Data was reanalyzed from an earlier study concerned with the person/environment (P-E) fit approach to stress. A new procedure involved regressing strain on E, P, and certain higher order terms, such as the square of E and P, their product, and other measures, supplemented by tests that explicitly evaluate the conceptual models underlying the fit measures used by the earlier researchers. The data were taken from the original article which used participants from 23 occupations, at 67 different sites. Occupations ranged from blue collar (forklift driver, machine paced assembler, delivery service courier and tool and die maker) to between blue and white collar (electronic technician, police officer, train dispatcher, and industrial supervisor), to white collar (air traffic controller, programmer, accountant, engineer, scientist, professor, administrator, and family physician). Two sets of measures were analyzed. Parallel items regarding four job dimensions including job complexity, role ambiguity, responsibility for persons, and quantitative work load comprised the first set. Seven indexes of psychological strain including job dissatisfaction, work load dissatisfaction, boredom, depression, anxiety, irritation, and somatic complaints made up the second set. The current authors conclude that the new procedure avoided problems with fit measures; clarified and elaborated on the relationship between E, P, and strain; and showed often significantly underestimated magnitude of this relationship.