# Thread: A bit of research: type frequency

1. ## A bit of research: type frequency

Are people equally likely to belong to each type, or are some types more common than others? I'm working on actually evaluating this.

What I need is the types of your family members. No need to state what relationship each person has to you, only their type. Relatives are preferred, spouses and in-laws are OK. Don't provide types of friends, because this may result in a bias toward a uniform distribution of types.

I have already gathered some information from the "Within the family" thread but I still need data from forum members belonging to the following types to conduct any analysis: ENFj, ISTj, ESFp, ISFj, ESTj, ISTp. Information from everyone else is quite welcome, as it will improve the accuracy of the results.

For those who are curious, I am finding the stationary distribution of a Markov chain, on the state space containing all 16 types, with the right stochastic transition probability matrix P. p_i,j is defined as the probability that a person of type i has a family member of type j. I assume that this is in fact a stochastic process with the Markov property, and also that after taking into account sufficient data, the chain will become irreducible and ergodic so that a unique stationary distribution exists.

2. I'd say the distribution is fairly even, although I would say more extroverts than introverts, and probably more rationals than irrationals.

3. How big is your sample? Your methodology is cool imho, I like it because you can actually run experiments (I think though they are still limited to a whole subset of the population, in other words you cannot take a given couple and try to predict the outcome of their mating in a stable way), but it definitely requires a shitload of data. I'm sure that if you are good at typing you can find it though. Speaking for myself - I have not found such a correlation, by visual approximation

4. I do think some types are more common than others, but I won't get into details.

5. Let me flesh this out in more detail.

A sample is probably not the right way to describe it here, as I will explain. Rather than performing a statistical test to determine the distribution of types in the population, which I can't even think of a way to do, I am fitting a probabilistic model to the observed density of intertype relationships.

Basically, to visualize how it works, consider all (or most) people being connected by family relationships. Each person has a type and many family relationships to people of other types.

By taking only family relationships I hope to avoid bias due to people selecting their friends, co-workers, and so on by basing their decisions in part on personality, and by extension type.

By estimating the probability of all people of types A knowing someone of type B and creating a smaller, random graph with probabilities estimated by data, I hope to approximate the actual real-world graph of interpersonal relationships. Then, say someone from every type seems to know a great number of ENFps. We can conclude in turn that ENFps are more dense in the population.

The Markov chain in particular provides a very easy method of calculating this long-run or "stationary" distribution of a random walker on the smaller, random graph, who moves at every time step and has a certain probability of going from say, INTj to ENTj through a family connection.

This is an unorthodox use of such a model, and there may be a more refined or correct way of doing it in graph theory, but I'm not aware of one. My intuition says it will be a fair approximation, if the pattern of the true graph is relatively homogeneous everywhere.

Each row vector of the transition matrix has its probabilities estimated from the data using a multinomial maximum likelihood estimation technique. In English, take the number of connections observed from INFp to each other type, divide by the total number of INFp connections observed. Some parts of the random graph will be more accurate than others due to more data being available, and their accuracy will "smooth out" the inaccuracy of the others in the long run. Because we're not interested in the transitions in the end, only how often the random walker visits the node of each type.

6. ENFj, INFj, INTj, ISTj (?)

7. you also have to trust everyones typings. : /

8. Thanks Logos. I had the first three from the other thread but the last one helps too.

As far as trusting people, I think that the individual bias will smooth out with large numbers of people providing data for each type's most frequently observed connections. The biases for each type are taken care of by the model.

9. ESE, SLE, don't know about my father.

10. Which sofware are you going to use to conduct the analysis?

11. INTj, ESTp, ESTj, INFj, INFp

12. ESTp ISTj ISFj ESFp

13. INTj, ISFj, ESTp, INFj

14. ISFp, ESTj, ESFj

15. Sounds interesting Ncassidy

ISTj, ISFj

16. This should be cool. I think the individual biases and typing errors will tend to even out, unless there are systemic typing errors among forum members.

Here's a fairly large set of types from among my relatives, using the 3-letter code (sorry). If the sample is too large, just take the first line, I guess.

