Referendum

To the Editors of the Crimson:

In your article on the undergraduate council referendum you managed to report both the figures and the question incorrectly. When I saw the referendum results divided into "Yes" and "No" with the percentages adding up to 100 I knew something was terribly wrong. Reading farther I found that the numbers you gave (the vote totals) simply do not agree with the percentages. Assuming the vote totals are correct, they can be used to deduce the actual results of the referendum. They suggest even stronger support for council involvement in divestiture. However, taking all errors into consideration, it seems those who want to see the council involved may be a small minority.

The true percentages (using the vote totals you gave) are, on the first question, 64 percent yes, 34 percent no, and 3 percent no answer (results do not add up to 100 due to rounding). On the second, 59 percent yes, 6 percent no, and 35 percent no answer. Eighty-three people voted on the second question but not the first. A startling 1018 voted only on the first. If you were told by council vice chairman Steven Smith that 38.2 percent voted No on the second question, he was doing his best to mislead you.

Were the results then distorted against divestment? A closer look suggests the opposite. The fact that almost four times as many voted no, Harvard should not divest, as voted no on question two (which The Crimson interprets as saying that the council should not divest) shows that something is very wrong. Indeed, far more misleading than Smith's report is The Crimson's incorrect paraphrase of the second question. Far more significant is the reason why 1018 did not vote on it, and why 55.3 percent did not vote at all. In the bold box on the top right corner of page one you tell us 61.8 percent voted that "the council [should] push for divestment." Nothing could be less true. The question was "[shall] the undergraduate council...support and encourage the majority view on the proceeding proposition through the appropriate means," a question perfectly suited to those who wish to see the council support divestment. I assume only the most fanatic would want the council to take a stand if the majority of undergraduates opposed divestiture. Thus by voting that "the...council shall support and encourage the majority view..." divestment supporters could accurately express how they feel. No doubt they wrote the referendum. Perhaps, given their partisan views, they never noticed the one-sided nature of the question. Perhaps they intended it to be loaded.

In any case, those opposed to council involvement were not nearly so lucky. They can probably be divided into two groups, those who support divestment but would rather the council not waste its time with the issue (like me), and those who do not want Harvard to divest at all. Would either group want to vote on a question that says "the...council shall support and encourage the majority view on the preceding question...?" Would they be able to cast a vote expressing their views? No way. Given the real question, "should the council support divestiture," they would say no. But this is not what they were asked. Given only a choice of whether the council should support majority views, even such people would be unlikely to vote no--especially not knowing the "majority view." Considering what a yes vote would mean they would be unlikely to vote at all. Many holding one view would fear their "vote" would suggest they support another.

It is people opposed to council involvement that were most likely not to "vote" on question two. It is people disgusted by the question, and people fed up with the council, doubting it can do anything for students, who did not "vote" at all. Such people comprise 71 percent of the undergraduate population.

What we have here is a sampling error of unheard of proportions. If, in my high school statistics class, I turned in such a poor poll, with such a poorly worded question, I would have failed. The effect a question can have on both the response rate and the nature of the response cannot be overemphasized. The council's referendum is proof. Its results are immensely uncertain, with a large bias favoring those who want the council involved. While 59 (the percentage of "Yes" votes on question two) is approximately the upper bound on the percent of undergraduates who want the council to support divestment, the lower bound is a mere 26.5 percent (100 percent minus 71 percent minus the 2.5 percent who "voted" "No" on question two equals 26.5 percent). The lower bound is probably closer to reality. Indeed, while as many as 64 percent (the percentage of "Yes" votes on question one) may support divestiture (a decline from an earlier referendum), the true percentage could be as low as 28.4.

I suspect the majority really do support divestment. The first question's results include a sampling error, but a smaller one, and question one was not biased. But I also suspect the majority do not want the council to support divestment. The only way to find out for sure is to ask an unbiased sample the real question. Why doesn't The Crimson step in where the council failed and, using a phone poll, find the real answer?

How important is this information? It depends on whether the council should be democratic, supporting every whim of its constituents, or republican (I'm talking political theory, not GOP), using its best judgment. This has never been decided, but this referendum suggests the council does not know best. Andrew F. Schmid '88