Testing the Etch-a-Sketch Hypothesis: Measuring Ideological Signaling via Candidates’ Use of Key Phrases∗

Downsian theory predicts that presidential candidates should shift toward the general electorate’s median voter after securing their parties’ nominations. Motivated by this largely untested hypothesis, we test the theory using candidates’ campaign speeches as data. We develop a model to identify ideological cues in political text. After performing validation and robustness checks, we fit the model using presidential candidates’ speeches from 2008 and 2012. The results show that Barack Obama, John McCain and Mitt Romney did indeed make substantively significant shifts away from the ideological extremes after securing their parties’ presidential nominations. Well, I think you hit a reset button for the fall campaign. Everything changes. It’s almost like an Etch-A-Sketch. You can kind of shake it up and restart all of over again. -Eric Fehrnstrom, Spokesman for Presidential Candidate Mitt Romney When a top advisor to presidential hopeful Mitt Romney compared the campaign to an EtchA-Sketch, the classic toy that allows children to erase a previous image with the flick of a wrist and begin anew as if on virgin canvas, news of the comment spread like wildfire on television and across the blogosphere, with “#etchasketch” quickly trending on Twitter. Within hours, parodies referencing Mitt Romney and Etch-A-Sketch were appearing on YouTube. The statement became news and a source of derision not because the tendency for candidates to move to the ideological ∗Prepared for American Political Science Association Annual Meeting, Chicago 2013 center once they have obtained their party’s nomination had come as a surprise to anyone. What was striking was the candor with which Romney’s advisor, Eric Fehrnstrom, actually admitted to it in such a memorable way in response to an interviewer’s query about whether the prolonged nomination battle had forced his candidate too far too the right to be able to appeal to moderate voters in the general election. Despite the criticism Romney faced after Fehrnstrom’s moment of candor, the “Etch-a-Sketch” serves a perfect analogy for what political scientists expect to happen during presidential campaigns. Classic Downsian median voter theory predicts that candidates will appeal to median partisan voters during the primary, before converging to the point of the median voter in the national electorate for the general election.1 Yet, despite the elegance and intuitiveness of the theory, empirically measuring the shift proves difficult. Typically, there is no opportunity for presidential candidates to demonstrate a shift in governing; sitting presidents who are running for reelection infrequently face a serious primary challenge, so they feel little pressure to make a special effort to appeal to primary voters. The main record of ideological positioning available during the nomination and general election phases consists of a candidate’s own words (crafted with a team of speech-writers). It is easy to point to anecdotal evidence or slips of the tongue in accusing candidates of flip-flopping. But how might one go about determining whether a systematic rhetorical shift toward the center has taken place? We model the manifestation of latent ideology through a speaker’s key phrases. In so doing, we take a categorical (typological) approach to the representation of ideology, rather than treating it as continuously measured. We start from the perspective that there are a small number of prototypical ideologies in contemporary American politics, exemplified by well-known opinion leaders (e.g., talk radio hosts, columnists, prominent politicians of established adherence to particular per1This oversimplifies a bit. Strictly speaking, extending the logic of spatial modeling to multicandidate elections under the plurality rule does not lead to the same expectations (Cox 1985, Feddersen, Sened & Wright 1990). Primary elections are not, in fact operating under the plurality rule and, due to the complex multistage process by which candidates tend to pull out at different points in the nomination process, it is not obvious what should be anticipated under standard spatial models. Nonetheless, whatever the most strategic quantile pivot point may be, that of the Republican and Democratic primaries would naturally fall to the right and left, respectively, of the median during the general election.