to know with reasonable certainty what is the inborn ability of the boy or girl who is entrusted to his guidance and tuition.

The question now arises: How is it possible to measure general intelligence? Here we are confronted with a difficulty which at first seems so great that its solution seems impossible. We have already said that general intelligence, which is innate and never learned, cannot manifest itself in independence of learning. From this it follows that any test for general intelligence that we may devise will be a test based on learning. However, it is not learning which we wish to test at all, but only the inborn ability which lies at the basis of all human attainment. Fortunately, there is a way out of the apparently insurmountable difficulty, and that is by devising tests which shall measure the attainments of the individuals in those things where all have had approximately the same opportunities to acquire skill and knowledge. If this is done, it will be found that within a group of individuals tested large variations in ability will be found. Now, since we assume that all of the members of this group have had the same opportunities, it follows that the variations in attainment are not primarily due to learning but to an inherited capacity to learn, in other words, to general intelligence.

It can be seen, from the above statement, that in framing tests to measure general intelligence we must be extremely careful to base the materials included in these tests on matters that are of common experience in the group tested, or perhaps we had better say, on matters for which there is an opportunity for common experience, since those of high intelligence will seize upon all opportunities to acquire worth-while experience while those of low intelligence will pass such opportunities by. It can further be seen that there can be no general test for general intelligence, that is, a test which will accurately measure the native endowment of all individuals chosen at random, since the materials for the test must be selected on the basis of common experience, and common experience cannot be indefinitely generalized. As the individual develops from infancy to maturity, experiences become more and more special and dependent upon group activities rather

than upon activities shared by all members of the human race. These latter would be very few and would be totally inadequate as the basis for a general intelligence test.

The first definite attempt to measure general intelligence dates back to the year 1905. At this time, the French psychologist Binet and the French physician Simon constructed a rough scale of thirty tests standardized by results obtained from normal children of the ages three, seven, nine and eleven years. The first quantitative scale by these authors appeared in 1908. In this there were fifty-six tests arranged according to ages from three to thirteen. For each age there were from four to eight tests. In 1911, a revised scale was published, in which some of the tests were discarded. In this revision there were five tests for every year, except one, from the ages three to ten. Tests were also added for the twelfth and fifteenth years and for the adult. The purpose of these tests was primarily to discover children of low mentality and separate them from those of normal ability.

I shall not at this time attempt in any detailed way to describe the nature of the Binet tests. Doubtless Doubtless many of you here are perfectly familiar with them, and those who have slight knowledge of the scale can find it presented in its best form by Terman, who has most thoroughly revised it, in his book, "The Measurement of Intelligence." The tests used, to apply Terman's description, are of the nature of "stunts" or "problems." They are based on the principle that all intelligent individuals have had an opportunity to acquaint themselves with the facts relating to these "stunts" and to acquire the ability necessary to solve these "problems," and if a child fail in the test suitable to his age it is because of a lack of original ability rather than because of a defect in learning. As has been said, the tests are arranged according to years, the test materials increasing in difficulty from year to year; in the Terman revision there are five for each year from three through ten, and also five for the ages of twelve, fifteen, and for the adult. The tests consist of such simple matters as the recognition of familiar objects, the repeating of digits, the enumeration of ob

2.

jects in pictures, counting, copying geometrical figures, executing commissions, ingenuity in rhyming and arranging words in sentences, abilities to draw comparisons and frame definitions and to make practical and esthetic judgments. Perception, memory, imagination, and reasoning ability are emphasized, largely, however, in the field of words and their uses.

Because they depend so much upon the verbal factor, these tests have been critisised. They have also been objected to because their findings are less exact for the early and late years of child life than they are for the middle ranges. They are good tests to apply between the ages of six and ten, but they are less capable of determining mental ability at three and at fifteen, and they yield only doubtful results in distinguishing between adults of superior and average mentality. They can be used to advantage in the middle grades of the elementary schools, but are of small value for the high school pupil. Another serious objection to these tests, if they are to be used in determining the intelligence of large 3. numbers of children, is that they must be administered individually. This requires a long time if a school system, or even a school building, is to be adequately tested. Further, their administration is not a simple matter. No one is competent to give these tests without thorough study of their nature and without training in technique. This ordinarily requires several months of intensive work on the part of the person who plans to give the tests.

