Recommendation Inflation

Brian Taylor

April 02, 2017

I am always suspicious of claims that "things used to be easier" in higher education, or "more straightforward" in the past. In reading C.P. Snow’s novels about academe in England, I’ve learned that intrigue has regularly hovered over coveted faculty appointments. That said, things certainly used to be different when it comes to writing letters of recommendation.

As a veteran faculty member, I am often asked to write such letters, and the most challenging ones to draft are for young scholars applying for tenure-track jobs. In an academic job market that remains highly competitive — with dozens or even hundreds of qualified candidates for each coveted position — I (as well as other senior professors) am not infrequently asked to write letters for more than one candidate for the same job.

In the first half of the 20th century, nearly all faculty appointments at selective institutions (in the United States, the United Kingdom, and other countries) came about through personal recommendations — in writing, in person, or by phone. The operating principle was the "old boys’ network" — and literally boys, since almost no "girls" were part of the network.

Letters of recommendation are not a place where one can expect candid professional judgments.
When there was an opening at an institution, or when a senior scholar had a promising student, relevant old boys would get in touch with one another and have a presumably frank discussion of a particular candidate’s strengths and weaknesses. Having read some correspondence from that era, I have been impressed by how candid the letters were — criticisms were at least as prominent as raves. In that sense, one can say recommendation letters in that period were truthful.

But it’s equally important to point out that scholars back then had their favorites. Having a candidate who (a) carried on your work, (b) agreed with your view of the field, and/or (c) was personally helpful to you, were undoubtedly fingers on the scale of a positive recommendation.

Certainly the old boys’ network needed to be exploded, and in the last several decades, it clearly has been. While sexism has hardly disappeared, the range and variety of candidates is much greater today, with female and nonwhite applicants at least in the pool, even when there has not been special encouragement for their candidacy. Jobs must be publicly advertised now. Further, in some places there are "sunshine" rules, such that either recommendation letters are made public or — more typically — letter writers are warned that confidentiality cannot be guaranteed.

With all of those changes have come efforts to ensure the fairness of recommendation letters — well-intentioned efforts that, unfortunately, seem to have decreased the usefulness of the letters.

One ploy — common for admission to highly competitive programs — is to ask a letter writer to rate the candidate on a percentile scale with respect to various traits like originality. An example: "In oral expression, as compared to other candidates, is this candidate in the upper 1 percent, the upper 5 percent, the upper 10 percent, etc.?"

Another ploy: Ask a letter writer to compare the candidate to other candidates (in the same cohort) who are explicitly named or who are chosen by the recommender.

Alas, such tactics are rarely effective. The quest for fairness and objectivity leads to "letter inflation." We are all familiar with grade inflation — the tendency over the decades to give students ever-higher grades (in many institutions of higher learning, yesterday’s C is today’s A-). With respect to letters, I’ve observed the same trend in the United States — recommendation letters often compete with one another for superlatives, with the result that neither absolute nor relative judgments can be rendered. Similarly the rank-ordering question rarely results in a low ranking for any candidate. Respondents either rank the candidate highly or claim they don’t know the other people named well enough to make an informed comparison.

Indeed, of the many letter writers whom I know personally or "on paper," only one of them is relatively candid about the flaws in a candidate. The rest seem to err on the side of excessive praise, or what I call "recommendation inflation."

In the light of all of those obstacles, what is left, if anything, of professional judgment? Faced with other letters that are likely to be laden with superlatives, as well as the prospect of public exposure of critical remarks (not to mention the possibility of a lawsuit filed by an unsuccessful job candidate), are there any principles to which letter writers should adhere in order to convey their assessment of a candidate in a reliable way? Here is what I would recommend:

  • When asked by a job candidate for a letter of recommendation, be prepared to say no, and to give reasons that are candid — although not, of course, gratuitously nasty. When I decline to write a letter, I often explain that I don’t know the candidate well enough to be helpful, that I have already agreed to write for someone else, or that I don’t think that the candidate is appropriate for the job. Better to be tough at the beginning than to find yourself in a quagmire.
  • Refuse to do rank orderings or checklists. Here’s the standard boilerplate that I use: "As a matter of personal policy, I do not complete ratings questionnaires as a portion of recommendations." Why that refusal? One almost never sees checklists that are not completely skewed to the positive — so much so that checking off "top 10 percent," rather than "top 1 percent," can be the de facto kiss of death for an applicant.
  • Be purely descriptive whenever possible. For example, when it comes to a discussion of the candidate’s research, put it in your own words and be explicit about the contributions of the applicant’s work as well as its limitations. Paradoxically, such candor may increase the credibility of the recommendation.
  • State in a clear, positive way the candidate’s strongest features. Search committees will be interested in how you see the applicant’s strengths. If possible, touch on the candidate’s less-strong features — or indicate areas on which you don’t feel competent to comment (for example, if you know the candidate’s research but not that person’s teaching, it is fine to state that).
  • If, for whatever reason, you cannot be explicit about a candidate’s weaknesses, be silent. Leave it to the readers of the letter to make inferences about what is not discussed. To avoid unintentionally harming the candidate, I always have a last line that reads, "Please let me know if I can provide any further information." If there is indeed a follow-up, you are free to say, "I am not comfortable commenting on that issue." Don’t lie.

As you can probably tell, this state of affairs in recommendation letters does not please me. I’d much rather be completely candid and have others be equally candid with me. (In that sense, I have sympathy for the normative behavior in earlier times, despite its obvious flaws.)

But that is not the world in which we live. If you’re writing measured recommendations and everyone else is writing superlative ones, that is unfair to your candidates as they are being treated in a way that jeopardizes their chances for a livelihood.

I come to the reluctant conclusion that, at least in the United States, letters of recommendation are not a place where one can expect candid professional judgments. Since these issues will not go away anytime soon, I’d be eager to hear others’ ideas about letters of recommendation, and how to exercise one’s professional judgment in a responsible way.

Howard Gardner is a professor of cognition and education at Harvard University.