By CAITLIN LUKACS AND JODI WESTRICK
Observer staff
Feb. 13, 2008

In the lead up to the New Hampshire primary, there was much speculation and prediction about whom the voters would choose as their candidates. Of course, this speculation was backed up by data from several polls given to perspective voters.

RealClearPolitics.com compiled polling results from numerous organizations in the days before the nation’s first primary. For the Democrats, prior to the Iowa caucus, and going as far back as December 2006, the majority of polls had Sen. Hillary Clinton in the lead, often by a two-digit margin. However, in the days following Sen. Barack Obama’s victory in the Iowa caucus, all but one poll showed the voters favoring him instead of Clinton. In fact, from Jan. 5 onward, Obama had at least a five-point lead over Clinton in every poll listed. A poll conducted by Reuters, C-SPAN and Zogby even had Obama with a 13-point lead over the other candidates. From Jan. 5 to Jan. 7, RealClearPolitics averaged all of the polls and found that Obama was favored by an 8.3 spread.

RealClearPolitics.com has done extensive polling regarding the 2008 presidential election.

When the Washington Post’s longtime political reporter David Broder addressed the staff of the Concord Monitor as part of a longstanding New Hampshire primary tradition, he suggested that the Democratic race might be over with an Obama victory. Broder went so far to say that Obama was doing the Republican Party a favor by pulling in the majority of the independent voters, making the Republican candidates work solely for Republican votes. The meeting was not the first place that Broder mentioned Obama’s ability to take hold of the Democratic race. He mentioned the possibility earlier in a Jan. 6 column. In it he wrote, “It may seem paradoxical, but New Hampshire is poised to close down the race for the Democratic presidential nomination and launch a wide-open Republican contest.” He concluded his piece by writing, “Any way you view it, the race is now Obama’s to lose.”

On the Republican side, prior to the Iowa caucus, Mitt Romney was the leader in most organizations’ polls until the end of December 2007, at which point Sen. John McCain took over the top spot in the majority of polls conducted. Despite the fact that Mike Huckabee won the Iowa caucus, he never appeared as a front-runner in any of the polls listed by RealClearPolitics.com. From Jan. 5 onward, only one poll showed Romney as the favored candidate with a four-point lead over his opponents. RealClearPolitics.com found that McCain had an average lead of 3.6 points in all polls conducted between Jan. 5 and Jan. 7.

Many people were surprised by the results of the New Hampshire primaries on the Democratic side. Clinton ended up winning the state with a lead of 2.6 points. This contradicted the results of all the polls conducted and stumped even the brightest pollsters. Conversely, the Republican primary did not cause much confusion as McCain won New Hampshire by a lead of 5.5 points.

In its coverage after the primary, CNN reported that exit polls showed Clinton won New Hampshire by a surge in votes from women and older voters. Tom Curry, a national affairs writer for MSNBC, wrote in a Jan. 9 article that Obama failed to attract voters in working class towns like Berlin, N.H., and Claremont, N.H. In Berlin, Clinton beat Obama “more than two to one,” and Obama came in third behind former Sen. John Edwards, according to Curry. Kathy Frankovic, CBS news director of surveys, noted in a Jan. 14 article that many white voters may have “over-reported their intention to vote for an African-American candidate” in pre-primary polls. With a state that boasts a high majority of white citizens, this could have been a downfall for Obama.

However, many believed the problem may not have been with the voters who turned out and the way in which they voted, but rather in the media’s coverage of the primary and a more centered focus on specific candidates of interest. This belief was not limited to just politicians or citizens – the media lashed out at its comrades, suggesting that a greater attempt at objectivity should be made in future primaries and elections.

In the days following the nation’s first primary there were numerous theories put forth to explain what went wrong in the Democratic pre-primary polls. For the most part, the polls accurately predicted the percentage of votes that would go to Obama. However, the problem, as stated by pollster.com’s Mark Blumenthal, was that the Clinton vote was “massively underestimated.” Blumenthal wrote that the trend estimate for Clinton was only 30.4 percent of the votes, while, in actuality, she received 39.1 percent.

