function that was tested with a view of detecting transfer, was the power to obtain high marks in portions of the Institute of Educational Research Intelligence Test. The children were given the test at the beginning and at the end of the school year.

Naturally they improved their scores on the second occasion as compared with the first. The question was whether some of the school subjects had improved their ability more than others.

From all the children acting as subjects two groups were selected who were taking the same school subjects, except that one took, say, Latin where the other took, for example, cookery. Only one subject was to be different in the two groups. In this way differences, if any, between the two groups may be supposed to be caused by the one subject which differs.

Suitable precautions were taken to insure carrying out the experiment as scientifically as possible. Perhaps the chief objections that may be raised would be that the test is artificial and unlike real life tests; that it was short; and that the tests had to be performed with a time limit at a certain place and on a certain occasion.

On the whole, Thorndike must be granted his claim to have shown that transfer is not easy to detect as it ought to be if it occurred on a wholesale scale.

The results showing the relation of studies to gains measured by the relative frequency of the study in groups making various amounts of gain, arranged in descending order from high value to low value were: French, chemistry, trigonometry, physics, general science, Latin, bookkeeping, physical training, arithmetic, geometry, algebra, history, music, shop, Spanish, English, drawing, business, civics, biology, dramatic art, stenography, economics, cooking, sewing. The tables show that there is a possible relation between the presence of French or chemistry in a pupil's program and gain in the test, and a negative relation between the presence of cooking and sewing in a pupil's program and gain in the test. In the case of shop, Spanish, or drawing, there is zero relation approximately, these studies being as frequent in the programs of children who gain little as in those who gain much. The

superiorities of French, chemistry and trigonometry were such as would be produced if taking these subjects increased a pupil's gain the respective number of points over what it would be if he had taken an average mixture of studies in place of the French or chemistry or trigonometry. The inferiority of dramatic art in stenography as shown in the same tables are such as would be produced if taking one of these studies instead of an average mixture of studies reduced a pupil's gain by so many points respectively. The meaning of his results may be made clearer by the following example. He finds that one year's study in either algebra or geometry as now organized does increase one's ability to think by a small degree. He finds that the gain is no larger than that due to the study of other subjects as they are now organized. Strangely enough, bookkeeping, biology, cooking and sewing increase one's ability to think as much or even more in some cases than does the study of mathematics or the classics. We shall let Thorndike give in his own words some of his conclusions and interpretations.

It should be noted that the facts. . . are not in opposition to the theory that general improvement of intellect results from school studies, and that studies vary in their value in this respect. The question is one of quantity. The traditional theory was (1) that the amount of general improvement due to studies was large; and (2) that the differences between studies in respect of it were large, so that the value of a study as training for the intellect was of very great importance relative to its value as special training in mathematical or linguistic or spatial or other thinking; and that (3) mathematics and languages gave much more of this general improvement than other studies did. The facts of Table XVI (not reproduced here), if corroborated by similar experiments, prove that the amount of general improvement due to studies is small; that the differences between studies in respect of it are small, so that the values of studies may be decided largely by consideration of the special training which they give; and that the languages have no claims to preeminence... (pp. 96-97).

By any reasonable interpretation of the results, it appears that the physical sciences are equal, if not superior, to languages and mathematics as now taught in our American cities, in respect

of mental discipline from any point of view. The social studies, including History and English, would probably make a better showing than they do if the examination tested thinking with respect to broad questions of human nature, morals, and community affairs, instead of some of its linguistic and numerical items. I doubt, however, whether the difference is wholly due to the content of the test. Something must probably be conceded to the more rigid and systematic and abstract nature of the thinking in mathematics, German, and physics as compared with that in English and History. As already suggested, the biological sciences probably suffered in a similar manner from the limitations of the tests (p. 98).

By any reasonable interpretation of the results, the intellectual value of studies should be determined largely by the special information, habits, interests, attitudes, and ideals which they demonstrably produce. The expectation of any large differences in general improvement of the mind from one study rather than another seems doomed to disappointment. . . . The chief reason why good thinkers seem superficially to have been made such by having taken such school subjects, is that good thinkers have taken such studies, becoming better by the inherent tendency of the good to gain more than the poor from any study. . . . When the good thinkers studied Greek and Latin, these studies seemed to make good thinking. Now that the good thinkers study Physics and Trigonometry, these seem to make good thinkers. . . . After positive correlation of gain with initial ability is allowed for, the balance in favor of any study is certainly not large. Disciplinary values may be real and deserve weight in the curriculum, but the weights should be reasonable (p. 98).

