What we can and can’t learn from the Iowa caucuses

If the Iowa caucus results were findings from a political poll, most people would see that they told us little about how the political winds were blowing. This is not just because Iowa caucus goers are unlikely to reflect the views of citizens nation-wide; there is a good chance that their preferences do not reflect the views of all Iowans. Caucus goers are a biased sample of Iowa’s citizens, and these “sample biases” shaped how the candidates fared. The most obvious bias is that only the most motivated Iowans are likely caucus participants. This benefits the candidates who most arouse their followers’ emotions and enthusiasm. Sanders’ results, particularly the overwhelming support he received from young caucus goers, bear this out. One might see in Trump’s poorer than expected showing a counter-example, but I doubt it. It is because of the passions he stirred that he did as well as he did. Entrance polls show him leading the field in few respects, but one area where he does lead is with first time caucus goers, 30 percent of whom backed him compared to 23 percent for Cruz and 22 percent for Rubio. (On the Democratic side Sanders trounced Clinton, 59 percent to 37 percent). No doubt it was the passions he (and Sanders) stirred that led many who had never before caucused to decide that it was something they had to do.

Trump may have only himself to blame for not winning. In hindsight, he appears to have made a strategic error by skipping the final debate, Not only did he miss one last time to rev up his followers, but he also allowed Cruz and Rubio to have more airtime to make their cases. Perhaps this decision contributed to his lagging Rubio and Cruz substantially, among the 45 percent of voters who made up their minds in the week preceding the caucuses—with most deciding during the last few days. These late deciders may also explain why Trump lost the expectation game. Even the polls closest to primary day used moving averages, so the trend away from Trump toward other candidates would not have fully registered.

Another reason why the Iowa caucus results do not necessarily reflect state voter preferences in the way a good poll would is because of the premium caucuses put on organization. In a well-designed poll, everyone whose views we want to represent has an equal chance of being queried, or we can weight respondents’ answers to mimic this condition. But not every Iowan had an equal chance of attending a caucus, even among those with political preferences of the same intensity. A well-organized ground game not only reminds people to caucus, but creates social pressures to encourage participation and provides services like transportation or child care that make caucusing easier. Moreover, the importance of a good ground game is amplified for precincts where smaller numbers of people caucus, as each person who is moved from non-participant to participant counts for more.

In terms of organization, the smartest thing Cruz did was to quietly cultivate evangelical pastors. White evangelicals constituted 62 percent of Republican caucus goers according to entrance polling, and fully a third of them supported Cruz. Trump and Rubio were each supported by only about a fifth of these voters. Although the proportion of Iowa Republicans who are white evangelicals is, no doubt, substantially higher than their 25 percent representation among all Iowans, their dominance among all Republican caucus goers most likely owes a good deal to church-mediated turnout efforts. Building on existing church organizational structures and courting evangelical pastors by itself can explain the edge Cruz enjoyed. Trump’s over-reliance on his charisma and under-investment in traditional field organization could only have aided Cruz and Rubio.

On the Democratic side, it may be organization that saved Clinton. She did about 7 percent better in sparsely populated counties than in densely populated ones, and although this may reflect the lesser radicalism of more rural Democrats. Her money and early start likely allowed her to build better get-out-the-vote organizations in areas with fewer and more scattered people. Interestingly, the one area where Marco Rubio led his party, and the densely populated area where Clinton did best, was the state capital, Des Moines. Voters in the capital are perhaps the state’s most politically connected. They may be more interested in choosing the strongest candidate to represent their party than voters in other areas, who are perhaps more prone to fall in love with personalities regardless of political viability. It is also likely that the mainstream party apparatus is stronger in Des Moines than in other areas. Both Clinton and Rubio may have been able to link to existing organizational networks more easily in Des Moines than in other counties and towns, and this may be why each did particularly well there.

Social networks complement organization in stimulating turnout. The chance that a person will attend a caucus is likely to increase if the person is embedded in a network of caucus goers. For a solitary individual considering caucusing, attending may be a daunting chore. If the person can recruit or is recruited by a like-minded friend or two, transportation can be shared, and attendance may seem more like a social event. Friends bringing friends is likely to have helped Sanders’ voter turnout in college towns and other areas where there are concentrations of young people, and would have aided Cruz’s efforts to raise participation among church-linked families.

It would be nice to know how Iowa’s Republicans and Democrats actually divide on their party’s candidates. With the exception of New Hampshire voters, they have, after all, been more exposed to the candidates and their campaigns than voters in any other state. But the skewed poll that is the Iowa caucuses does not tell us what Iowa voters think, except perhaps by coincidence. Ordinary polling could shed more light on Iowans’ opinions, except recent polls were geared toward predicting the votes of caucus goers. This means their samples and weighting decisions were skewed to mimic the biased sample of likely caucus goers. So whether one looks at the pre-caucus polling or the caucus results, there is little reason to think one has an accurate picture of who Iowa’s Republicans and Democrats in fact favor. The Iowa polls and caucus results tell us even less about national voter preferences. Only a few states are as white as Iowa or will hold Republican primaries in which 62 percent of those voting are white Evangelicals.

