The primary purpose of the experiments summarized here was to search for a psychophysical procedure exhibiting maximum reliability and independence of variables which are irrelevant to visual functions. The results of all the experiments indicate that the following procedural conditions will maximize reliability and validity of the data: a discrimination indicated by forced choice not by the usual phenomenal report of "Yes"and "NO"; B- forced choice to involve temporal intervals rather than special locations; C- target luminances to be grouped into blocks of the same magnitude rather than to be randomised, D- as few target luminances to be used in the psychophysical series as practicable, E- subjects to be taught cues for discrimination by notifying them of the correctness of their responses; and F- subjects to be given reasonable extensive experience in threshold measurement.
Abstract