Can we do a better job in selecting candidates who will complete their psychoanalytic training? In 1990 the Admissions Committee of the Houston-Galveston Psychoanalytic Institute (HGPI) decided to take a systematic look at its experiences as a young institute.
There was nothing new in that interest; institutes have been trying to pick candidates who will succeed ever since analytic training was formalized. However, the HGPI undertook a methodical review of admission procedures and data to see what it could discover. At monthly meetings, members of the Admissions Committee reported in detail the recorded data on candidates who had been selected for admission from 1973, when the institute began its training program under the auspices of the New Orleans Psychoanalytic Institute, until 1984, a date selected arbitrarily. We developed lively discussions around the records, wondering whether we could have been more accurate in predicting the course of the various candidates by emphasizing one or another aspect of the candidate's autobiography, educational history, or interview behavior.
We reviewed all the available published papers pertaining to candidate selection and discovered that they were largely anecdotal. The usual approach was, “What qualities make for a good psychoanalyst?” That was not our focus. We wanted to concentrate on whether we could better predict success in training at the time of the admission evaluations, accepting the premise that we were in general agreement as to the desirable qualities of our prospective candidates.
We began the study of the thirty-one candidates selected for training from 1973 through 1984. Sixteen completed their training; fifteen did not, for a variety of reasons. Our committee studied data collected at the time of screening for selection from various viewpoints: educational background, reference letters, autobiographical essays, reports of three interviewing psychoanalysts, and the recorded discussions of the Admissions Committee. To help us organize these data, we developed a rating profile on a five-point scale that included all these variables. Since this was a retrospective study, we had to evaluate records that were compiled by a variety of analysts who did not have research in mind and who recorded data they felt to be important. Consequently, the records were by no means uniform. In this study we did not scrutinize the records of applicants not accepted for training.
In the first phase of our study — having previously made serious efforts to standardize our rating procedures — we asked the seven members of the committee to complete a rating profile on each candidate. To our delight, we found we could readily differentiate between the candidates who completed training and those who did not. We had a variety of formulations about what the autobiographies revealed and how to read the usual unhelpful reference letters. We realized, however, that we were a very biased group, in that many of us knew the candidates we were rating — even without seeing the names on the records.
We reluctantly faced the need for a more objective evaluation of those records by a panel of analysts experienced in selection procedures and not affiliated with our institute. We were fortunate in gaining the willing participation of three experienced training analysts from three different institutes: Cornelis Heijn of PINE, Edward Schiff of the Cleveland Institute, and Eugene Trunnell of the St. Louis Institute. We sent them a series of identical records and met with them to help them develop a uniform method of rating the twenty-four variables on our worksheet.
When we and they were satisfied that they rated in a reasonably reliable manner, we sent each of them copies of the thirty-one candidate admission records. We removed any identifying data from each of the records — names, place of birth, schools, residency programs, and names of those writing reference letters. (All three raters wanted all thirty-one files at once. They returned their ratings to us within a couple of months. A true labor of love!)
With the help of Gerald Adams, an educational psychologist experienced in statistical evaluation of scientific research data, we examined what our outside psychoanalytic educators had reported. The data were subjected to a variety of sophisticated statistical analyses too complex to report here. We have learned that most analysts did not understand or appreciate these numerical data but wanted the results in words. This is what we shall present, recognizing that we do have the data to lend support to our conclusions.
We are aware we are dealing with the evaluation of a small number of candidates accepted for training. There are specific ways of dealing with small samples statistically, and we have tried to take that into account. We are also aware that these data are from our institute only. However, we asked other institutes if they had undertaken a similar study, and none had, though all expressed an interest in learning of our findings.
The three outside raters could indeed distinguish between the two groups of candidates using our rating scale. Using all the variables, they predicted correctly 76 percent of the time.
We were disappointed to discover that the ratings of the written autobiographies of the candidates did not vary between the two groups. Our original hypothesis was that the candidates who described a richer emotional life, close family relationships, friends, etc., would be the more successful. Data from reference letters also washed out.
The single collection of variables that was most helpful, with some 65 percent accuracy, was the raters' evaluation of the three interviewers' reports done at the time of the admission interviews. Interviewers reported good contact with the candidate, wrote a good description of the experience, regarded the applicant as analyzable, and recommended acceptance of the applicant; in each case, the recommendation of acceptance followed logically from the body of the interviewer's write-up.
In summary, our data substantiate objectively what most psychoanalytic educators have believed all along. Interviews done and reported carefully and thoughtfully by three psychoanalysts at the time of application are the best predictors of the applicants' completing psychoanalytic training.
In reviewing our work, we decided to modify some of the instructions to applicants. In the request for written autobiographies, the applicants should be instructed to include a description of individuals who have been important in their lives and an assessment of the impact of these relationships. We found that some applicants, even quite introspective ones, did not believe we wanted a truly personal story.
Requests for letters of reference should ask for comments on areas of particular interest, strengths, and problems of the candidate. In reporting their findings and opinions, interviewers should be asked to guard against global positive judgments for particularly promising candidates.
Anyone interested in learning more about this research should feel free to write Edwin C. Wood at 5300 San Jacinto, Suite 150, Houston, TX 77004-6886 or Betsy Comstock at Route 2, Box 39, Penial Road, Troyon, NC 28782.
Edwin C. Wood, M.D., is a training and supervising analyst for the Houston-Galveston Psychoanalytic Institute and chairs its Admissions Committee. He is a clinical professor of psychiatry at the Baylor College of Medicine in Houston, Tex. Betsy Comstock, Ph.D., is retired from Baylor, where she was a professor of clinical psychiatry, and from HGPI, where she was on the faculty.