A Further Study of the Rorschach Test Applied to Delinquents: M. I. Pescor. Public Health Reports, LVI, No. 9, 1941.

'This is the third and final article dealing with the relationship of various factors to the Rorschach test as applied to a group of four hundred seventy-six prisoners admitted to the United States Northeastern Penitentiary.' The first article dealt with the age factor, the second with the marital status. This paper takes up many other factors and studies statistically their interrelationships. Pescor's results lead him to a rather sceptical attitude towards the Rorschach test in general. To quote his summary: '(1) The use of the Rorschach ink-blot test in a study of Federal delinquents at the United States Northeastern Penitentiary is reported. (2) The test was applied under rigidly standardized situations and scoring was made as objective as possible. (3) Under these conditions the test was found to be statistically unreliable except in measuring original, total and form responses. (4) A high degree of positive correlation (0.60 or higher) occurred between the following total scores: (1) Total responses versus original, detail, form, animal, and analysis of cards by the general to detail method; (2) popular unweighted versus animal; (3) original versus detail, form, and general to detail method of analysis; (4) whole versus form and analysis of cards by the general method only; (5) detail versus form, animal, and analysis of cards by the detail method only; (6) form versus animal; (7) animal anatomy versus detail to general method of analysis. The only highly significant negative correlation occurs between popular weighted responses and original. (5) Correlations among sundry extrinsic factors and various intrinsic test factors revealed only six coefficients above 0.50, namely, educational grade status versus total detail responses; mental age versus total detail, color, and motion responses and mental age versus analysis of cards by the general to detail and detail to general methods. (6) Applied as a measuring instrument, comparable to psychometric t