tice too long on one subject. They ought to be tested on ability to investigate, to collect material and organize debates, and on skill in presenting material to an audience in such a manner as successfully to persuade and convince."

"Abolish the special rebuttal speech. This forces the speaker to introduce rebuttal into his main speech. A few years of debating under this plan does much to develop flexibility. It is used exclusively by the Eastern Intercollegiate Debate League.'

"Eliminate championships everywhere. Have speakers present their discussions before real audiences, audiences whose opinions they definitely set out to influence."

"Let us have more variety in types of debates; more willingness to try new sorts and arrangements."

MEASUREMENT AND ANALYSIS OF AUDIENCE OPINION

HOWARD S. WOODWARD
Western Reserve University

AUDIENCE opinion is an elusive thing. Shy at times, at times

graciously pliable and accommodating; sometimes stolidly indifferent, emotionally effervescent sometimes; sometimes vociferous and belligerent; intelligent, informed and self-expressive, or uninformed and averse to exposing its ignorance; generally revealing a high state of inertia when asked to use pencils as well as minds,-an inertia that urging and cajolery can only partially overcome. The following is a part of the record of a persistent and extensive effort to get this elusive thing on record and evaluate it. To throw, by experimental findings, some illumination upon the behavior of public opinion in response to public discussion in the form of debating.

I. Statement of Conditions and Methods

This is a record that covers a period of three years, beginning in 1924 and ending in 1927. The number of audiences was 118; the number of individual voters, 3540; the number of questions, 8; the method of securing the expression of opinion was by use of a shiftof-opinion ballot; the speakers were nearly all students at Western Reserve University. These debates, carried on by the twenty or

thirty men who are active in the work each year, are called "Forum Debates," to differentiate them from the intercollegiate debates. Several intercollegiate debates are included in the record because the conditions differed from those of the "Forum Debates" only in that the debaters represented two schools instead of one. The groups were all Cleveland audiences and those of nearby towns; in no cases were they campus audiences or audiences actuated by patriotic interest in the home school. All debates were open to the public but consisted largely of members of the organizations under whose auspices the debates were held and friends of such members. There was a great variety of these organizations, local lodges of the Masonic order, the Knights of Pythias, the Maccabees, church clubs of divers denominations, Kiwanis, Chambers of Commerce, ward political clubs, factory clubs, a few high schools.

For two years before this record begins there were similar programs of debates. This was a period of experiment in an effort to find a ballot which would secure a maximum number of votes from each audience and a minimum of error that would invalidate the ballots. This involved testing the various classifications of opinion to be recorded on the ballots by the audience. Many revisions of the wording of instructions on the ballots and of the typographical forms also became necessary.

In the fall of 1924 the form of ballot was adopted which, with only one change, has been in use ever since. (Form A, page 109.) This change provides for sex and age data instead of the name of the voter. These ballots are 7x9 inches and space is provided for a brief statement of reasons for the attitude that is voted. Whenever conditions are such that it can be done, the audience is urged to vote before the debate, and again after it is over. This is usually done by a member of the instructional staff or by the management. The audience is also urged to mark the sex and age blanks. This is all that is requested of the audience. Generally the desirability of giving "Reasons" is mentioned but not emphasized. The back of the ballot is a form providing for "Comments on argument and presentation;" but no effort is made to secure its use by the audience. Many people do use it and give criticism that is profitable to the speakers. The whole form is so designed as to require a minimum of oral persuasion by the management and of activity

by the audience. The ballot serves its fundamental purpose if it bears two X's, registering the opinion of the voter.

The shift is tabulated after the debate by the management and announced as a decision of the debate at the close of the forum period. The next day the manager who is in attendance takes the ballots to the debate library and records and analyzes them on a form provided for the purpose. (Form B, page 110.) The ballots are then put into a binder and left in the room for study by the debaters, and the record sheet is there posted for their perusal.

