Shadow Play
New member
- Joined
- Oct 28, 2018
- Messages
- 236
Introduction
I was going to post this in the (currently ongoing) thread called "MBTI is dead", but the post got so long that it warranted its own thread.
I don't agree that MBTI is invalid; not completely invalid, at least. It is an attempt to systemise varying temperamental gut pulls in individuals by describing four different types of "two kinds of people", as well as the ways in which those various "two kinds" overlap within each individual. We all know people who are outgoing or recluse, down to earth or speculative, detached or personable, and hard working or lazy. There is even statistical data - compiled in the MBTI Manual - that shows correlations between types and career satisfaction or various personality traits. MBTI has been independently confirmed to share four out of five factors with the Big Five.
Identical twins have a higher likelihood of testing with similar type preferences, which potentially demonstrates Jung's hypothesis about twins demonstrating genetic personality traits (though he didn't frame it in those terms, since it was the 1920s).
It doesn't warrant the comparisons with Horoscopes it receives. After all, hardly anyone would expect two people to share personality traits in common because they were born around the same time. No statistical data would show this unless it was heavily skewered to support that conclusion.
However, MBTI is not without its limitations.
The Problem with Preferences
MBTI is a typology system oriented around the dichotomies, not functions, and there are a number of aspects where the MBTI has diverged - for the better, in my opinion - from Psychological Types. Despite of this, MBTI pays lip service to Jung's type theory through jerry rigging a function stack to each type, claiming in bad faith that they reflect in one's dichotomy preferences. Proper, "actual" functions as patterns of cognition described on typology forums are not falsifiable, since they cannot be adequately tested for. Thus, quasi functions are devised which are based on two or three letter dichotomy combinations. These are erroneously passed off as functions. The continuous lip service is hindering the credibility of MBTI.
In response to criticism of bimodal distribution, proponents of the MBTI claim that, instead of being binary letters, the dichotomies represent four spectrums of preference which overlap with Big Five factors. It makes sense in theory, but the types themselves do not reflect the strength of an individual's preferences. For example, one ENFP could test with strong N, F, and P preferences, but have a borderline E preference, while another ENFP could test with moderate E and P preferences and borderline N and F preferences. Those two ENFPs are technically the same type, and they would show up as such if compiling results for statistical data, but their labels would not reflect the significant MBTI-related differences in their own types.
Reynierse, who concluded that the functions are a category mistake, proposed a workaround for this, suggesting that dichotomies be arranged in order of preference. This would allow for NFPE and EPNF typings, for instance. The problem is it only shows the order of preference, not the strength of preference. It's all fine and dandy to use "x" if a preference is close enough to the middle, but the official MBTI does not promote the use of "x" (even though it doesn't explicitly deny that possibility), and tests do not allow for middle preferences to reflect in one's results.
Test Limitations
The test itself has its limitations. Many of the items do not constitute logical opposites, which is something that's acknowledged in the MBTI Manual.
The items in the official test are the ones that passed muster after testing against thousands of subjects, with those items showing the clearest preference for their respective dichotomies. The problem is those items have binary Y/N responses, meaning that for some responses, the subject might just pick a random response without feeling strongly about the one they chose. This can lead to results which, although showing a dichotomy preference, do not adequately reflect the strength of that preference. For example, a subject could choose mostly N responses, but if they are ambivalent about the majority of them, they could test as having a strong N preference despite only having a mild N preference.
Accuracy of results requires the subject to answer in "shoes off" mode, but some subjects could potentially be projecting their ideal self or some other self (assuming the self is more than a psychological construct) when taking the test. Alternately, because of a lack of validity scales to assess exaggerated or socially accepted responses, it's possible for a subject to fake their responses.
Retest Reliability
The NEO-PI-R, the name of the standard Big Five test, has the following retest rates for each factor: N .92, E .89, O .87, A .86, C .90. This leaves a retest rate of approximately 60% for the four corresponding facets combined, excluding Neuroticism because it lacks a directly corresponding dichotomy. The average retest rate for individual MBTI dichotomies is approximately 83%, which is slightly lower than the average for the Big Five facets, but not much lower. However, the retest rate consistency overall for subjects is approximately 50% within nine months, while the retest rate after nine months is approximately 36%.
On the one hand, it doesn't make sense to criticise MBTI for a lack of reliability while praising the Big Five for its reliability, since 60% isn't much better than 50%. On the other hand, we're looking at no more than half of subjects producing consistent test results within a short term timeframe.
