The Measurement of Attitude
Chapter 7: Some Further Problems in the Measurement of Attitude
L. L. Thurstone and E. J. Chave
AN ALTERNATIVE METHOD OF CONSTRUCTING THE ATTITUDE SCALE
In the present experiment the scale values of the opinions were determined by what we may call the sorting method. It is a variation of the method of equal-appearing intervals of psychophysics. Several hundred subjects sort out the statements into eleven piles which seem to them to be equally spaced on a subjective continuum from extreme pro to extreme anti on the issue in question. This procedure requires the participation of a large group of subjects who during the sorting do not express their own attitudes but merely sort the statements into successive equal-appearing intervals according to the meaning that the statements imply to the readers. Later, when a group is to be studied by means of the scale, the subjects actually vote on the statements in accordance with their own convictions and attitudes.
We shall mention here in passing the possibility of determining the scale-values of the statements without the rather laborious sorting process. It may be possible to scale the statements directly from the voting records of a large group of subjects provided that a considerable range of attitudes is represented in the group of subjects used for this purpose. The principle involved is that if two statements are close together on the scale, then the people who vote for one of them should be quite likely to vote for the other one also. If the statements are very different, spaced far apart on the scale, then those who vote for one of the statements should not be very likely to vote for the other one also.
It might be possible to reverse this reasoning. We might then
( 91) be able to infer the scale separation between two statements in terms of the number of subjects who indorse both statements, n1.2, the number who indorse the first, n1, and the number who indorse the second, n2. It is certain that the intrinsic popularity of a statement must be taken into consideration because two statements may belong to the same point on an attitude scale and yet differ widely in the relative frequency with which they are indorsed. This factor of relative popularity is accounted for by the total number of subjects who indorse each statement.
The index of similarity that we have described for the criterion of irrelevance incorporates these three factors in a rather crude way. In a later publication we hope to describe a procedure for scaling opinions directly from the records of voting and, if the procedure shows satisfactory internal consistency of the data, the sorting method here described may be superseded. The sorting method rests on fewer assumptions that can be questioned, as far as we can see now, and it was therefore chosen for our first experiments in the measurement of attitude.
THE VALIDITY OF THE SCALE
The scale must transcend the group measured. One crucial experimental test must be applied to our method of measuring attitudes before it can be accepted as valid. A measuring instrument must not be seriously affected in its measuring function by the object of measurement. To the extent that its measuring function is so affected, the validity of the instrument is impaired or limited. If a yardstick measured differently because of the fact that it was a rug, a picture, or a piece of paper that was being measured, then to that extent the trustworthiness of that yardstick as a measuring device would be impaired. Within the range of objects for which the measuring instrument is intended, its function must be independent of the object of measurement.
We must ascertain similarly the range of applicability of our method of measuring attitude. It will be noticed that the con-
( 92) -struction and the application of a scale for measuring attitude are two different tasks. If the scale is to be regarded as valid, the scale values of the statements should not be affected by the opinions of the people who help to construct it. This may turn out to be a severe test in practice, but the scaling method must stand such a test before it can be accepted as being more than a description of the people who construct the scale. At any rate, to the extent that the present method of scale construction is affected by the opinions of the readers who help to sort out the original statements into a scale, to that extent the validity or universality of the scale may be challenged.
Until experimental evidence may be forthcoming on this point, we shall make the assumption that the scale-values of the statements are independent of the attitude distribution of the readers who sort the statements. The assumption is, in other words, that two statements on a prohibition scale will be as easy or as difficult to discriminate for people who are "wet" as for those who are "dry." Given two adjacent statements from such a scale, we assume that the proportion of "wets" who say that statement A is wetter than statement B will be substantially the same as the corresponding proportion for the same statements obtained from a group of "drys."Restating the assumption in still another way, we are saying that it is just as difficult for a strong militarist as it is for a strong pacifist to tell which of two statements is the more militaristic in attitude. If, say, 85 per cent of the militarists declare statement A to be more militaristic than statement B, then, according to our assumption, substantially the same proportion of pacifists would make the same judgment. If this assumption is correct, then the scale is an instrument independent of the attitude which it is itself intended to measure.
