This is a good starting point for finding out what is wrong with myers-briggs (the most important paragraph is the third if I am not mistaken). Sure it usually seems predictive to you but this is at least partially due to the Forer effect.
The in thing right now (and for the last two decades) is the Big Five/OCEAN.
I/E is obviously a thing. S/N has big correlations with g so it’s obviously tracking something. F/T is perhaps less obvious statistically(?) but introspectively and anecdotally-observationally still pretty clear, and P/J is the most questionable and confusing part of MBTI so I won’t defend it. Hypotheses and conceptual frameworks shouldn’t be pet causes.
Data: Wikipedia claims E/I is very correlated with E, S/N is very correlated with O, F/T fairly correlated with A, J/P fairly correlated with C and somewhat correlated with O, and Neuroticism isn’t measured in MBTI. So this backs up your claim that P/J doesn’t measure any concrete “thing”.
Clicking through the citation gives that N is not well-correlated with anything in men (a tiny bit with E/I), and somewhat correlated with the F/T in women. Also F/T has a small effect on extraversion in men, but it’s S/N and J/P which has the effect on women.
As far as I know S/N as measured by MBTI does not have big correlations with anything and it doesn’t even correlate with other scales designed to measure the same type of thing(intuitions etc.). Same for F/T.
“High correlations with comparable scales of other instruments” seems to be an odd metric to demand—is there some reason this would imply a flaw in Myers-Briggs rather than a flaw in the other instruments or simply that they’re measuring different things?
The second paragraph’s criticism that Myers-Briggs scores aren’t bimodal is something I find even more baffling. I wouldn’t have expected any accurate test of non-sexual human behavior to show bimodal results, but I would expect the results of any test to be oversimplified into “you’re more/less X than average” categories.
The Forer effect probably is a significant factor in people’s appreciation of Myers-Briggs, though. And the criticism of its dependence on honest self-reporting hits close to home: when I took a few versions of the test long ago, I found that where I fell on the J/P scale seemed to be heavily determined by what fraction of the questions were phrased as “do you try to X” versus “do you typically X”.
is there some reason this would imply a flaw in Myers-Briggs rather than a flaw in the other instruments or simply that they’re measuring different things?
This is the usual modus tollens/ponens question: just pointing out inconsistency (low correlation) doesn’t tell you who to favor. In this case, the argument for rejecting MBTI rather than the others would go something like ‘it has a highly questionable origin and does not seem to measure anything interesting; the other scales have good theoretical justifications in their areas or were derived directly from the data, have demonstrated various forms of usefulness like predicting relevant behavior, and are less likely to be collectively wrong than MBTI uniquely correct’
As for measuring different things, well, then you get into other things like the lower psychometric reliability of MBTI compared to Big Five—if a instrument is not reliable, then it may be measuring nothing of interest.
“High correlations with comparable scales of other instruments” seems to be an odd metric to demand—is there some reason this would imply a flaw in Myers-Briggs
This is a good starting point for finding out what is wrong with myers-briggs (the most important paragraph is the third if I am not mistaken). Sure it usually seems predictive to you but this is at least partially due to the Forer effect.
The in thing right now (and for the last two decades) is the Big Five/OCEAN.
I/E is obviously a thing. S/N has big correlations with g so it’s obviously tracking something. F/T is perhaps less obvious statistically(?) but introspectively and anecdotally-observationally still pretty clear, and P/J is the most questionable and confusing part of MBTI so I won’t defend it. Hypotheses and conceptual frameworks shouldn’t be pet causes.
Data: Wikipedia claims E/I is very correlated with E, S/N is very correlated with O, F/T fairly correlated with A, J/P fairly correlated with C and somewhat correlated with O, and Neuroticism isn’t measured in MBTI. So this backs up your claim that P/J doesn’t measure any concrete “thing”.
Clicking through the citation gives that N is not well-correlated with anything in men (a tiny bit with E/I), and somewhat correlated with the F/T in women. Also F/T has a small effect on extraversion in men, but it’s S/N and J/P which has the effect on women.
This abstract follows the Wikipedia excerpt:
http://www.sciencedirect.com/science/article/pii/0191886996000335
As far as I know S/N as measured by MBTI does not have big correlations with anything and it doesn’t even correlate with other scales designed to measure the same type of thing(intuitions etc.). Same for F/T.
“High correlations with comparable scales of other instruments” seems to be an odd metric to demand—is there some reason this would imply a flaw in Myers-Briggs rather than a flaw in the other instruments or simply that they’re measuring different things?
The second paragraph’s criticism that Myers-Briggs scores aren’t bimodal is something I find even more baffling. I wouldn’t have expected any accurate test of non-sexual human behavior to show bimodal results, but I would expect the results of any test to be oversimplified into “you’re more/less X than average” categories.
The Forer effect probably is a significant factor in people’s appreciation of Myers-Briggs, though. And the criticism of its dependence on honest self-reporting hits close to home: when I took a few versions of the test long ago, I found that where I fell on the J/P scale seemed to be heavily determined by what fraction of the questions were phrased as “do you try to X” versus “do you typically X”.
This is the usual modus tollens/ponens question: just pointing out inconsistency (low correlation) doesn’t tell you who to favor. In this case, the argument for rejecting MBTI rather than the others would go something like ‘it has a highly questionable origin and does not seem to measure anything interesting; the other scales have good theoretical justifications in their areas or were derived directly from the data, have demonstrated various forms of usefulness like predicting relevant behavior, and are less likely to be collectively wrong than MBTI uniquely correct’
As for measuring different things, well, then you get into other things like the lower psychometric reliability of MBTI compared to Big Five—if a instrument is not reliable, then it may be measuring nothing of interest.
This is how you typically meassure test validity.