Suggested Criteria for Writing Attitude Statements

Charles C. A. Wang[1]

The measurement of attitude by using attitude scales has in the past two or three years attracted wide-spread interest. Investigators in many fields are not only endeavoring to measure attitudes with ready-made scales but are also constructing various scales of their own. The successful construction of an attitude scale, however, requires more than knowing merely the mechanical steps of the procedure. In fact, the crucial task which first confronts one in constructing an attitude scale for a given issue is the collection of attitude statements on that issue. This is not a technical part of the construction method, but the success or failure of the scale depends much upon how well the initial list of statements is compiled and edited.

The purpose of this paper is to suggest a number of practical rules for writing attitude statements. These suggestions naturally are of special interest to those who construct attitude scales.

( 368)

Aside from the fact that care must be taken to cover adequately the entire range of attitudes so as to prevent a break in any part of the scale 'gird to include a variety of arguments so as to add reliability to the scale, the statements must be edited with regard to sentence construction and choice of words. The criterion of ambiguity and that of irrelevance (1, p. 44)valuable as they are In the selection of statements for the final scale, should not be depended upon too much in compiling the initial list of statements. Much time can be saved and statistical labor avoided if the experimental list does not. contain too much worthless material.

The criteria suggested below are by no means to be taken as final but have been compiled from work done on statements for a series of attitude scales. They may be considered as an elaboration of the informal criteria given in Thurstone and Chave's monograph. Whenever possible, actual sorting ' data are presented in order to illustrate the ambiguous effect of violating the rule under discussion. Since normally a good attitude statements is distributed in no more than three or four consecutive piles, the significance of these sorting data is apparent.

1. An attitude statement must be debatable. It must represent only an opinion which has no general acceptance. Thus, a universal truth or a statement of fact should never be used in an attitude scale. To illustrate:

Bad: Unions are organized to protect labor.
Better: Unions are desirable for protecting labor.

The first statement is one with which everyone would agree. Its endorsement indicates not any kind of an attitude but the acknowledgment of a fact. If it were included in an attitude scale, it would be endorsed by people who are opposed to unions as well as by those who are friendly to them. For a similar reason:

Bad: It is hard on the children to have the mother working.
Better: Women with children should not work.

2. All statements on a given issue should belong, as nearly as can be judged, to the same attitude variable. That is, they must be not only relevant to the issue but belong to the linear continuum that is being measured. As violations of this rule constitute the most common type of faulty statements, it is well that every attitude statement be considered in terms of this criterion. Examples:

Statement: In an ideal society there would be no law. (From a scale on attitude toward law, where the variable being measured is from complete respect to utter disrespect for law.)

Sorting data:

Pile   I   II   III   IV   V   VI   VII   VIII
Frequency   5   6   11   10   7   10   3   2

Statement: We are governed by laws, not by men. From the same scale as above.)

Sorting data:

Pile   II   III   IV   V   VI   VII   VIII   IX   X   XI
Frequency   1   3   5   1   3   0   15   15   10   1


Statement: Total abstinence from liquor can be accomplished only by education. (From a scale on attitude toward prohibition, where the variable being measured is from complete approval of, to violent opposition to, prohibition.)

Sorting data:

Pile   II   III   IV   V   VI   VII   VIII   IX   X
Frequency   1   8   3   6   9   8   4   1   1

The wide scatter in each case shows that the statement was ambiguous with reference to the principal issue.

3. An attitude statement must not be susceptible to more than one interpretation. It must contain no word or phrase which can be construed to mean different things by different individuals. A few illustrations follow:

Statement: Birth control legislation is a disgrace to our civilization.

Sorting data:

Pile   I   II   III           IX   X   XI
Frequency   10   19   9           4   13   7

The ambiguity of this statement is apparently caused by the fact that the statement can be endorsed by people who oppose as well as by those who endorse birth control, although with different interpretations.

Statement: Anyone who makes others suffer should suffer in return. (From a scale on attitude toward the treatment of criminals.)

Sorting data:

Pile   I   II   III   IV   V   VI   VII   VIII   IX   X   XI
Frequency   5   7   1   1   3   1   2   5   13   13   3

The ambiguity here seems to lie in the possibility that the term "anyone" may refer to criminals or to society.

Statement: There can be no compromise with the evolutionists. Sorting data:

Pile   I   II   III   IV   V   VI   VII   VIII   IX   X
Frequency   24   7   7   4   3   1   1   0   2   1

It is interesting to observe the consistency of the judgments by noting the results on a counter statement as shown below:

Statement: There can be no compromise with the enemies of evolution.

 Sorting data:

Pile   I   II   III   IV   V   VI   VII   VIII   IX   X   XI
Frequency   2   3   1   0   2   0   2   5   5   11   18

To endorse one of these statements does not clearly show to which end of the scale the subject belongs because the statements can be interpreted in different ways.