The audience gets the statement of the question for discussion from a program which is distributed to it. This is a program printed for the season, not a special program for each debate. The program carries statements of all questions being debated, the list of names of all the students, from twenty to thirty in number, who are doing the speaking, the names of the student management and the instructional staff, the intercollegiate debate schedule, and a "catechism" relative to the "Forum Debates." (Form C, page 110.)

The form is set up and the type held throughout the season, changes being made as the list of questions, the list of speakers and the intercollegiate schedule change. Several printings are made in the course of the season.

Usually at these "Forum Debates" there is a critic judge. He is generally an alumnus member of the Western Reserve Chapter of Delta Sigma Rho. This judge is in attendance to decide the debate on the basis of effective debating and to sit down with the speakers at the close of the discussion for a "post mortem." He is provided with a form for his notes of criticism. This is brought to the debate library the following day, fastened into a letter file folder and placed at the disposal of the whole squad. Sometimes there are two, occasionally three such judges. These decisions furnish the basis of comparison with the audience decision which is discussed later.

This in brief is a statement of the conditions under which were secured the data I am submitting, and a suggestion of the methods employed throughout.

II. Success of the method in "getting out the vote" Audiences use this shift-of-opinion ballot to an extent that

makes it highly serviceable as a measure of audience opinion on public questions prior to discussion of them and as a means of evaluating the effect of debate on these questions.

The total attendance in the three years under consideration was 8135 and the number of valid ballots returned was 3540, a voting percentage of 43.3. The record is really much more favorable than it would appear from this summation. Of the 118 audiences, 52 voted less than 50% and 66 voted over 50%. Of the 52 audiences voting less than 50%, only 10 voted less than 30%. Of the 66 audiences voting over 50%, 17 voted between 60% and 70%; 6, between 70% and 80%; 8, between 80% and 90%; and 1 audience voted 100%. In general the size of the vote is in inverse ratio to the size of the audience, which conforms to normal expectation.

The voting percentage would be still higher were it not for the fact that these attendance figures represent counts made at the peak of attendance or estimates that are always liberal, and the further fact that only ballots voted both before and after the debate appear in the count. The voting percentage would be materially increased if the attendance figures showed only the number of persons present from beginning to end of the debate. Again, in some cases conditions were such that it was not advisable for the chairman to urge voting or even to call attention to the ballots and allow time for their use. It was in spite of all these and at times other untoward conditions that 66 of the 118 audiences voted over 50% and that none voted less than 20%.

III. Use of the audience vote to decide debates

Whether the change of opinion expressed by the audience is or is not a good method of deciding contest debates is of small significance. It seems hardly profitable to talk about just another method of adding forensic scalps to the war belt. However, in the 118 debates under discussion the tally of the change in audience opinion has been used as a decision. In 94 of these debates there were also judges (as noted above) who rendered decisions on the comparative effectiveness of debating. This method has seemed generally to be considered better for evaluating debating skill. Anyone who holds this presumption may draw what comfort he can from the record of 60 agreements between audience and judge and 34 disagreements, as here indicated.

If agreement with the judges is to be taken as the criterion of success, the case for audience judgment is even better than the figures indicate. There were instances of disagreement in which there were three judges, and the audience voted with the minority. On several occasions when there was disagreement the audience vote was a tie, while in a number of other cases a change of one or two ballots would have reversed the decision. It may well be that the minority judge and the audience better evaluated the work of the teams than did the other two judges. Many a time, also, a tie or a close vote is more just than a 1 to 0 decision.

But since when have judges become infallible? Who will say that a third of the time the audiences were wrong because they disagreed with the judges? In our superior wisdom the members of our instructional staff will testify that on a number of occasions of disagreement the audiences rendered the decision that better evaluated the effectiveness of debating! One suspicion that our observations tend to confirm is that judges as well as audiences are affected in their decisions by preconceived opinions or prejudices on the question. I do not maintain that observation of all this machinery at work and the statistics so far assembled prove a thing. But it is my feeling that in the long run debaters will as often get justice at the hands of audiences as judges, and I think much may be said for some method of measuring debating success by the expression of its effect on the opinion of the group of listeners.

Intercollegiate debating before home audiences is not under discussion. But personal observation, and evidence brought to