05 June 2012

"The Development of the MMPI": Notes to Rod Buchanan's Excellent Essay in the Journal of the History of the Behavioral Sciences

This essay by Roderick D. Buchanan offers numerous important insights into the technology of psychometric testing. As I was reading this essay, I took fairly extensive notes from the work. These may well interest regular readers. Where appropriate, I provide a bit of context. My comments below are in bold italics.
Buchanan begins by surveying the context of interwar American psychology and psychiatry.
"[The MMPI] became the most researched personality test in the world. Its widespread application in medical, clinical, industrial, and correctional settings in the U.S. rivalled the spread of the Binet intelligence scales in the first decades of this century." (148)
What was available before the MMPI?
"The earliest tests of personality - standardized inventories said to measure attitude, emotions, and various forms of psychopathology - assumed that the important components of psychological functioning were present in all individuals, differing between individuals only in degree. Responses to inventory items operationalized the measurement of these personality components or traits; responses were combined in some way to indicate the magnitude of the trait in question. The magnitude of a trait for a particular individual was then compared with the distribution of scores of a group of respondents."
Psychologists and psychiatrists seem not to have been overly fond of these personality inventories - Buchanan claims (but cf comments by Danziger here).
"They were widely regarded as difficult for the patient to comprehend and vulnerable to "faking." Their deviation samples were usually confined to well-educated, non-clinical respondents, such as college students. Their scoring categories reflected detachment from day-to-day clinical practice and bore little relation to existing psychiatric classification systems. In essence, these tests betrayed the non-clinical context in which they were constructed." (149)
In a later essay (see here), Buchanan described how projective tests represented a real competitor. But in this essay, Buchanan merely acknowledges the distance between both forms of psychometrics:
"Projective tests were difficult to adjust to the diagnostic concepts which the more conservative psychiatrists understood and respected." (150) And we should probably note that at Minnesota those conservatives were J McKinley and A B Baker, whose positions were in "neuropsychiatry" and "neurology" respectively. Baker would be worthy of a biography by the way.
Minnesota was a peculiar place in the mid-1930s.
"During the 1930s, the state psychiatric hospital system was still expanding. In 1936, the Minnesota State Legislature provided funds for the establishment of a psychiatric unit in the University of Minnesota's Hospital. There, a rationale for a new diagnostic instrument was formulated in the late 1930s by Starke R. Hathaway and J. C. McKinley." (150)
The test was really established within a psychiatric context - a point with significant implications but one that falls outside the scope of this study.
"From the outset the purpose of the project was represented in general psychiatric terms. As Hathaway summarized it, "the entire venture (the MMPI) began because Dr McKinley and I wanted to condense those long psychiatric interviews, which were very expensive for the patient." Hathaway recalled that "at the beginning, I did not have a definite number of scales in mind; however, I knew we had to include schizophrenia and depression." (150)
The interviews were long. In their manual on neuropsychiatry it seems clear that Hathaway, Baker, and McKinley were sometimes taking multiple hours for the first interview. But that doctrine of efficiency did not dominate as the sole justification:-
"McKinley and Hathaway outlined a quasi-industrial model for the test's utilization and pointed to increases in professional efficiency that would accrued with the test's use in mental health practice." (150)
There was a further claim - namely that Hathaway didn't like the various shock treatments.
"This would prove to be an extremely effective selling point, not only for the test but also for the professional services of the psychologists using it." (151)
So how did McKinley and Hathaway collaborate to make the test?
"They relied on clinical experience, psychiatric textbooks [there own?], directions for case-taking in medicine, neurology and psychiatry, and on earlier scales of personal and social attitudes when drafting the items." (152)
Curiously the content of the questions that ultimately appeared in their inventory had not immediate import:-
"The actual content of items was, for the purposes of scale construction, unimportant. Hathaway recalled that "no item was ever eliminated from a scale because its manifest content seemed to have no relation to the syndrome in questions." "(152)
Together Hathaway and McKinley created most of the scales. There was one exception:
"The last of the MMPI's ten basic clinical scales was the Social introversion scale. Developed independently in 1946 by Lewis E. Drake at the University of Wisconsin, the scale was found to be so useful by those at Minnesota that it was routinely scored and added to the MMPI's basic clinical scales." (153)
Amazingly, however, a great deal of the more "superfluous" items were kept in the inventory:-
"Nearly one third of the final 550 items were never included in the MMPI's ten clinical scales or three validity scales. Apart from adding considerable length to the test, some of these filler items were to become the most notorious, because they had unusual or very personal somatic, sexual or religious content. These items could not be removed, however, because they might alter the psychometric properties of the test." (154)
Overtime the psychometric quality of the test engendered a practice of assuming that:-
"In effect, mental "illness" came to be defined as clinical scale T score of around 70 or more. This was a means of classification that psychiatrists, directing psychologists, could actively control. It was an administratively efficient means of classification with a claim to precision based on empirical science." (156)
But rapidly clinical psychologists pulled away from this potentially harmful relationship between experts. And thus:
"The test was reinterpreted as a measure of psychological character types. The psychiatric legacies inherent in the test's origins were systematically removed or obscured. Interpretation of the test became a distinctly psychological skill and technical inscriptions generated by the test's use took on sophisticated non-medical language. New research possibilities opened and innovative interpretive strategies were elaborated. The MMPI became a technological product that could help build clinical psychology as a profession based on a specialized expertise." (157)

No comments:

Post a Comment