After the second Democratic debate, several media organizations took polls to see how the candidates were faring. They all conducted their polls in the time frame of June 28-July 1.

Three polls – ABC/Washington Post, Quinnipiac, and CNN – interviewed Democrats and independents leaning Democratic across the country. USA Today/Suffolk University interviewed likely Democratic caucus voters in Iowa only.

The ABC/WP poll is the real outlier among the three national polls. It shows the greatest divergence from the other polls with respect mainly to Bernie Sanders and Kamala Harris, and to a lesser extent with Joe Biden and Elizabeth Warren.  (The fifth place candidate, Pete Buttigieg, receives 4% support in all three national polls, 6% in the Iowa poll.)G

Sanders trails Biden by only 6 points in the ABC/WP poll, and leads both Harris and Warren by 12 points. The other two national polls, however, show Harris in second place, with Warren and Sanders essentially tied for third. Biden is in first place in all three polls, but with 7 points smaller support in Quinnipiac and CNN than in ABC/WP.

One possible explanation for the different results of ABC/WP is that this poll first asked respondents to name their favorite candidate without reading the whole list of names. 

After recording the answers, the interviewers then read the whole list of candidates and again asked who the respondents would vote for. 

The initial open-ended question favors candidates with higher name recognition – which would mean both Biden and Sanders. And having indicated their preference for one of those candidates in the first question, it is unlikely respondents would then change their response for the second question. 

Had the whole list of names been read right from the beginning, reminding respondents of candidates who are not as well-known as Biden and Sanders, it is possible ABC/WP would have obtained results similar to the other two national polls.

The important point to remember is that any preference at this point in the campaign is at best a tentative one. Voters tend to wait until closer to an election before making a final decision. And since the three national polls are interviewing voters across the country, many of them have hardly begun to pay close attention to the campaign.

Major fluctuations in the national polls are to be expected, especially once the Iowa Caucuses and subsequent state primaries are under way.

The Iowa Caucuses

We all know there is no national primary, and that national polls are conducted mostly because they are cheaper than state polls and give the illusion of insight into what might happen in the primary season.

For genuine inklings into the Democratic campaign, however, it’s important to look at the early state contests. What happens in Iowa can have a major effect on what happens in New Hampshire, and then in Nevada and South Carolina – all leading to Super Tuesday with many larger states chiming in, to include California and thirteen other states.

But even in the early states, polls in this time period may be interesting, but not predictive. As reported in USA Today, the poll in Iowa finds only one quarter of the electorate firmly committed to their first choice. 

Still, the results should give pause at least to Sanders. He is trailing Warren by 4 points, Harris by 7, and Biden by 15. He cannot afford to come in fourth, especially behind Warren, whose home state – like Sanders’ – borders New Hampshire. A poor showing in Iowa would almost certainly hurt Sanders in New Hampshire, giving Warren a boost.

The utility of polls like these in the early states is not to predict outcomes, but rather to indicate to the candidates where they have to focus their efforts. This Iowa poll should suggest to candidates where they may need to adjust their campaigns.

For us news junkies, the Iowa horserace is much more informative about the coming dynamics of the primary season than the national horserace – a mythical contest, created by media organizations who want to comment on the primary horserace, but don’t have the resources to spend money and time polling in state contests where results really matter.