Home Topic Model M(ichae)L News Research




Using Randomized Experiments for Causal Inference: Do Appearances Matter?

You were assigned to a treatment condition with a photograph:

Bio

I am a Ph.D. candidate in American politics and political methodology. My dissertation develops methods of text-based machine learning to assess political coverage and better understand the ability of the media to monitor elected officials on behalf of voters. I am currently working on using text-based machine learning to provide real time analytics of media coverage across the political Internet. My other ongoing research examines changes in ideology over time.

One project draws on over 1 million individual poll respondents to measure public attitudes towards the New Deal at the state level from 1936-1952. Another project scales the text of present day campaign websites to measure whether candidates moderate their campaign message between the primary and the general election. My past research has examined whether citizens vote based on candidate appearance (they do) and whether people who donate to presidential campaigns have any close friends who donated to the opposing party’s candidate (they do not).

Other visitors to the site saw a control condition without a photograph:

Bio

I am a Ph.D. candidate in American politics and political methodology. My dissertation develops methods of text-based machine learning to assess political coverage and better understand the ability of the media to monitor elected officials on behalf of voters. I am currently working on using text-based machine learning to provide real time analytics of media coverage across the political Internet. My other ongoing research examines changes in ideology over time.

One project draws on over 1 million individual poll respondents to measure public attitudes towards the New Deal at the state level from 1936-1952. Another project scales the text of present day campaign websites to measure whether candidates moderate their campaign message between the primary and the general election. My past research has examined whether citizens vote based on candidate appearance (they do) and whether people who donate to presidential campaigns have any close friends who donated to the opposing party’s candidate (they do not).

Researchers in many fields have found that people routinely make costly decisions based on facial inferences, even when those inferences are shown to be uninformative. In politics, observational studies have shown a relationship between candidate appearance and election results, but candidates who work hard to get elected may also work hard to get high quality campaign photos. To rule out this and many other alternative explanations, my past research has randomized whether candidate photographs appear on mock ballots in order to show that exposure to candidate appearance influences voters.


"Face Value? Experimental Evidence that Candidate Appearance Influences Electoral Choice." (with Douglas J. Ahler, Jack Citrin, and Gabriel S. Lenz) Polical Behavior (2017).

According to numerous studies, candidates’ looks predict voters’ choices—a finding that raises concerns about voter competence and about the quality of elected officials. This potentially worrisome finding, however, is observational and therefore vulnerable to alternative explanations. To better test the appearance effect, we conducted two experiments. Just before primary and general elections for various offices, we randomly assigned voters to receive ballots with and without candidate photos. Simply showing voters these pictures increased the vote for appearance-advantaged candidates. Experimental evidence therefore supports the view that candidates’ looks could influence some voters. In general elections, we find that high-knowledge voters appear immune to this influence, while low-knowledge voters use appearance as a low-information heuristic. In primaries, however, candidate appearance influences even high-knowledge and strongly partisan voters.


So if you have a private jet and would like to hire me to fly your plane, please learn from voters and do not use low information heuristics. Because despite what it may look like, I have no idea how to fly a plane! On the other hand, I do know how to design an experiment.