The way politics works these days, the first thing they did was to take a poll, which showed that the people of Arkansas were looking for some good old basic values in their candidates for office.

That's where Ray Strother applied what he called his "cocoon" theory. He would take the results of Bill Hamilton's poll and spin a warm, protective cocoon around Jim Guy Tucker, a bright young Democrat who wanted to be governor last spring.

Their poll gave some promising guidance. Arkansans, for example, weren't too keen on the longish hair of Bill Clinton, the former governor who also was running in the primary. And they weren't keen on his wife's professional use of her maiden name. They longed for tradition.

Strother, a political film-maker based in Washington, took it from there. He produced a television campaign film limning an appealingly basic Tucker: Short hair, homespun, tough on crime, straight talker. His wife, emphatically identified as "Mrs. Tucker," editorialized on Jim Guy's great way with the kids at home.

Line drive into the bleachers, right? Wrong. Tucker ran third in a three-man race, getting only 23 percent of the vote. Clinton ended up with the nomination. So much for the polls; at least so much for the polls at that time and place.

Well, you win some and lose some. William R. Hamilton, a pioneer pollster even though he's been at it only 19 years, and Raymond D. Strother, the film-maker, have gone on to other candidates and other campaigns. They'll likely be more remembered for the wins than for the losses.

Serious candidates in the 1980s don't even consider running without the expensive help and advice of campaign consultants, film-makers and pollsters. Somehow, and their own self-promotion plays a part here, they have become bigger than life. Ergo: Consultants are surgeons, film-makers are image and idea builders, pollsters are oracles.

Polling in American politics goes back at least to the early 1800s. But in its modern incarnation it is not much more than a quarter-century old and, while the techniques are becoming increasingly sophisticated, there is about it an aura of pseudo-science and large dollops of mumbo jumbo that pass for wisdom.

The political opinion sampling is done by several hundred organizations around the country but only a relative handful dominate the industry. The best-known can be counted on two hands. They tend to work only for one party. They tend to be selective about taking on the "right" kind of candidate. Increasingly, they seem to have insinuated themselves into the candidates' influential strategy-and-tactics circles.

Their names include Bill Hamilton, Peter D. Hart and Patrick Caddell, all located here, who work for Democrats. Republican stars are Richard B. Wirthlin of California, pollster for Ronald Reagan; Robert Teeter of Detroit and V. Lance Tarrance, based in Houston. All have worked for presidents and presidential candidates, senators, members of Congress, governors.

New Yorkers Richard Dresner and Arthur Finkelstein, although operating nationally, are a special case. Neither is much admired by the brethren; Dresner because he works for candidates of all stripes and Finkelstein because he works mainly for the far right and is a director of the National Conservative Political Action Committee, which the establishment pollsters thoroughly dislike. Finkelstein, as they say, can laugh all the way to the bank about this: NCPAC has paid him more than $260,000 for his polls since 1975.

Dresner was a senior vice president of the Harris Survey before he helped set up the firm of Dresner, Morris & Tortorello Research. He is aware of his critics, but they don't much bother him.

"It is a 'closed' industry," he said. "and it is hard to get clients. . . . It's a very difficult business to crack into. Once you gain respect, you keep it. . . . It is a matter of whether we are respected by candidates and whether we feel they are honest, decent people."

All of these pollsters take themselves very seriously and so does American society, lay and political, ever intrigued at knowing a final score before a game is played. The public and press listen to their assumed insights with varying degrees of care, reverence and awe.

University of Virginia political scientist Larry J. Sabato sharply criticized pollsters in his book last year on the consulting business. "They are the vox populi," he wrote, but the voice seems to come from Mount Olympus."

Sabato added: "The average political poll is almost certain to be flawed in at least a couple of respects, and the sooner this is accepted and understood by candidates, press and public, the healthier and more realistic will be the perceptions of the polling consultant's role in the election campaign and beyond."

Peter Hart, who began 18 years ago as a lowly coder for Louis Harris, one of the sires of the modern polling industry, doesn't agree.

"A poll is not a substitute for brains or insight," he said. "For all the sophisticated polling, Dick Wirthlin working for Reagan or Pat Caddell for Jimmy Carter, it didn't lower the prime rate by 1 percent or get our hostages out of Iran. . . . So it is not a substitute. As with any invention in our society, it is a more efficient way of planning campaigns and making the most of one's resources. It still goes back to a candidate's ability to make judgments, deal with issues."

