Page 1

POLLS AND THE POLITICAL PROCESS-PAST, PRESENT, AND FUTURE BY G E O R G E G A L L U P I n this address, delivered at the 1965 Conference of the American Association for Public Opinion Research, Dr. Gallup gives what many consider his most forceful reply to critics of the polls.


HIS is the thirtieth anniversary of modern public opinion polls. During the three decades since 1935 an effort has been made to gauge opinion on all political, social, and economic issues that have concerned the public. As a consequence, mountains of data are available for those who wish to examine the views of the American people on specific issues, or to judge the reliability of modern polling methods. Professor Samuel Stouffer of Harvard once said that modern sampling polls "represent the most useful instrument of democracy ever devised." Whether you share his measure of enthusiasm or not, the truth is that the concept of polling public opinion has captured the imagination of people the world over. Polling organizations are now operating in all the major democracies, and, at least in some forty nations, polling procedures are now regularly employed. Governments are themselves making greater and greater use of sampling methods for administrative purposes. Our own government carries on one of the most extensive polling operations in the world. While it does not refer to these efforts as poll taking, the process is the same, and had not polls on political issues blazed the way, their use for administrative purposes would likely have been delayed many years. Time has permitted polling methods to be tested in the cruelest of all possible ways-by election returns. Not once, but fifteen times in as many national elections, polls have come under the scrutiny of critics, many of them hopeful that forecasts would be on the wrong side. IMPROVEMENTS IN METHODOLOGY

Elections are by no means the perfect test of the accuracy of polls. (The efforts of machines to get out the vote still cannot be predetermined, to cite just one problem.) Nevertheless, you must agree



that the record is extraordinarily good. Our final poll results, published the day before each one of these fifteen national elections, have an average deviation from actual election returns of only 2.9 percentage points. And in the years since the fateful 1948 election-which shook all of us out of our early complacency-the average deviation has been only 1.9 percentage points. Looking ahead, it is hard to see how this record can be improved very much. I n the words of the musical comedy girl who sings about Kansas City, "We've gone about as fur as we can go." I do not mean to imply that we cannot improve our present procedures in many ways. Scarcely a year goes by that we do not make important changes. We are still an infant science and the developments of the next thirty years, I am sure, will overshadow those of the last thirty. We used to say that no great social good comes from telling the public a day in advance how an election is likely to turn out. While there is still truth in this, I am certain that had we not put our necks on the chopping block every two years since 1935, we would not have made the progress we have in developing polling methods. And not the least of the benefits that have accrued from election polling is the proof offered the public, and all those interested in the social sciences, that behavior can be predicted accurately on the basis of opinion, attitude, and past action data. And, may I ask, is there any other field of social science where human behavior can be, or has been, predicted with the same degree of accuracy?

We have come a long way in three short decades. At the same time, I must admit that while we have done some things well, we have done other things badly. For example, many social scientists, many people in public office, many political writers, are wholly uninformed about our methods and the function we try to serve. I n many instances, these persons are not only uninformed but actually misinformed. After thirty years, we can take little satisfaction from having to answer the same criticisms that were voiced during our early days. When these criticisms come from the uneducated it does not bother me. When they come from college teachers, from newspaper columnists, and from members of Congress, I am disturbed as well as annoyed. T h e generalization that irritates me most is the phrase: "The polls show." Since there are at least thirty kinds of polls, ranging all the way from feed-bag, bubble-gum, street-corner, post-card, barber-shop, movie-patron, on to those which claim to be accurate but have never been tested, it makes no sense to me to put all polls in the same bag



and to parrot the phrase: "The polls show." I t would make just as much sense to say "editors think," "social scientists believe," "political writers say." NO SECRETS IN OUR PROCEDURES

We have admittedly done a poor job in our public relations. Although members of our staff have published four books and countless magazine articles and brochures; although our election methods have been described in detail in more than one issue of the Public Opinion Quarterly, and although we have issued supplementary documents for those who wish the technical and statistical detail of how we operate in each presidential year, we still read that our methods are secret. This has never been true, even from the day we started. Could it be that it is easier for a writer or scholar to say that our methods are secret than to take the trouble to find out what they are? Certainly, our election record is no secret; it is contained in the thirty-year files of every newspaper that has carried our reports during these three decades. Unfortunately, many critics do not have the objectivity or the intellectual honesty to report, or refer to, this record. If there is any reference to election accuracy, it is usually in a way that leads readers to believe that, apart from the current election, the only other one we have ever covered was the election of 19481 THE BANDWAGON MYTH

No amount of factual evidence seems to kill the bandwagon myth. Our early experience indicated no evidence of a bandwagon movement among voters in national elections-at least none that we or anyone else could either detect or measure. More often than not, the candidate who is lagging far behind does better than expected. Now, after thirty years, the volume of evidence against the bandwagon theory has reached staggering proportions, and yet many writers continue to allude to this theory as an accepted fact. T h e workings of the laws of probability, as they concern the Sampling method, are still a mystery and likely will remain so for many years-at least until our high schools and colleges recognize the importance of teaching students the simple workings of these laws which affect every person so intimately in his daily life. But until that goal is realized, we will still have to go through the ritual of answering those who wonder how it is possible to reflect public opinion with a high degree of accuracy without interviewing every adult in the United States, or, at least, many millions!