The experimental test for this assumption consists merely in constructing two scales for the same issue with the same set of statements. One of these scales will be constructed on the returns from several hundred readers of militaristic sympathies and the
( 93) other scale will be constructed with the same statements on the returns from several hundred pacifists. If the scale values of the statements are practically the same in the two scales, then the validity of the method will be pretty well established. It will still be necessary to use opinion scales with some discretion. Queer results might be obtained with the prohibition scale, for example, if it were presented in a country in which prohibition is not an issue.
TWO POSSIBLE TYPES OF ATTITUDE SCALE
The present experiments have been confined to one type of attitude scale. When we started to solve this problem of measuring attitude we found that the scale could logically be constructed along either one of two rather different lines. One of these types was chosen as preferable and the decision was probably correct as far as one can tell as yet. For certain kinds of attitude material it is conceivable that the alternative type of scale would be the preferable.
The two types are illustrated by the two hypothetical diagrams of Figure 24. The upper diagram represents the type of scale with which we have here been working. The ordinates rep-resent the probability of indorsement of a statement while the base line is the scale itself. The interpretation is that any particular statement is most likely to be indorsed by the people who are scaled at the median of the distribution of indorsements for the statement. In other words, as we proceed from one end of the scale to the other by class-intervals, the probability of indorsement for any particular statement increases to a maximum and then decreases again to zero, as shown in the diagram.
Statements differ in intrinsic popularity as shown by the fact that the two hypothetical curves in the upper part of Figure 24 are of unequal area. Even if a statement and a person are scaled at the same point on the scale, it does not therefore follow that the person will necessarily indorse that statement. But when this condition obtains, then the probability is a maximum that the state-
( 94) -ment will be indorsed. The probability of indorsement of this particular statement is lower for people in other class-intervals than the one in which the statement is allocated. This type of attitude scale may be called the maximum probability type.
The alternative form of scale is shown in the hypothetical diagram of the lower part of Figure 20. It is best illustrated by one or two examples. Let us suppose that the issue in question is the desirability of capital punishment. Let the base line represent a series of crimes ranging from minor offenses at the left end of the
( 95) scale to the most serious crimes imaginable at the right end of the scale. Let us also assume that these crimes have been allocated to the continuum by sorting on the basis of relative seriousness. Now if we asked the subjects to check each crime in the list which they consider serious enough to deserve capital punishment, we should expect to find that as the seriousness of the crime increases the probability of checking for capital punishment would also increase. Very likely the proportion of such judgments would be zero for the minor offenses. It would rise as the seriousness of the offenses rises but there would be no maximum. The curve would be asymptotic probably to the level of unity. If some subjects refuse absolutely to indorse any crimes for capital punishment, then the curve will approach unity but will not reach it. It is also possible that the curve will approach a level below unity. It will not again fall because if the group indorses capital punishment for a crime of a specified degree of seriousness, certainly we expect the same group to indorse capital punishment just as often, or possibly more often, for all crimes that they judge to be more serious. On account of this characteristic we have called this type of scale the increasing probability type.
Another example of the same type of scale with a different sort of base line can be thought of for the prohibition question. Suppose that the base line represents the percentage of alcohol that should be allowed. The curve for such a scale would begin supposedly at unity and it would fall toward the base line if the X-scale is one of increasing percentage of alcohol. This is clear be-cause any one who indorses, let us say, 10 per cent of alcohol will almost certainly indorse a lower percentage as legally allowable. The scale-value and the Q-value of a statement so scaled would be determined by methods analogous to those already described.
Still another example of this type of scale would be a list of justifiable provocations for war. Let us suppose that these provocations were listed in rank order of seriousness by the sorting method described. Now if each subject is asked to check those
( 96) provocations which he considers sufficiently serious or aggravating to justify the declaration of war, then it is reasonable to expect that the person who checks a relatively trivial cause as a suitable provocation for declaring war will also check the more serious situations in the same way. The resulting curve would be a rising probability curve which reminds one of the integral of a frequency distribution which may or may not be normal or symmetrical.
It is possible that certain issues may lend themselves to measurement by the
increasing probability type of scale but it is probable that most issues will be
better described if the scale is intentionally constructed so that a person is
more likely to indorse the opinions at some one part of the scale than at any
other part. Such a scale is the maximum probability type.