Signal Extraction Problems: Recommendation Letters
| Nicolai Foss |
Some kinds of recommendation letters need careful interpretation. A letter written for a student to help him or her study abroad usually doesn’t need much interpretation. But a letter written by a colleague for a colleague to a colleague is a different matter. One reason is that writers of recommendation letters differ. Some express themselves very directly, others more indirectly. The same words mean different things to different people. “Solid research” may mean “boring and unimaginative” to one person, but may mean, well, “solid” to another person.
Another reason is that recommendation letters can be used strategically, such as to get rid of unwanted persons. I confess that on one occassion I have written a glowing recommendation letter for somebody that I thought my School would be best served by getting rid of. The person in question was very smart, but obnoxious. What I did in the letter was to strongly praise the person’s scholarly qualities, but I didn’t mention anything about personal qualities (which was no doubt to the benefit of the relevant person). That way, the letter wasn’t dishonest (I think).
The highest praise that academics give each other is something like, “(S)he is very smart — and nice, too.” Is this what we should look for, that is, we should be suspicious if the “nice” part is left out? What other indications are there? How does the O&M readership handle the recommendation letter signal extraction problem?