Eight days before the New Hampshire primary, Ronald Reagan's poll showed George Bush beating him by a solid 9-point margin. Three days later, the same poll showed the race dead even. On voting day, Tuesday, Reagan scored by a 27-point landslide over Bush.

What happened in the final week of the New Hampshire GOP contest was a dramatic demonstration of the volatility of the American electorate this year and of the way a series of seemingly accidental incidents can trigger a trend that could well decide a nominee for president.

The view of Reagan's pollster, Richard Wirthlin, who opened his data to The Washington Post yesterday in an interview, is that the critical event in the Reagan surge was not the widely publicized confrontation with Bush in Nashua last Saturday night but the League of Women Voter's debate in Manchester three nights earlier. That was widely, but mistakenly, viewed as a rather boring, no-win affair.

Wirthlin's view is supported by the tracking data of Dick Bennett, who polled New Hampshire for Rep. John B. Anderson, and, at least indirectly, by some of the information collected by television networks in interviews with voters leaving the polls Tuesday.

The NBC-Associated Press "exit interviews," for example, showed that 30 percent of the Republican voters decided on their candidate preference in the last week. Of that number, 49 percent voted for Reagan, 21 percent for Sen. Howard H. Baker Jr., 12 percent for Anderson and only 14 percent for Bush.

Wirthlin's data indicate even greater volatility, with 48 percent of the Republicans in New Hampshire changing their minds about their candidate choice between Jan. 28 and Tuesday, and 33 percent switching in just the last week.

Bush was propelled into the favorite's role in New Hampshire by his victory over Reagan in the Jan. 21 Iowa caucuses. A week after Iowa, Bush had a 6-point lead on Reagan in New Hampshire, and that margin built to 9 points by Feb. 18, as Bush's campaigning and advertisements hammered home the message to New Hampshire Republicans that Iowa had given him the magic momentum" that could bring the party victory.

But, Wirthlin said, "The polling also showed the same kind of fuzziness about Bush's image that we picked up in the summer of 1976 about Jimmy Carter. A 'halo' had been placed on his head as a winner, also showed the same kind of fuzzi- [TEXT OMITTED FROM SOURCE] they really should be for him. Giving voters additional knowledge about that kind of candidate can strip away a lot of support very quickly. And that is what the Wednesday night debate did."

In a survey taken the day after that debate, Wirthlin found that 37 percent of the Republican voters had seen or heard about it, and 86 percent of them thought there had been a clear-cut winner. Reagan was named as the winner by 33 percent, Bush by 17 percent, Anderson by 14 percent and Baker by 12 percent.

Among those who had followed the debate, Reagan soared 15 points against Bush, and overall in the GOP electorate, he closed the gap to 2 points.

"I knew at that point we would win it," Wirthlin said.

There were several factors about the League debate that made it particularly beneficial to Reagan, Wirthlin said. The Republicans who watched tended to be disproportionately older and more conservative voters -- Reagan's core constituency.

The questions of gun control, abortion and the Equal Rights Amendment, where Bush and Reagan have contrasting, if not opposing, positions, allowed Reagan "to send the cues that helped solidify the base among those watching," Wirthlin said.

In handling a range of other domestic and foreign policy questions, Reagan appeared to most of those interviewed at least as competent as Bush, thus helping to remove "the halo" of unique leadership the Iowa victory had conferred and also to reduce doubts about whether Reagan's age made him too feeble to be president.

In answering a final question from the audience on the ethnic story he had told a few days earlier on his press bus, Reagan showed both anger and embarrassment and, as Wirthlin interpreted his data, "shared a greater range of his emotional makeup with Republicans than he had done since the dramatic moments at the 1976 convention."

By Thursday, Reagan had erased his deficit to Bush in a whole variety of personal-attribute questions, ranging from leadership and competence to caring about people.

While Reagan was using the Wednesday debate to achieve critical objectives of his comeback campaign, Bush was adhering strictly to his front-runner's goal of "keeping my cool" and "not making any mistakes." Reagan was, in effect, playing in an emotional and substantive vacuum.

That same contrast was made even more vividly in the Saturday night debate in Nashua, where Reagan escorted the four other Republican contenders onto the platform, urged that they be allowed to participate, and appeared as the champion of free speech, while Bush squirmed uncomfortably in silent embarrassment.

When Wirthlin ended his daily telephone surveys Sunday afternoon, barely 12 percent of the Republican voters knew what had happened in Nashua. But by that time, Reagan had surged to a 45-to-24 percent lead over Bush, and Wirthlin was able to advise his client that it was all over but the shouting.

The pollster said he is convinced that as film clips from the Saturday night debate were shown repeatedly on Sunday evening and Monday morning television news shows, the pictures of Reagan bellowing, "I paid for this microphone, Mr. Breen," and humorously allowing that "I agree with George Bush" that Reagan is not too old to be president, all contributed to the picture of a dynamic, commanding and appealing human candidate, contrasted with a stiff, formal and uncommunicative one.

Then, Wirthlin said, the Bush campaign handed Reagan two more favors in the final 48 hours. Recognizing the damage Bush was suffering from the Nashua debate issue, the Bush campaign rushed out a new radio ad, attempting to explain the incident and giving it additional publicity.

And Bush's campaign manager, ex-governor Hugh Gregg, made the same decision he had made four years before when he was running the Reagan campaign: he sent his candidate out of the state to free his organization workers for get-out-the-vote efforts.

"Hugh Gregg knows New Hampshire," Wirthlin said, "but he does not understand the dynamics of voter decisions. When large numbers of voters are making up their minds or changing their minds, the absence of the candidate sends a message that he does not care what they do. On Monday and Tuesday, New Hampshire Republicans saw Ronald Reagan campaigning in an overcoat in their state. They saw George Bush jogging in shirtsleeves in Houston."

The result of these seemingly accidental events, none of them apparently of great significance in itself, was a landslide victory for Reagan that is sure to give him a "halo" of added support in the coming contests in Massachusetts, South Carolina and Florida.