Voters to get last word in battle of the polls

US pollsters are on the spot. Today's presidential election, which might be called ''the ultimate poll,'' will be the tantilizing moment of truth for more than half a dozen of the major poll takers across the United States.

The pollsters, using differing methods, all agree that President Reagan appears to be on his way to another victory. But that's where the agreement ends. Some look for ''a major blowout,'' with Mr. Reagan winning one of the biggest victories ever. Others look for him to win by a far smaller margin.

In other words, some of the polls are going to be very, very wrong. Among the election-eve estimates:

* Reagan will probably win a 24- or 25-point landslide of historic proportions (polls by NBC News and USA Today).

* Reagan should win a solid victory of 18 or 19 points, but not so big as Richard Nixon's 1972 trouncing of George McGovern (polls by Gallup, CBS/New York Times, and ABC/Washington Post).

* Reagan's lead is suddenly narrowing, but he should still win by 10 to 12 points (polls by Louis Harris and the Roper Organization).

All this has left some voters befuddled. But G. Donald Ferree Jr. of the Roper Organization reminds everyone that polls are ''a quasi-science and a quasi-art.'' Every poll has some built-in inaccuracies, and those flaws appear to be showing up clearly this year.

James Shriver, an analyst with Gallup, says this year's wide-ranging polls point to the need for a scholarly investigation of the whole subject after election day.

Mr. Ferree emphasizes that there are always bound to be some differences in the polls.

The same organization, using the same methods, could run two polls on the same day and still get slightly different results, he says.

Mr. Shriver says the wide spread in the polls probably results from the difference in methods used by various organizations.

There are always variations, such as: How is the question worded? When is the poll taken? How many people are contacted? Does the poll include only registered voters, or only those who say they will definitely vote, or the entire adult population? How is the result ''weighted'' - that is, how is a balance struck between the number of women, men, blacks, whites, rich, poor, etc.? This final factor can be the trickiest of all.

Each poll has its own methods. Gallup, in its final poll released over the weekend, showed the race at 59 percent for Reagan, 41 percent for Democratic challenger Walter F. Mondale. It reported interviewing 1,985 people. In fact, Gallup interviewed about 3,500 people. But its statistical experts then used 10 special questions that were asked of each person to screen out more than 1,500 people as probable nonvoters.

Further, every poll concedes it can have at least a ''margin'' of error. That could account for some of the differences.

For example, Shriver notes that Gallup estimated Reagan's lead at 18 points. Harris showed it at 12. Each poll has a 3 percent error margin, either plus or minus. If one adds 3 points to Harris, takes 3 points from Gallup, each then shows 15 percent.

So the results, says Shriver, do appear in some cases to be within the margin of error.

The pollster who may be looking forward to today's results with the greatest excitement and anticipation is Gordon S. Black, whose surveys appear in USA Today. Mr. Black, along with the NBC poll, has consistently showed Reagan with the largest lead. His final poll was Reagan 60, Mondale 35, undecided 5.

Black says his polls indicate that this election is a political turning point for the US.

''I think it's the most important election since 1936. It may be a historic landslide that could produce a realignment in the electorate. It's been headed that way for months,'' he says.

Black takes issue with those who say that this is a ''negative'' election - one in which voters are motivated by being against someone, rather than for someone. He says this is one place that the Harris poll, which has consistently given Reagan some of the lowest numbers, has missed the point.

''This is as affirmative an election as we have had in years,'' he says. Black adds that both his poll and that of Gallup have shown that about 80 percent of Reagan's voters are very positive in their attitude toward the President, rather than simply voting against Mr. Mondale.

In one sense, the Republicans missed a great opportunity this year, Black says. If the GOP had known how strongly Reagan would run, and if they had known how many people would like to vote Republican in congressional races, the GOP might ''have won control of the House.''

Instead, the GOP left many Democratic incumbents unchallenged. Polls-eye view of Campaign '84. How surveys have tracked the presidential race since Labor Day. (Not including 'undecided' or 'will not vote' responses)

Reagan- Mondale- Margin Bush Ferraro ABC News/ Washington Post Sept. 7-11 56 40 16 points Sept. 22-Oct. 2 55 37 18 points Oct. 8-9 56 41 15 points Oct. 12-16 54 42 12 points Oct. 19-20 52 42 10 points Oct. 22-23 54 42 12 points Oct. 29-Nov. 1 57 39 18 points

CBS News/ New York Times Sept. 12-16 57 35 22 points Sept. 30-Oct. 4 59 33 26 points Oct. 9 58 38 20 points Oct. 14-17 54 41 13 points Oct. 21 56 40 16 points Oct. 23-25 56 37 19 points Oct. 31-Nov. 2 58 37 21 points Gallup Sept. 7-9 55 40 15 points Sept. 21-23 57 39 18 points Sept. 28-30 56 39 17 points Oct. 15-17 58 38 20 points Oct. 26-28 57 40 17 points Nov. 2-3 59 41 18 points Harris Sept. 5-9 55 42 13 points Sept. 21-25 55 42 13 points Oct. 8-9 54 42 10 points Oct. 12-14 53 44 9 points Oct. 22-23 56 42 14 points Oct. 26-28 58 41 17 points Nov. 1-3 55 43 12 points USA Today Sept. 4-11 57 35 22 points Sept. 28-30 60 36 24 points Oct. 14-16 61 36 25 points Oct. 24-26 59 36 23 points Nov. 1-4 60 35 25 points SHIRLEY HORN - STAFF