But there is one feature that raises suspicion. The Italians and the Poles are too close to the Negroes. They stand much nearer to them in intelligence, according to these figures, than they do to the white Americans. Can this be so—at least, can it have racial significance? Are these Mediterraneans, descendants of the Romans, and these Alpines, so large a strain of whose blood flows in the veins of many white Americans, only a shade superior to the Negro? Scarcely. “Something must be wrong” with the figures: that is, they contain another factor besides race.

A little dissection of the lump results reveals this factor. The northern Negro far surpasses the southern in his showing. He gets ten times as high a proportion of individuals into the above-average grades, only half as many into the below-average. Evidently the difference is due to increased schooling, improved earning capacity, larger opportunity and incentive: social environment, in short. So strong is the influence of the environment that the northern Negro easily surpasses the Italian in America.

Negroes, 5 northern states, 4,70546513
Italians, 4,00763361
Negroes, 4 southern states, 6,8468614(.3)

Evidently the psychological tests are more a gauge of educational and social opportunity than of race, since the Italian, although brunet, is of course a pure Caucasian.

This conclusion is reinforced by another consideration. The type of test first used in the army had been built up for reasonably literate people, speaking English. Among such people it discriminated successfully between the more and the less fit. But the illiterate and the foreigner knowing no English failed completely—not because their intelligence was zero, but because the test involved the use of non-congenital abilities which they had not acquired. A second set of tests, known as Beta, was evolved for those who were obviously ineligible, or proved themselves so, for the old style of test, which was designated as Alpha. The illiteracy of the subjects given the Beta test was in most cases not an absolute one. Men who could not write an intelligible letter or read the newspaper or who had had only half or less of the ordinary grammar school education, together with aliens whose comprehension of English remained imperfect, were put in the group of “illiterates” or badly educated. Separating now the literates from the illiterates among a number of racial, national, or sectional groups, we find:

Alpha Test: Literates
Englishmen, 374 5 74 21
White draft generally, 72,618 16 69 15
Alabama whites, 697 19 72 9
New York negroes, 1,021 21 72 7
Italians, 575 33 64 3
Negroes generally, 5,681 54 44 2
Alabama negroes, 262 56 44 (.4)
Beta Test: Illiterates
White draft generally, 26,012 58 41 1
Italians, 2,888 64 35 1
New York negroes, 440 72 28 0
Poles, 263 76 24 (.4)
Alabama whites, 384 80 20 0
Negroes generally, 11,633 91 9 (.2)
Alabama Negroes, 1,043 97 3 (.1)

It must be borne in mind that the two groups were not set apart as the result of tests, but that the two tests were devised to meet the problem of treating the two groups with reasonable uniformity. The point was to find the excellent man, and the unfit man, with the same degree of accuracy whether he was literate or illiterate. When found, he was assigned to the same grade, such as A, or D—, whether his examination had been Alpha or Beta.

Now let us observe some of the figures. The New York negro is nearly on a par with the Alabama white, among literates, and a bit ahead of him among illiterates. Approximately the two groups come out the same; which means that bringing up in a certain part of the country has as much to do with intelligence, even in the rough, as has Caucasian or colored parentage.

The literate negroes of the draft, irrespective of section, slightly surpass the illiterate whites.

In every case the literate members of a race or nationality make a far better showing than the illiterate.