Shadow function prioritization: a model

[narticus]

Lenore Thompson’s Personality Type: An Owner’s Manual opened up the MBTI to me in a way not previously experienced or understood. I still refer to it often. My one major disagreement is with her prioritization of the shadow functions. Her theory would order functions for an ISTJ thusly:

Si Te (Ni Fe | Ti Se) Fi Ne

I have a very small set of data from myself and four friends from the cognitiveprocessesDOTcom. Comparing the S/N split and T/F split, I found that in seven of the ten cases, the pair of i/e functions of one type both scored higher than the opposing pair (for example, both Ss scoring higher than both Ns or vice versa). In two of the other ten, a ‘sandwich’ pattern existed (e.g. SNNS). In only one of the ten comparisons did the LT model hold (Ti > Fi > Te > Fe) for an INTP friend.

Granted this is a miniscule sample size, but to me the results make sense. If, for example, my dominant function is Ni, then my comfort with abstract information supersedes my comfort with concrete information, regardless of i/e attitude. Returning to the ISTJ example, I would order the functions:

Si Te (Ti Se | Ni Fe) Fi Ne

For each of the five people for whom I had data, I ordered the functions in LT’s projected order as well as my own. Returning to the ISTJ friend the scores are below.


LT BB
Si 44.4 Si 44.4
Te 38.2 Te 38.2
Ni 19.6 Ti 37.3
Fe 26.8 Se 29.1
Ti 37.3 Ni 19.6
Se 29.1 Fe 26.8
Fi 27.9 Fi 27.9
Ne 17.1 Ne 17.1

I counted any time that the scores increased when moving from one function to the next one on the list (indicating that a function was out of order), and the size of the jump (an approximation of how far out of place the function is). In each of these two lists, there are two jumps in scores, indicating two functions ‘out of order’. In the case of the Thompson model, the two jumps sum to a magnitude of 17.7, while in my model, the two jumps sum to only 8.3, suggesting that my model was a better predictor of order. For the five samples, results are below. (Pair ordered: Thompson, Me)


INTJ (3,2) (31.5,22.4)
ISTJ (3,2) (37.2,32.4)
INTP (1,1) ( 3.0,10.9)
ISTJ (2,2) (17.7, 8.3)
ESFP (3,3) (19.0,24.0)

The jumps were equal in three of five cases, with my system creating fewer jumps in the other two. In four out of five, the magnitude of jumps also favored my system, the lone exception being a case in which both predictions were quite accurate.

Five data points is certainly not enough. If any of you are willing to provide your own data (please include both the type you think you are as well as the type that the results suggested, if the two differ). PMPM me your results. When I have accumulated a more credible amount of data (or when data stops coming in), I will try to provide more results to the board.

 

[Totenkindly]

Five data points is certainly not enough. If any of you are willing to provide your own data (please include both the type you think you are as well as the type that the results suggested, if the two differ). PM me your results. When I have accumulated a more credible amount of data (or when data stops coming in), I will try to provide more results to the board.

There's a number of scores for the cognitive function test on INTP in this thread.

Also here on MBTIc is this thread.

 

[narticus]

Caveat: Everyone is unique, and any personality models will have abundant exceptions. The information below is a look at trends observed from one set of data. If the model below doesn't work for you, c'est la vie.

I have a lot more data now. I don't have time (yet) to do all the math in the original post with the new data, but I can throw together a more abbreviated analysis with almost as much truthiness.

For each member in the sample, I looked at the pattern of the T/F split and the N/S split, though without the i/e attitudes (that's the part with much more math that will come later). There are three possible patterns: AABB (e.g. Ni Ne Si Se), ABBA (e.g. Ni Si Se Ne), or ABAB. The middle two letters are part of the shadow functions. My theory is an AABB pattern. Lenore Thompson's is ABBA. I haven't seen anyone with an ABAB theory.

Looking at splits:
58.8% AABB
22.8% ABAB
18.4% ABBA

Looking at both pairs for each person:
38.6% had AABB patterns for both T/F and N/S
21.1% had one AABB pattern and one ABBA pattern
19.3% had one AABB pattern and one ABAB pattern
8.8% had ABAB patterns for both T/F and N/S
8.8% had one ABAB pattern and one ABBA pattern
3.5% had ABBA patterns for both T/F and N/S

Though the data supports my AABB patterns over LT's ABBA, it is likely that some questions do a good job of distinguishing between, for example, T and F, but not Ti and Te. Thus the test itself might have bias toward the AABB pattern.

On the other hand, switching to Socionics for a moment, the 7th and 8th functions (the opposite i/e attitude of the MBTI's dominant and auxiliary functions), or the 'id block', are thought to be strong, frequently used in private life, but do not provide the sense of fulfillment that the 1st and 2nd functions (same as dominant and auxiliary in MBTI) do. This supports, at least in terms of ability, an AABB pattern. Under the Socionics model, I am proficient with both Ni and Ne, but I get more enjoyment from using Ni.

Based on my own experience (and thus just my opinion), the Socionics concept rings true in my own life as well as that of my friends. For example, if creating a tool to solve a problem, it's more important to me that I get to the answer quickly (Te) than to take the time to clearly define each term and each step (Ti). I can do the defining parts quite well, but I find it tedious. I work with a bunch of INTPs. They can crank out an answer quickly if they have to, but they'd rather clear up any amiguity before answering if time allows. In both cases (me: INTJ, and them: INTP) it's more important to get the correct answer (objective T) than any subjective consideration.

 

[nightning]

Thank you for doing the mathematical analysis. That's the general pattern I thought make sense.