Commentary: Not all attitudes are created equal.

LaPierre begins his article with the cogent statement that ‘By definition, a social attitude is a behaviour pattern, anticipatory set or tendency, predisposition to specific adjustment to designated social situations, or, more simply, a conditioned response to social stimuli.’ Alas, if that statement were only regularly honoured. Instead, as LaPierre goes on to note, ‘But by derivation social attitudes are seldom more than a verbal response to a symbolic situation’. Let us note that LaPierre was writing those statements 75 years ago. Nothing has changed, and social psychologists continue to observe and fret about the fact that ‘attitudes’ do not necessarily relate to ‘behaviour’. As LaPierre’s opening statement makes clear that that idea is nonsense because attitudes are behaviour. Or, perhaps a bit more accurately, attitudes are inferred from behaviour, and only from behaviour. The whole issue hinges on one’s ideas about just what constitutes ‘behaviour’ and whether one wants to divide behaviour into the two categories of ‘what people say’ and ‘what people do other than say things’. LaPierre was on the right track 75 years ago, and Donald T. Campbell made the case explicit and compelling 46 years ago. It is unfortunate that LaPierre’s thinking has been so regularly misinterpreted and that Campbell’s thinking has been so regularly ignored—or, perhaps to be fairer, that Campbell’s thinking has been so little known. Campbell’s (should have been) seminal paper was a longish chapter in a distinguished, but often dense, multi-volume work, and in volume 6 at that. In his first major paper, published in 1950, Campbell did write ‘In the problem of assessing social attitudes, there is a very real need for instruments which do not destroy the natural form of the attitude in the process of describing it’. That statement, and the article for which it was the introduction, give clear evidence of Campbell’s view of attitudes as response dispositions and not simply as verbal acquiescences. I first heard about the remarkable study by Richard LaPierre, due to become a classic, in an undergraduate class in social psychology. The time was the spring of 1950, the place was Ohio State University and the instructor for the class was Donald T. Campbell (Assistant Professor). The LaPierre study, which was well and enthusiastically described by Campbell, but which we were not required to read, made a strong impression on me. I doubt, however, that I fully understood its implications at the time. That Campbell regarded the study as important on several grounds must have been obvious to me, however, as I remember my own many citations of it in subsequent years. In the summer of 1950, I was called into service as a reserve in the Marine Corps and ended up in Korea. Although I was able to return to Ohio State University by the fall quarter of 1951, Campbell had already left Ohio State University—to my disappointment—for the University of Chicago. I completed my graduate study at Ohio State University and, in 1956, went off to Pennsylvania State University as an Assistant Professor. In my teaching at Pennsylvania State University, I certainly made use of my knowledge of the LaPierre study in my own classes, but I confess that I still had not actually read it. Then, by a stroke of good fortune, in 1958, I received an offer of an assistant professorship at Northwestern University where, mirabilis dictu, I found myself a colleague of Donald T. Campbell! Even in retrospect, I can scarcely believe my good fortune. Under the influence and tutelage of Campbell, I began to develop my already strong interest in research methodology and measurement. Specifically, though, I expanded my understanding of the difference between psychological constructs and the indicators on the basis of which we infer them. In the early 1960s, I began to be involved in the discussions and work that, in 1966, culminated in the publication of ‘Unobtrusive Measures’. The development of that book, the product of a collaborative team led by an organizational psychologist teaching in a journalism school, and including a social psychologist, a sociologist and a somewhat apostatic clinical psychologist, led me to a much deeper and elaborated understanding of the epistemology underlying measurement in the social sciences. It was in the course of developing the thinking underlying unobtrusive/non-reactive measures that Published by Oxford University Press on behalf of the International Epidemiological Association