Pearson has approached the same problem in another way (_164_) (_167_). He has used the results of the psychological tests applied by Norsworthy to children in New York in special classes and institutions for feeble-minded compared with those in the regular school classes, and the results of Jaederholm obtained with the Binet tests applied to 301 children in Stockholm in the special classes compared with 261 others selected from the regular classes. He found that “70.5% of normal children fall into the range of intelligence of the so-called mentally defective; and 60.5% of so-called mentally defective children have an intelligence comparable with that of some normal children” (_167_, p. 23). On the statistical assumption that those in the normal classes would distribute according to the Gaussian normal probability curve he estimates that, with the Binet tests, among those in the special classes “10% to 20%, or those from 4 to 4.5 years and beyond of mental defect, could not be matched at all from 27,000 children” (_164_, p. 46). Another 20 to 30% could be intellectually matched by those in the regular classes having from 3 to 4.5 years of mental deficiency, but they would be matched very rarely. On the assumption that 1% of the children were feeble-minded, not more than about two children in a thousand of this regular school population would be expected to be 3 or more years retarded and thus overlap those of like deficiency in the special classes (_167_, p. 30). Considering the results of Norsworthy's study he says on similar assumptions: “It seems, therefore, that a carefully planned psychological test, while not sufficing to differentiate 50 to 60% of the mentally defective from the normal child, would suffice to differentiate 40 to 50%” (_164_, p. 35). Again we come back to the estimate that psychological tests may well be expected to select nearly half of the children at present found in special classes for retarded pupils. Moreover, a considerable part of the overlapping of intellectual deficiency in the regular classes with that in the special classes which he found may be accounted for by the inadequate methods of selection of pupils for the special classes by teachers or examiners who have used no objective tests. Some who were left in the regular classes should undoubtedly have been transferred to special classes and vice versa. There seems to be nothing to indicate that less than half of those properly sent to special classes would be of clear or doubtful intellectual deficiency. If the tests served to select even a smaller proportion of those assigned to special instruction, the “school inefficients” as Pearson calls them, their value as an aid to diagnosis would be demonstrated.