706 



MEMOIRS NATIONAL ACADEMY OF SCIENCES. 



[Vol. XV, 



preceding August draft of negroes. The best indication that verbal instructions raise beta 

 scores is contained in a comparison of the median alpha and beta scores of each draft: 



Draft. 



August 



September . 



Men tak- 

 ing alpha. 



776 

 401 



Median 

 score. 



26.4 

 14.4 



Men tak- 

 ing beta. 



954 

 1,999 



Median 

 score. 



18.7 

 18.6 



Although the September draft negroes make a median alpha score only 53 per cent of that 

 made by the August draft, their median score in beta with verbal instructions is exactly the 

 same as the median score of the August draft in standard beta. Opportunity was lacking at 

 Camp Sevier to try beta with verbal instructions on a negro draft of even average quality. 



Camp Pike (November 1, 1918) suggests the adoption of the following plan when negroes 

 are to be examined: (1) During rush periods examination alpha to be abandoned except for 

 negroes in specially selected units; and (2) equivalents of letter ratings in beta to be revised 

 in order to get a better distribution as follows: 



The grade of D— covers two groups. Negroes making scores from 15 to 19 should be 

 interviewed and examined further if time permits, and those making from to 14 points should 

 be given individual examinations. 



Camp Meade (November 1, 1918) and Camp Gordon (November 5, 1918) were both of the 

 opinion that beta is not a satisfactory measure for negroes. The former holds that "too large 

 a percentage of the negroes who should make high scores fail in beta." The latter submits 

 data which indicate "that beta is poor at differentiating grades of intelligence of negroes in 

 the low ranges." 



Other camps are better satisfied with the standard beta as a test for negroes. Camp Dix 

 (November 5, 1918) forwarded data which "show quite clearly that the intelligence of the 

 negro recruit is measured quite as accurately as the intelligence of the white recruit of the same 

 mentality." The same report holds that the group examinations are unsatisfactory for 

 measurement of the lower grades of intelligence of both negroes and whites. This observation 

 of course applies to a very much larger proportion of negroes than of whites and, as the report 

 goes on to state, "the fact that 48.8 per cent of the negro beta subjects are rated as D — 

 indicates the extent of the failure of the beta examination to adequately measure all the 

 negroes." 



Camp Travis (May 6, 1918) reported that beta is "a splendid examination for negroes. 

 Every negro should be given beta." In the same report, however, beta is criticized on the 

 ground that it "is more of an ' Auf gabe ' test than a test of ability to work out problems that 

 are understood." It is not stated to what degree this criticism applies to beta when used 

 with negroes. Another report from Camp Travis (October 8, 1918) submits data which show 

 that negroes and whites fail in individual examination in the same ratio that they fail in beta. 

 "If the results of individual examinations are accepted as a satisfactory check, the group test 

 appears to be as reliable in grading negroes as in grading whites." The percentage of negroes 

 who failed in the group test was 50 per cent, that of the whites 4 per cent. 



It appears definitely from camp reports that beta is unsatisfactory for use with negro 

 drafts because its use results in such a large percentage of failures that they can not be recalled 

 for individual examination. The need of a group test which will differentiate between the 