4. Avoid "double-barreled" statements. While it is sometimes necessary, in order to express neutrality or a mild attitude, to balance one idea with an opposite idea, experience has shown that this type of statement usually causes either high ambiguity or a break at the neutral pile, i.e., resulting in a bimodal distribution. To illustrate:

Statement: Athletic conditions are bad, but officials are trying to improve them.

( 370)

Sorting data:

Pile   III   IV   V   VI   VII   VIII   IX
Frequency   2   7   19   19   42   9   2

Statement: Birth control would be all right if we could prevent people from taking immoral advantage of it.

Sorting data:

Pile IV V   VI   VII   VIII
Frequency   2   17   2   39   3

These distributions show that the subjects had difficulty in deciding whether the real attitude is favorable or unfavorable to the issues involved.

5. An attitude statement should be short. It should rarely exceed fifteen words in length. Most statements should be much shorter. A long statement can usually be reduced to a shorter one without altering its essential point. For example: Instead of saying,

The Bible represents the sacred word of God and should be respected by everyone as such" the statement can be reduced to merely, "The Bible is the sacred word of God."

In writing attitude statements, it is well to try to shorten the length of each sentence written. In doing so, one usually also avoids the violation of many of the other rules here mentioned.

6. Each attitude statement should be complete in denoting a definite attitude toward a specific issue. Do not assume that the issue in question can be understood without specific reference to it. Thus, do not use "laws" to mean prohibition laws, or "it," "they," etc., unless the reference is perfectly clear within the statement in which they are used.

7. Each attitude statement should contain only one complete thought. Too many ideas in one statement cause confusion in interpreting the attitude and thus increase the chance of high ambiguity. The following ex-ample illustrates the effect of violating this rule:

Statement: Prohibition laws and enforcement are merely political games and there is no moral issue in keeping or breaking them.

Sorting data:

Pile   IV   V   VI   VII   VIII   IX   X   XI
Frequency   3   27   129   66   30   27   12   6

Another example of violation of this rule is the following:

"The church was established to serve a useful purpose but it has outlived its time; therefore, it is doing more harm than good."

The remedy for this type of statement is to break it up into two or more shorter statements, thus:

a. The church serves a useful purpose.
b. The church has outlived its usefulness.
c. The church does more harm than good.

8. Avoid grouping two or more complete sentences as one attitude statement. Do not transplant quotations by the paragraph en bloc, but rewrite

( 371)  them into one single sentence or several separate statements. For example, instead of quoting:

"War is the concentration of all human crimes. Under its standard gather violence, malignity, rage, fraud, perfidy, rapacity, and lust. If it only slew men it would do little. It turns man into a beast of prey."

as one attitude statement, it would be better to rewrite this into three separate statements, thus:

a. War is the concentration of all human crimes.
b. War is degenerating.
c. War makes man a beast of prey.

9. An attitude statement should be clear-cut and direct. Avoid statements which are not directly an attitude but from which an attitude is to be inferred, unless the inference is clear and unquestionable. To illustrate, as a statement of attitude on patriotism, the following is ambiguous as the sorting data demonstrate:

Statement: The makers of American policy are too idealistic and should not get mixed up in European affairs.

Sorting data:

Pile   I   II   III   IV   V   VI   VII   VIII   IX
Frequency   5   2   6   4   5   5   7   5   6

The wide scatter indicates that the subjects tried to guess at the attitude of the statement. Illustrating further, the following statements, as attitudes toward war, are likewise too indirect to be satisfactory:

"Military instruction should be separated from colleges and universities."
"Patriotism is the imperative end of education."
"Loyalty to our country comes before world-brotherhood."

On the other hand, some of the best attitude statements are those which infer unmistakably an attitude but do not explicitly assert it. Examples of this type are "Might is right" as an attitude favorable to militarism,. and "We do not need more babies but better ones" as an attitude favorable to birth control.

10. Use with care and moderation such words as "only," "mere," "just" (in the sense of only), "merely," etc. Many statements containing one or another of these words have been, found to cause ambiguity or bimodal distribution, even though on deliberation the affect may be clear. Examples:

Statement: Only by taking the money out of football can it be made really amateur.

Sorting data:

  Pile   I   II   III   IV   V   VI   VII   VIII   IX   X
  Frequency   6   16   11   20   12   2   11   2   2   1

Statement; In my thinking God merely means an ideal.

Sorting data:

Pile   II   III   IV   V   VI   VII   VIII
Frequency   1   4   17   19   4   25   7

Statement: I respect only those laws which represent the wishes of the majority.

