Committee Selection by Statistical Sampling

Abstract

Statistical ideas are sometimes useful in applications ordinarily regarded as nonstatistical. In this note I suggest a statistical sampling procedure for the selection of a committee when it is desired that the committee somehow be of a larger group and when it is natural to be suspicious that the appointing authority might consciously or unconsciously slant the composition of the committee. The procedure incorporates both judgmental and random elements and has the incidental merit of assuring equal probability of selection to each member of the larger group, although it is a probability sampling design only in the weak sense of, say, a systematic sample with a random starting point in the first interval. The idea arose out of recent events at the University of Chicago. In the backwash of several student disruptions during the last four years, there have been two studies of university disciplinary procedures. Recommendations made as a result of the second study are being considered by the academic governing body of the university at the time of this writing (February, 1970). In both of these studies consideration was given to the use of some form of statistical sampling in the selection of members of disciplinary committees that would be assigned to judge participants (or alleged participants) in possible future disruptions. I was involved in the first of these studies and am now involved in the review of the second. As a result I have had the opportunity both to evaluate the case for random selection and to propose a more elaborate selection scheme that may be of statistical and practical interest. The purpose of this note is to expose this scheme to technical criticisms and suggestions for improvement and to consideration for possible use in other applications. I would be especially interested to know if any such scheme has been tried and, if so, what the experience has been. In candor I must report that I am reasonably confident that the proposal will not be adopted at the University of Chicago. The reason for rejection, if this forecast is correct, will probably have nothing to do with reservations about the technical feasibility of the proposal for achieving a reasonably committee but rather will stem from doubts that a representative committee is really desirable. Many feel, for example, that only people with a certain combination of sensitivity, judicial temperament, tough-mindedness, compassion, coolness under fire, and the like, should be considered for a job that is arduous, frustrating, and even hazardous. In this note, however, I will simply assume that is desired. The proposal therefore may be of interest whenever this goal is of prime importance, whether for a disciplinary committee or some other kind of committee, whether in an educational institution or some other kind of organization. For simplici;ty and concreteness I shall convey the suggestion in termns of the following example, which differs in some respects from the local situation at Chicago that originally gave rise to it. There should be no difficulty in adapting the approach to different circumstances. Suppose that it has been decided that a disciplinary committee should be made up entirely of faculty members, that it is desired that the committee be in some sense representative of the views on discipline for disruptions held by the faculty at large, and that the achievement of this representativeness appear credible to all groupsfaculty, administration, students, and especially the students being judged. (If the disciplinary committee were to include students, the method of student selection would be an obvious adaptation of the method of faculty selection to be described.)