A Factorial Study of the Supreme Court.

The object of the present study is to ascertain by factorial methods whether the voting records of the individual judges give evidence of any groupings in the Supreme Court. If these can be found, it should be possible to make prediction of the votes of individual judges on some particular issues. Ideally it might be supposed that the judges make completely impartial and impersonal legal judgments about legislation. No group of human beings can ever be expected to perform that way. It can be taken for granted that, no matter how the judges are selected, there will be differences among them as to their loyalties, their political and social philosophies and their attitudes on economic issues. It can be taken for granted that these differences in attitudes and points of view will have an effect on their individual decisions. There is increasing interest in the application of scientific method to the study of social phenomena. One of the requirements of a scientific method is that it shall be relatively free from the value judgments of the investigator. The interpretation of scientific work by any method is always essentially subjective. The gain with scientific method is that it restricts the range for subjective interpretation to that which is not denied by empirical demonstration. No scientific hypothesis can ever be proved to be correct. At best, it can only be shown to be plausible. It is our belief that social phenomena can eventually be studied with the rigor that characterizes the older sciences, and that social phenomena are not necessarily more complex than physical phenomena. This paper is an exploration in scientific method on the problem of identifying the blocks or subgroups within a larger group as in a legislature, council, or court, in terms of the voting records of the members. The present study was made on the voting records of the judges during the 1943-1944 and 1944-1945 terms of the Court on 115 cases in which there were at least two dissenting votes. The cases in which the Court was unanimous evidently could not give any indication of differences among the judges and hence those cases were eliminated. A correlation coefficient was computed for each pair of judges. With nine judges there were thirty-six such correlations and these are shown in table 1. A positive correlation coefficient shows that the two judges tended to agree in their votes, whereas a negative coefficient shows that the judges tended to disagree. These coefficients are of some interest. A fourfold table was prepared for each pair of judges. The majority judgment of the court was 628