Useless for Self-Improvement
Despite being sold as a tool for personal growth, MBTI has limited use for self-improvement. Whether or not it needs to be a tool for self-improvement is a matter of opinion, and this hangs on the assumption of there even being a meaningful self, but the dichotomies serve mainly to outline temperamental gut pulls that supposedly lead to various "two kinds of people". It is possible to adjust to one's temperament to some extent, such as making a deliberate effort to develop punctuality if one has a habit of lateness. However, considering studies have shown identical twins have a higher likelihood of sharing similar type preferences, this would suggest personality is at least as much nature as it is nurture. If that's the case, there is only so much anyone can do in adjusting against their own temperaments.
Furthermore, if a system really is to offer meaningful improvement, it must be honest about the shortcomings of the thing it aims to improve. MBTI emphasises the strengths each type has to share, but in doing so, it avoids focusing much on the negative aspects that can come with type preferences.
The Big Five does not share either of those problems. It neither aims to be a system of self-improvement, and nor does it go out of its way to emphasise the positive traits for each factor. Instead, the Big Five merely aims to describe different sorts of individuals using psychological terms, even if those terms might seem loaded with negative adjectives. It lacks an agenda.
Only Four Dichotomies?
More importantly, can we be sure the dichotomies are even the most accurate labels for our differing temperaments? What if there are other temperamental spectrums that haven't been factored by either the MBTI or the Big Five? I disagree with the criticism that those typologies don't explain all of human personality - and nor should they explain all of human personality. But I suspect there are underlying methodological issues which limit identifying core factors. Factor analysis lacks a universally recognised basis for pinpointing a solution when there are different numbers of factors. This means factor analysis is dependant on the analyst's own interpretation of data.
Although MBTI claims to be a soft science, extraversion and introversion are its only confirmed factors at this point in time. The rest of the dichotomies could well be multiple factors rolled into one, which means an individual could be strongly P in some respects and strongly J in others, or there could be other factors that only show up through dichotomy correlations - such as S/N and J/P correlating with one another. The MBTI Step II aims to compensate for this by outlining five facets for each dichotomy, but this assumes each dichotomy necessarily has the exact same number of facets. This is a shortcoming the Big Five shares with the MBTI, since it assigns six facets each to the five factors.
I was going to post this in the (currently ongoing) thread called "MBTI is dead", but the post got so long that it warranted its own thread.
I don't agree that MBTI is invalid; not completely invalid, at least. It is an attempt to systemise varying temperamental gut pulls in individuals by describing four different types of "two kinds of people", as well as the ways in which those various "two kinds" overlap within each individual. We all know people who are outgoing or recluse, down to earth or speculative, detached or personable, and hard working or lazy. There is even statistical data - compiled in the MBTI Manual - that shows correlations between types and career satisfaction or various personality traits. MBTI has been independently confirmed to share four out of five factors with the Big Five.
Identical twins have a higher likelihood of testing with similar type preferences, which potentially demonstrates Jung's hypothesis about twins demonstrating genetic personality traits (though he didn't frame it in those terms, since it was the 1920s).
It doesn't warrant the comparisons with Horoscopes it receives. After all, hardly anyone would expect two people to share personality traits in common because they were born around the same time. No statistical data would show this unless it was heavily skewered to support that conclusion.
However, MBTI is not without its limitations.
The Problem with Preferences
MBTI is a typology system oriented around the dichotomies, not functions, and there are a number of aspects where the MBTI has diverged - for the better, in my opinion - from Psychological Types. Despite of this, MBTI pays lip service to Jung's type theory through jerry rigging a function stack to each type, claiming in bad faith that they reflect in one's dichotomy preferences. Proper, "actual" functions as patterns of cognition described on typology forums are not falsifiable, since they cannot be adequately tested for. Thus, quasi functions are devised which are based on two or three letter dichotomy combinations. These are erroneously passed off as functions. The continuous lip service is hindering the credibility of MBTI.
In response to criticism of bimodal distribution, proponents of the MBTI claim that, instead of being binary letters, the dichotomies represent four spectrums of preference which overlap with Big Five factors. It makes sense in theory, but the types themselves do not reflect the strength of an individual's preferences. For example, one ENFP could test with strong N, F, and P preferences, but have a borderline E preference, while another ENFP could test with moderate E and P preferences and borderline N and F preferences. Those two ENFPs are technically the same type, and they would show up as such if compiling results for statistical data, but their labels would not reflect the significant MBTI-related differences in their own types.
