In reply to a more recent inquiry, 360 say that during the past year they have demanded additional educational and professional preparation.

One means of improving teachers has not received the attention it should-a salary schedule based upon qualifications and efficiency, rather than upon experience. Promotion on experience alone does not always reward the teachers who are making the most improvement, and does not always tend to call forth the best effort on the part of teachers. It may, however, help retain teachers in service for a greater number of years, and should, up to a certain point, be a factor in salary schedules. Advanced education and professional training are often not rewarded. It has been said that in formal salary schedules the premiums usually given are too low for education and too high, relatively, for experience in teaching. The tendency at present, however, is strongly toward schedules based upon experience, education, training, and success, with the teachers grouped into three or four classes.

Nearly all superintendents who are promoting largely on education and efficiency testify to the practical value of such schedules. The superintendent of schools at Owensboro, Ky., says:

One of the far-reaching movements of the past year was the organization of the system for the employment and promotion of teachers and fixing their salaries. The new schedule, which classifies teachers according to experience, preparation, and success, rather than upon experience alone, is having the effect of arousing teachers to the necessity of making professional preparation for the responsible work they have to do. Never before in the history of the schools have so many teachers attended summer schools, colleges, and universities. More teachers' periodicals are being read than ever before, and teachers are taking every advantage to improve themselves in the technique of their profession.

The following rules regarding the promotion of teachers on merit are typical:

Houghton, Mich.-Teachers are classified into A, B, and C grades. The superintendent ranks them, C below 75, B 75-90, A 90-100. Anyone who ranks C for the period of one year is not retained in the service; teachers ranking A and B receive the regular increase until the first maximum is reached. They remain at this salary for one year, during which, if their grade is A, they are permitted to participate in an annual increase until the final maximum is reached.

Ottumwa, Iowa.-Teachers are divided into three groups, A, B, and C, on the basis of their efficiency. Those ranking A receive $5 a month increase each year until they reach the maximum of their grade. Those ranking B receive an advance of $2.50 a month increase each year until the maximum of their grade is reached. Those ranking C may be elected without an increase of salary.

Astoria, Oreg.-A teacher may be given a double or triple increase at the end of the year upon recommendation of the superintendent.

Freeport, Long Island, N. Y.-The matter of salary increase is based upon merit and efficiency. Teachers are formally rated by the superintendent twice each year. Cripple Creek, Colo.-The salary schedule provides for progressive classification for each succeeding year until the maximum is reached. A teacher may advance one more class each year or she may remain more than one year in the same class. A

teacher may also be demoted to a lower class if her work does not show continued interest and improvement.

Bristol, Va.-After a teacher has entered the system she is sure of an annual advancement until a maximum is reached, provided she makes a success grade of 90. All teachers must attend a summer school at least once every five years.

Beaver Falls, Pa.-No teacher is advanced in salary after she has ceased to advance in efficiency and successful experience; and fitness for increase is based upon an efficiency record kept by the superintendent. A definite statement showing wherein a teacher fails to meet the requirements is furnished to each teacher who is not given the increase in salary provided by the schedule.

Dayton, Ky.-Teachers are awarded an increase in salary for attendance at summer school.

Moline, Ill.-Teachers drawing a schedule salary for the third year are granted $5 a month increase upon presentation of satisfactory evidence of summer-school work or extended travel.

MEASURING THE EFFICIENCY OF A TEACHER OR A SCHOQL.

A few years ago, if a superintendent of schools had been asked how he measured the efficiency of his schools or of his teachers, he would have answered only in the vaguest of terms, usually saying that the eighth-grade pupils had been well prepared for high school and that the high-school pupils in college were doing creditable work. To-day many of the superintendents in the smaller cities are attempting to judge the efficiency of their schools by other and more reliable standards. Although only a beginning has been made, the results as reported are uniformly encouraging.

One of the ways of judging the work of a school is by means of standardized tests, such as the Courtis tests in arithmetic, the Thorndike and Ayres test for measuring the quality of handwriting, the Hillegas and Rice scales for measuring English composition. Of 756 cities reporting, 389 are making use of standardized tests for testing teaching efficiency. The following extract from a letter by D. C. Bliss, superintendent of schools at Montclair, N. J., is typical of many received at this bureau regarding the measuring of efficiency:

It may be stated, with little fear of contradiction, that no system of testing which can be devised can ever take the place of the superintendent's personal and frequent visits to the classrooms of the teachers under his care. The most elastic test becomes arbitrary unless there is some knowledge on the part of the examiner of the difficulties with which the teacher is contending.

