Officers. A year's schooling was worth roughly 9.46 points on intelligence score most people. If that is the case, the results of the tests ought not to be influenced by schooling. Schools are not the whole of education by a long shot, and the number of years a boy spends at school is no absolute measure of the quality of the education he has received. Nevertheless, the number of years spent at school is a rough measure of the amount of his formal education. Now, in the army: Officers had come on the average almost through the third year of college. Their median schooling was 14.7 years. The native-born white draft had come on the average almost through grade Their median schooling was seven. 6.9 years. The foreign-born white draft had come on the average almost through grade All the whites in this table are less than one and a half points apart on the intelligence scores per year of schooling. The Northern negroes are only two thirds of a point behind the native whites. The Southern negroes, on the other hand, score only about half as many points per year of schooling as the officers or the foreign whites. If we may assume that each year of schooling is as good as any other, these differences may represent measures of different native abilities. But that is to assume that practice in taking written examinations like the army tests is not cumulative. Yet until we can discount this factor of practice, we are in no position to reach a conclusion as to what these differences represent. The correspondences are, of course, true only for the average of large groups. The figures do not mean that you can predict any man's score by multiplying the number of years of schooling by approximately 8.9. A man's capacity to learn and a school's capacity to teach are not so uniform as that. For each particular man the score will vary in accordance with factors which we cannot separate, such as his native endowment, his physical and emotional development, his infantile development, and the quality of the schooling he has received. But the figures do mean that in large groups the average scores will correspond fairly closely with even so crude a measure of education as the time spent at school. The psychologists, of course, noted this fact. They found a positive correlation ranging from +.65 to +.81. There were, then, three conceivable ways of interpreting these facts. The psychologists could have said that in general the more schooling, the better the score. They could have said that while schooling alone did not determine the score, it was certainly a powerful influence. Either interpretation would, however, have knocked out their claim that the tests measure native ability pure and simple. They have, in fact, chosen to argue not that schooling affects the scores, but that the scores indicate how many years of school each group was capable of completing. They assert that virtually all men stay in school as long as their native ability enables them to stand the intellectual strain. Therefore, those who never went to college are inherently incapable of going to college. Those who never went to high school have not the capacity to go to high school. Consequently, the men's scores did not rise and fall with the amount of their schooling. They had more or less schooling, corresponding to better or worse scores, because they were born superior or inferior men. It is not necessary to reply that every person is capable of completing a college course in order to offer this theory as a blue-ribbon exhibit of the determination to hold a theory in the teeth of the facts. Nobody in his senses thinks that schools are the whole of education, or that education is all powerful. It is only necessary to say that education makes some difference in the result of an intelligence test to destroy the claim that the tests measure native ability pure and simple. To refute the argument that the amount of schooling affects the score, the thoroughgoing testers offer what Dr. Carl C. Brigham describes as "a crucial test." They compare the scores of 660 officers who had never gone beyond the eighth grade in school with the alpha scores of 13,943 native-born recruits, all of whom had gone beyond the eighth grade. The median scores are as follows: The difference in median scores between officers with eight years schooling or less and men with eight years schooling or more is about ten points. The difference between all officers, with an average schooling of nearly 15 years, and all men, with an average schooling of nearly seven years, is over 79 points. And the third thing to note is that we have not the vaguest notion who these 660 officers were, or what they and the recruits with whom they are compared did with their time between the day they left school and the day they were tested. When we have noted all these things, we may very well admit as an hypothesis that a difference of native ability existed, though how much difference we have no way of knowing. The "crucial test," then, seems to me a considerable failure as a proof that the scores are not affected by education. It is a complete failure, in fact, if the scores are in any appreciable degree affected by education. Those of us who reject this dogma do not claim that native ability is no factor in the result. We claim that it is an unknown factor. If it is not the only factor, we win our case against these mental testers who have committed themselves to the task of proving a negative. They say education does not affect the scores at all. We say it affects the scores in some unknown degree. One of the curiosities of