18. Summary and Abstract of Study on Transfer [Abstracts from The Reorganization of Mathematics in Secondary Education. A Report by The National Committee on Mathematical Requirements under the auspices of the Mathematical Association of America, 1923, Chap. ix.]

The present investigation has proceeded along the following lines:

(1) A study of the experiments and discussion of psychologists on the subject of transfer was made. (2) A set of statements of what are thought to be logical inferences from these experiments and discussions was drawn up. (3) The opinions of forty of our leading educational psychologists with regard to

the truth of these statements was solicited. (4) An analysis of the replies and an attempt to formulate conclusions from them has been made. . .

...

The principal facts, drawn from the results of these experiments, from which later inferences have been made, are the following:

1. A very great majority of the experiments show transfer of training.

2. Under constructive criticism certain defects in the experiments have been pointed out, and consequent improvements made in their technique. For example: It has become the general practice to form upon the basis of a first test two equal groups of practically the same ability, one of which, the practice group, is trained, while the other, the control group, is not. Both groups take a final test, the results of which are compared in a variety of ways, for the purpose of finding to what extent, and in what ways, either group has surpassed the other. The school room as well as the laboratory is now used as the place of experiment, and the young person as well as the trained adult as the subject of experiment. The material used in the training has changed also. The tendency now is to use classroom exercises in place of the more or less artificial materials formerly employed.

3. Many of the experiments have dealt with mental functions which would not be likely to show much transfer under the best of conditions.

4. The more objective the relation between the matter of the training and that of the tests, the more noticeable the transfer, especially in cases where the training was narrowly specific.

5. Negative transfer, or interference due to practice, is sometimes found.

6. Carefully conducted experiments directed toward the conditions of transfer show certain concomitants of specific training to have been operative in bringing about transfer, among which are: (a) The development of an ideal. (b) Understanding of the method of procedure. (c) Conviction of the worthwhileness of the training. (d) Maintenance of a high level of attention.

7. The relation of abilities of students in school subjects is sufficiently close to make it not impossible that their training in one subject is influencing their performance in others.

[3] A Formation of Inferences.-From the general considerations listed above the following seven statements were formulated and submitted to psychologists for their consideration. The psychologists were asked for any further information they would

be willing to give, and it was hoped that a free expression of opinion on their part in regard to any debatable point would tend to obviate some of the objectionable features of a ques tionnaire.

I. Transfer of training is an established fact, and may be positive, negative, or zero.

II. The true amount of transfer from one field to another has not yet been found by experiment, on account of one or more of the following handicaps:

(a) The maturity or previous training of the subjects tested.

(b) The absence from the training of the factors most favorable to transfer.

(c) The inadequacy of the tests to measure the traits sought.

III. It is a reasonable inference that a substantial amount of transfer to some related field would be found by an adequate test, if in teaching children, emphasis were placed upon the trait which the selected subject was most capable of developing, and if the factors controlling transfer were present in the training.

IV. Even if no great amount of transfer of training to any one field should ever be found by experiment, it would still be true that if small amounts of transfer of a valuable trait extended to a large number of fields, the sum total of all these small amounts would be a very valuable educational asset.

V. Negative transfer or interference may take place when in the training of a certain trait, auxiliary habits are cultivated, which have to be broken down before the trait can function in a new situation.

VI. Zero transfer may occur when the habits tending to interfere, and those tending to transfer, just offset each other.

VII. There are elements of situations so fundamental in their nature that they occur again and again in connection with almost anything else. Special training with these elements has general value.

[4] The Judgment of Psychologists and General Conclusions. -Of the forty psychologists to whom the statements were sent twenty-seven replied. Three of these did not give permission that their names should be used. The names of the remaining twenty-four, together with their replies in full, will be found in Section 5 of this chapter. [Some not reproduced here.]

In the following table an attempt has been made to show roughly the nature of the responses to each statement. The figures in the body of the table indicate numbers of men; totals