* *

*

To support the superintendent's personal opinion, however, there may well be documentary evidence of a teacher's success or failure. This comes wholly from the progress of the class, for the demand is not for a knowledge of psychological theory, but for the power to achieve results. At the beginning of the year the class should be tested in one or two fundamental subjects, such as English or practical arithmetic, with the purpose of determining its measure of attainment at the time the new teacher for the year assumes charge. This gives a standard for comparison. Whether the class improves upon this standard or falls below it as the months go on, the responsibility for the success or failure belongs to its present teacher. It is not necessary to test in all subjects. It is generally safe to assume that a teacher who is doing good, solid work in the few subjects most needed by pupils who are going out into a world of

practical living is the kind of teacher who carries her faithfulness and ability into all departments of her work. Unless the superintendent sees from his visits to the room that things are going radically wrong, a class need not be tested frequently. Ordinarily, a test at the beginning of the year, one at the end, and one at some time during the intervening months will be sufficient.

Testing for efficiency does not imply a long, formal examination of the old-fashioned type, with the results marked in per cents and a teacher's ability estimated by the place her class attains in the scale for the whole city. With the varying personnel and environment of the different classes, such an estimate is manifestly grossly unfair to the teacher. The fairer way is to use the standardization methods employed by Dr. J. M. Rice, Dr. Thorndike, or Mr. Courtis, establishing from the results obtained all over the city an average standard for that city, to the attainment of which a class, and through the class the teacher, may rightly be held. In this way the teacher's work is judged on its own merits above or below the established standard, not by comparison with the work of another teacher in a school in which the conditions may be many times easier than those she has to contend with.

In one instance, at least, a grammar-grade supervisor has worked out a standard test in English from the composition of the pupils under her supervision. Miss Gertrude Earhart, supervisor of grammar grades at Boise, Idaho, who has formulated such a scale, says:

In an attempt to standardize the work in composition in the grammar grades, the "Hillegas scale for measuring composition" was first used. As the use of any language scale was new to us, we were somewhat inept in evaluating the work of our pupils, particularly the younger grades, according to this scale. In the hope of helping the teachers to something more definite, a language scale from the compositions of our own pupils was made. A composition from each pupil in the fifth, sixth, seventh, and eighth grades of our schools was read and placed in one of several groups. After all the papers were classified in groups for each grade, the groups for the sixth grade were ranked beside the fifth-grade groups. As would be expected, the lowest group of sixth-grade papers was better than the lowest group of fifth-grade papers, and the best group of sixth-grade papers ranked above the best fifth-grade papers. The seventh-grade papers were ranked beside the sixth, and the eighth beside the seventh in the same manner, each grade ranking higher at both the lower and upper end of the scale than the grade a year behind it. The distribution of compositions for each year followed the normal curve of distribution.

This attempt to make a continuous scale which would cover work from the beginning of the fifth to the end of the eighth year was unsatisfactory, however, as it was hard to balance so much skill in form, punctuation, etc., in a fifth-grade reproduction exercise, as worth so much thought, facility in expression, etc., in an original eighthgrade exercise, so the idea of using a continuous scale for the four grades was abandoned and a scale of six or seven groups was made for each grade. The lowest group for each grade was numbered 1, as we did not expect ever to have any lower classes; the second group, 2; and so on, so as to leave room for indefinite expansion at the upper end of the scale. These scales were then typed and put in the hands of each teacher, and she graded the papers of her class according to the scale. Once a month papers graded by the scale were collected, the results tabulated by building and by grade, and the tabulation given to the teachers. These results showed a marked increase from month to month in the number of papers in the upper grades of the scale. This improvement was doubtless due in a great measure to clearer ideas in the teacher's mind of what could be expected from her grade.

That much of the work was accomplished last year, and this year we have gone a step further and have attempted to have the pupils of the sixth, seventh, and eighth

grades evaluate their own work according to the scale before it is handed to the teacher. The teacher then examines the paper and places her grade below that of the pupil. The pupils, contrary to all prediction, have shown very fair judgment in grading their own papers and have made great efforts to raise their grades.

While the scales we use are open to criticism at numerous points and are unsatisfactory to us in certain respects, yet we feel that by their use we have been able to do more definite work in improving the quality of composition written by our pupils, and we have been able to differentiate the kind of work done in the grades, so that each grade has some definite things to accomplish.

Practical schoolmen are, however, pretty well agreed that standardized tests, valuable as they are, are but one means of testing teaching and a poor substitute for the observance of actual classroom work. Only recently have any attempts been made to set up welldefined standards for judging a teacher's work, and as yet there is no agreement among schoolmen as to the points upon which a teacher should be graded or the relative value of these points. Among the standards that superintendents are using are those formulated by Dr. Frank McMurry in his "Report on the quality of classroom instruction in the New York City schools." In the Ohio State survey the bureau of municipal research used a card containing points on which teachers were judged. As this card is in use among some of the smaller schools, it is printed on the following page.

1 See, for a critical view of these standards, the "Reply of the district superintendents of New York to the report of the Hanus inquiry," edited by J. S. Taylor, New York, 1914.