Statements As Opinion Indicators
Ohio State University
PUBLIC OPINION cannot be measured directly. An individual can observe the changes or stability of his opinion on a certain topic without using a means of observation. The opinions of a group of individuals, however, cannot be inspected without the help of a symbol. The symbol serves the purpose of mediating the group opinions to the investigator. It also allows a comparison of an individual's opinion with the opinion of another individual which is an essential characteristic of measurement. A symbol mediating an opinion may be called an opinion indicator.
Several opinion indicators may be distinguished, such as a number on a line, a word, a phrase, or a verbal statement. ‘A number on aline' has been used in rating scales, so has a word or a phrase. The latter two have also been used in the ranking method and the method of paired comparison. However, the verbal statement is the most widely used indicator of opinion. It, of course, can hardly be used in the rating method and the method of paired comparison; but nearly every investigator using the questionnaire method, the ranking method, and the method of equal-appearing intervals has employed statements to represent the opinions in question.
There is some disagreement among writers interested in attitude measurement as to whether the statements do represent attitudes. Statement for some seems to be too far removed from the deep seated dispositions to act. A verbal symbol according to these writers does not necessarily indi-
( 551) -cate what the actual disposition of a person is toward a certain issue or what the subsequent action will be. On the contrary, there is almost unanimous agreement between writers that statements do represent opinions, the more changeable and perhaps more rational sentiments. The writers would admit at the same time that in spite of the changeability and superficiality of an opinion it is very effective in influencing public affairs.
Upon an examination of a number of experiments the writer found that verbal statements are very indiscriminately used by the various investigators in this field. In-sufficient care is exercised in preparing and selecting the statements. No distinction is made between the various forms of statements. Several forms of statements are used in a single measuring instrument disproportionately. As an example of the indiscriminate use of statements we will examine a scale recently constructed by Thurstone for measuring attitude toward the movies. The scale consists of seven different forms of statements. Out of the 40 statements that constitute the scale 25 statements indicate the general function of the movies. Six statements are of the personal function type. Three statements are in the personal general function form, and three statements in the personal conduct form. One statement is causal, one is a personal general statement, and one is of the `should be done' type. All of Chave's  45 statements about the church are ire the personal form but one. The personal forms in his scale may be subdivided into several forms, such as the general function form, the general statement form, the conduct form, and the personal function form.
( 552) Vetter's  180 statements with a few exceptions are practically all in the `what should be done' form. The writer's scale  is not different in this respect from other statement scales because it was constructed before he ever thought of the advisability of discriminating between the different forms of statements.
The objection that statements do not represent continuous degrees of opinion even if scaled, and that such scales are not unidimensional, but at least bidimensional, is partly due to a lack of uniformity in the forms of statements used in scaling. Each form of statements represents a different trend of opinion and, therefore, it can be argued that a scale has as many dimensions, as many forms as are represented by the statements in the scale. If only one form is used throughout, the chances for its being a multidimensional scale are greatly diminished.
An analysis of the various forms of statements will be given below. A number of samples will be taken from current literature to illustrate the forms. The selection is rather arbitrary, being guided largely by the types of statements employed by the writers. Although our analysis will, on the whole, be made from the point of view of scaling, the illustrative samples will not only be taken from scales, but also from other instruments of measurement using statements as opinion indicators.
The forms of statements may be divided into three major groups : the impersonal form, the personal form, and the question form. By `impersonal' is meant a statement in which no pronoun is used or if it is used it does not appear in the singular first person. A `personal' statement is one
( 553) in which a pronoun is used singularly in the first person. By a `question' form is meant a statement with an interrogation point. Each of these three main divisions can be subdivided into several forms.
The impersonal form may be subdivided into the following forms : the general, the causal, the functional, and the `should be done' form. By a general form is meant one in which no specific reference is made to causes, functions, or conduct. "If one really loves his own country he will not love other countries," and "Japan's attitude in her relation with the United States in the last five years has been finer than our attitude toward her," are samples of general statements with an element of comparison in them." It is impossible for the world to become one great society," and "There is no conceivable justification for war," are purely general statements. A statement of this type is probably the farthest removed from an indication of what the individual would actually do. The use of such statements should, therefore, be discouraged.
Causal statements include a reference to the cause of the issue in question: e.g., "The most frequent cause of war is the rivalry of nations for possession of territory, markets, concessions, and spheres of influence." Statements of this type probably cannot be used alone for constructing a scale. A statement of cause does not have a very direct bearing upon the unfavorable or favorable opinion of the person endorsing it. But they can be used to express the medium position.
