Two weeks ago, freshman Tom Jawetz spoke to a woman over the telephone who did not know the difference between liberals and conservatives.
The woman, Jawetz said, had to ask her husband for the answer. In the background, Jawetz heard him reply, "Conservatives are what we are ... liberals are those flower children."
For the past five months, Dartmouth students have been polling New Hampshire voters, gathering information for the nationally recognized WMUR-Dartmouth polls.
As voters head to the New Hampshire Republican primary booths tomorrow, the media and political pundits will be scrutinizing the latest poll results, to best predict the outcome of the election.
The fifth and final WMUR-Dartmouth poll, conducted by the Rockefeller Center for the Social Sciences, released Friday, shows Pat Buchanan neck and neck with Kansas Senator Bob Dole.
Each poll required the Dartmouth students working on the poll to telephone about 2,500 New Hampshire residents for four nights in a row, said consultant Tami Buhr, whom Director of the Rockefeller Center Linda Fowler hired to assist her with the polling efforts.
National attention
Organized by Fowler, Dartmouth's pre-primary polls have received national attention in the weeks leading up to the election.
The New York Times, the Cable News Network, the Washington Post, the Los Angeles Times and a multitude of other news organizations have featured results from the WMUR-Dartmouth College poll.
The poll caught the nation's eye when the results of the first poll were submitted to the Political Hotline -- an on-line daily compilation of political news to which most national news organizations subscribe -- explained Carl Cameron, a political reporter for WMUR.
"I hope it makes us look good," Fowler said. "I think it's a good thing to have the Dartmouth name out there."
Fowler said so many media sources have used the poll because "we have more in-depth information."
Keating Holland, CNN's polling editor, said, "We consider the [WMUR-Dartmouth poll] as accurate as any other major polls in New Hampshire right now."
Holland said among members of the national media, Dartmouth's poll was more popular than the University of New Hampshire's poll.
Morton Kondrake, executive editor of Roll Call, a biweekly capitol hill publication said the Dartmouth poll has a good reputation.
Cameron said the WMUR-Dartmouth poll was considered to be more accurate than other New Hampshire polls.
"There was a belief from the outset that the Dartmouth poll was widely considered to be the most representative of the New Hampshire electorate," he said.
Cameron said the Dartmouth-WMUR poll's ability to detect Forbes' rise and fall nearly a month before other polls as a testament to the poll's accuracy.
Those at the College also seem pleased with the polling efforts.
"It's good for Dartmouth to engage, to be involved in" polling, Government Professor James Shoch said.
Jim Brennan '96, political programs intern for the Rockefeller Center also works as an intern for FOX news in Manchester, where, he said, "people mention [the WMUR-Dartmouth poll] all the time."
Nuts and bolts
Fowler said she decided to establish a poll at Dartmouth "for WMUR to monitor change in public opinion" about Republican primary candidates.
Fowler said her secondary reason for sponsoring the poll "was a scholarly one ... to generate data on how voters' attitudes change over time to campaign stimuli."
She said WMUR, a Manchester based television station, approached the College and asked if it would be interested in helping to gather poll data.
"This is one of those things they needed polling done for, and I was interested in compiling a data set that scholars could use to study the effects of media on voters," Fowler said.
Cameron said while WMUR had previously coordinated its polling with the University of New Hampshire, it decided to work with Dartmouth this time because the of the College's international reputation.
"Anybody who knows anything about academia, knows about Dartmouth," he said,. "and anybody who knows anything about politics knows about WMUR."
Buhr, an alumnus of the University of Iowa and now a graduate student in Harvard University's government department, plans to use data gathered from the polls for her dissertation about the ways the news media affect political campaigns.
Buhr, Fowler said, "establishes the procedures for drawing the samples, directs the students who do the telephoning, analyzes data and writes up analyses."
Fowler said before each poll, she and Buhr sat down together and drafted several questions.
Buhr said the poll included several stock "horse race" questions, such as "Who would you vote for if the election was today?" plus several open-ended questions designed to gain more insight into voters' sentiments and to supplement data for Buhr's dissertation.
Each poll also contained questions particular to current political events at the time of the poll's release. For example, Buhr said in October student pollsters asked New Hampshire residents what they thought about the possible candidacy of General Colin Powell.
Each night of each of the five four-day polls, Buhr said she printed the questions and explained them to student callers, who showed up for work around 4:30 p.m.
Buhr said she gave new surveyors a 45-minute training lesson before setting them to work. On the phones until 9 p.m., students earned $8 an hour plus all the pizza and soda they wanted, she said. She employed about 70 student callers during the five polls.
Buhr said once a student caller contacted someone, they first asked the respondent about any political lean.
After a poll participant told the student caller about his or her political alignment, either Democrat, Independent, or Republican, the student caller would begin asking questions from one of two lists of questions, one for Democrats, and another for Independents and Republicans.
Buhr said she was primarily interested in the views of Republicans and Independents, since they would be the most likely to vote in the Republican primary. These potential voters answered a more lengthy questionnaire.
Students usually spent about 15 minutes speaking with a person, although some phone conversations have lasted up to an hour, she said.
Buhr said students called about 2,500 random numbers, which she purchased from Survey Sampling Inc., based in Fairfield, Connecticut. Buhr said each of Dartmouth's polls garnered about 700 useful interviews.
Of those interviewed, about 40 to 45 percent of respondents considered themselves to be Independents. Approximately 30 to 35 percent said they were Republicans and 20 to 25 percent said they considered themselves Democrats.
Crunching numbers
After students conducted the telephone interviews, other students transferred the more quantitative and "yes/no" responses onto sheets of scanning paper, which were scanned by computers to generate raw numerical data, Buhr said.
Buhr took the rest from there.