None of this makes the Iowa caucus vote politically irrelevant, and the candidates can take important lessons away from what happened there. Cruz’s victory suggests that opposing corn-based ethanol is no longer a kiss of death in that state—conservative voters hate government subsidies more than they love corn. Trump should have learned that charisma based on bombast will only take him so far. Without a strong ground-based organization, which will require him to dig rather deeply into his own pocket, he is unlikely to survive primary season.

Trump may have also lost in a bigger way. To the extent his success was based on his confidence and the appearance of success, the air may be leaking from his balloon. Should his apparent lead in New Hampshire disappear, he could disappear along with it. Cruz has a different problem. He is not going to find many states where his pitch to evangelicals and his ability to organize them will carry him as far as it did in Iowa. Iowa Republicans have a strong track record of preferring Republican candidates who go nowhere nationally. If opposing corn-based ethanol is not a kiss of death, winning Iowa may be. Rubio, however, must be pleased at the result. Iowa’s caucus goers have given him a strong boost toward emerging as the Republican establishment’s consensus candidate. But unless he follows up with a strong showing in New Hampshire by emerging from a pack in which his pre-Iowa support was not much different from that enjoyed by Bush and Kasich, he may find that Iowa counts for nothing. Doing well in New Hampshire may become more difficult for Rubio than it now seems. Given his Iowa performance, it is likely that in the lead up to New Hampshire, Cruz, Bush and maybe Kasich will be training their guns mainly on him. He is now a more immediate threat to their ambitions than Donald Trump.

Perhaps the candidate who can learn the most is Clinton, if she and her advisors will listen. She can say she won Iowa. She ended up with the most delegates, and her advantage is substantially enhanced when superdelegate preferences are counted. But she knows she is locked in a race. Not only is Sanders not going away, but his Iowa performance will increase his ability to secure funds and volunteers. Moreover, Clinton must get her groove back without coming down hard on Sanders. She will need the energy of his enthusiastic young supporters, as well as their votes, to win in November. Right now, she and Sanders are in the comfortable position of knowing that the voters who prefer one, by and large find the other a more than tolerable alternative. This can change if attacks on each other get too strong or personal. Harsh attacks on Sanders will simply elevate one of Clinton’s strongest negatives— personal perceptions of the candidate. They may also do little good since perceived weakness on honesty and integrity is her strongest specific negative.

The key takeaway for Clinton lies in the relationship between the candidate qualities that Iowa caucus goers said most mattered to them, and how they voted. Differences are astoundingly stark. Clinton prevailed by margins of 77 percent to 17 percent among voters whose greatest concern was who “can win in November,” and 88 percent to 9 percent among those who said “has the right experience.” Sanders prevailed with caucus goers most concerned with who “cares about people like me,” and who is “honest and trustworthy,” leading by 74 percent to 22 percent and 83 percent to 10 percent respectively. Clinton’s task is to change the dominant campaign narrative so that it plays to her strength. She would be particularly wise to focus on who “can win in November,” for it allows her to attack Sanders’ weaknesses without attacking him personally or distancing herself from his concern for inequality or his most popular policies. Forcing the narrative in this direction seems particularly promising because Clinton led Sanders among Democrats who are likely to care most about which party prevails in November, by 56 percent to 39 percent. Sanders was able to tie in Iowa only because he enjoyed a 69 percent to 29 percent advantage among voters who labeled themselves as independents or something else. In later primaries, the proportion of independents choosing between Clinton and Sanders is unlikely to be as high. Clinton also has room to show that she cares about people as much as Sanders does, but this involves embracing Sanders’ positions more than opposing them. For example, she would be wise to aspire for universal health care and be open to including a public option, rather than disparage Sanders’ proposals in blunderbuss fashion.

Sanders’ task, of course, is the opposite of Clinton’s, but in some measure he has the more difficult row to till. There is little he can say to make it appear that his experience, particularly in the foreign policy arena and in fighting terrorism, matches Clinton’s. Also, he could look mean spirited if he tries to suggest that Clinton does not care about ordinary people, and it would be out of character if he went after Clinton on honesty and trustworthiness. He may also want to do neither because he would certainly prefer a Clinton presidency to whatever a Republican triumph would yield. This leaves electability as the area where he might be able to improve voter perceptions, but to focus on this puts at center stage an issue that Clinton wants to discuss and has a substantial lead on. There is, however, nothing like winning primaries to deal with electability without discussing it. Winning in New Hampshire is, however, not likely to be of much help. Not only is Sanders so far ahead in the polls that it will be hard to win the expectation game, but he hails from New Hampshire’s neighbor, Vermont. These are states that many voters west of the Alleghany and South of the Mason-Dixon Line see as indistinguishable. So, a victory in New Hampshire will prove nothing—at least to voters outside of New Hampshire trying to read the tea leaves. But if he wins, or even comes close in South Carolina and Nevada, voters in later primaries may see the electability issue as having been put to rest. They would be wrong, but that is a topic for another day.