The Measurement of Attitude

Chapter 4: The Experimental Attitude Scale

L. L. Thurstone and E. J. Chave

Table of Contents | Next | Previous


A final list of 45 statements of opinion was selected from the original list of 130 opinions. The selection was made with consideration of the criterion of ambiguity, the criterion of irrelevance, the scale-values, and by inspection of the statements. The statements were so selected that they constitute a more or less uniformly graduated series of scale-values. The scale-values rep-resented by the 45 statements in the final list are shown in Figure 17, the purpose of which is to show that the 45 statements repre-

figure 17

sent a more or less evenly graduated series of scale-values. The two upper horizontal lines constitute the scale proper and between these two lines there are small cross lines. Each of these cross lines represents the location of one of the 45 opinions. There are forty-five such cross lines. The class intervals on the scale and their numerical designations are indicated immediately below. Inspection of the figure shows that the graduation is by no means perfect but it is probably as close as is necessary. Four opinions were selected from each of the eleven class-intervals of the scale. In addition, one extreme statement that scaled at 11.8 or beyond was included.

The experimental scale as finally presented to the several hundred subjects for actual voting together with the accompanying instructions is given here. The questions on the title page of the

( 60) blank were inserted for the possibility of correlating facts obtained from the replies with the scores on the attitude scale. For our present purposes these items are of secondary importance al-though in the practical use of a scale of this kind these correlational items probably would constitute an essential part of the investigation. As a by-product of our present investigation, we have tabulated the various facts about the voters in order that separate frequency distributions may be constructed for subjects of various classifications.


This is an experimental study of the distribution of attitude toward the church. You will be asked to read a list of statements about the church and to indorse those that express your own sentiment. Let your own experience with churches determine your indorsements.

1. Name*

2. Group

3. Underline your classification: Freshman, Sophomore, Junior, Senior, Graduate, Faculty, Unclassified

4. Department of major work:

5. Do you attend church frequently? Yes No (Underline one)

6. Are you an active member of a church? Yes No (Underline one)

7. Before turning this page write a brief statement indicating your general attitude toward the church as you know it.

8. Write an X somewhere on the line below to indicate where you think you belong.

Strongly favorable    Neutral   Strongly against

to the church the church

* You need not sign your name, if you prefer to give your opinions anonymously.

( 61)

Check () every statement below that expresses your sentiment toward the church. Interpret the statements in accordance with your own experience with churches.'

(40)   1. I think the teaching of the church is altogether too superficial to have much social significance.

(50)   2. I feel the church services give me inspiration and help me to live up to my best during the following week.

(110) 3. I think the church keeps business and politics up to a higher standard than they would otherwise tend to maintain.

(103) 4. I find the services of the church both restful and inspiring.

(54)   5.When I go to church I enjoy a fine ritual service with good music.

(28)   6. I believe in what the church teaches but with mental reservations.

(')  7. I do not receive any benefit from attending church services but I think it helps some people.

(13)   8. I believe in religion but I seldom go to church.

(69)   9. I am careless about religion and church relationships but I wouldnot like to see my attitude become general.

(96) 10. I regard the church as a static, crystallized institution and as such it is unwholesome and detrimental to society and the individual.

(93) 11. I believe church membership is almost essential to living life at its best.

(21) 12. I do not understand the dogmas or creeds of the church but I find that the church helps me to be more honest and creditable.

(7) 13. The paternal and benevolent attitude of the church is quite distasteful to me.

(19) 14. I feel that church attendance is a fair index of the nation's morality.

(34) 15. Sometimes I feel that the church and religion are necessary and sometimes I doubt it.

(114) 16. I believe the church is fundamentally sound but some of its adherents have given it a bad name.

(65) 17. I think the church is a parasite on society.

' The statements of this list are numbered consecutively as shown. The number in parentheses before each statement refers to the original list of 130 statements.

( 62)

(64) 18. I feel the need for religion but do not find what I want in any one church.

(38) 19. I think too much money is being spent on the church for the benefit that is being derived.

(11) 20. I believe in the church and its teachings because I have been accustomed to them since I was a child.

(87) 21. I think the church is hundreds of years behind the times and can-not make a dent on modern life.

(89) 22. I believe the church has grown up with the primary purpose of perpetuating the spirit and teachings of Jesus and deserves loyal support.

(113) 23. I feel the church perpetuates the values which man puts highest in his philosophy of life.

(51) 24. I feel I can worship God better out of doors than in the church and I get more inspiration there.

(128) 25. My experience is that the church is hopelessly out of date.

(33) 26. I feel the church is petty, always quarreling over matters that have no interest or importance.

(95) 27. I do not believe in any brand of religion or in any particular church but I have never given the subject serious thought.

(125) 28. I respect any church-member's beliefs but I think it is all "bunk."

(74) 29. I enjoy my church because there is a spirit of friendliness there.

(41) 30. I think the country would be better off if the churches were closed and the ministers set to some useful work.

(101) 31. I believe the church is the greatest institution in America today.

(27) 32. I believe in sincerity and goodness without any church ceremonies.

(75) 33. I believe the church is the greatest influence for good government and-right living.

(72) 34. I think the organized church is an enemy of science and truth. (56) 35.I believe the church is losing ground as education advances.

(24) 36. The churches may be doing good and useful work but they do not interest me.

(110) 37. I think the church is a hindrance to religion for it still depends upon magic, superstition, and myth.

(107) 38. The church is needed to develop religion, which has always been concerned with man's deepest feelings and greatest values.