Of a slightly more sophisticated order is the criticism that polls interfere with the election process. This comes from the political fraternity who are annoyed that ordinary voters should want to have a say

in choosing candidates. After thirty years of observing the workings of the electoral process, I am strongly convinced that this is the weakest and least defensible aspect of democratic government here in the United States. T h e whole electoral system needs to be reformed-beginning with the selection of candidates, on through the primaries and the conventions, to the campaign itself. We must make changes if we ever hope to select and elect our ablest citizens to public office and if we conduct election campaigns befitting a civilized nation. But we will never get any help in making these changes from machine politicians, who have a vested interest in the status quo. They have always resisted change, even of the most innocuous sort, in our election procedures. Another criticsm, voiced from time to time, is of the same order. This criticism is based upon the assumption that those in elective office will follow poll results blindly, even against their better judgment, in order to get re-elected. T h e simple truth is, as anyone can ascertain by examining poll results over the last thirty years, that legislators do not follow poll results. T h e quality of government, I might add, would be appreciably better if they did. Legislators are not particularly responsive to public opinion, chiefly because they become entrapped and enmeshed in the power struggle and archaic rules within their own establishment. INTELLIGENT LEADERSHIP DEMANDS FACTS

We are often told that the function of leadership is to lead. Not poll results, but that inner voice alone, should be heeded. This is an attractive and appealing concept of leadership and one which has intrigued mankind from the earliest days. Unfortunately, it fits perfectly such eminent leaders as Adolph Hitler, Benito Mussolini, and Premier Stalin. All three had supreme contempt for the views of the mass of people. Hitler, you may recall, described the common people as "mere ballot cattle." This is not the kind of leadership we want. I n a democracy we demand that the views of the people be taken into account. This does not mean that leaders must follow the public's views slavishly; it does mean that they should have available an accurate appraisal of public opinion and take some account of it in reaching their decision. As Bryce pointed out many years ago, the people are better fitted to de-



termine ends than to select means to those ends. T h e task of the leader is to decide how best to achieve the goals set by the people. THWARTING PRESSURE GROUPS

Pressure groups and their lobbyists continue to have great influence over legislation. T h e usual claim that they represent a large bloc of voters gains them a ready audience and influence with legislators. Were it not for public opinion polls, these claims, usually completely unfounded, would go unchallenged. All through the years, beginning with the famous Townsend plan, which had Congressmen frightened to death until the myth of the Townsendites' voting strength was exploded by poll results, down to the present time, survey results have whittled these claims down to size. At the present time, one of the most effective pressure groups in America-the makers of firearms and ammunition-is trying desperately to stop any legislation that would require guns to be registered. T h e only possible hope that legislation may be passed at some time in the future is the evidence that, as of this time, a large majority of Americans-even those who own guns-favor registration. Similarly, by a sizable majority, citizens of the United States believe that the distribution of birth-control literature should be legal. Even a majority of Catholics hold this view. But a powerful pressure group has been able to prevent the passage of this legislation in many states. Without polls it would be easy for this group to claim that all Catholics oppose the distribution of birth-control information. POLLS HELP GOVERNMENTS BE CREATIVE

Polls help the democratic process by refuting the claims of pressure groups. They help in another important way not foreseen in our early days; polls help governments to be creative. Every institution grows by trying new ideas, as new needs and interests arise. If governments are to move forward, they must accept new proposals that meet with public acceptance. T h e legislative area of government is the logical place for initiating legislation the aim of which is to meet the changing needs of the people more quickly and more efficiently. But as Senator Joseph Clark, in his monumental book, Congress, the Sapless Branch, says: "Congress compensates for its inability to act creatively by exercising its negative power to the hilt." He adds: "Since the founding of the Republic, Congress has rarely initiated anything, rarely faced up to current problems, even more rarely resolved them." Failure to act in these situations has left the responsibility for new legislative proposals to the executive branch of government. T o meet



this situation, Presidents, especially in earlier years, resorted to the device known as the "trial balloon." A new idea was leaked to the press from "persons in high authority." If the proposal met with a cold reception from the press, the trial balloon was promptly shot down and the President disclaimed any connection with it. Only if it was warmly received did he claim credit for it. Apart from moral considerations, this is a crude and inefficient way to test the public's acceptance of any proposal or idea. T h e modem public opinion poll, on the other hand, is admirably suited to this very important and necessary task of discovering how ready the public is to move forward on any front. Actually, a surprisingly large part of the space on our ballot forms is devoted to new proposals-our own trial balloons. During the last three decades we have tried out hundreds of proposals on the public, many of which are widely approved but may have to wait for years until Congress catches up with the people. THE FUTURE

And now about the future. As students, scholars, and the general public gain a better understanding of polls, they will have a greater appreciation of the service polls can perform in a democracy. In my opinion, modern polls are the chief hope of lifting government to a higher level, by showing that the public widely supports the reforms that will make this possible, by providing a modus operandi for testing new ideas, and by helping to reduce what Senator Clark calls "Congressional lag." Polls can help make government more efficient and responsive; they can improve the quality of candidates for public office; they can make this a truer democracy.


and to parrot the phrase: "The polls show." I t would make just as much sense to say "editors think," "social scientists believe," "politica...