Paul E. Meehl was an intellect and Renaissance man who made enduring contributions in a diverse array of subject areas in psychology and beyond. In psychology, he was best known for his work in measurement, philosophy of science, biological bases of schizophrenia, statistics, and actuarial approaches to assessment and prediction. He also published in areas as diverse as political science, extrasensory perception, psychology and law, and religion.
Meehl was a relentless analyst, and his formidable critical faculties led him to some radical conclusions. His theory of schizotaxia (an inherited predisposition to schizophrenia-spectrum disorders) and his dislike of statistical significance tests as a basis for theoretical inferences have since gained wide acceptance. Other conclusions, such as his inclination to think that “there is something to telepathy” and his “skepticism about the received doctrine of organic evolution,” place him well outside the scientific mainstream and into the area of the “strictly taboo.” Meehl’s open-mindedness and probing intellect led to illuminating commentaries on virtually every subject that captured his interest. As a result, he probably authored or coauthored more “classic” conceptual papers (and certainly in more diverse areas) than anyone in the history of psychology.
Perhaps his most influential contribution was to measurement theory. In a 1955 article coauthored with Lee J. Cronbach, Meehl revolutionized the process by which psychologists provide evidence for validity of measures used in research and practice. A major contribution of this article was the acknowledgment that, for most psychological measures, validity cannot be established by a straightforward comparison with a criterion, because normally no unambiguously valid criterion measure can be found. For example, one would like to validate a pencil-and-paper measure of depressive tendencies by showing that it differentiates depressed from nondepressed individuals—but there is no perfectly accurate method of differentiating these two groups to serve as a criterion measure. Cronbach and Meehl explored the implications of this revelation for measurement theory and laid the groundwork for a new approach to construct validation that is central to the postpositivist philosophy of science.
Meehl was a unique figure in psychology in another way. He was a first-rate philosophical and mathematical thinker who also identified strongly as a clinician. He practiced from a psychoanalytic orientation during his early professional decades, then incorporated perspectives from rational-emotive and brief psychodynamic therapies in his mature years. Psychologists who take the trouble to read his original papers on applied psychology (rather than relying on textbook or reviewer summaries) are likely to be pleasantly surprised by Meehl’s astute clinical “ear.” His writings reveal a keen appreciation for the virtues of clinical experience as a basis for theory and practice, and offer wise counsel on applied matters including training and certification, assessment and case conceptualization, and the role of science in informing the practice of psychology.
With a foot in each camp, Meehl was fascinated and sometimes surprised by the tensions between those psychologists who identified primarily as scientists and those who identified primarily as practitioners. Members of each faction were granted exalted social status in their respective domains—practitioners in their clinical settings, and scientists in the halls of academia—and each group could be sharply critical of the other, frequently employing unflattering emotion-driven characterizations. Meehl had little patience with such ad hominem attacks, and worked diligently to impress each camp with its limitations as an exclusive basis for claims about valid theory or optimal practice in psychology.
Limitations of Clinical Experience
As a clinician, Meehl was acutely aware of the power conferred on psychological practitioners by their social and institutional status. Psychologists are empowered to restrict individual liberty, to challenge and often change clients’ construals of their life events, and even to influence law and social policy (via “expert testimony” in courtrooms and beyond). They are also, like all human beings, subject to perceptual and mnemonic biases that can cause them to draw faulty conclusions from their experiences. The dangers of overconfidence in “clinical wisdom” (particularly when the stakes are high) was the theme Meehl chose to return to in his Centennial Award address to the Society of Clinical Psychology (Division 12) of the American Psychological Association in 1997. He stated that clinicians must and should rely on their clinical experience for guidance in many professional contexts, but they should be honest with themselves and others about the limitations of this basis for knowledge claims and for behavior.
A case in point is the diagnostic and predictive decisions made by clinicians in a variety of clinical and legal contexts. Psychologists often rely on vague “clinical impressions” or on idiosyncratic synthesis of impressionistic and objective data in making such predictions. In 1954 Meehl compared such clinical (or impressionistic) approaches to prediction with a statistical (or actuarial) approach in which multiple pieces of information were combined using a (usually quite simple) mathematical formula. His scientific masterpiece, Clinical Versus Statistical Prediction: A Theoretical Analysis and a Review of the Evidence (and more recent reviews), demonstrated unequivocally that actuarial methods virtually always yield predictions as accurate as, and quite often more accurate than, less systematic approaches.
