Preview (the tl;dr version)
This is pretty long post, so for the tl;dr crowd, I'm including a preview at the beginning:
Douglass Wilde's book gives an quantitative way of calculating function-attitudes (cognitive functions, or "modes" in Wilde's terms) which seems to match experience better than tradition MBTI type dynamics. So, here's a screen shot of a webpage I made that allows you to estimate your preferences and see the results given his model:
Give it a try and post your results below... it seems to fit the people I know well better than the TD model (see "Personal Note" below).
Douglass J. Wilde's Jung's Personality Theory Quantified attempts to address some of the problems of the traditional MBTI "type dynamics" model (abbreviated by Wilde as "TD"). In the typical MBTI TD model:
- The dichotomies (E-I, S-N, T-F, J-P) are dichotomous. Strength of preference isn't important and, at most, may reflect some aspect of type development.
- The four letter MBTI code entirely determines the ranking of one's function-attitudes (or often just called "functions", Wilde uses the term "modes"). So, for example, an INFP 's function-attitudes in rank order are: Fi Ne Si Te. Every person of a given personality type, in this model, has the same dominant, auxiliary, tertiary and inferior. So, all ESTJs would have dominant Te, auxiliary Si, tertiary Ne and inferior Fi.
However, the standard type dynamics model has failed to be empirically validated. For example, ESTJs and ENTJs ("Te-doms") don't actually test higher in Te than ISTJs. Plus, as Reynierse points out, the effect of whole type can be entirely explained by the effect of a single preference or the additive effects of a dyad or triad of preferences. Therefore, MBTI type dynamic theory (TD) seems to be no more than a set of rules that cannot be validated against empirical data, and lacks any solid theoretical underpinnings.
Wilde had previously published Teamology, which describes a technique for composing teams at work such that each function-attitude is represented. Wilde's research claims to empirically show that having a diverse set of cognitive-functions ("modes", in his terminology) on a team results in better performance. Since it's typically impossible to have every function convered, Wilde came up with a method for assigning people based on their ranked, preferred function-attitudes. So, for example, if your team lacks someone whose strongest function is Te, you might assign a "Te representive" who does not have Te as their strongest function attitude (it might be their second or third strongest preference).
Wilde had already come up with an algorithm for assigning team members to teams, so when Reyneirse's "The Case Against Type Dynamics" was published (see this thread for summary of Reyneirse's approach), Wilde saw an opportunity to provide a quantitative means of determining function-attitudes that addressed some of Reyneirse's critiques of TD, and better fit observed strengths in practice.
Dispensing with Attitude Balance
Wilde proposes that attitude balance (which claims that the dominate and auxiliary function must always have opposite orientation) was never a hard and fast rule, and was based on a misreading of Jung. Given his arguments, it's easiest to quote him here:
Myers herself mentions that some people are "pure introverts" or "pure extraverts", which gives added credence to the standard TD model being insufficient to describe the mode preferences of some.Originally Posted by Wilde, loc2046-2058
Wilde starts by providing a very simple means of quantifying strength of preference. He basically assigns a percentage strength of preference based on the number of responses. So, someone who answered all "E-I" questions with a "prefers extraversion" answer would be 100% E and 0%I, where as someone who only gave extraverted answers to half the E-I questions would be 50% E and 50% I. This makes is possible to use standard MBTI scores (or any other similar instrument) to provide a standardized, quantitative "strength of preference" percentage
Decoupling the Attitudes
Wilde then proposes decoupling the attitudes into two separate, independent two dimensional scales. I'll just quote Wilde here:
So, Ej-Ij only affects judging and Ep-Ip only affects perceiving. To calculate Ej-Ij and Ep-Ip he just averages the two scores (although note that the P/J is reversed in the formula for the introverted versions). So, for example, Ej = (E + J) / 2, while Ip = (I + J) / 2.Originally Posted by Wilde Jung Quantified, loc 817
Wilde then proposes representing these two scales on a two dimensional grid. This yields a grid with four corners:
- Ep (Extraverted Perception) - Exploration
- Ej (Extraverted Judgment) - Control
- Ij (Introverted Judgment) - Appraisal
- Ip (Introverted Perception) - Focus
Each of these pairs has set of qualities (as Reyneirse would predict). Wilde gives simple, starter description just by mixing and matching descriptions from the underlying preferences. For example for Ij he gives (and remember, Ij is highest for IPs):
Wilde notes that clearly better descriptions are possible, but his are just starter ones.Originally Posted by Wilde Jung Quantified, loc 817
Back to Modes/Function-Attitudes
Wilde then maps S-N onto Ep-Ip and T-F onto Ej-Ij to calculate the modes/function attitudes. This makes some sense, given that Te (organization) is Thinking + Control (the Ej descriptor given above), while Ti (analysis) is Thinking + Appraisal (the Ij descriptor given above). So, for each value, he calculates an average (so S-N gets averaged with Ep-Ip and T-F with Ej-Ij).
