Principal Bias In Rating Teachers

Abstract

IN PREVIOUS paper Cole (l)has reported the findings of a five year study conducted to deter mine if material obtained from certain measures of personality was useful in predicting success in ele mentary and secondary teaching. The measure of success dealt with in that paper was the rating given each teacher by a classroom observer. It was noted in that paper that a rating was also given each teacher by the principal, but that such ratings were so severely skewed to the favorable end as to make the data unsuitable for analysis. The purpose of the present paper is to report on the results of an at tempt to evaluate the validity of this skewing, and then to evaluate the relationship between principal ratings and the prediction of teaching success which had been from the personality test data. At the time of the collection of the original data, each principal was interviewed and asked to place the teacher under discussion into one of five quintiles, ranging from the top to the bottom 20 per cent of other teachers of comparable experience, known to the principal. Upon the accumulation of the principal responses, it was apparent that the quintile placements were not only generally skewed toward the high end, but that this was particularly true in certain school districts which played an im portant role in the study. Teachers in over a dozen school districts were evaluated during the study. These districts varied greatly in size and in proximity to the training in stitution making the study. Analysis of the data it apparent that the skewing of principal rat ings was particularly evident in certain large key districts, located close to the training institution involved here. This is documented in Table I. District A is a large district immediately adja cent to the teacher training institution. District is also a large district, fairly close to the training school. All of the other districts partici pating are lumped in the third column. study of the predictions which had been for the teachers in the various districts, or their performance in practice teaching, gave no basis for believing that superior prospects had entered dis trict A, although there was some support for the be lief that those entering district B were more prom ising than average. There was no basis, either from principal interview material, from interviews with the teachers themselves, or from knowledge of curriculum planning and philosophy to believe that teachers entering certain districts had been tailor made in terms of training for that particular dis trict. Beyond this, moreover, it was apparent from a study of the qualitative remarks by the prin cipals during the interviews that in both districts and B there were often what appeared to be large discrepancies between what a principal said and the quintile ranking he eventually assigned. In certain instances the nature of the criticism were suchthat it would have been anticipated that a low quintile rat ing would be forthcoming, only to have it emerge otherwise as the interview progressed. The same person did ail of the interviewing throughout the study, and was not associated either with any school district, or with the depart ment where the teachers had been trained. The principals did understand, however, that the person represented the training institution in question.