My central concern in this work is with the deep, abiding, and highly problematic confusion that has become, by now, part of the very fabric of mainstream thinking about the proper use of statistical methods in psychological research. It is a confusion that has persisted for decades, and one that the discipline as a whole has thus far simply refused to acknowledge. The confusion itself can be stated quite simply: it consists of the notion that statistical knowledge about variables defined only for aggregates of individuals entitles scientifically authoritative claims to knowledge about the individuals within those aggregates.
The consequences of this confusion are serious and far-reaching for scientific psychology, and are both epistemic and socio-ethical in nature. In the epistemic domain, the confusion routinely leads psychological investigators to overstate what they may justifiably claim to know about individuals on the basis of their research findings. The result is bad science. In the socio-ethical domain, the same confusion routinely leads psychological investigators to overstate what they may justifiably do, or endorse doing, under the banner of scientifically licensed interventions, in the lives of individuals . The result is bad professional practice.
The pervasiveness of this confusion within contemporary scientific psychology would be troublesome enough were it a phenomenon of but recent vintage. What makes matters profoundly more worrisome is the fact that thoughtful and trenchant critiques of invalid interpretations of aggregate -level statistics have surfaced periodically in the literature of the discipline dating back at least to the 1950s—and comparable critiques can be found dating well back into the nineteenth century if one consults scholarly literature outside of psychology (cf. Porter , 1986). Astonishingly, however, while those critiques have never been successfully rebutted, neither have they been duly heeded within the mainstream of psychology as guidelines to corrective methodological practices. Instead, the basic assumption that population-level research findings can warrant knowledge claims about—and hence justify systematic interventions in the lives of—individuals within those populations have continued to prevail. The long-entrenched methodological canon has thus endured as if the confusions, invalidities, and dubious intervention justifications resulting from and sanctioned by that canon had never been challenged to begin with, much less compellingly defended in the face of the challenges. In this respect, mainstream psychology has proven itself to be, in a word, incorrigible. It is difficult to imagine an intellectually more worrisome state of affairs for a putatively scientific field.
In subsequent chapters of this book, major historical developments that have contributed to the emergence of these conceptual difficulties within psychology will be discussed at some length. In this chapter, however, some preliminary observations of an historical nature are in order so as to facilitate the reader’s orientation to the material that follows.
Some Orienting Historical Context
The General and the Aggregate
In 1955, psychologist David Bakan (1921–2004) published a brief, two-page commentary in the journal Perceptual and Motor Skills in which he averred the following:
The failure to distinguish between general -type and aggregate -type propositions is at the root of a considerable amount of confusion which currently prevails in psychology. There are important differences in the research methods appropriate to these two types of propositions. The use of methods which are appropriate to the one type in the establishment and confirmation of the other, leads to error . (Bakan, 1955, p. 211)
It seems, however, that Bakan’s (1955) altogether valid argument fell on deaf ears, so that more than a decade later he found reason to make the same point again, this time embedded within a broader discussion of fallacious thinking that psychologists were routinely indulging in the course of null hypothesis significance testing (Bakan, 1966). In this latter article, published in the high-profile journal Psychological Bulletin, Bakan cast some light on the historical sources of the confusion between aggregate -type and general -type propositions that he had identified in the earlier article. In that connection, he emphasized that at its founding in the latter part of the nineteenth century, experimental psychology was clearly aimed at establishing the validity of
propositions concerning the nature of man in general—propositions of a general nature, with each individual a particular in which the general is manifest. This is the kind of psychology associated with the traditional experimental psychology of Fechner, Ebbinghaus , Wundt , and Titchener . (Bakan, 1966, p. 433, emphasis in original)
Here, as before, Bakan (1966) was entirely correct.1 The general experimental psychology founded by Wundt and prosecuted by him and the other luminaries Bakan mentioned was—and was explicitly referred to at the time as—an individual psychology. This was true even as it was recognized that the overarching objective of the discipline was knowledge of the general laws presumed to govern various aspects of mental life. As I have pointed out in other discussions of this point (see, for example, Lamiell, 2015, 2016), any apparent contradiction between an investigative approach whereby experimental findings are defined for individuals even as the scientific quest is for knowledge of general laws is clearly resolved once one appreciates the meaning that the early experimentalists attached to the notion of ‘general.’
In German, the language of the country where experimental psychology was first formally established, the word for ‘general ’ in the sense relevant here was (and remains) allgemein. That word is a contracted form of the expression allen gemein, which means ‘common to all.’ The allgemeine Gesetze or ‘general laws’ of mental life that psychology’s founding fathers sought to discover experimentally would thus be laws found to be common to all of the investigated individuals and, at least presumptively pending subsequent experimental outcomes, to non-investigated individuals as well.2 What scientific psychology’s first experimentalists were not in search of were empirical regularities found merely to be ‘true on average’ for collections of individuals.
What must also be appreciated is the fact that the individual psychology of the early experimentalists was not a discipline concerned in any way at all with individuality . Precisely because the knowledge sought was that of laws that would prove generalizable across individuals —i.e., from one individual to another and so on—there was no programmatic interest in any empirical indicators of personal idiosyncrasies, i.e., characteristics that, by definition, would not be generalizable across individuals . In this connection, Bakan (1966) aptly cited the anecdote mentioned by E.G. Boring (1886–1968) in his History of Experimental Psychology (Boring, 1950), according to which the suggestion to Wundt by his student, James McKeen Cattell (1860–1944), that he be permitted to study individual differences in reaction times was abruptly dismissed by Wundt as ganz amerikanisch or ‘so American.’
It was into just this breach that William Stern (1871–1938) stepped in 1900 with the publication of a book proposing a new subdiscipline for scientific psychology that he called ‘differential’ psychology (Stern, 1900). In that work, Stern boldly proclaimed ‘individuality ’ as the ‘problem of the twentieth century’ (“Problem des zwanzigsten Jahrhunderts”; Stern, 1900, p. 1). What Stern meant by ‘Problem’ in this particular context is better rendered in English by the word ‘challenge,’ for what he wished to convey was the idea that in order for scientific psychology to be viable on through the twentieth century (and, presumably, beyond), some accommodation would have to be made to the possibility—indeed, the certainty—that any given individual’s psychological ‘doings’ would be scientifically graspable only partly in terms of laws common to all, and would therefore have to be viewed partly in terms not applicable to all—and possibly not even to any—other individuals .3 Stern’s view was that a subdiscipline of psychology devoted to the study of nonrandom individual and group differences in any psychological domain would highlight the need for, and thus serve to facilitate the eventual realization...