A statement in the function form is one expressing some-thing about the function of the issue or object of opinion. In the statement, "The League of Nations is the best possible solution of world problems, and absolutely prevents future wars," the `absolutely prevents future wars' is the function of the League of Nations. This is one of the two most frequently used forms of statements. All degrees of opinion can be expressed by it. A statement of the function of the object of an opinion is at the same time a statement of the reason of the opinion. In the statement, "War takes only the most physically fit, and thus results in national suicide of manpower," the reason for the opinion expressed is that war results in national suicide of man-power.
The fourth form refers to statements as to what should be done about the object of opinion. Examples of this form are: "We should observe and teach complete chastity in sex outside of wedlock," and, "Compulsory education should extend to four years of secondary or vocational school, adapted to the child's capacities." This is the other form of the two most frequently used forms in opinion measurement. Statements of `what should be done' seem to be more directly related to conduct than any of the above three forms.
The second main division of statements, the personal form, may be subdivided as follows : general form, causal form, general function form, specific function form, `should be done' form, and the conduct form. The personal forms were not used as frequently as the impersonal forms; our quotations will, therefore, be confined to fewer authors. In case the statements had to be modified to fit the form,
( 555) no reference is given. Statements falling under this general division are very similar to the ones cited above and will be passed by with fewer comments.
The general form may be exemplified by the following statement : "I believe in what the church teaches but with mental reservations." The causal form is illustrated by : "In my opinion the modern wars have been caused largely by diplomats who have played with human lives as pawns in a game of chess." The following statement is in the general function form: "I feel that church attendance is a fair index of the nation's morality.”
The statement, "Moving pictures bore me," is an illustration of the specific function form. This function form is mentioned here for the first time. Obviously, it does not have a corresponding form under the impersonal division. The `should be done' form may be exemplified in the following statement: "I believe that for the liberty of op-pressed nations wars should be fought."
Another addition in this division is the conduct form. It refers to the conduct of the individual endorsing the statement: e.g., "According to my first feeling reactions I would willingly admit members of each race (as class, and not the best I have known, nor the worst members) to one or more of the classifications under which I have placed a cross (X) to close kinship by marriage, to my club as personal chums, etc." In this statement the endorser is given a chance to indicate what he would do to races such as Canadians,
(556)Czechoslovaks, and Germans. The statement, "I like to see movies once in a while but they do disappoint you sometimes," although it has the element of function in it, also tells that the endorser goes to movies once in a while. An opinion about the church is expressed by reference to conduct in the following statement: "I believe in religion but I seldom go to church."
The question form is not used much, and never in scaling. The following sample is taken from Moore's questionnaire: "Do you believe in the principle of a minimum wage for workers?" The question form division may be subdivided into several subforms, similar to the ones used in the two preceding divisions, such as general form, function form, and the `should be done' form. General forms are such as, "How long do you expect to live?" or, "Do you favor any form of trial marriage?"  The `should be done' form is exemplified in the following statement: "Do you favor the early entrance of the United States into the League of Nations ?" Other forms may be devised, such as: "Does prohibition raise the morale of the nation?" in the function form, and: "Would you be glad to fight on the battlefield if your country would ask you to?" in the conduct form.
The question form is probably not an appropriate form for scaling. It does not represent the opinion as well as the non-question forms. The weight of the opinion lies in the response rather than in the question. Suppose a person answers "Yes," that word is the center of his opin-
( 557)-ion. In the non-question form the center of the opinion is in the statement itself. The statement, rather than the answer, is his opinion. An advantage of the question form is that it stimulates more than the non-question form to reveal a spontaneous and a genuine opinion.
Three major types of indicators were discussed in this paper: the impersonal type, the personal type, and the question type. Each of these types can be subdivided into various forms, such as the general form, the function form, and the conduct form. Most investigators made no distinction between the above forms. Two or more have been used to obtain a composite score. However, it is very likely that reliability of a scale would vary with the form of statements, or at least with the degree of uniformity in the use of the forms. Statements of function of the issue or statements of what should be done, are probably more re-liable indicators of what the endorser would do than the general form of statements. The personal conduct form would probably be the best indicator of what the individual would do.
It would be of value to construct several scales on the same issue using
different forms of verbal symbols for each. The scores on such scales could be
compared as to reliability and validity. Until this is done, some of the
assertions made above about the value of the forms of statements and their
combinations will naturally remain incomplete.