This, in turn, yields the following formulas for the functions:
- Se = (Ep + S) / 2
- Ne = (Ep + N) / 2
- Si = (Ip + S) / 2
- Ni = (Ip + N) / 2
- Te = (Ej + T) / 2
- Ti = (Ip + T) / 2
- Fe = (Ej + F) / 2
- Fi = (Ij + F) / 2
This both makes intuitive sense (given how things were built up), and is as simple as mathematical calculations get. One can also render the same formulas starting from individual preference strength scores:
- Se = (((E + P) / 2) + S) / 2
- Ne = (((E + P) / 2) + N) / 2
- Si = (((I + J) / 2) + S) / 2
- Ni = (((I + J) / 2) + N) / 2
- Te = (((E + J) / 2) + T) / 2
- Ti = (((I + P) / 2) + T) / 2
- Fe = (((E + J) / 2) + F) / 2
- Fi = (((I + P) / 2) + F) / 2
This approach produces some interesting qualities:
- Opposing functions are opposed. That is, your Te and Fi scores are inverses of one another. This means 100% Te means no Fi in this model. This contradicts many interpretations of the TD model (but not Thomson's in this respect).
- Non-opposing functions are not opposed. Therefore, it's possible to have both relatively high Te and Ti. Or high Te and Fe. This fits better with results from cognitive function tests than the tradition TD model.
- One's strongest two functions may share attitude/orientation. For example, an ESTJ whose preferences are 100% with the exception of being 75% J, calculates as having higher Se than Si, making Se his or her strongest perceiving function.
Expanded Type Codes
Given these added combination, Wilde proposes an expanded type code that adds asterisks to represent some of the new variations. Note that large swaths of results require no asterisks, which helps explains how the TD model survived as long as it did as a rough approximation.
So, in Wilde's expanded type codes, someone's whose top two function share the same orientation has a type code with an asterisk following E or I. So an E*STJ leads with Te and Se, which an I*NFJ leads with Ni and Fi.
Someone whose top two functions share the same judging or perceiving preference get an asterisk after the judging or perceiving preference in their type code, respectively. So, an ENFP who leads with Ne and Ni would be an EN*FP, while an ISFJ who leads with Fe and Fi would be an ISF*J.
These expanded type codes give the possibility of a somewhat overwhelming 64 types.
Wilde goes on to propose a quantitative approach to Beebe's archetypes. This approach has some complexity and I haven't taken the time to absorb it at this point. Still, it does seem to have the potential to redeem Beebe somewhat. Personally, I have found Beebe's proposed mode to archetype mapping to be a mismatch for my personal experience (although of course the experience of others may vary).
On a Personal Note
Using an approximation of my MBTI Step 2 scores with this system, my highest functions in order are Ne, Ni, Fi, Ti. This is a pretty good match for what I perceive via introspection (except that I don't think I'm that strong on Ni). I score significantly stronger on "N" than "F", and this model accurately reflects that. I also consider Te to be my weakest judging function, and that's reflected as well. It also yields my Si being lower than my Se, which seems accurate as well.
My partner's highest functions on this system come out to be Si, Ti, Te, Se. This is also a much better fit for him than the standard TD model. He clearly leads with Si and Ti (which breaks the standard TD model), but doesn't seem to have a history of trauma or invalidation that would explain any kind of "type falsification."
My mother shows a lot of social adroitness and group dynamic awareness, and in this system she should have high Fe. She finds objective reasoning (Thinking) pretty draining, though, and that's predicted by this model as well.
In summary, I think this is a promising quantitative approach. It's mathematically very straight-forward, but it yields result that seem to match with personal experience (which the standard TD model only sometimes does, in my experience). It helps highlight the importance of some of Reyneirse's dyads (previously ignored by the MBTI and TD), and makes intuitive sense. I'd be curious how other people fare if they plug in their numbers (or gut-level approximation of the same).
Consequently, I created a web-page which creates a graph and results of Wilde's formulas [url=http://dl.dropbox.com/u/1896920/20120729-1800/wilde/WildeCalc.html]available here[/u].
More specifically I'd be interested in:
- The text of the "results" box for folks from the above, after they've estimated their preference strength. (Note that if you type in percentages you have to hit enter/return... a bug.)
- Does this Wilde's model match the results from cognitive function tests better than the standard type dynamics model?
- Does Wilde's model match your subjective impressions better?
I'm less interested in hearing that this model is different from normal MBTI type dynamics (because yes, it clearly is).