Questions Confy Guide

Now that the Confidential Guide has made its annual appearance, some evaluation of its claims to be representative of student opinion is in order. With admirable and characteristic directness, its editors inform us that the CRIMSON "is generally acknowledged to be a reputable source." This, however, hardly constitutes documentation of their further claim to have "carefully collected" freshmen views on the courses under exhumation. The elements of reputability in a newspaper are not the elements of reputability in a public opinion poll--as any reasonably sophisticated concentrator in Social Relations will tell you. The misuse of the public opinion poll during the 1948 election demonstrates demanding the most devoted care. It is not built on the simple mechanical order of, let us say, a bologna slicer, and confusion of the two can cause no end of difficulty.

The Guide claims to present an accurate picture of student opinion. This claim is not alone explicit; it is implicit in the phrasing of its reports, which cannot fail to remind us of the measured and solemn, yet assured, judgements of the great nineteenth century German historians: "Professor X was thought by '53, on the whole, to be an insensible dullard. Some, however, found him a towering intellect and an insipiring teachers, etc. etc." But what does "on the whole" mean? And is "some" ten, twenty, thirty or forty precent? We are never given any forthright statement of proportions. I do not wish to challenge the good faith of the CRIMSON's writers, but surely, all will agree that phrases like "on the whole" are open to the most varying sorts of interpretation, by editor and reader alike.

Off Base

There remains a more disturbing difficulty. The Guide does not inform us of its sampling base and its readers do not know which proportion of the members of each course actually returned questionnaires. The public opinion polls consistently over-estimates the Republican vote because their interviewers fail to contact enough working class respondents. We can suppose that omission of some part of the membership of a course may bias the results in a similar way.

Personal conversation with CRIMSON editors, whom I believe to be no less reputable than the historic institution for which they work, discloses that this indeed may have happened this year. A somewhat critical report on an unusually able young member of the faculty was submitted on the basis of questionnaires completed by less than half the members of his course. Among the missing were thirteen out of a total of fourteen "A" students, whose reports might have altered the report's entire tone. This sort of bias, of course, can work in both ways. A course with which I was associated received gratifying high praise. Those teaching it would be only to happy to believe that the Guide's hosannas were voiced by the massed phalanx of student opinion. But the questionnaires returned came from less than half our students. Possibly, inclusion of the others would have given us cause for anger and/or shame. (Quite possibly, since those who slept through the course might reasonably be expected to sleep through the poll). Further, I was informed that the questionnaires were not distributed at Radcliffe until well into the examination period--when many students had already gone home.

Responsibility

I do not bring up those points to make an attack on the continuance of the Guide. It is, in my opinion, one of the better features of the local scene, a tribute to the maturity of the Harvard undergraduate and a useful if painful source of information to the faculty. But with the responsibility of presenting student opinion to the University community goes the responsibility of making an accurate and complete survey of that opinion. Norman Birnbaum   Teaching Fellow in General Education

The Confidential Guide attempts to sample all members of the courses it lists; actual return on the polls runs about 30 to 50 percent. The comparison with national public opinion polling seems hardly valid: in such surveys, a small error can cause a faulty prediction; the Confidential Guide makes no attempt to predict. It tries only to present student opinion on courses, and it has frequently revised its methods to make the sampling of those opinions more accurate.