Do public opinion polls have any useful role in determining American policy in Central America? Plenty of people say no, notably Irving Kristol in a recent Wall Street Journal article. Kristol says his first commandment for a president would be "thou shalt not permit a pollster on the premises." He believes that Reagan is not being Reagan on this issue, because of interpretations--or misinterpretations of the polls. But despite such bellows of outrage, polls will always be used--and misused-- in a country where leaders are chosen in elections.

They are not usually misused in the ways people think. Few politicians mechanically take a stand because a poll says it is popular; they have more convictions than that. Nor do many politicians change positions on issues in response to polls. Rather, they avoid issues on which their stands are not popular and stress those on which their stands are. And when polls are misused, there is often a penalty. Jimmy Carter's pollster, Pat Caddell, took satisfaction throughout 1980 in pointing out that each of the president's responses to the hostage crisis, even those that were contradictory, had public approval. But when Election Day came, and the president's policies failed to bring the result the public wanted, the voters threw him out--their approval ratings over the previous 12 months notwithstanding.

So consider the limitations of polls:

They are just pictures of public opinion at one point in time--and a fuzzy picture, because what they try to measure is subtle and complex, while the measuring device, the percentage response, is misleadingly simple and precise.

Polls mix responses to matters that people know and care little about and matters they care very deeply about.

4 On every difficult policy question, public opinion is inconsistent. Ordinary people are maddeningly unwilling to resolve those inconsistencies: that's what they hire public officials to do.

6 Polls cannot predict the future. You can ask people how they will respond to hypothetical events. But their replies are usually bad predictions.

So it makes no sense to look at poll results as hard data--like the latitude and longitude of El Salvador. Interpreting polls--like governing--is an art, not a science. Poll results must be considered in context, and with their limitations in mind.

So what do the polls on Central America tell us?

First, that President Reagan faces a public that is especially skeptical on this issue. The Washington Post-ABC News poll reports that more Americans say they believe the news media on Central America than the administration; overwhelmingly (67-19) they reject the contention that recent naval manuevers were routine. This skepticism reflects a general tendency to disbelieve leaders today. But it also reflects wariness of another Vietnam. By a 55-40 percent margin, Americans say Central America will not turn out like Vietnam. But 40 percent consistently respond negatively to the president: on the creation of the Kissinger commission, on U.S. military exercises, and on their likely effect on chances for peace. Fully 48 percent disapprove and only 33 percent approve of the Reagan Central American policies generally.

These are much more negative responses than Lyndon Johnson got in 1964 and 1965. Then the public approved almost every initiative he took by a wide margin. Only a few Americans in 1964 and 1965 expected a long war in Vietnam, and virtually no one expected a communist victory. They expected something much like Korea--then a more recent memory than the Gulf of Tonkin Resolution is now. Now they seem to expect-- and fear--a war much like Vietnam. President Reagan, unlike his predecessors, cannot count on a quick surge of public approval for any step of escalation. Even approval from a quick, limited military success is likely to dissipate quickly: remember the Mayaguez.

In the face of the vivid historical experience of Vietnam, the Great Communicator has not been able to change public opinion on Central America at all. To question after question the results in Post-ABC polls in March 1982, May 1983, and August 1983 are statistically identical. Some observers have made much of the fact that the public finds it hard to say which side we are on in Nicaragua or El Salvador. But those are quiz-show questions; the public understands quite well the danger that we may be called on to act militarily against a communist- allied government there. On that question, opinion is not changing, as it is on remote issues the public has thought and cared little about. Opinion, on all sides, seems rock-solid, as it is when it is based on deep-seated feelings and searing personal experiences.

All of this says that there will be a tremendous short-term political cost for an administration that escalates in Central America--regardless of the facts. But the poll results do not determine the verdict of history. The final response to Vietnam was far different from how the Americans of 1964 and 1965 predicted they would respond, and the final response to Central America may be different from current poll results. Remember that the same Americans who dread military escalation in El Salvador were in many cases strongly opposed to the Panama Canal Treaty and voted against those who supported it. What the polls seem to be telling the politicians is that they are likely to get into political trouble at home no matter what they do in Central America--unless they are lucky enough to be able to win a war without fighting it.

The polls do not tell Ronald Reagan to follow a particular policy in Central America. Rather, they tell him that it is not in his interest to make it a central focus of politics. The president has been a lucky politician, but if El Salvador dominates the news next year, his luck may run out.