operation. This group was used for comparison with the operated group of thirty-five. For statistical purposes the groups should have been paired according to life age. It was possible to pair only thirteen in this study, a group too small for results to be significant.

The Standard-Binet test was given subsequent to diagnosis. always preceding the operation, usually by one or two weeks. All of the mental tests were given in the school buildings, and all made and scored by the writer, thus minimizing such factors as excitement and non-uniformity in method. . .

Summary.-1. There was a mean increase of 3 months (average 1.1) in mental age found to be in favor of the operated group. No other investigation has found any change in mental age as revealed by a mental test. . . .

2. The results show no reliable increase in I. Q. one year after the operation.

3. The non-operated group gained as much or more in scholarship than the operated group.

4. The teachers' estimates of other mental traits revealed no difference one year after the operation. This may mean that these were not adequate measures or that the changes in the traits were too small to be noticed.

61. The Significance of Unwillingness to Be Tested [KNIGHT, F. B., "The Significance of Unwillingness to Be Tested," Journal of Applied Psychology, June, 1922, Vol. 6, pp. 211-212.]

As a pedagogical device in introducing the "mental test" to teachers of a county in Iowa, I had the teachers take Thorndike's College Entrance Examination, Part I. At the time of taking the test the teachers were given the choice of signing their names or not as each saw fit. . . . However, those teachers who signed their names to the tests would be told what scores they made and how their scores compared with scores made by college students, stenographers, soldiers, and the like.

There were 65 teachers in the group, 39 signed their names, 26 did not. The average score of the signed papers was 100. The average score of the unsigned papers was 75. Only 4 of the 26 unsigned papers equalled or exceeded the average of the signed group. Only 5 of the 39 signed papers were as low or lower than the average of the unsigned tests. We have very definite contrast in performance between these two groups.

Although we have too few cases upon which to base a certain conclusion it is evident that unwillingness to sign one's name to a mental test may be an unconscious confession of probable inability to do well. Willingness or even desire to know how you compare with others is associated with relatively good

scores.

62. Nordic Intelligence

[BAGLEY, W. C., "The Army Tests and the Pro-Nordic Propaganda," Educational Review, April, 1924, pp. 179–187.] (Adapted.)

Disputes have been waged concerning the intelligence of different racial stocks. Dr. Carl C. Brigham, who propagated the theory that the Nordic races are the superior, based his argument on the results of the Army Intelligence Tests, which were given to both native and foreign-born men during the late war.

Dr. William C. Bagley of Columbia University attacks the position of Dr. Brigham. His arguments are as follows:

He denies that the differences in intelligence scores are due exclusively or even primarily to native intelligence. For one thing, it is essential to Dr. Brigham's proof that all men, who attended school for eight years, should have an equal amount of education. But the number of days in a school year is known to vary for the different states; in 1900 a school year in North Carolina contained 70.5 days, while in Rhode Island it contained 191 days, almost three times as much. In New York, seven years in the urban schools produce the same development as eight years in rural schools. It is, therefore, unscientific to argue that if two men report the same number of years of schooling, the difference between their intelligence scores must be due to native intelligence.

Secondly, Dr. Brigham admits that the table giving the proportion of Nordic blood is only an approximation to the truth.

Thirdly, the very source used by Dr. Brigham-the Army Intelligence Tests-proves that education counts more than native intelligence. For while the whites as a class are superior to the negroes, the literate negroes of Illinois surpass all the literate negroes from the South and even make a higher average than the literate whites, "the purest Nordic

stock in America," according to Dr. Brigham, from nine Southern States. The literate negroes of New York surpass the literate whites of five Southern States. The literate negroes of Pennsylvania surpass the literate whites of two southern states, and all the literate northern negroes combined made a higher average than the literate whites of Mississippi, Kentucky, and Arkansas.

To conclude: Intelligence tests measure, not native intelligence, but intelligence as developed by environment and education. A low score does not indicate defective stock and irremovable racial handicaps, but chiefly lack of education. Hence the folly of the present talk about "superior" and "inferior" races and stocks.

63. Intelligence Tests in Colleges and High Schools [COLVIN, S. S., "Intelligence Tests in Colleges and High Schools": An Address given before the Second Educational Conference at the Ohio State University, 1923. Abstract printed in The Ohio Teacher, September, 1923.]

