Mark Blumenthal | November 12, 2007
Topics: 2008 , The 2008 Race
If you haven't had a chance to read it yet (and you should), my colleague Charles Franklin has taken a closer look at the trend lines in our Democratic presidential charts (as well as his own more "sensitive" estimators) in light of four new national polls and three new surveys from New Hampshire, all released in the last week. He concludes that the Democratic race has narrowed slightly in New Hampshire but sees less change nationwide:
I think the evidence favors a view that Clinton's problems in New Hampshire may be more specific to that state rather than reflecting a more widespread change in her campaign's fortunes
I want to add a few thoughts about why that may be, with a closer look at the three newest surveys from New Hampshire.
Charles made the point that the three new polls in New Hampshire - from Marist College/WNBC, Rasmussen Reports, and the University of New Hampshire/Boston Globe - show a bit more change since the recent debate than is evident in our trend lines. Let's take a look at the "apples to apples" comparison he mentioned.
All three organizations were in the field last week and had conducted prior surveys in mid-September or early October. All three showed similar 5-6 point declines for Clinton and increases for Obama ranging from 2 to 7 percentage points. The Clinton decline is statistically significant for the Rasmussen survey, but just misses given the smaller sample sizes on the Marist and UNH surveys. Still, given the consistency of the pattern across the three surveys, we can have reasonable confidence that Clinton's support has fallen in New Hampshire since late September, something not evident in the recent national polls.
Now consider how the results look by education. Both Marist and UNH provided tabulations of vote preference by education that look remarkably similar. Virtually all of the decline for Clinton and the gain for Obama occurred among college educated Democrats. The race is virtually unchanged among non-college Democrats (a category that includes those with only some college education but no degree).
This pattern is a big clue that the change has occurred among those voters that pay the most attention to campaign news. Well educated voters are far more likely to pay attention to news about politics and typically possess more political knowledge.
The table does not show it, but earlier surveys by the University of New Hampshire showed that most of Clinton's gains between July and September also came from college educated voters (particularly college educated women). Ron Brownstein noticed that pattern and, using data provided by Gallup, reported the same thing among Democrats nationwide.
So why a bigger change in New Hampshire than nationwide in recent weeks? One answer is that at this stage, New Hampshire voters are paying more attention to the campaign than voters elsewhere. They have been inundated with millions of dollars of political advertising and, given the many in-person visits from candidates, are seeing more political coverage from their local news outlets. So they know an election is near and are following it more closely.
Four years ago, the Pew Research Center conducted parallel surveys in early December among voters nationwide as well as in Iowa, New Hampshire and South Carolina. As the table below shows, Democrats in New Hampshire were more than twice as likely to say they following the campaign "very closely" (33%) than Democrats nationwide (15%).
So we have a pattern here that neatly connects some of the points I made in two recent posts ("Three Campaigns" and "A Hits B, B Hits A, C Wins?". Most of the political news of the last week has gone unnoticed by the vast majority of ordinary voters nationally who are not paying great attention to the campaign. But in New Hampshire (and probably in Iowa as well) the pool of attentive voters is greater and so recent news has more potential to change voter preferences.
Also, many of the campaigns are advertising heavily there. Barack Obama's campaign has started running television ads in New Hampshire in late September, and voters there now know a bit more about him than those elsewhere. Also, as Ben Smith reminds us, "both McCain and Romney have been attacking her in television ads in the state" which may be a factor as well.
The consistency of the two most recent New Hampshire polls makes me more confident that the changes we are seeing are real, at least for now, especially since they seem to be occurring among the most attentive voters that are most likely to react to news stemming from a debate.