TABLE 6.-BLEWETT'S GRADUATES IN SOLDAN

HOW DO BLEWETT PUPILS GET ALONG IN THE SENIOR HIGH SCHOOL? Here is Soldan's June, 1921, Graduating Class:!

Of the pupils who continue through Soldan, the last table indicates how successful our rapid-progress A and B-pupils are. Few C-pupils complete senior high school; perhaps none in normal time, as all of these pupils have done. It should be noted from the table that the top 17 pupils were all Blewett pupils, the Blewett median score was 82 per cent.; the non-Blewett median 76 per cent. Only three Blewett pupils averaged below 73 per cent. Of the 137 Blewetts, 107 or 78 per cent. graduated in the upper two-thirds of the class; of the 52 nonBlewetts, 19 or 38 per cent. graduated in the upper two-thirds.

It may be only an opinion, but I am quite satisfied that homogeneous grouping is one very important factor in this success.

INITIAL GROUPING OF PUPILS

CHESTER ROBBINS, Principal of High School, Bridgeton, N. J.

It is the purpose of this paper to discuss methods of providing for the initial grouping of pupils by ability. It is assured that pupils are

to be grouped according to their presumed ability to do the academic work of the school. Search will be made for the plan which will predict most accurately the educational achievement of pupils.

The following four bases of grouping will be discussed: (1) marks of pupils in elementary school; (2) achievement tests; (3) mental tests; (4) a combination of the first three. School publications are publishing results of investigations and experiences of schoolmen with these various plans. Some of these results will be used in this paper.

The use of the pupil's past school record is based upon the assumption that the future achievement of a pupil can be predicted by his past performance and that teachers' marks are a fair measure of his performance.

According to Freeman (School Review, Vol. 29, Page 742) investigations have shown that there is a fairly close correlation between the pupil's relative standing in one school or one grade and his relative standing in the succeeding school or grade. He cites results of investigations which show that 50 to 75 per cent. maintain approximately the same relative standing as they advance from one school to another. According to Briggs (Junior High School, Page 148), Fretwell has shown that marks of pupils in grades 5 and 6 are good for prognosis of ability to do junior high school work, that marks for grades 1 to 6 are better, and that selected tests are best.

There are some objections, however, to the exclusive use of marks. A junior high school may recruit its members from schools of varying standards and types of children. In this case school marks alone would be unsatisfactory. Studies have been made of the variability of teachers' marks which throw considerable doubt upon their accuracy in measuring the actual achievement of children. Marks are the personal opinion of the teacher indicating what she thinks the pupils have achieved. While the opinion of the teacher who has had an opportunity to observe a pupil and learn to know him should not be discarded, objective tests of achievement are needed as a check upon her judgment.

The validity of the use of achievement tests for prognosis rests upon the assumption that they measure achievement and what a pupil has achieved is prophetic of what he will accomplish. Superintendent D. C. Bliss, (Journal of National Education Association, April, 1922, Page 15) after four years of experience with this plan and after careful checking of scholastic records with prognostic records, states "Our experience with tests for prognostic purposes proves them to be most valuable in indicating the placement of pupils.”

Breed and Breslich (Intelligence Tests and Classification of Children, School Review, Vol. 30, Page 223), however, conclude that "measures of arithmetical ability did not constitute a very satisfactory index

of the later success of these pupils in mathematics." It must also be remembered that these tests do not measure all the abilities which lead to achievement. They measure specific abilities. As Freeman points out (Bases on Which Students Can Be Classified Effectively, School Review, Vol. 29, Page 744). "They do measure fairly accurately ability of the academic type under the conditions of a spurt in which speed is an important factor. They fail to measure, in anything like the same degree, the pupil's staying power, his ability to organize large masses of material, and his independence and initiative. They often do some injustice to the comparatively slow, but accurate, and perhaps profound student. On quite another ground, they sometimes do injustice to the nervous student who is easily flustered under test conditions. Pupils who do not make an especially high score in the test, because of such factors as these may do much better work in the course of their school work."

It must also be remembered that these tests were designed to measure and compare the progress of groups of pupils. This they do reasonably accurately. But in case of individuals they do not measure accurately. Therefore it is not wise to judge on the basis of achievement tests alone any more than upon teachers' marks exclusively.

Teachers' marks are a measure of what the teacher thinks the pupil has accomplished, and achievement tests are a more objective measure of the same thing. Both deal with the past. Intelligence tests indicate to us what the pupil should do. They help in the search for those pupils who should do better than they have done. It is just this type of pupil whom grouping by abilities should stimulate to do better work. Teachers' marks and objective achievement tests might not have revealed to us the possibilities of this type of pupil.

