Wednesday, April 8, 2009
A Rose By Any Other Name
Are the "poles" presented in various measurements of psychological type truly quantumly different? Is introversion just very low extraversion? Is the quantum difference present with some preferences and not with others? Would we say that Sensing is just very low Intuiting? Are the three functions--energy, perception, and judgment--so very different that we need to reconceputlaize how these are presented and studied?
Subscribe to:
Post Comments (Atom)
This is an important and multi-faceted topic. The typological assumption is, at least for psychologists concerned with measurement issues, arguably the preeminent criticism of the MBTI instrument, the most commonly used measure of "type." It underlies the objections to reliability and validity of the instrument. If one tests reliability using continuous scores, as many researchers have done, including authors of the manuals from Isabel Myers and on, reliability meets generally accepted standards. If one measures reliability by consistency of classification, particularly for four discrete measurements (the four domains), the results are not impressive.
ReplyDeleteAnswering one item differently can change type classification, and we know that "low preference" scores are associated with lower consistency over repeated measures-- i.e., low test-retest reliability.
So, from this perspective, by reducing scores to a binary outcome (dichotomy) instead of a continuous score, we give up precision, especially at the boundaries.
I see no logical reason why type and continuum cannot coexist, though I recognize there are also theoretical issues at play here. Right/left handedness, a favorite analog for type preference dichotomies, is far more observable than psychological phenomena. Even though it's relatively easy to reliably classify someone as right or left-handed, there are still degrees of handedness and people who are more or less ambidextrous. Baseball switch-hitters come to mind. (I will also note parenthetically that it's confusing to distinguish a "strong" hand preference from a "clear" one. As with type, that distinction is lost on many.)
Why not report both type and continuous score as a regular practice? And look at the differences in people who fall into the preference categories along the same pole. That can be done with all measures of type, and there is precedence within the type literature.
One last tangential point in defense of type reliability: type tools measure four things, and typically combine those four measures into a single result. The reliability of such a combination cannot exceed the product of the four separate reliabilities.