Response to Guttman & Levy's article ‘on the definition and varieties of attitude and wellbeing’

Guttman and Levy have prepared an extravagant critique focused mainly on the 1980 Andrews-McKennell article in this journal. The clearly stated purpose of that article was to report a “series of explorations into the affective and cognitive components of some of the more widely used measures of perceived well-being”. Guttman and Levy ignore this. They proceed on the mistaken impression that we were (or perhaps should have been) embarking upon a definitional exercise to relate the concepts of attitude and wellbeing. Yet the reason we did not cite their article on that topic was precisely because it did not address in a direct or focused way the topic that concerned us. Their critique consists of an entirely irrelevant reanalysis of some attitudinal data by Ostrom, together with a tissue of recondite definitional and methodological issues of little consequence either for the objectives or the conclusions of our research. Their dismissal of our work as ‘scientific retrogression’ rests on an a priori definition of science that fits their own methodological style but excludes that of many other prominent researchers. Their comments reflect an attempt at methodological imperialism. We defend our independence — and that of other investigators — to use promising new methodologies other than the particular approach advocated by Guttman and Levy. (Their denunciation of the new methods of structural equation modeling is not shared even by the authoritative reviewer they themselves quote.) In addition to Guttman and Levy's specific criticisms, our Response addresses general methodological issues such as the status of structural modeling and the testing of structural models. In a concluding section we identify areas that merit further research.