Armed with statistical numbers and anecdotal information, she spent a couple of days following each polling interpreting all of the information.
Buhr said she employed carefully-planned equations to weigh respondents' answers to most accurately reflect how New Hampshire residents would vote in the Republican primary.
"Those who have a higher probability of voting, their responses weigh more," Buhr said. For example, Buhr would count more highly the opinion of a voter who said he or she was a Republican voter, because that person was more likely to actually vote in the Republican primary.
Pizza and Partisanship
Sitting around an oval table in the Rockefeller Center, several student callers took a break, swallowed down bites of hot pizza and talked about their experiences on the phone lines.
Jim Brennan '96 served as a liaison between Buhr and the student callers. Brennan said he found the whole experience interesting. "It's neat to hear side-bar comments that don't show up on in the statistics," he explained.
Cracking a smile, Brennan said one surveyed resident said he did not like the staunchly conservative Republican candidate Pat Buchanan because Buchanan was "too liberal" for the poll participant.
Some student callers spoke to respondents on the other side of the political playing field.
Jennifer Potash, a junior visiting from Smith College this term, said she enjoyed talking to a woman born in 1920 who was a "diehard Democrat."
"She kept politics a life-long interest," Potash said. "She had voted in every election since she turned 18."
Potash said she was a government major and she found the polling to be a great chance to learn more about the political process.
Poller Josh Brann '99 had a different experience with another older respondent.
"The very first person I called was a lady who was 90 or more years old," he said. "She talked to me for 45 minutes about everything you could possibly imagine."
Brann said the old woman said "all kinds of crazy things" and he thought "I hope they're all not like that."
Brann, who said he enjoyed the calling, said students often found amusing messages on answering machines.
Skye Gurney '99 said she has learned a lot through her polling experiences. Buhr said she depended a lot on Gurney, who she hired as an assistant. Gurney said she helped Buhr by answering pollsters' questions and doing clerical work.
"I really like it," Gurney said. "This poll has probably motivated me more to keep in touch with current events."
A drama major, Gurney said she was peripherally interested in politics but did not see herself getting involved with government or the media.
She said her drama classes helped her as a poller, because, "You have to do a lot of acting while you're on the phone," she said. "You have to cater to people."
Gurney said one of the biggest challenges of polling people was to convince them to answer the questions.
"A lot of people are really opposed to taking a poll," she said. Those respondents not wanting to talk attribute their reluctance to a lack of time, a desire for privacy or because they do not feel competent enough about politics to speak, she said.
'Something close to gospel'
Sometimes the importance placed on poll results by the media, results in a skewed interpretation of the data.
Fowler, Holland and Kondrake each lamented that so much attention is paid to polls during the campaigns, so much so that the polls often overshadow other aspects of the campaign.
Polls are very good "snapshots" of voter opinion during the course of elections, Shoch said. But, he added, "I think [polls] are accurate for the moment they're taken ... peoples' opinions of candidates change very volatilely."
Kondrake said polls are usually accurate. He said, "Everybody believes in them, the press believes in them, the voters may not like them, but they believe them too."
Kondrake said often the press accepts poll results as "something close to gospel, and they really shouldn't."
Polls are just good to "concretize hunches" about the status of candidates at the particular moment they are conducted, Kondrake said. Polls do not portray "the whole picture," he said.
Fowler said the media sometimes puts its own spin on poll results.
"It introduces a perspective on the way reporters cover the primary," Fowler said. "I think [polls have] too much of an effect because the press spends all its time reporting about poll results rather than reporting about the issues."
While the media seems to obsess over polls, Kondrake said polls play a much smaller role in voters' minds.
"Polls focus more attention on the front-runners who have a chance to win," Kondrake said.
He said candidates who do not fare well in polls will suffer because voters reading the polls will be less likely to vote for that candidate.
Kondrake used Indiana Republican Sen. Dick Lugar's struggling presidential campaign as an example of the importance the media places on poll results. When people read that Lugar unpopular in the polls, Kondrake said, they were less likely to support him and his approval ratings waned.
But most experts agreed that polls have little effect on what voters actually decide while inside voting booths.
Kondrake pointed out that before polls, people were still able to communicate with others to determine the popularity of candidates.
Voters have always had access to information about public opinion, so polls could not actually affect voting patterns much more than before the media began using them.
"There has always been, in effect, public opinion," he said. "Now it's more accurate."
Holland said, "Most people are not as stupid, not as sheep-like" as cynics may believe. Voters have a good track record."
Government Professor Dennis Sullivan added that "most people who are aware of the polls are committed enough to their own positions not to be affected" by poll results.
Buhr said the Dartmouth poll showed people found the most useful aspects of learning about candidates during a campaign were not polls.
The College poll found that voter's most useful sources of campaign information, in order of descending influence, were: televised interviews of candidates, general news coverage of events, conversations with other people and finally advertisements.
But for the candidates, polls are of utmost importance, Kondrake said.
"Polls are their intelligence source on how what they're doing is playing on the public," Kondrake said. He said candidates would often consider poll reports and "sometimes alter messages in order to appeal to the public."
"Their tactics and strategy are affected a lot by what the polls show," Kondrake said.
He said candidates run their own polls and depend on them.
For instance, presidential candidate Sen. Bob Dole R-Kan. learned through his campaigns polls that the flat tax issue, a major component of Forbes' platform, was not very popular with people and hence, "a vulnerability," Kondrake said.
He said Dole's campaign analyzed the flat tax and publicized how it would raise taxes on the middle class.
Polls also help campaigns to target their advertising.
"I heard that [Steve] Forbes figured out the best predictor of whether people would vote was if they watched the CNN show Inside Politics ... then he bought up all the time he could on CNN" to broadcast his advertisements," Kondrake said.