a work like Dr. Brigham's is that he uses the army data for almost every kind of comparison except one. He makes comparisons between officers and men, between whites and negroes, between native and foreign born, northern Europeans and southern Europeans, but none between the States of the Union, and none between sections of the country. For some reason the army editors also left that comparison alone. Yet the sectional differences are very striking. Using the median alpha scores for white soldiers worked out by Mr. Herbert B. Alexander of Stanford University, I have grouped the States according to geographical sections. If what follows seems to present some invidious comparisons to the reader, I can only say that I do not interpret these figures, as the mental testers do, as indications of native superiority and inferiority. Pacific. New England. 79.1 67.4 61.9 61.4 59.4 56.2 47.6 44.3 44.1 West North Central. (omitting Del., W. Va.) (omitting Florida) How do these scores compare with the school systems of the different groups of States? groups of States? The recognized measure for state school systems is known as Ayres' Index. This index measures ten facts about the schools: the per cent. of school population attending school daily; the average days attended by each child of school age; the average number of days schools were kept open; the per cent. that high-school attendance was of total attendance; the per cent. that boys were of girls in high school; the average annual expenditure per child attending; the average annual expenditure per child of school age; the average Now, in this list the New England States rank third on the tests, second in education; the East North Central States fifth in both; the West North Central fourth in tests, sixth in education; the Mountain States second in tests, fourth in education; the Pacific States first in tests, third in education; the MidSouth Atlantic States seventh in both; the South Atlantic ninth in tests, tenth in education; the East South Central tenth in tests, ninth in education; the West South Central eighth in both. The one group in which the correspondence is not pretty close is the Mid-Atlantic group, consisting of New York, New Jersey, and Penn New England.. East North Central. South Atlantic. East South Central. West South Central. except New England and the FarWestern groups. These figures are based on the school systems of the year 1900. They may measure roughly the efficiency of the schools in that year. But suppose we ask ourselves which school systems were most progressive relative to the others? For it is a fair assumption that a school system which is improving is likely to be better in quality than one which is just holding its own. If you rank these groups of States in the order in which, according to the material furnished by Ayres' Index, they made the greatest gains from 1890 to 1918, you arrive at the following result: Rank in Tests Rank in Improvement 1890-1918 Now, this would seem to argue that better schools on the whole produce better scores in the intelligence tests. But if it does, the claim of the testers that they measure the original endowment of the racial stock collapses. Therefore, in order to overcome the correlation between schooling and scores, it becomes necessary for the tester to argue that superior stock is responsible for high scores and good schools, inferior stock for poorer scores and poorer schools. Before we accept any such farfetched conclusion, it is well to exhaust the common-sense explanations. Schools cost money. Is there any correspondence between the rank of these groups of States in intelligence scores, school improvement, and per capita income? less wealth are all the results of inferior stock. The intelligence scores, remember, are for the white draft only, the educational index and income are, of course, for both negroes and whites. Probably, if negro schools were omitted from the figures for the Southern States, their educational index would improve somewhat, but not enough, I believe, to alter the general result. On the other hand, separating negro and white schools in the South would almost certainly tend to make a closer correspondence between the intelligence test of the Southern negro and his educational opportunities. The testers are all the time talking about superior and inferior stocks. Their explanation of the whole business is in these terms. They claim that officers as a class are biologically Rank in Tests Rank in Educational Rank in Per Capita Improvement The really striking thing about this table is the great difference in per capita income between the three groups of Southern States and all the others. They all fall below $500, whereas no other group, except the Mid-South Atlantic is below $600. That may be an explanation of why their school systems are the poorest, and this in turn may be some part of the explanation of why their intelligence tests are the poorest. In order to explain away these figures, your radical tester has to argue that poorer scores, poorer schools, and 10 superior to the native white men, the native white men biologically superior to the foreign-born whites from Ireland and southeastern Europe, the recent immigrants superior to the negroes, the Northern negroes superior to the Southern. They speak of racial superiority when they mean superiority in the intelligence tests, for they are determined to believe that the tests measure the quality of the race rather than a mixture of race, opportunity, and education. Their determination to believe this requires some of the most ingenious |