Reynierse, who concluded that the functions are a category mistake, proposed a workaround for this, suggesting that dichotomies be arranged in order of preference. This would allow for NFPE and EPNF typings, for instance. The problem is it only shows the order of preference, not the strength of preference. It's all fine and dandy to use "x" if a preference is close enough to the middle, but the official MBTI does not promote the use of "x" (even though it doesn't explicitly deny that possibility), and tests do not allow for middle preferences to reflect in one's results.
Test Limitations
The test itself has its limitations. Many of the items do not constitute logical opposites, which is something that's acknowledged in the MBTI Manual.
MBTI Manual said:In writing items, every effort was made to make the responses appeal to the appropriate types, for example, to make the perceptive response to a JP item as attractive to P people as the judging response is to J people. The result is that responses may be psychologically rather than logically opposed, a fact that annoys many thinking types. Item content is less important than that the words and form of the sentence should serve as a "stimulus to evoke a type response."
The items in the official test are the ones that passed muster after testing against thousands of subjects, with those items showing the clearest preference for their respective dichotomies. The problem is those items have binary Y/N responses, meaning that for some responses, the subject might just pick a random response without feeling strongly about the one they chose. This can lead to results which, although showing a dichotomy preference, do not adequately reflect the strength of that preference. For example, a subject could choose mostly N responses, but if they are ambivalent about the majority of them, they could test as having a strong N preference despite only having a mild N preference.
Accuracy of results requires the subject to answer in "shoes off" mode, but some subjects could potentially be projecting their ideal self or some other self (assuming the self is more than a psychological construct) when taking the test. Alternately, because of a lack of validity scales to assess exaggerated or socially accepted responses, it's possible for a subject to fake their responses.
Retest Reliability
The NEO-PI-R, the name of the standard Big Five test, has the following retest rates for each factor: N .92, E .89, O .87, A .86, C .90. This leaves a retest rate of approximately 60% for the four corresponding facets combined, excluding Neuroticism because it lacks a directly corresponding dichotomy. The average retest rate for individual MBTI dichotomies is approximately 83%, which is slightly lower than the average for the Big Five facets, but not much lower. However, the retest rate consistency overall for subjects is approximately 50% within nine months, while the retest rate after nine months is approximately 36%.
On the one hand, it doesn't make sense to criticise MBTI for a lack of reliability while praising the Big Five for its reliability, since 60% isn't much better than 50%. On the other hand, we're looking at no more than half of subjects producing consistent test results within a short term timeframe.
Useless for Self-Improvement
Despite being sold as a tool for personal growth, MBTI has limited use for self-improvement. Whether or not it needs to be a tool for self-improvement is a matter of opinion, and this hangs on the assumption of there even being a meaningful self, but the dichotomies serve mainly to outline temperamental gut pulls that supposedly lead to various "two kinds of people". It is possible to adjust to one's temperament to some extent, such as making a deliberate effort to develop punctuality if one has a habit of lateness. However, considering studies have shown identical twins have a higher likelihood of sharing similar type preferences, this would suggest personality is at least as much nature as it is nurture. If that's the case, there is only so much anyone can do in adjusting against their own temperaments.
Furthermore, if a system really is to offer meaningful improvement, it must be honest about the shortcomings of the thing it aims to improve. MBTI emphasises the strengths each type has to share, but in doing so, it avoids focusing much on the negative aspects that can come with type preferences.
The Big Five does not share either of those problems. It neither aims to be a system of self-improvement, and nor does it go out of its way to emphasise the positive traits for each factor. Instead, the Big Five merely aims to describe different sorts of individuals using psychological terms, even if those terms might seem loaded with negative adjectives. It lacks an agenda.
Only Four Dichotomies?
More importantly, can we be sure the dichotomies are even the most accurate labels for our differing temperaments? What if there are other temperamental spectrums that haven't been factored by either the MBTI or the Big Five? I disagree with the criticism that those typologies don't explain all of human personality - and nor should they explain all of human personality. But I suspect there are underlying methodological issues which limit identifying core factors. Factor analysis lacks a universally recognised basis for pinpointing a solution when there are different numbers of factors. This means factor analysis is dependant on the analyst's own interpretation of data.
Although MBTI claims to be a soft science, extraversion and introversion are its only confirmed factors at this point in time. The rest of the dichotomies could well be multiple factors rolled into one, which means an individual could be strongly P in some respects and strongly J in others, or there could be other factors that only show up through dichotomy correlations - such as S/N and J/P correlating with one another. The MBTI Step II aims to compensate for this by outlining five facets for each dichotomy, but this assumes each dichotomy necessarily has the exact same number of facets. This is a shortcoming the Big Five shares with the MBTI, since it assigns six facets each to the five factors.
Last edited: