts. We are pleased to have him guest today, with a provocative post questioning the new conventional wisdom about the lack of effectiveness of political television commercials – like those we’ve already seen in the Scott Brown-Jeanne Shaheen contest. Also keep an eye out for the launch later this month of Matt’s blog HashtagDemocracy.

Political scientists do love puzzles. They are the grist of our academic research mills. But sometimes puzzles that question academic orthodoxy, especially orthodoxy arrived at through seemingly airtight statistical analyses, are left unexamined.

Take for instance the orthodoxy that political commercials don’t matter. This scholarly assertion has grown over the ivy covered walls of academe and blossomed into a new conventional wisdom, voiced in a recent article by Sasha Issenberg in The New Republic, that Senate “Democrats should not be too worried about the inbound negative ads” attacking them this year.

That’s one puzzling whopper of a recommendation that may have taken the findings of academic studies a bit further than is warranted.

The contemporary scholarship, on which this popularized don’t worry, be happy wisdom about television ads is founded, begins with a series of natural experiments published in 2007 and 2008. These studies found presidential political commercials neither persuade nor mobilize voters. Statistical questions about the “as if” randomization requirements of the unit of analysis (media zones) and the measurement of ad exposure (gross rating points) used in these natural experiments aside, the fact is, as Goldstein and Ridout noted a decade ago, “presidential elections …[are] the place where one is least likely to find campaign effects in general and advertising effects in particular.”Moreover, the data on which experiments are based comes from only the closing months of the campaign.

There is simply no parallel between the cases studied by these presidential campaign natural experiments and the reality on the ground in 2014 Senate campaigns.Midterm campaigns for Senate bear little resemblance to national presidential contests.And the negative campaign ads confronting Democratic and Republican Senate candidates this cycle began months ago. For example, the Senate Majority PAC launched a $150,000 ad buy against Scott Brown before he announced his candidacy in New Hampshire, Americans for Prosperity spent $600,000 attacking Sen. Jeanne Shaheen beginning in mid March, and the Ending Spending Action Fund has already run ad attacking Shaheen and praising Brown. In fairness, Issenberg acknowledges “it’s nearly impossible to measure the effect of any individual TV or radio spot” but then ignores the implications of this statement. (Measuring gravity was a tricky business too. But Galileo did not deny its existence and test his theory by jumping off the Tower of Pisa!)

Issenberg and other advocates of “the new science of Democratic survival” are also convinced that persuasion is not useful or necessary in 2014. Democratic Pollster Mark Mellman and others have questioned the sagacity of this strategic recommendation, calling it a losing strategy. Mellman points out that even a 2.5 percentage point increase beyond presidential Obama levels – a highly unlikely event — would lead to a pick-up of just two House seats by Democrats in 2014. Issenberg admits there is a risk of focusing on mobilization writing “the risk is that November arrives and Democrats are so unpopular that the unreliable (voters) they need to mobilize and Reflex voter they need to persuade are too far out of reach.” I am certain there are many Republican strategists hoping Democrats do just that. The puzzle for me and any political scientist and strategist worth his or her salt is: Why would one party ever take that risk? This isn’t bean ball, and it is not an experiment. It is a contest for the future of the U.S. Senate.

The “new science” also seems firmly rooted in the past. With the findings of numerous natural experiments in hand showing direct mail and door-to-door are effective campaign tools, Issenberg and others recommend that the recipe for Democratic survival in 2014 is, of course, door-to-door contact and direct mail. (Can you say Drunkard’s Search Syndrome?) While there is little doubt these two techniques can be effective, it is also true that the calendar year is 2014 not 1970. Online advertising and social media are highly targetable and effective campaign tools. Ignoring their utility is a serious mistake especially when you consider the demographics of the Unreliable Voters that Issenberg argues are key to Democratic fortunes in the midterm elections. The Unreliable Voters are young. Sixty-one percent of them are 18-49 years of age. The younger members of this cohort are simply much more likely to be reached, persuaded, and motivated online or through social media than through snail mail. But if you still must question the effectiveness of online media, just review the case of how it was used by Governor Scott Walker’s campaign to win his hotly-contested recall battle in 2012.

Polling this year in Senate elections in states as varied as Arkansas and Minnesota has already demonstrated the power of broadcast and online commercials to move voter preferences. The argument for mobilization over persuasion is statistically untenable and unproductive – if winning is your bottom line. And online advertising is arguably a better tool for reaching the large cohort of young Unreliable Voters than direct mail. Using science and statistics to better guide campaigns is a smart and necessary pursuit. If you don’t do it, your opponent certainly may. But extrapolating the lessons learned from natural experiments, into a scientifically-supported campaign dogma seems dangerous. Issenberg cautions that “the new truths about voter behavior…may not reveal themselves until all the votes are counted.”  The alternate hypothesis is that the new truths are only half right and a grand natural experiment that counts on them exclusively is an exercise in folly.

 

This entry was posted in New Hampshire Politics, U.S. Politics and tagged , , , . Bookmark the permalink.

One Response to Political Scientists Love Puzzles

  1. Ed Lyons says:

    Dear Profs,

    Neat! Thanks for this kind of post! It’s great to have big doses of vitamins in the political information diet, in order to make up for all the empty calories out there. :-)

    Ed

Leave a Reply

Your email address will not be published. Required fields are marked *

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong>