A study of present practice in that important part of school administration having to do with the measurement of teaching efficiency reveals the greatest diversity of method. The study also discloses a surprising number of large cities having no method at all in determining the relative specific efficiency of their teachers.

Discussions of the "ideal teacher" are numerous in educational titerature. Every educational essayist has described that mythical creature in glowing terms. The pages of the reports of the National Education Association are full of her virtues. There is much of inspiration and suggestion in these discussions but little or nothing of scientific value, little of which a superintendent could make practical use. Each writer or speaker has described the qualities as they appeal to him, usually emphasizing one or two virtues to the exclusion of others. The great bulk of the writing on the subject of teaching merit has been of the subjective, a priori type without any basis beyond the theory and experience of the one who was writing.

Objective measurement of educational products and processes has not yet reached the point where we can rely upon such measurements to give us all or even most of the information we need about our teachers. We are still dependent, to a very great extent, on the judgment of supervisors and others whose business it is to have this information. This does not mean that we are to be without exactness in our judgment of teaching efficiency, or that we have to belong to the ranks of those who judge by "general impression."

Any conclusions to which we may come later in regard to the relative importance of the various qualities of merit and in regard to other problems involved are based on the experience and work of between 40 and 50 school men and women who have tried out and criticized for us the scheme just described. In the correlations to be discussed later, ratings of 424 teachers from 39 schools were used. These 39 schools were representative of 27 cities, all but 8 of which are in Illinois. They represent populations varying in size from less than four thousand to two million.

The rating device which wehave described in the preceding chapter was designed primarily to obtain for us information on which to base an accurate estimate of the value of the qualities of teaching merit there set forth. Our idea has been to make it as easy as possible for school officials to give us correct information. The rating officer is not worried by any schedule of values or what he thinks is the relative importance of the qualities under consideration. He has before him in each case an actual teacher whose various qualities he is judging in terms of Excellent, Good, Medium, Poor, and Very Poor, with variations.

The following points are apparently most important as resulting from our study of the problem up to this point.