05/27/2008 06:50 pm ET Updated Dec 06, 2017

Last 3 Primaries: Why So Few Polls?

A reader sent the following email over the weekend:

We have consistently seen a flurry of polls ahead of every democratic primary contest -- however, since Oregon and Kentucky polls have all but gone silent... the upcoming states, Puerto Rico, Montana, and South Dakota have been virtually unreported in the public. What do you attribute this to? It seems almost like some polls have always had a second agenda, and with the coordinated silencing of polls at a time when there is also a coordinated push to reshape the media focus on the contest as being over, seems very suspect. Given your excellent nonpartisan analysis of the numbers and meaning behind the polls, I wonder if you could shed some light on the lack of polling at such a crucial time in the nominating process?

"Coordinated silencing?" I don't see it.

The following table shows the number of polls we logged for both the the final week before each primary (the last column on the right) and the full month before the final week (second column from the right). Obviously, there has been only one poll in the last month in any of the three remaining primary states, but that represents a considerable tapering off since the Pennsylvania primary.

05-27_poll count.png

Here are four reasons for the fall-off:

1) The number of delegates at stake in each state has also trended down considerably since Pennsylvania, and with it the budget priorities of pollsters and the national news media. Notice that the number of delegates up for grabs in Montana and South Dakota combined (47) is a fraction of those that were at stake in either Pennsylvania (188) or North Carolina (134). Also, national media outlets sponsored just 9 of the 124 polls logged in the table above (all from PA, NC and IN).

2) Puerto Rico is an obvious "outlier" of sorts, given that polling there requires Spanish speaking interviewers and (hopefully) a knowledge of the island's culture with respect to both politics and telephone surveys. Mainland pollsters with a lack of experience in Puerto Rico are understandably reluctant to field surveys there.

3) Many of the polls we have seen for previous primaries are sponsored by local newspapers or television stations. Larger states have more media markets and thus more media outlets and more competition among them. Those factors tend to translate into more polling in bigger states.

4) The remaining polls without media sponsors come mostly from pollsters or public relations firms that release polls for their "marketing" value, to generate traffic to their websites, or both. Many of these outlets are now shifting their resources to general election polling. Many have already spent more than they ever anticipated on primary polls. Most presumably see less of a return (in terms of marketing or traffic) from polls in those final states.

Altogether, the fall-off seems mostly evidence of a shifting market forces, not any coordinated effort.