4.

In order to overcome this difficulty in the administration of tests, psychologists have recently been working to devise group tests to measure general intelligence. The group tests can be administered to a large number of children or adults at the same time. They require very little training in the technique of administration and they are scored according to very simple and definite rules, thus largely excluding the factor of individual judgment in assigning grades to those tested. A great impetus was given to the work of determining the intelligence of individuals in groups by the tests worked out by a committee of the American Psychological Association and applied to the draft men

in the army. Two sets of tests were ultimately devised,

the Alpha Tests, for the purpose of measuring the intelligence of the literates in the army, and the Beta Tests, for the purpose of determining the innate ability of those who could neither read nor write, a surprisingly large number in all. We are here primarily interested with the Alpha Tests, because from these and similar tests are now being compiled group scales for measuring intelligence of children and adults who have a fair knowledge of and skill in the use of words.

I shall not discuss the army tests or in any way comment on their results, except to say that they proved effective for the purposes for which they were designed, separating very readily into groups men of varying degrees of intelligence. Professor Dodge, who follows me, will give you an account of the tests and their significance. Largely because the results of these tests showed the possibilities of their application to various groups in civil life, Professor Thorndike has recently drawn up and standardized sets of psychological tests designed to determine the varying degrees of intelligence among high school seniors and college freshmen, one purpose of which is to decide whether men seeking admission. to college have sufficient native ability to carry successfully to completion a college course. These tests have recently been given at Brown University to about 300 first-year men, with the view of ascertaining as far as possible the varying degrees of intelligence of these men, so that they may be adequately advised in regard to their college work and helped to determine their careers after graduation.

These tests require three hours for administration and consist of five parts: a fore-exercise of ten minutes, in which the students get some acquaintance with the general nature of the tests to be subsequently given; two tests of thirty minutes each, identical in character with the tests in the fore-exercise, but differing in content; a fourth test more difficult in character, occupying an hour of actual work; and a final test, resembling in many ways the one previously given, occupying forty minutes. The fore-exercise and the two following tests are each composed of thirteen sub-tests,

the first of these being a directions test; the second a test in the arrangement of words in such an order as to make sense; the third is a test in the fundamental operations of arithmetic; the fourth, in arithmetical problems; the fifth, a range of information test; the sixth, a synonym and opposites - test; the seventh, a bestreasons test; the eighth, a number-completion test; the ninth, an analogies test; the tenth, a number-checking test; the eleventh, a test distinguishing between true and false statements; the twelfth, a fact-and-conclusions test; the thirteenth, an identification test of geometrical forms, numbers and hieroglyphics. In the fourth main test, the first two sub-tests are silent reading tests; rather difficult paragraphs, such as a student might meet in his general or technical reading are given, and the subject is required to answer questions based on these paragraphs. The next test consists of a number of pictures in which certain parts that are missing are to be supplied. That following is an analogies test, with pictured objects instead of words. The next test is also an analogies test, in which three geometrical figures are given and the fourth is to be supplied which bears the same relation to the third as the second does to the first. The next is a simple algebraic test, and this is followed by an information test relating to mechanical facts and principles. The last test in this series is an extensive range of information test, consisting of sixty items. The final main test has three parts, the first two being rather - difficult reading tests, and the last a test in supplying missing words in mutilated sentences.

Another group test which is now available for use is the Otis Group Intelligence Test, consisting of ten parts and requiring an hour for administration. It is designed to be given to children and adults of all grades of school attainment above the fourth. It cannot be given in the lower grades because it requires a fair reading knowledge in order to comprehend it. This test has a number of parts similar to the Thorndike Tests just described. There are two direction tests, one with letters of the alphabet and the other with geometrical figures, beginning with very simple problems and ending with those of considerable difficulty. There