There is no mental test that has ever been devised, or that ever will be devised that measures innate mental ability. Examine any mental test and you will find that it is a test of what has been acquired either in school or outside of school. Therefore you cannot rely upon mental tests to tell you the difference in mentality between groups of widely different education, experience and environment. But where these other factors are constant, the tests show reasonably well what are the probable differences in mentality. If you are going to do anything with mental tests worth while, you cannot take general standards throughout the country. When you take several tests and find out exactly what they mean in relation to the environment and experience of those tested, these tests serve at least four very useful purposer

1. They tend to classify pupils according to their learning ability. 2. They show what is probably the educational achievement of a pupil-whether he is likely to go through high school and whether he is suited to go through college.

3. They help determine who shall enter schools which have a restricted number who can be admitted.

4. They are also extremely important in determining who shall stay in college.

Two concrete illustrations of the use of an intelligence test that has been checked up against known facts will show some of the things that can be accomplished by them and some of the things that cannot reasonably be accomplished by them.

At Brown University, I worked with intelligence tests to determine fitness of students for college and the desirability of students continuing in college. In 1918 I gave the Army Alpha tests to a fairly large group of students. Later in the year, I gave what is now known as the Brown University Psychological Examination. These tests had some value in determining the success of college students. The next year in the fall, prospective freshmen were given the Thorndike examination for college freshmen and high school seniors and the Brown University Psychological Examination. They are of a general correlation. Besides this we have a system of getting all sorts of information in regard to a student. We gradually get a collection of facts that follows him through his course as long as he stays.

We have been using tests of the same sort over a number of years, so we know exactly what these tests mean. A student scoring in the lowest tenth of the psychological scores has not more than two chances in ten of being able to graduate and will very rarely get a mark above the average. He is a poor college risk. Those that score in the next tenth have not more than one chance in four of getting through with a decent record. Two out of ten that do get through generally have qualities that compensate for those lacking.

The upper tenth has some of the poorest students in college, because they are lazy and have no purpose, but most of them do extremely well. In the last four years only one man won honors when his score was below median. Most of the honor men are in the highest part of the intelligence scale. Poor psychological records are a more certain indication of failure than the grades at the end of the first semester. High college grades at the end of the first semester are a better indication of further success in college than high psychological scores. The former measures many factors, the latter only one.

64. Uses of Intelligence and Achievement Tests in 215 Cities

[DEFFENBAUGH, W. S., "Uses of Intelligence and Achievement Tests in 215 Cities," City School Leaflet No. 20, Bureau of Education, Washington, March, 1925.]

For what purposes are intelligence and achievement tests used in the elementary and secondary public schools of the country? This is a question often asked. In order to gain as definite information as possible regarding the uses made of such tests, the Bureau of Education, with the advice of several well-known authorities on tests, prepared a questionnaire containing 21 possible purposes for which the tests may be used, and sent it to all superintendents of schools in cities of 10,000 or more population, with the request that they indicate the various purposes for which they are using group intelligence, individual intelligence, and achievement tests in the elementary, junior high, and high schools. Replies were received as follows: 35 from cities of 100,000 or more population, 65 from cities with between 30,000 and 100,000 population, and 115 from cities with between 10,000 and 30,000 population, or 215 cities in all. In preparing the summary tables, the size of the cities was disregarded, since the purposes for which the tests are used are practically the same, irrespective of the size of the city.

The superintendents were requested to indicate the purposes for which tests are used extensively, slightly, or not at all. In the tabulation the classification of extensive or slight use has been disregarded; thus the data in the following tables simply show whether the tests are used at all for the purposes listed.

The per cent of the cities reporting the use of the tests for each purpose is given. For example, 64 per cent of the 215 cities in Table I use group intelligence tests in the elementary schools for the classification of pupils into homogeneous groups and 62 per cent for the purpose of supplementing the teachers' estimate of the pupils' ability.

The purposes for which the tests are used are ranked, beginning with the purpose for which the test is most used.

Although only 35 per cent of the superintendents replied to the questionnaire, the data contained in the following tables at least show the purposes for which the tests are used in 215 cities. If all the other cities using tests had replied, it is doubtful whether the ranking of the purposes would be materially altered.