Public Opinion

To the Editors of the CRIMSON:

Three polls have been distributed in the Harvard community during the past few weeks-the Young Progressives' Peace Poll, the Harvard Liberal Union follow-up poll, and the Yearbook Poll. Taken together, the three are guilty of practically every error that can be committed in opinion polling. We as Social Relations concentrators are exasperated at the carelessness and inaccuracy of the polling methods.

The first of these questionnaires, the YP poll, was apparently constructed to prove a point, not to measure attitudes accurately. In attempting to prove this point, the investigators broke the first rule of polling. By inserting a "rider clause" at the beginning of a question ("Desiring peace, I believe..."), the investigator forced the student to subscribe to a belief which was irrelevant to the main question. It is not surprising that the poll was declared invalid by "a faculty member who asked that his name be withheld."

Though the HLU poll was more carefully constructed, it seemed to take the form of a political weapon designed to repudiate the YP poll, rather than that of an impartial gauge of public opinion. In effect the HLU was rebuking the YP's for swearing, because swearing sounds like hell.

The Yearbook poll was longer; therefore, it afforded a greater variety of errors. Here are a few of the most obvious broken rules:

1) Unless more than one answer is requested, suggested answers should not overlap, but should be mutually exclusive. Question 44: "Do you expect your wife to be a...(1) Career Woman? (2) Mother? (3) Social Partner? (4) Other?"

2) Suggested answers should exhaust all reasonable alternatives. Question 33: "If Harvard continues to have losing football seasons, would you favor dropping football? (1) Never (2) No (3) No opinion."

3) Questions should be clearly and unambiguously worded. Question 48: "Would you be interested in serving in some political office in...(5) Non-political civil service?..."

4) The Investigator's choice of alternative answers should not be blased. Question 51: "If the presidential election were held tomorrow and you could choose from among these men, would you vote for...(here seven Republicans and one Democrat are proposed for selection)...?"

We do not question the sincerity of the Yearbook staff; undoubtedly they made honest "efforts to avoid loaded questions and vague alternatives." Instead we feel that their poll exemplifies the attitude that "anybody" can construct a poll. It appears that the Yearbook investigators did not seek intelligent advice. If they did, they did not follow it.

We hope that future amateur pollsters can learn the following lessons from these failures: A valid poll is a difficult instrument to construct. There are certain rules, however, which an investigator must follow if he desires accurate and honest results. Morton D. Goldberg '52   Neil J. Smelser '52