Some of the more popular explanations are that there was a last minute swing in Democratic voters from Obama to Clinton that was not caught by the polls, that there was a non-response bias in the polling, and that there were large numbers of independents polled who ended up voting for a Republican candidate.

The idea of the last minute swing was deemed to be a plausible one by both Mark Blumenthal and the Director of Surveys for CBS News, Kathy Frankovic. Blumenthal cited the exit polls stating that 17 percent of Democratic voters didn’t decide on a candidate until the day of the primary, and 21 percent of voters made up their minds in the three days prior to Jan. 8. This could have accounted for the late shift in polling numbers. Frankovic described the electorate as “fluid” and said that voters “were likely to change their decisions right up to the last minute.”

This could be the reason that pre-primary polls showed Obama ahead by a comfortable margin, but then the outcome of the election ended up as the complete opposite.

Professors Robert S. Erikson of Columbia University and Christopher Wlezjen of Temple University wrote a polling analysis for pollster.com in which they too suggest that pollsters were wrong because of last-second voter excitement. They said that people who were originally deemed as unlikely to vote by pollsters could have been inspired at the last minute to get out and support Clinton.

There are, however, polling analysts who disagree with this theory. Andrew Kohut, President of the Pew Center, is one such analyst. He said that the gap between the two candidates was too small to be caused by such a swing. Additionally, Kohut wrote, this does explain the magnitude of Obama’s lead in pre-election polls.

Another theory, that there was a non-response bias, is based on the idea that Clinton supporters were less willing to be interviewed. Kohut took this idea even further by adding race and socioeconomic class into the equation. He wrote that white people from higher socioeconomic backgrounds who are better educated are more likely to respond to surveys than whites who are from lower socioeconomic backgrounds and are less educated.

A third explanation for the pre-polling not matching the actual outcome of the primary is that independents were over-polled and they ended up voting for a Republican candidate. While this is a popular theory, many polling analysts do not believe that it is correct. Blumenthal cited exit polling data that showed that 60 percent of independents voted Democratic while 40 percent voted Republican. This matches the pre-polling data presented by CNN, WMUR and UNH.

Additionally, many major media figures spoke out against the difference in the pre-primary polling and exit-polling numbers, saying that the media may have gotten ahead of itself. On Jan. 14, Katie Couric reported in her “Notebook”: “For a long time pollsters have warned that surveys are an inexact science and media critics have warned that those of us without crystal balls are smart to avoid predicting the future. But until last week, those warnings were largely ignored, as was the implicit understanding that journalists are not supposed to play favorites. Hopefully this renewed objectivity is a change that’s here to stay.”

NBC apologized on-air for its inaccurate polling numbers, suggesting that the media’s attention may not have been for the greater good, allowing for some biased reporting to occur.

CBS News’ Bob Schieffer addressed concerns that the news media was to blame for incorrect polling numbers in New Hampshire.

However, CBS News’ Bob Schieffer conducted a somewhat humorous interview with himself to figure out what went wrong. Schieffer made a standard media defense. “Well, first I would like to put that in context,” he said in his interview. “I think the more important question is ‘Did we get it all wrong?’ and the answer is ‘No.’ Yes, we got it wrong on Senator Clinton, but we got it right on Senator McCain’s victory and we clearly said Dennis Kucinich was going nowhere even though his wife got that new tongue ring. So, we’re batting more than .500 here, and in baseball that’s great.”

Largely, the results in New Hampshire show that voter turnout is unpredictable. While the youth vote has been portrayed as a “decider” in many elections, it may not have worked as well in this primary as the other votes – such as those of women or the working class – might have. Polls can indicate certain things, but as Couric said, it is impossible to predict the future without a crystal ball. The question many might be asking in the future may have to do more with the validity of polls rather than which voters turn out.