Theories of voting abound, but most assume it’s a rational act.
We might consider each candidate’s positions on issues and vote for the one closest to us, at least on the issues most important to us.
{mosads}Or perhaps we take a shortcut, deciding that most of the time one party or the other represents our views and values better than the other, and then consistently support that party. We may err occasionally, but mostly we’re right, and we selected our preference in largely rational fashion.
Maybe we just vote for the candidate we “like” better. It’s hardly the most rigorous process, and we may not even be conscious of how we’re making up our minds, but it’s not irrational— our decisions bear some relationship to our preferences.
However, academics have long asked, what if our reasons for choosing a candidate really are irrational? What if they make no sense at all?
Two highly respected scholars recently ignited a small intellectual firestorm by concluding that purely random events — a series of New Jersey shark attacks in 1916 — had a statistically and substantively significant impact on support for Woodrow Wilson’s presidential campaign.
Now, that would stretch the bounds of rationality to, if not beyond, the breaking point.
Others cast doubt on the findings, leaving them in limbo, at best.
Long before shark attacks became the rallying cry for those concerned about voter irrationality, Stanford professor and jazz drummer Jon Krosnick directed attention to a bloodless form of irrational voting behavior.
Krosnick and others produced evidence that the order in which candidates’ names appeared on the ballot affected citizens’ votes, and even election results.
One of Krosnick’s early studies, of 1992 Ohio races, found statistically significant name-order effects in 48 percent of them. Another 40 percent of those contests also featured name-order effects that didn’t achieve statistical significance.
In a meta-analysis of all published studies on the topic by a variety of scholars, which included 1,086 tests of name-order effects, Krosnick found candidates did better when listed first than when they appeared farther down the ballot in 84 percent of cases.
Obviously, some analyses contested his conclusions, but these were few and far between.
All the evidence Krosnick adduced might lead us to bemoan, bewail and lament the irrationality of a few American voters, but in the end, all we could really do is shake our heads and wag our fingers at the foolish choices some make.
But what if government exacerbated this irrational behavior by the way it ordered names on the ballot?
Marc Elias, this generation’s preeminent voting rights attorney, took Florida to federal court, arguing that its law giving consistent first-place position to candidates of the governor’s party violated the 1st and 14th Amendments.
The judge agreed, strongly.
Interestingly, and foolishly, so did the Republicans who, in contesting the suit, argued that they “stand to be most directly harmed by a change” in the ballot order rules.
Another Stanford professor, Jonathan Rodden, provided further evidence of ballot-order effects in a situation clearly revealing the GOP’s malevolent intent.
North Carolina also placed the names of those candidates who were members of the governor’s party atop the ballot.
However, in 2016, after losing reelection, the outgoing Republican governor, knowing this law helped the incumbent gubernatorial party, signed a law replacing this practice with an alphabetical ordering, purposely trying to improve his party’s prospects, while accidentally creating a natural experiment.
Rodden compared the same precincts in 2016, when Republicans were always listed first, and in 2018, when Republicans were listed first in only half of the precincts, and found the change in ballot order had meaningful effects on outcomes.
Democrats’ vote share increased 1.5 percentage points more in precincts where Republicans were listed first in 2016, but not 2018, than in precincts where Republicans were listed first in both years. In open seats there was an 8 percentage-point difference.
With strong empirical backing, and a favorable court ruling in hand, Elias has brought suit on the same issues in Arizona, Georgia and Texas, which are governed by similar rules and by Republicans seeking to benefit from irrationality.
Mellman is president of The Mellman Group and has helped elect 30 U.S. senators, 12 governors and dozens of House members. Mellman served as pollster to Senate Democratic leaders for more than 20 years and as president of the American Association of Political Consultants.