( 372)

Sorting data:

Pile   II   III   IV   V   VI   VII   VIII
Frequency   3   4   12   44   0   22   2

These distributions show that by qualifying a statement with such words as "only," "mere," etc., it becomes difficult to interpret the attitude.

11. Avoid colorless expressions or statements lacking affect. While it is not necessary that every attitude statement must be emotionally toned, it should always represent some clearly defined conviction. Thus, do not say:

"The unions (or anything else) are all right."
"Churches Are beautiful buildings," . Etc.

To illustrate further:

Statement: Intercollegiate athletics are a necessary evil. Sorting data:

Pile   I   II   III   IV   V   VI   VII   VIII   IX   X   XI
Frequency   2   4   5   8   31   4   34   5   5   1   1

Since this type of statement expresses no conviction in one way or another, it can be endorsed by a person who may belong to any part of the scale.

12. Whenever possible, write an attitude statement in the form of a simple rather than a complex or compound sentence. The simple kind of statement reduces the chance for a wrong interpretation. Thus:

Bad: Women have always had enough to do to look after their men, and they should not need more.
Better: Women work enough without outside employment,

13. When a statement cannot be made in the form of a simple sentence, write it as a complex rather than a compound one. For example:

Awkward: Some code of ethics is necessary for the guidance of our conduct and the Bible may well serve that purpose.
Better: Since we need an ethical code for guiding conduct, the Bible is indispensable.
Best: The Bible is necessary for guiding conduct.

14. It is usually better to use the active rather than the passive voice. Thus,

"Punishment does not deter crime"

is preferred to

"Crime cannot be deterred by punishment."

An exception to this rule, however, will be discussed in the next paragraph.

15. In general, use the term of the issue as the subject of a statement. This is desirable in order to secure proper emphasis and attention. Hence it is permitted even in violation of Rule 14. For example, if the issue is on attitudes toward public Office, it is better to say, "Public officials are con-trolled by crooks,'," than to say, "The crooks control public officials." On the other hand, if the issue is on attitudes toward the social influence of crime, it would be better to use the second statement in preference to the first.

16. Avoid high-sounding words, uncommon words or expressions, technical terms not ordinarily understood, etc. When a scale is being prepared for use in a specific age, school, or sociological group, the vocabulary of

( 373) that group should be borne in mind. In any case, it is better to write statements in the simplest and most precise language possible.

In addition to the foregoing criteria, there may be mentioned several general rules, based largely upon good usage in English. These rules improve sentence structure although they are not necessarily concerned with the scale values or the Q-values of the statements.

1. Avoid a negative expression whenever a positive one can be substituted. Thus, use "disagree" instead of "not agree" "difficult" instead of "not easy," etc. Exceptions, of course, are permitted when the negative effect is desired.

2. Avoid double infinitives, especially in a short statement. For ex-ample, instead of saying, "To work on Sunday is to be immoral,"say, "Working on Sunday is immoral." Usually, in a case of double infinitives, at least one of them can be changed into a present participle.

3. Do not use redundant phrases. To illustrate:

Bad: We should not knock but boost our public officials.
Better: We should boost our public officials.

Bad: Communism should be absolutely banished at all costs.
Better: Communism should be banished absolutely.

By eliminating the redundancy, these statements are improved by their shorter length, yet lose none of their effectiveness as attitudes.

4. Avoid excessive use of such phrases as "I think that ... ; "I believe that ... "I feel .... etc., to precede a statement. Simply make the direct statement. Such prefices are excess baggage.

5. Avoid double negatives. Statements such as the following have some-times been found to result in high ambiguity:

"I don't believe in disobeying the law."

"I do not dislike the negroes."

This double negative ,type of statement lacks the force of directness with which every attitude should be expressed.

In conclusion, the distinguishing feature of an attitude statement lies mainly in that it expresses an attitude. To be sure, the technical requirement of the statement is that it belongs to the attitude continuum being measured, but that is adequately provided for by objective criteria. Other-wise, an attitude statement differs little from other statements, at least so far as language structure is concerned. It should therefore follow all the rules of good English (except occasionally for special sociological groups), with the added feature of being simple but precise, short but complete. The criteria suggested in this paper are discussed in detail only because it is hoped that, in doing so, they would help in obtaining initial statements that are more likely to produce satisfactory attitude scales.


1. THURSTONE, L. L., & CHAVE, E. J. The measurement of attitude. Chicago: Univ. Chicago Press, 1929. Pp. 96.

The University of Chicago
Chicago, Illinois


  1. Thanks are due the authors of several attitude scales for the use of certain illustrative statements. I am especially indebted to Prof. L. L. Thurstone through whom I obtained the original sorting data on these statements.

Valid HTML 4.01 Strict Valid CSS2