Sorry Ian, I don't see how this pays off as an idea. if I say 40 times in a row that Ian Ayres is my best student ever, is it not possible that he just might have been? Am I missing something?

Here in the UK, where letters of recommendation are - it seems - now accessible to failed candidates for jobs, I am hearing increasingly of institutions/individuals in higher education limiting their observations on students applying for jobs to fixed details in the public domain (transcript data, when attended, GPA [or equivalent for us]) in case a failed student sues for a 'bad' letter of recommendation. Its an acute issue - how do you estore faith in private recommendations given in good faith if the threat of suit is there?

In a sense, this idea has already been implemented, to the bane of students nearly worldwide (though, ironically, not at YLS) - it's called a grade.

If you have in mind something more along the lines of "direct-to-the-employer" recommendations, then this is an interesting idea, but I'm not sure I see the ultimate benefit, and it might have some unwanted side effects.

Presumably, the group of students for whom a professor is writing recommendations will contain a disproportionate number of that professor's "Top X%" students. Thus, the fact that half of a particular professor's recommendations laud the target student as being top-knotch is not necessarily a reason to be skeptical. Furthermore, as a technical matter, there's nothing inconsistent with, throughout the years, labeling different students one's "best student ever."

As for downsides, I think such a system would foster unwarranted distrust of professors' recommendations in general. That would go doubly for professors who simply decided to keep their recommendations semi-private, a perfectly reasonable choice given that the underlying information concerns the student, not the professor. Somewhat along those lines, though I admit I am not an employment lawyer, I believe that employment recommendations are at least partially privileged under defamation law, and would worry about the effects of turning a one-to-one communication into a broadly-available, public statement.

A counterproposal: If we are concerned with quantitatively analyzing a professor's opinions of his or her students, why not make that professor's grading history publicly available? Thus, an employer (or graduate program) would not have to decipher an obscure phrase like "one of my top students," and instead would be able to see that the professor awarded the student one of only 5 A's in the past three semesters, in which the professor had 150 students.

I would not give a general recommendation for any student. Each recommendation would have to be tailored to the students strong points. If you are looking for ways to verify the ethical background of a professor, why not increase the scope and include everyone? Seriously. Between credit reports, bankruptcy, ...., there are many resources for creating a "Character Composite Score Card" for each person. So if you wanted to find out about the parents and the boy that wants to date your daughter, you can quickly pull up a general rating as to trust risk, criminal risk, credit risk, propensity to lie, automobile accident risk, health risks, ...