The Measurement of Public Opinion (Report from the Round Table on Political Statistics at the Second National Conference on the Science of Politics, Chicago, September 8 - 12, 1924)
A. N. Holcombe
The round table on political statistics at the Chicago Conference dealt with topics of a different nature from those considered by the round table on the same subject the previous year at Madison. It was impossible, therefore, for this round table to begin where its predecessor left off. While its predecessor's work was continued by the round table on municipal administration, the round table on political statistics made a fresh beginning.
1. The Meaning of Public Opinion. Before even a beginning could be made, it was necessary to come to some agreement concerning the meaning of terms. Some members of the round table believed that there is no such thing as public opinion; others believed in its existence but doubted their ability to define it with sufficient precision for scientific purposes. Others again, more sanguine or perhaps more credulous, believed that the term could be defined, but were of different minds concerning the kind of definition that should be adopted. To reconcile these differences, it was decided first to consider the nature of opinion in general in the hope that a common understanding of the meaning of public opinion would emerge from the discussion.
To this end each member of the round table was asked to prepare a definition of the term "opinion." From these definitions nine questions concerning the meaning of the term were derived. After further discussion, it was decided that for the purposes of the problem before this round table the essential points in a definition of opinion could be narrowed to three: (1) opinion need not be the result of a rational process; (2) it need not include an awareness of choice; and (3) it must be sufficiently clear or definite to create a disposition to act upon it under favorable circumstances.
On the question when is opinion public, the round table was unable to come to a definite conclusion. The main points of disagreement were as follows: (1) whether there is and must of necessity be a single public opinion, or whether there may be a number of public opinions upon a given question; (2) whether opinion is public because of the subject-matter to which it relates or of the kind of persons who hold it; (3) what part of the public must concur in an opinion to make it public opinion; and (4) must there be acquiescence by those who do not concur. After some discussion of these paints, it was agreed that an exact definition of public opinion might not be needed until after the technical problem
( 124) of measuring the opinions of the individual members of the public had been disposed of. It was decided therefore that the round table might well proceed to consider the problem of measuring opinion, especially that relating to political matters, and avoid the use of the term public opinion, if possible.
2. Methods of Measuring Opinion. Twenty-three methods by which opinion might be measured were suggested for consideration. These methods were arranged under four heads and then discussed, first by four committees of the round table and afterward by the round table as a whole. The first group of methods included all those based on the study of official election returns. The second group included those utilizing fair samples of the bodies of opinion to be measured, collected deliberately but unofficially. The third group included all methods utilizing voluntary or spontaneous expressions of opinion. The fourth included those which utilize the data that may be derived from the proceedings of legislative bodies and the acts of public officers of all kinds, possessing any representative character.
3. Direct Votes Upon Measures and Candidates. The most comprehensive as well as the most accurate method by which opinions upon political questions are measured, is direct voting at public elections upon measures or candidates. It is the only method which has been used in such a way as to furnish records readily available for statistical analysis, and it is the only one which considers the whole adult population. However, even voting upon measures, which obviously is a more exact indication of opinion upon public questions than voting upon candidates, has certain definite limitations: For example, a categorical answer must be given to questions which are often complex, and thus a misleading simplicity is obtained. Moreover, it fails to measure the intensity of opinions, and takes no account of the opinions, if any, of those who do not vote. Two definite proposals for further study of the problems involved in this method of measuring opinion were agreed upon: (1) the supplementing of the study made by Professor Merriam and Dr. Gosnell concerning the causes of nonvoting by a somewhat similar study of the reasons for voting; and (2) the further investigation of the comparative proportions of those entitled to vote who vote at special elections upon referendum proposals, whether constitutional or statutory, and at general elections, with a view to ascertaining the amount and causes of the difference, if any, and the significance of such differences.
4. The Questionnaire as a Method of Measuring Opinion. The second group of methods for measuring opinion includes particularly the taking of straw votes and the use of questionnaires. Practically all the discussion of this method was devoted to the problems connected with the use of the questionnaire: The over-representation of the interested, the literate and the well-to-do, the fairness of phrasing, the possibility of plural voting, the presence or absence of discussion, the maturity of the answer, the durability of the opinions so measured, the desirability of secrecy. Considering all these difficulties, certain members of the round table endeavored to frame a model questionnaire suitable for measuring opinion on a problem of contemporary interest. It is hoped that these members will be able to carry on such experiments with this questionnaire as will more accurately determine the usefulness of this method.
5. Propagandist Organizations and Spontaneous Expression. The third group of methods dealt with what may be termed unsolicited or spontaneous expressions of opinions. Among these are public hearings, voluntary party enrollment, lobbying, petitions, circulation of the press, public meetings, editorials, campaign contributions, speeches of public men, private research groups, letters to the press, the behavior of crowds, and propagandist organizations. It was found that certain of these, for example, letters to the press and private research groups, are of little importance, as being at best incapable of exact measurement. Others, as campaign contributions, although they may be of more significance as a means of measuring opinion, are not yet capable of very meaningful measurement. Still others, such as public hearings and the extent of the circulation of the press, while considered to be of considerable significance and capable of more or less accurate measurement, were not considered very carefully because of lack of time. It was deemed advisable to give nearly all the available time to the consideration of propagandist organizations. Their rapidity of growth, and their distribution, while susceptible of measurement, are of varying significance.
Membership lists, if accurate, and if made public, are of considerable importance. The number and character of the activities of such organizations, whether active or quiescent, whether, for example, they have speakers' bureaus, hold public meetings, issue pamphlets, and the like, may indicate with some exactness the extent to which an organization is really an index of opinion, an index, that is, not merely of the number of persons holding certain opinions, but of the intensity with which they hold them. Closely related to these questions is that of the extent to
( 126) which the activities of propagandist organizations stimulate their members to individual expressions of opinion through public or private letters, telegrams, petitions, etc. A pattern of the nature and activities of various propagandist organizations might perhaps be worked out so that it would furnish graphically and with a fair measure of exactness an indication of the extent to which these organizations are really to be relied upon in determining the extent and intensity of opinion. The round table was unanimously of the opinion that the public should know the facts concerning the membership and finances of all organizations designed to influence opinion on public questions.
6. The Analysis of Opinion of Official Representatives. The fourth and last group of methods under consideration for the measurement of opinions included those dependent upon an analysis of the opinions of official representatives of the public. These methods could utilize such material as the resolutions and other public acts of legislative bodies, recommendations and veto messages of chief executives, decisions of courts and administrative tribunals on questions of public policy, and the platforms of political parties. It was decided that at present a general answer to the question, "To what extent do these acts conform to the opinions of the electorate?" is impossible. Many people believe that the chief executive more correctly speaks for his constituency than does the legislature. Yet one of the tests of the accuracy of the executive's statement of what the people desire is the reception of his recommendations by the legislative body. It seemed advisable, then, that in developing these methods, a beginning be made with the legislative branch of the government. One problem for such a method might well be the determination, if possible, of the relation between the attitude of representatives toward measures in legislative bodies and the important characteristics and interests of the communities which they represent. This would involve detailed analysis of the environmental factors of all kinds, economic, social, racial, religious, which influence the opinions of individuals, a laborious undertaking, but one promising the development of a technique which would enable the political scientist, not merely to measure opinion when duly formed, but even to prognosticate it before it is formed.
At this point in its deliberation the round table was forced to adjourn by the expiration of its time; and the members separated in the belief that a beginning had been made from which the next round table on political statistics could make substantial progress.
A. N. HOLCOMBE.