FOR IMMEDIATE RELEASE: Wednesday
January 9, 2008
The New Hampshire Pre-election Polls
Contact: Dr. Lee M. Miringoff, Dr. Barbara L. Carvalho Marist College 845.575.5050
So, what exactly happened last night in New Hampshire? Did Hillary Clinton have a stunning comeback in the closing hours of the campaign or were the pollsters and pundits alike just dead wrong all along? Well, for answers, we at the Marist Poll took a look at the numbers.
For starters, the contest on the Republican side played out true to expectations. Bolstered by independent voters, Senator John McCain beat former Massachusetts Governor Mitt Romney. McCain has always been a good fit with New Hampshire Republican primary voters with his “Straight Talk Express.” He won the state eight years ago against George Bush only to falter in South Carolina.
|
|
Election Results 96% Reporting |
Average of Last 5 Polls Reported on NH |
Final Marist Poll |
|
McCain |
37 |
32.6 |
35 |
|
Romney |
32 |
28.2 |
31 |
|
Huckabee |
11 |
11.2 |
13 |
|
Giuliani |
9 |
9.8 |
5 |
|
Paul |
8 |
7.6 |
8 |
|
Thompson |
1 |
2.0 |
4 |
Poll average
compiled by pollster.com
|
Pollster |
Dates |
N/Pop |
Giuliani |
Huckabee |
McCain |
Paul |
Romney |
Thompson |
|
1/6-7/08 |
500 LV |
11 |
13 |
26 |
5 |
30 |
2 |
|
|
1/6-7/08 |
600 LV |
13 |
14 |
31 |
9 |
24 |
1 |
|
|
1/5-7/08 |
859 LV |
9 |
10 |
36 |
9 |
27 |
2 |
|
|
1/5-7/08 |
1549 LV |
8 |
10 |
32 |
8 |
31 |
3 |
|
|
1/5-6/08 |
492 LV |
10 |
13 |
31 |
10 |
26 |
1 |
|
|
1/5-6/08 |
477 LV |
5 |
13 |
35 |
8 |
31 |
4 |
|
|
1/4-6/08 |
500 LV |
9 |
11 |
34 |
5 |
27 |
2 |
|
|
1/4-6/08 |
600 LV |
8 |
13 |
35 |
7 |
27 |
5 |
|
|
1/4-6/08 |
776 LV |
8 |
13 |
34 |
8 |
30 |
3 |
|
|
1/4-6/08 |
409 LV |
8 |
9 |
38 |
7 |
29 |
2 |
Poll summary
compiled by pollster.com
But, what happened on the Democratic side of the ledger? Editorially, the polls missed the mark. In other words, they touted poll watchers in the wrong direction suggesting Senator Barack Obama was positioned to win.
Interestingly, with few exceptions, most polls were accurate in measuring Obama’s level of support in the state. The Marist Poll had him receiving 36% of the vote and eight of the polls conducted over the last days of the campaign had him within the statistical range of his final vote. What they underestimated was the support Senator Hillary Clinton would garner from New Hampshire Democratic primary voters.
All of the Democratic candidates were well liked in New Hampshire [Gallup Poll]. During the final days of the campaign, Democratic primary voters were deciding whom to vote for among candidates they liked with nearly four out of ten making up their minds in the last three days according to the exit polls. They weren’t voting against any of them.
As pollsters, we can re-evaluate our models and recalculate the numbers. But, what the weekend polls found was an Obama lead as primary day approached. What they do not reflect is what was apparent here in New Hampshire. The context of the campaign was changing. The last hours of the campaign were a media feeding frenzy over Clinton’s show of emotion when responding to a voter’s question on Monday morning. Video of her “emotional” moment was everywhere. It was played over and over with unrelenting commentary. Hillary Clinton was again the victim.
New Hampshire has a tradition of voting for women. Democratic Primary voters also like the Clintons. If the pollsters and media pundits erred, it was not in their weekend numbers but in not polling Monday and missing the impact of the unrelenting media coverage that characterized the Clintons as finished.
Unfortunately, few pollsters polled here in New Hampshire on election eve and no one released Monday only poll results. How New Hampshire voters were evaluating the race and the factors they were weighing in the last hours of the campaign were never measured. Even the exit polls, designed in advance, would not capture the final mood of the voters about the campaign. As a result, explanations about what happened here are fueled more by conjecture than by the numbers.
One such example is the suggestion of the Bradley Effect. The idea is that in bi-racial contests support for the African American candidate is overstated. The view suggests that white voters lie to pollsters about their choice because it is politically correct to support the African American candidate but in the privacy of the voting booth cast their ballot for the white candidate. Many years ago in a New York City mayoralty race David Dinkins, the Democrat and an African American, was poised to trounce Republican Rudy Giuliani. On Sunday, most polls including the Marist Poll showed Dinkins with a double-digit lead. Dinkins won, but narrowly.
Pollsters and pundits struggled to come up with reasons including the Bradley Effect. But that was not an accurate explanation of what happened. On the Monday before the election a story of corruption broke out about Dinkins. It had not been the first. The Marist Poll surveyed voters through election eve. The results found in the midst of such intense media pounding about the corruption scandal many voters had, in fact, become disillusioned with Dinkins. His lead had collapsed in our Marist Poll from 13 percentage points on Sunday to barely 2 percentage points by Monday night. This matched the eventual election result.
Here in New Hampshire, for the most part, Obama’s support was not overstated in the polls. But the polls missed the magnitude of the support for Clinton. Although the pollsters in New Hampshire cannot have Monday night back, we plan to re-contact in the next few days the voters we spoke with over the weekend to glean whatever additional insights we can.
|
|
Election Results 96% Reporting |
Average of Last 5 Polls Reported on NH |
Final Marist Poll |
|
Clinton |
39 |
31.0 |
28 |
|
Obama |
37 |
38.4 |
36 |
|
Edwards |
17 |
18.2 |
22 |
|
Richardson |
5 |
5.4 |
7 |
|
Kucinich |
1 |
---- |
3 |
|
Biden |
0 |
2.6 |
---- |
Poll
average compiled by pollster.com
|
Pollster |
Dates |
N/Pop |
Clinton |
Edwards |
Obama |
Richardson |
|
1/6-7/08 |
500 LV |
34 |
15 |
39 |
4 |
|
|
1/6-7/08 |
600 LV |
31 |
20 |
40 |
4 |
|
|
1/5-7/08 |
862 LV |