Meehl’s conclusion—that idiosyncratic methods are more likely to impede than enhance predictive accuracy—has frequently been construed as a broad indictment of clinical judgment per se. This was not Meehl’s interpretation of his findings. He set great store by expert clinical observations and noted that the actuarial models in some studies he reviewed incorporated at least some data derived from clinician judgments into the prediction algorithm. In fact, more of Clinical Versus Statistical Prediction was devoted to defending clinicians’ unique inferential activities than to criticizing their predictive abilities. Recent reviewers note that many of the studies reviewed by Meehl focused on prediction of broad life outcomes (e.g., Will this person be successful in the army? Will he or she become manic at some point in the future?). Meehl’s cautious optimism regarding the validity of clinical judgment receives stronger empirical support when researchers examine moderate levels of inference typical of clinical practice (e.g., judgments that a patient is perfectionistic or sensitive to rejection).
In summary, Meehl believed that when no conclusive scientific data are available expert judgment is a necessary and sound basis for many clinical decisions. However, clinical wisdom, like other conclusions based on anecdotal evidence, is sure to lead to incorrect conclusions at least some of the time, and needs to be distinguished from scientific findings (or “credentialed knowledge”). Meehl believed that practitioners are acting irresponsibly and even unethically when they behave as if their expert opinions constitute credentialed knowledge.
Limitations of Scientific Research
The scientific method permits a stronger basis for knowledge claims (including claims about effective practices) than expert judgment does. It involves systematic observations and measurements that are replicable in principle, thus reducing the contribution of motivated cognition and other biases to the findings on which conclusions will be based. This does not mean that researchers always follow the scientific conventions that maximize scientific validity (e.g., complete and accurate reporting of a priori hypotheses and research procedures). Nor does this mean that these conventions are rationally derived. Meehl recognized that the scientific method as currently practiced in psychology is imperfect, and with his colleague David Faust, developed a methodology (termed cliometric metatheory) for identifying scientific practices that could be demonstrated empirically to enhance the validity of disciplinary findings.
Meehl lamented the tendency of psychological researchers to be preoccupied with methodology but oblivious to the banality of many fashionable theories. In a 1978 article Meehl wrote, “I would take Freud’s clinical observations over most people’s t tests any time” (p. 817). Social sciences (to a much greater extent than physical sciences) are fraught with conceptual complexities that researchers may be intellectually or methodologically ill equipped to handle, such as the stochastic (probabilistic) nature of causation and the necessity for open (evolving) concepts. Finally, Meehl believed psychologists’ use of statistical significance testing as their primary method of theory testing was incompatible with postpositivist philosophy of science, and a chief culprit in the disappointingly “slow progress of soft psychology.”
Science and Practice
The implications of these critiques for the relation between science and practice are complex. Meehl was disheartened by what he saw as some clinicians’ unwarranted faith in the authority of conclusions drawn from their own clinical experience, and their seeming disinterest in weighing evidence for alternative interpretations, even when these were grounded in scientific findings. As Meehl wrote in a 1987 essay, the “general scientific commitment not to be fooled and not to fool anybody else” (p. 9) is a desirable characteristic of clinicians as well as researchers.
Meehl acknowledged that few clinicians fulfill the scientist-practitioner aspiration of producing independent research (beyond the dissertation), and was supportive of scholar-practitioner (Psy.D.) training models that (ideally) train students as intelligent and critical consumers of research produced by others. Meehl was pessimistic about the value of most “soft” psychology research for either practice or theory, but identified five “noble intellectual traditions” he considered to have enduring value. Interestingly, three of these (psycho-metrics, applied learning theory, and behavior genetics) have emerged from scientific findings, whereas the other two (descriptive clinical psychiatry and psycho-dynamics) constitute practice-based knowledge.
- Cronbach, L. J., & Meehl, P. E. (1955). Construct validity in psychological tests. Psychological Bulletin, 52, 281-302.
- Meehl, P. E. (1954). Clinical versus statistical prediction: A theoretical analysis and a review of the evidence. Northvale, NJ: Jason Aronson.
- Meehl, P. E. (1978). Theoretical risks and tabular asterisks: Sir Karl, Sir Ronald, and the slow progress of soft psychology. Journal of Consulting and Clinical Psychology, 46(4), 806-834.
- Meehl, P. E. (1987). Theory and practice: Reflections of an academic clinician. In E. F. Bourg, R. J. Bent, J. E. Callan, N. F. Jones, & J. McHolland (Eds.), Standards and evaluation in the education and training of professional psychologists: Knowledge, attitudes, and skills (pp. 7-23). Norman, OK: Transcript Press.
- Meehl, P. E. (1989). Paul E. Meehl. In G. Lindzey (Ed.), A history of psychology in autobiography (pp. 337-389). Palo Alto, CA: Stanford University Press.
- Meehl, P. E. (1990). Toward an integrated theory of schizotaxia, schizotypy, and schizophrenia. Journal of Personality Disorders, 4(1), 1-99.
- Meehl, P. E. (1992). Cliometric metatheory: The actuarial approach to empirical, history-based philosophy of science. Psychological Reports, 71(2), 339-167.
- Meehl, P. E. (1997). Credentialed persons, credentialed knowledge. Clinical Psychology: Science and Practice, 4(2), 91-98.