( 63)

(36) 39. I believe the churches are too much divided by factions and de-nominations to be a strong force for righteousness.

(48) 40. The church represents shallowness, hypocrisy, and prejudice.

(127) 41. I think the church seeks to impose a lot of worn-out dogmas and medieval superstitions.

(14) 42. I think the church allows denominational differences to appear larger than true religion.

(90) 43. I like the ceremonies of my church but do not miss them much when I stay away.

(100) 44. I believe the church is a powerful agency for promoting both individual and social righteousness.

(73) 45. I like to go to church for I get something worth while to think about and it keeps my mind filled with right thoughts.


We have given numerical designations to the successive class-intervals of the scale. The unit of measurement is defined by the number of equal-appearing intervals into which the original list of 130 statements was sorted by the first group of subjects. Since the subjects were asked to sort out the 130 statements into eleven piles subjectively equally distant from each other, the unit of measurement is thereby defined. The origin is arbitrarily as-signed. We could have placed the origin in the middle of the scale, but that would necessitate dealing with negative class-intervals and nothing is statistically gained thereby. We therefore assigned the origin to one end of the scale, the extreme pro-end. In most mental tests the high and low scores represent performances that can be described as good or bad, but in the present instance there is, of course, no such possibility. We have no right to say that a person who is very much devoted to his church is in any sense better than a person who has no such affiliations. Nor can we say that one person scores "higher" than another except in the arbitrary sense that one end of the scale is called zero and the other end eleven. It is a matter of indifference which end is chosen for the high numerical scores. What we are here concerned with

( 64) is merely the description of one aspect of the attitudes of people about the church, an aspect which can be thought of as a linear continuum. We have no interest in any implications that one score is better than some other score in a moral sense or that one score is higher than some other score in the sense of relative value or achievement. These considerations, important for the unbiased construction of an attitude scale, of course leave any user of such a scale entirely free to make his own moral interpretations of the scores.

It is to be expected that many of the users of an attitude scale will have as their motive the influencing of people toward a chosen end of the scale. Our object is to produce a tool, as objective as possible, by which to describe attitudes. In the absence of objective tools for the description of social phenomena, the conclusions of any investigator are always subject to the challenge that he has reported his facts with his own bias, either intentionally or unintentionally. To the extent that social phenomena may be de-scribed with devices that are free from the investigator's own bias, to that extent we shall be able to make sound inferences, to just that extent shall we be able to separate our facts from our own personal desires.

Having adopted an origin of measurement which is assigned arbitrarily at one end of the series of available opinions, and having a unit of measurement, the equal-appearing interval as de-scribed, we can proceed to ascertain the mean scale-value of all the opinions that any individual subject indorses. This mean scale-value of all the opinions which a subject indorses we call his score.

In the clerical work of scoring the blanks there are two alter-native procedures which will probably give substantially the same results. We may assign the scale-value to each of the statements that a subject has indorsed and then calculate their arithmetic mean. This is legitimate to the extent that each class-interval of the scale is equally represented by opinions. Since we have the 45

( 65) statements in the experimental scale graduated more or less evenly, we feel justified in allocating each subject to the scale at the mean scale-value of the opinions that he has indorsed. This seems reasonable because there is approximately the same number of opinions available for him to check in each class-interval.

Another procedure which is statistically equivalent is to assign a rank order number to each of the 45 statements in the scale and merely to calculate the average rank order of the statements which he has indorsed. This procedure is equivalent to the previous method in so far as we have the 45 opinions evenly spaced throughout the scale. In the blank the 45 statements were presented in random order, not in the order of their scale-values. This was done in order to encourage the subject to read all of the statements.


In order to test the reliability of the experimental scale it was divided into two parts. The usual procedure of assigning alternate items to the two forms A and B was slightly modified because that procedure would give one of the two parallel forms a slightly higher mean scale-value than the other. In order to make the two forms truly parallel, as far as that was possible with the material at hand, we arranged all the opinions of the scale in rank order ac-cording to their scale-values. Successive pairs were then marked off. The first opinion in each pair had, of course, a slightly lower scale-value than the second. In the odd numbered pairs the first opinion with the lower scale-value was assigned to form A of the scale, and in the even numbered pairs the second opinion with the higher scale-value was assigned to form A of the test. The others were assigned to form B. In this manner we obtained two forms, A and B, each half as long as the experimental scale, and so pre-pared that the average scale-values of the two forms were practically identical. The odd statement scaled at 11.8 was included in both forms, A and B.

The blanks of two hundred Freshmen were used for the pur-

( 66) -pose of determining reliability. Each blank was given two scores, one score for the opinions that had been assigned to form A and a second score from the same blank for the opinions that had been assigned to form B. The two sets of two hundred partial scores were then correlated. The correlation between the two sets of scores was 0.848. When this correlation between the two halves of the scale is interpreted by means of the Spearman-Brown formula, the estimated reliability of the whole scale is 0.92, which is quite satisfactory. Earlier in the study the same procedure was applied to the blanks of one hundred subjects in which the correlation between the two halves of the scale turned out to be o.89, which is comparable with the correlation of o.85 between the two halves of the scale for the larger group of two hundred subjects. If the correlation of 0.89 between the two halves for the one hundred subjects is interpreted by the Spearman-Brown formula, the reliability of the whole test is estimated to be 0.94.


  1. Since the completion of this monograph the index of similarity has been developed in a better and more correct form. Its complete description will be published shortly.

Valid HTML 4.01 Strict Valid CSS2