Familiarity Breeds Contempt for Romney, Liking for Paul

I noted here that Romney garnered a lower proportion of all votes cast in the 2012 Iowa caucus than he did in 2008 (though only 0.5% fewer votes in absolute number). Alex Seitz-Wald made a parallel observation regarding the Missouri, Colorado and Minnesota contests. In each of these states Romney attracted a dramatically lower absolute number of voters in 2012 than in 2008.

But looking at just one repeat candidate’s performance in consecutive elections can be misleading. The number of votes a candidate attracts in one primary election cycle versus the next could change because one year had unusually good or unusually bad weather or because a state moved its primary to a more versus less exciting spot on the calendar. Further, the same absolute number of votes has a different meaning each election cycle depending on how many other candidates are still in the race and on whether it is a high turnout or low turnout year in general.

There is however a way to roughly control for all these factors in assessing Romney’s performance in 2012 relative to 2008, because Ron Paul is also a repeat candidate. Whatever generic external positive and negative factors have affected Romney’s primary vote totals will also have affected Paul’s: The sun shines — or fails to — on all candidates’ voters equally.

The chart below presents the percent change in number of votes received by Paul and Romney in each of the nine states that have so far voted. The comparison with Paul make Romney’s 2012 run look even less impressive than it does already. Paul’s growth in number of votes from 2008 to 2012 beats Romney’s in all states, and in no case did he decrease his number of votes (his smallest gain was 1.5% in Nevada). In states where a generic factor helped all the candidates net a higher absolute number of votes, for example South Carolina’s 35% higher turnout in 2012 than in 2008, the bump Romney got in relative performance still lagged Paul’s. And Romney’s post-Florida run of decreasing his absolute number of voters from 2008 to 2012 — which he extended today in Maine — can’t be blamed on generic external factors: Ron Paul increased his vote totals in all five of those states.

Note: Maine results imputed at 7pm EST based on 95% of returns. Romney’s Iowa change does not show on the chart because his number of votes in Iowa in 2012 was almost equal to his 2008 total

Author: Keith Humphreys

Keith Humphreys is the Esther Ting Memorial Professor of Psychiatry at Stanford University and an Honorary Professor of Psychiatry at Kings College London. His research, teaching and writing have focused on addictive disorders, self-help organizations (e.g., breast cancer support groups, Alcoholics Anonymous), evaluation research methods, and public policy related to health care, mental illness, veterans, drugs, crime and correctional systems. Professor Humphreys' over 300 scholarly articles, monographs and books have been cited over thirteen thousand times by scientific colleagues. He is a regular contributor to Washington Post and has also written for the New York Times, Wall Street Journal, Washington Monthly, San Francisco Chronicle, The Guardian (UK), The Telegraph (UK), Times Higher Education (UK), Crossbow (UK) and other media outlets.

8 thoughts on “Familiarity Breeds Contempt for Romney, Liking for Paul”

  1. The question is whether this comparison shows that Romney is doing badly or that Paul is doing well. I can’t find a source saying how much money Ron Paul raised for his run in 2008, but if he has more money now that could explain why he has improved relative to Romney.

  2. I suspect that Keith’s observations are somewhat misdirected. I’ve noticed that in every caucus/primary since Nevada R-money got a lower total than he did in 2008, irrespectively of whether he won or lost the state. But, more importantly, every state other than South Carolina, so far, has had a noticeable drop in primary/caucus voters overall. R-money’s vote drops are a symptom of the general malaise, not just a statement about Romney. Paul was a novelty in 2008. He’s an established …er… something now. His followers get more loyal the more ridiculous his candidacy looks. There is no assumption that he’s going to win the nomination, so there is no hand-wringing over the Paul vote–people either vote for Paul or they don’t. Plot another set of numbers against these two–the totals. And the fourth set should be the campaign spending (2012) in each state–or some index indicating campaign investment (time and money, as opposed to one or the other). R-money invested heavily in Florida and NH and, to a lesser extent, Nevada, but he came in very late in Iowa (and most of the investment was the PAC), quickly lost SC and virtually ignored the other three states, hoping to conserve funds while holding the line (big mistake, of course, but that’s not the issue). Paul put up big stakes in Iowa, NH, SC and Maine. He spent very little time and money on other states. And the little bumps he got were corresponding to the campaign investment.

    Another set of dynamics is important. A top-tier candidate–as R-money was in 2008, prior to being wounded by Huckabee and finished off by McCain–should not be expected to rack up significant increases in totals from year to year. A marginal candidate in one year is more likely to win over a higher percentage in another year, even if he is still not the top-tier candidate. Add two other factors–strong anti-incumbency sentiment on the Right and the general weakness of the field, making choices more equitable–and this second-turn bump can become rather significant. Paul had nowhere to go but up–if he got lower numbers than in 2008, he would not still be running, but he’s been running for four years just like R-money and it’s paid off in a couple of the more libertarian/fascist-inclined states. In the other states, it hasn’t really paid off (like Nevada). But if you combine all these factors together, it’s easy to see how they add up in the states where there is a Paul bump and that some parts are missing in the rest of the states.

    Finally, negative campaigning pays off, but in an odd way. R-money and Gingrich have been mostly attacking each other. Santorum has also swung some dirt and received a bit in return. But Paul has seen surprisingly little attack campaigning against him, certainly no attack TV ads. So he adds by subtraction–from everyone else. If one of the other candidates started hammering Paul consistently in every state, his numbers would drop like a rock, down to the die-hard minimum. R-money is not just unlikable to Dems, he’s been hit very hard on the right as well.

    None of these factors have anything whatsoever to do with people liking Paul in relation to exposure and disliking R-money in relation to exposure. The duration of the campaign is just a random correlation with all the other factors. The other factors are causal–exposure is not.

    Having said that, it is a general rule of thumb that people who’ve met the candidate personally are more likely to vote for him. When the expected returns are fairly low, this kind of exposure can have a significant effect. When the expected returns are high, the effect is almost negligible. So, while there is a general correlation to exposure, it’s not quite the same kind of exposure Keith identifies.

  3. ShadowFox wrote: I’ve noticed that in every caucus/primary since Nevada R-money got a lower total than he did in 2008, irrespectively of whether he won or lost the state. But, more importantly, every state other than South Carolina, so far, has had a noticeable drop in primary/caucus voters overall. R-money’s vote drops are a symptom of the general malaise, not just a statement about Romney.

    But this is just the point of the comparison, if it were truly a “general malaise” it would be pulling down Ron Paul’s vote totals, but it isn’t.

  4. Nice work. But the use of a line graph puts the emphasis on the 2012 timeline and the possible trend, rather than the intertemporal comparison which is your main point. Why not use histograms, or a spiderweb chart?

    1. You are right. I have gone with a different format, although the nice thing about the line chart is that is shows values close to the zero line which a histogram does not (at least in powerpoint). Sure there are better styles, but given that I am changing planes as I see your comment, this will have to do.

Comments are closed.