The tables of Minneapolis school children were prepared by Mr. Andrew J. Lein and of delinquents by Miss Lydia B. Christ, to whom I am much indebted. CHAPTER IX. COMPARISON OF THE SCHOOL TEST AND THE BINET TEST There has been considerable discussion of the question whether psychological testing should be expected to conform to the ranking of pupils in school. This discussion however, does not attack the question in which we are especially interested, _i. e._, how to get the best information from both. If the school level were measured by the _progress_ made in school by passable work and not by the school _position_ attained often merely through age or size, Binet would be right in expecting that in general they would correspond among groups of children in the public schools. Agreement with real school progress could, therefore, be taken as a criterion of a good series of tests, as it has been by Binet and Bobertag. On the other hand Meumann and Abelson were right in objecting to the proof of the value of tests by agreement with the school level, if they limited their objection to tests applied to exceptional children and to using school _position_ as a final test of school level. Lack of correspondence with our group of delinquents is, of course, no indication of a weakness in the Binet scale. In numerous instances they had been promoted in school because of age without doing passable work. The reader should also see the evidence of the teacher's bad judgment of a pupil's ability assembled by Terman and by Terman and Knollen (_196_). Terman has calculated the correlation between intelligence quotients determined by the Binet scale and the teacher's estimates of scholastic or of general ability. These gave coefficients of .48 and .45. Doll has found for Goddard's data on school children that the correlation of school grades is closer with life-age than with test-age, .84 as compared with .73 (_12_). This indicates an influence of life-age upon promotion. In a school for deficients Burt found the correlation of teachers' estimates with Binet ages was .55, with mental retardation or excess .59, with intellectual quotient .48. He quotes McIntyre and Rogers as finding coefficients about .5 for similar calculations with normal school children in Scotland (_85_). Starch has shown that measured by the combined ability in reading, writing and spelling a third of the pupils are in a grade behind and a third are in a grade ahead of their ability (_186_).