ABSTRACT
My most sustained line of investigation, with publications extending over more than 40 years, has had to do with methods for appraising the accuracy of psychological or educational measurements and of inferences from them. Goldine Gleser's collaboration made possible an intellectual climb to a level I would never have achieved unaided. This essay 1 is centered on events that culminated in our Psychological tests and personnel decisions (utility theory, 1957; 2d ed., 1965) and Dependability of behavioral measurements (generalizability theory, 1972). It mentions some motivational influences, some stimuli that moved thoughts into one channel rather than another, some false starts and misjudgments, and some of the work left undone. This display of hindsight is neither modest nor penitent; rather, I hope to offer to persons entering mathematical psychometrics a realistic perspective on a career.