Group tests are now being used for the classification of pupils according to ability. Economy of time, effort, and expense is an essential feature of any method of providing for initial grouping in many schools which are limited by lack of funds to buy tests and clerical help to score tests and make elaborate and lengthy computations. If a test can be devised which in thirty-five minutes can give us the information needed to classify pupils accurately, this criterion will be satisfied.

Breed and Breslich found in their study of Intelligence Tests and Classifications of Pupils (School Review, Vol. 30, Page 223), that the "Otis test actually classified the pupils as well as either of the other means employed," The Otis Group Intelligence Tests, Terman Vocabulary, and Chicago Reasoning Tests were used to classify pupils in the Washington Junior High School, Rochester, N. Y. Superintendent Brooks (Some Uses for Intelligence Tests, Journal of Education Research, Vol. 5, Page 238) states that although he had used achievement

tests for grading, "If I had any more preliminary grading to do, I should do it with the group intelligence tests supplemented by the BinetSimon individual test wherever there was any doubt of the accuracy of the group tests." Justification of the exclusive use of intelligence tests for classification of pupils depends upon (1) how accurately they measure intelligence and (2) how accurately they predict educational achievement. Intelligence is not the only factor which makes for scholarship. Interest, industry, perseverance, will, and health all affect his school standing. It must also be remembered that group tests measure groups fairly accurately, but are unfair in individual cases. Breed and Breslich (Intelligence Tests and Classification of Pupils) conclude that the intelligence tests failed to classify according to intelligence from onefifth to one-sixth of the pupils. They further found that pupil displacement for the ninth grade, when divided into three classes, according to the tests, was 51 per cent. The claim they make for the intelligence tests is that they provide a preliminary classification more economically than any other means tried and otherwise as satisfactorily.

Great caution must be observed in drawing too hasty conclusions from a single test. A pupil, classified in a poor section by one test, might be classified in a good section by another test. Doubtful cases wherever possible should be given several group tests or better an individual test. There is considerable reaction against testing because of its abuse in the hands of over-enthusiastic and sometimes untrained friends. The movement which promises much must be saved from falling into disrepute because of its well-meaning but overzealous friends.

A preliminary classification by any one of the previous plans would probably be better than classification by the alphabet. However, the information given to us by all the methods should be utilized. The correct placement of every pupil should present itself to us as a problem. In the solution of this problem we should have the attitude of the social case worker. All available knowledge should be sought and the decision made upon the basis of that knowledge for the best interest of the individual.

The justification of using several judgments rather than one judgment is that the composite of several expert opinions is more likely to be true than one opinion. Teachers' marks and test marks should be used as checks upon each other.

At the meeting of the National Association of High School Principals it was recommended that a group intelligence test plus a silent reading test plus an English composition test be added to the teacher's subject tests in classifying high school pupils.

It may be argued against the plan of a composite ranking that it requires too much time, effort and expense. The writer has faith that

the results will be the best justification of the time, effort and money expended.

It cannot be too strongly urged that all the above plans provide only for a tentative grouping. The highest percentage of prediction of achievement by school records was 75 per cent.; coefficients of correlation between intelligence tests and achievement tests range between .40 and .60; intelligence tests show an error in classification as high as 50 per cent.; and different intelligence tests do not give the same relative ranking to pupils. Even the composite ranking is not perfect. The pupil's best performance rather than his central tendency may be the best index of what he can do. Until we can more accurately predict, we cannot expect to have a permanent classification.

Often school officials take pride in the few changes in classification which are necessary to be made. This may be a good or bad indication of the success of their plan. It may mean that the teachers have succumbed to routine and are no longer studying each individual. It must not be forgotten that grouping by abilities is only one phase of the movement to provide for individual differences. If teachers fail to study their individual pupils and pupils lose the stimulus of the possibility of improving their classification, grouping by ability loses its spirit and becomes a mere form.

VARIATION IN TREATMENT OF DIFFERENT GROUPS J. J. BREHM, Principal of Camp Curtin Junior High School, Harrisburg, Pa. During the year 1920-1921, Camp Curtin Junior High School of Harrisburg made some attempt at the homogeneous grouping of pupils. At the beginning of the second semester of this year pupils of the eighth grade were regrouped, in so far as their course elections would permit, on the basis of work done during the first semester. To be more exact a certain average was agreed upon up to which pupils must measure in order to be eligible to the two select groups, one academic and one commercial, for which it was considered possible to arrange, in view of the number of pupils in each of these courses from which selections could be made.

After a few weeks' trial it was decided to give these two groups an opportunity to do two years' work in one and a half year's time. Within the first six weeks five were again transferred to more slowly moving classes; the remainder, excepting a few losses incident to moving out of the district, have stayed together, and by the sixteenth of June this year will have creditably completed two years' work in three semesters.

And this attempt at homogeneous grouping has been of advantage not only to the bright pupil but to the slow as well. Plans of teachers