Writing Letters of Recommendation

Are you the right recommender? When the person wanting to be recommended contacts you, do the following:

• Discuss whether you would be an appropriate person to supply the recommendation. You should do this in as neutral but friendly way as possible. See items "No" and "Yes" for tactics if you think those labels are the answers to the question.
   
  

  No. Often a request for a recommendation is made without considering if the individual's letter would be effective. A common mistake is asking someone too junior: assistant professors should not make recommendations for tenure; usually theorists should not recommend experimentalists, and vice versa; the recommender should be known to at least one of the persons who will be making a judgement based on the recommendation.
   
  If you don't qualify as an effective referee, explain why and counsel the individual in finding effective references.
   
  If you can't write a positive letter? Try to escape being a references. The discussion topics in the paragraph above may provide a way out. Avoid telling the individual that you can't write a positive letter. After all, your negative opinion may be based on incomplete information. No one needs to be turned down roughly.
   
  Do not promise you will write a letter with no intention of doing so. This will only hurt both you and the individual in question.
   
  

  Yes. To assure your letter will be effective, find out who will read your letter and read bullet below:

  
  
• Secure information you need to write an effective letter. Often the official request for the letter (for example, in the case of a tenure decision) will not provide all the information you need. Ask the individual to send you the follow:
   
  Vita. This should include a complete education and employment history; concise descriptions of research achievement and plans; any awards, honors or fellowship; if relevant, a description of teaching and course development; if relevant, any service on college or national committees; and other contributions such as editing journals.
   
  List of publications. This should include complete titles, full list of authors of each paper, and should distinguish between refereed and unrefereed publications (i.e., most conferences papers).
   
  Publications. Have the individual send you some current papers and, in the case of tenure and award letters, papers of which the individual is especially proud.

Who will read the letter? A tenure letter will be read by every member of the department - those who know you and, more importantly, those who don't. You must structure the letter so that it makes arguments that anyone, even a dean or provost, can understand. It should have the four items discussed below.

1. Introduction. Explain in what way(s) you know the individual. Use effective `short stories' that bring individual alive for readers of the letter.
2. Answer the questions! If the letters asked specific questions, answer them if you can. The trickiest question to answer is to to rank the individual against others of similar qualifications. The best institutions will give you a specific list. This question also tests you; for example, if you don't know some of the persons on the list it implies you are not an expert. You may need to do homework on the other names. Generally a safe form of answer is one that explains the strong points of the various named individuals.
3. Evidence for your arguments. Don't say the individual's work is outstanding; cite specific achievements in direct language, understandable at the provost level.
4. Conclusion. Finally summarize the positive arguments for recommending the candidate for an assistant professorship, tenure or an award.

Beware weak endings. End the letter with the conclusion above. Avoid writing some sentence at the end that mentions even a potential weakness. It will be read as a warning that the real message of the letter is: don't appoint, don't tenure or don't award.

Your letter is probably not confidential. Assume your letter will be read by the person being recommended. Accordingly, simply omit discussion of weak points; the omissions will be noted. For the dangers of letter writing, see Why you can't trust letters of recommendations.