Deficiency and Delinquency: An Interpretation of Mental Testing

by James Burt Miner

How closely the degrees of immorality are associated with the degrees of deficiency remains one of the most important problems to be answered authoritatively by the correlation of these traits when properly measured. That the greater degrees of immorality and of deficiency are on the whole associated and not opposed we have good reason to believe, but there are undoubtedly examples in which the degree of immorality or delinquency is out of proportion to the degree of deficiency. The fact that certain instances are found of moral imbeciles without corresponding intellectual deficiency, which has been noted by Stern (188, p. 75) and by Anton ([67]), does not of course determine the direction of the tendencies. We must base our deductions as to the danger of delinquency among lower and higher grades of deficients on our knowledge of the general tendencies. Are morons, relative to their numbers, more dangerous to the community than lower grade deficients? We must not make the absurd deduction that because morons are most numerous they are most likely to be delinquent and should therefore be most carefully isolated or supervised.

B. The Correlation of Deficiency and Delinquency.

Modern statistical methods afford the ultimate quantitative tool for determining the cause of delinquency, whether or not we also require that the data should be assembled under experimentally controlled conditions. The rapid strides which have been made in answering this fundamental question of criminology may be judged by noting the treatment of it in such a work as Goring's compared with the impressionistic literary style which has prevailed. Illustrations of particular cases, opinions subconsciously formulated by experts from wide experience in dealing with delinquents, even the votes of the majority of leaders in the field, give way before the acid test of measurement of tendencies in human traits just as poorer methods succumbed in the Middle Ages in the realm of the physical sciences. Quantitative determinations can no longer be brushed aside with a smile on the supposition that statisticians are the biggest liars. They must be answered by better data or more refined methods. The form of the discussion of social questions has changed. Correlation is a powerful new weapon for attacking these problems which promises to go far beyond the range of earlier blundering methods.

While partial correlation affords an ideal approach to answering the question of causation, it has been used only to a very limited extent. The necessary data for comparing the closeness of relationship of various suggested causes of delinquency are not available and too few who are interested in social problems have appreciated the significance of the method. We should, therefore, lay especial emphasis on the measurement of the correlation of deficiency and criminality by Goring. He laboriously assembled the only data which are sufficiently extensive to allow much reliance to be placed upon their statistical reduction. In his use of correlation, moreover, he acted under advice from the main center for this work at the Galton Laboratory in London.

If those who were “mentally defective” under Goring's designation were always convicted of crime and none of those who were not defective were ever convicted of crime, the measure of the relationship between criminality and deficiency would be expressed by a correlation coefficient of +1.00. If there were no relationship whatever between deficiency and criminality the coefficient would be 0.00. If the deficients were never convicted of crime and the non-deficients were always criminal the coefficient would be -1.00. Intermediate degrees in the relationship of these tendencies would then be represented by decimals which would be either positive or negative, depending upon whether the traits were associated together or were opposed. The coefficient which he found for the male population was +.6553, which was much higher than that for any other constitutional or environmental factor which he measured.

In calculating this correlation Goring regarded 10% of the criminal male population as defective. He found that this was in agreement with the common tendency in English convict prisons to class officially about this portion of the criminals as defectives and needing care. He also assumed that 0.46% of the non-criminal male population in England and Wales was defective, the proportion suggested by the report of the Royal Commission on Feeble-mindedness. By a careful computation he calculated that 7.2% of the males either have been or will be convicted of crime before they die. He then constructed the four-fold table on the basis of these estimates as applied to the 948 convicts whom he examined as to their mental condition. The coefficient was then calculated by Pearson's method for a four-fold table. This method assumes that the mental ability and the tendency to criminality are distributed normally in the population and that the difference in numbers between the criminal and the non-criminal, deficient and non-deficient are not too great. In case the percentage of defectives among the criminals were taken as 20% instead of 10% the correlation would be increased to .79.

Using the same four-fold method we may calculate the correlation between deficiency and juvenile delinquency among Minneapolis boys. It is necessary to make a good estimate of the proportion of boys who annually become delinquent in Minneapolis for the first time, and of the proportion of these boys who are correspondingly deficient. Fortunately these comparisons can be made fairly accurately on the basis of the reports for the year 1915 and of our tests of juvenile delinquents. We may use a minimum and a maximum estimate of deficiency among the delinquents corresponding to those that tested below borderlines which represented the lowest 0.5% and the lowest 1.5% of the population of corresponding ages. We need to assume that the frequency of tested deficiency among the boys found delinquent would correspond within these limits to the frequency among the Glen Lake group. The indices for the amount of school retardation in these two groups (Table XIV) indicate that this is a liberal estimate. We must also assume that the proportion of juvenile delinquents for the year 1915 may be regarded as typical for a series of years. The number of new cases of boys in juvenile court in 1915 was within 18 of the median number for the last four years. The result of these estimates is Table XIX for the minimum estimate of deficiency. A similar table for the maximum estimate of deficiency would be the same, except that the proportion of all boys of these ages who were deficient would be 1.5%, and of the delinquent group, 7.3%.

The computation of the correlations by Pearson's tetrachoric r shows the relationship between juvenile delinquency and deficiency among boys to be .16, P. E. .07, on the minimum estimate of deficiency. On the maximum estimate the correlation is .29, P. E. .05. In order to make a closer comparison between Goring's calculation and my own I have recalculated the correlation for his group on the assumption that 0.5% of the general male population were deficient and that 1.29% would be convicted felons of the type among which he found 10% to be deficient. This brings the minimum correlation for his figures to .59, P. E. .03.

TABLE XIX

Four-Fold Correlation Table for Juvenile Delinquency and Deficiency in Minneapolis (Minimum Estimate).