It’s recommendation season at colleges across the country. Last year, when asked by a Yale senior majoring in English to recommend her for work after graduation with a notable public service organization, I was happy to oblige — until I opened the online recommendation form that required me to answer every question about her numerically.

Let’s pause here. As someone whose professional life has been devoted to the study and teaching of those soft and subjective humanities subjects ­— dramatic literature, theater and film — I may be suspected of social-scientific Luddism. But no, I take reliable statistics seriously, on issues of health, environment, education, population, standards of living and many more. I sit up when reading that 35 percent of adult Americans are clinically obese, especially when I’m told why; or that more than 40 percent of Bangladeshis live below the international poverty line; or that 50 percent of the world’s agricultural land is threatened by encroaching salt water. When my blood is tested, I want to know the cholesterol numbers. But can I swear that my student falls in the top 5 percent, or the next 10 percent, of all undergraduates I’ve taught in the last 30 years, or even five, in this or that category?

Yes, the grade sheets help. But can I be sure that my standards haven’t shifted, however subtly, over the years? Could I have mellowed? Or how do I compare an A grade in, say, “Modern American Drama” with an A in acting Shakespeare? And as to more personal assessments like consistency of preparation, quality of seminar commentary, listening to classmates, can I really trust my necessarily subjective memory — comparatively — any more than that of any other past experience? When I read those supposedly clinching testimonies from colleagues, like “This is one of the three best and brightest students I’ve taught in 37 years at Yale, Stanford, Penn, etc.,” I’m no longer so quickly bowled over.

This move toward quantifying essentially volatile human factors has been coming at us steadily for a while. But last year’s recommendation form was new to me. In times past you could get away with crossing out all the comparative questions and substituting “see attached,” i.e. a regular letter in continuous prose. Not this time. These statistical questions had to be answered. And that creates its own problem, since when you genuinely (if mistakenly) trust your judgment that the student belongs not in the top five percent but only in the next 10 percent, do you risk damning him or her by saying so, if your reader dismisses all but the very top-liners? Several guessing games are going on here, because you’re also trying to predict the “honesty” of your fellow recommenders and how skeptical of a perfect score its unknown readers will be.

I despaired of doing both the right and the best thing by my student. Then I hit on a ruse. At the foot of the recommendation form was a space for “any additional comments.” So I took the plunge and, having answered every statistical question with “Not able to judge,” I filled in the final space with a page of hopefully nuanced evaluation that recalled for me what I trusted would impress others about this thoughtful, articulate, well-read, understated, appreciative, compassionate, responsible, good-humored, hard-to-describe human being.

And you know what? Seconds after pressing “send,” I was told that the recommendation had been “successfully submitted!” Better still, my student got the job — though she opted for another public service outfit instead.

Murray Biggs is a professor of theater studies. Contact him at murray.biggs@yale.edu.