Another of the heavy hitters, Robert Teeter, who did polling for Reagan and George Bush in 1980, Gerald R. Ford in 1976 and Richard M. Nixon in 1972, also rankles somewhat at criticism from the Sabatos of the world.

"We've really gotten away from notion of mystique in polling after the Kennedy election in 1960," Teeter said. "There was a perception it was done with incense over a computer. We are providing a fundamental kind of information and politicians have followed the business world to some degree on this score, in that they operate better with more information."

Teeter continued, "We've learned to use the sample and to interview better by telephone. The sophistication of computers and computer software has allowed analytical techniques to become more sophisticated. . . . So I think it has helped the process. The more information that is fed into a process, the better it is.

"But the whole question is, has it become manipulative? Are the candidates tools? Automatons? That doesn't happen. It isn't our role," he said. "By the time someone is campaigning at the gubernatorial or Senate or presidential level, their motivation has to be higher. Ninety-nine percent of the candidates have become good users of information. So the net is, it has helped the process."

Yet there are other puzzling things about this, more intangible but there nonetheless. For one thing, nobody really knows how much a political poll becomes a self-fulfilling prophecy. For another, polling has somehow taken a part of the inherent suspense and serendipity out of political process.

"We find it's not fun to watch politics anymore," Hamilton said. "Pollsters know what is happening. . . . It takes a little of the intensity from October, when people are busting their butts for a candidate. That's lamentable."

Hamilton, who works in rolled-up shirt sleeves, takes his business seriously and doesn't fret too much about the critics who see the pollsters as a pernicious influence on the political process.

"A poll is a communications link," he said. "It is a feedback system of information from voters. It can be done in a positive way, or in a nefarious, devious way. A poll is a like a little town hall meeting, it's the only way the candidate can get a feeling of what people want."

The idea is that after the polls sample that public feeling, the candidate answers the voters through television and radio ads and direct mailings, all of which, of course, have helped lift the cost of campaigning to historic highs.

"Our industry is doing nothing more than mirroring the impact of technology on our society," Hamilton said. "Television is more important than it used to be. It's the only way we can communicate. . . . I'll bet a greater proportion of Americans saw George Washington, but Americans today feel they know Ronald Reagan. Through television, they got a feeling for the man.

"I lament the whole move of society toward instant information. . . . The tube has built instant gratification. Ours is a new industry in the last 20 years. . . . But I don't see us as the 'bosses' we have been proclaimed. . . . If talents and technology are equal, then the only difference is the candidate."

Tarrance, who set up his own business in 1977 after working for years in Republican campaign circles, worries about the rap that he and other pollsters are no more than money-grubbing, cynical con men.

"My own brother told me he didn't know how I could stand to be in this business, that I was manipulating voters. I sat him down and showed him and explained it and he is less cynical now."

Tarrance said: "The Kennedy-Nixon debates of 1960 were a hallmark for strategy and preparation. That was the real revolution that started people saying we had better use TV. This came about in the 1960s at the same time that major advances were being made in the behavioral sciences and computers. By 1970, the candidate who was going to win was out hiring the best consultant he could two years before the election. Certain candidates are taken seriously today only by which consultants they hire."

The advances of technology, which give a pollster the ability to more quickly get the drift of public opinion at any given time, cannot be underestimated as a campaign influence. Washington media consultant Robert D. Squier relied heavily on polls to help him guide last year's Virginia gubernatorial campaign of Democrat Charles S. Robb.

Early polling indicated that Robb would be in trouble on key Republican issues. Squier and others confected a strategy to respond to that as well as neutralize Republican Marshall Coleman by putting him on the defensive.

The first Robb television ad was what Squier termed "the drug pusher spot." The fight was on after Coleman was pushed into denying he was soft on drugs as attorney general.

What does it all mean? "We are relatively insignificant," said Matthew (Matt) Reese, a dean of the political consulting business. "I did all I knew how to do for Birch Bayh in 1980 in his losing Indiana race for reelection to the U.S. Senate and I did a good job. But there was a tide and we couldn't stem that tide," he said.

You pays your money and you takes your choice. Babe Ruth also played in Yankee Stadium and he, too, occasionally was insignificant.