LETTERS OF RECOMMENDATION FOR SURGICAL RESIDENCIES - WHAT THEY SAY AND WHAT THEY MEAN

Screening of applicants and selection of interns for surgical residenc y programs continue to rely heavily on letters of recommendation from surgical faculty. To define important aspects of letters useful for sc reening and selection, letters written on behalf of housestaff with kn own performance and those of new applicants were circulated to a sampl e of 120 academic surgeons. Eighty letters, written on behalf of 42 st udents, were divided into 10 groups of 8; each group was sent to 12 in dividuals. Half of the letters were ''open'' (full letterhead and sign ature), and half were ''text only'' (i.e., no identification). Candida te names were omitted. Each faculty reviewer was asked to rank the can didate represented by each letter, for their program, on a scale of 1 to 5, and indicate key negative or positive phrases on individual lett ers. Reviewers were also asked to indicate the importance they general ly ascribe in evaluating such letters to the school of origin, academi c rank of the letter writer, personalization (of letter to reader), fo rmal vs informal reference to the candidate, and mention of interest i n specialty training. A response rate of 60% was obtained. Faculty rev iewers stated explicitly that school of origin, personalization of the letter, and the academic rank of the writer are important factors inf luencing their ranking of candidate letters. The scores they gave to ' 'open'' and ''text only'' letters, however, do not differ significantl y (P > 0.05). Letters were scored similarly, indicating that school of origin or writer may not be as critical as implied or believed. Overa ll scores for letters of candidates selected for a single program over a 3-year period were virtually identical, differing significantly (P> ( 0.05) only from letters written on behalf of candidates ranked at th e bottom of the applicant pool. Interestingly, scores given to letters of mediocre candidates, those ranked somewhere between those assessed as potential interns and those ranked at the bottom of the applicant pool, did not differ significantly from those written for candidates s elected for the program, suggesting that other, less explicitly recogn ized factors are at work in the ranking process. The letters may be of a more complex nature than presently believed. The commonly used phra se, ''If I can provide any additional information, please call...,'' w as almost uniformly identified as a strong negative comment, while ref erences to fund of knowledge, motivation, work habits, and willingness to put in long hours were viewed very positively. This study suggests that for a document so critical to both departmental screening and se lection processes, the letter of recommendation has little specificity , and therefore little predictive power. Further study, focusing on de coding the content of recommendation letters, as well as correlating l etters with actual performance in residency, is indicated. (C) 1994 Ac ademic Press, Inc.