IEE, SEI, IEE, EIE, EII
LSE EII LSE LSE EII LSE
LSE EII LSE EIE
SLE LII ESI SLI SLE IEE LSE IEE ILI IEI
ILI ESI ILI ILI SEE
ESE SEI SEI SEI SLI
ESI SLI SLI SEE

17. The types in my biological family: ESFj, INTj, INTp, INFp, ISTj.

18. Some other families around:

ESI mother SLI father SLE son IEI daughter
SEE mother LSI father LSI son EIE daughter LSE daughter
LSE father SEI mother ESI daughter SLI son

19. My family: INTj, ESFj, ENFp

Relatives
ENFp, ISFp, ESTp, ENFj
ENTp, INTj, ESFj, INTp

20. my two cents:

immediiate: ESTp (myself), ISFj, ENTj, INTp, ESTj, ENFj, ESFJ.
others: ISTj, ESFp, ENTp, ISTp.

21. these typings should all be considered suspect. i will bold the ones i am definitely sure of. take them as you will.

LIE, ESI, ILI

LSI, SLI, two kids that are too young to be typed.

ESE, alpha NT, SEE

other random relatives in no particular order
SEE, SEI, ESE, ESE, IEE, LIE, LSE, ESI, ILI

22. Immediate:
Me - ILE
Sister - EII
Father - EII
Mother - LSE

Extended:
Brother-in-law: ESE
Grandfather (M) - LSE
Grandmother (M) - SLI
Grandfather (P) - ESE
Grandmother (P) - EIE

Aunt - SLI
Uncle- EII
Son - IEE
Daughter - ILI

Aunt - EIE
Uncle - SLE
Son - IEE
Son - SEE

Other families I know:
Son - SLE
Mother - ESE
Father - ILE

Son - LII
Father - LSI
Mother - ESE

Son - LII
Son - SEE
Father - EII
Mother - ILI

Son - IEI
Mother - SEI
Father - LSI

(my favorite family EVER)
Son - SEI
Mother - IEI
Sister - SLE

Son - IEI
Mother - ILI
Father - ILI

Son - ILE (x2)
Daughter - IEI
Mother - IEE
Father - ILI

Mother - ESE
Father - LII
Daughter - SLE
Son - SEI

Son - ILE
Father - LSI
Mother - ESE

23. ISFj, ENFj

24. I'm keeping the data in Excel and I'll be solving the matrix equation either with Matlab or R or by hand, nothing sophisticated.

Also ladies and gents, some of you have not included your own type, which I need separately from the family members to be able to put your information in the right row. If it's in your signature I got it.

25. Me: probably ENTp

Father: ENTp

Mother: ESFj

Sister: INFj

Brother in law: ENFp

Father's side:

Grandmother 1: ISFj

Uncle 1: ESTp

Uncle 2: ISTp

Uncle 3 (deceased): ISFj

Cousin: ESTp

Grandmother 2: ISTj

Aunt: ENFj

Cousin: ENFp

Aunt: INFj

(her husband) ESTj

cousin: ENTp

cousin: ESTp

Uncle: XSTx (beta)

(his wife): INFp

cousin: ENFj

26. ENTJ, ESTP (dad), ISFJ (mom)

I am 100% sure about these typings. I have yet to type my sister accurately as I rarely get to see her.

27. i could probably give much more accurate typings of people like my teachers, friends and other random people, because i make a point of seeing my family (aside from my immediate family, which essentially consists of me and my dad) as little as possible.

28. My family:

Males:
ENFp
ESFj
INTj

Females:
ENTp
ENFj

Other families I know:

Males:
ESFj
ISFp
ESTp

Females:
ESTj
ESFp
----------------------
Males:
ENTp
INFp

Females:
ESFj
----------------------
Males:
ESTj
ISTj

Females:
INFj
ISFp
----------------------
Males:
ENTj

Females:
ISFj
(children unknown, too young to type)
----------------------
Males:
ESTp
ESTp

Females:
ESTj
----------------------
Males:
ESFp
ESFp
ESFj

Females:
ISFj
ISFp

29. mother: LSI
father: ILE (possibly LII)
me: LII
my sister: ESE
my brother: ILI

maternal grandmother: ESE
mother's sister: ESI
mother's half-brother: EII
his wife: IEE

father's brother: SEE/SLE

father's sister 1: ESE
her husband: LSI
their son 1: SLE
their son 2: SEI
his wife: ESE

father's sister 2: SEI
her husband: ILE (possibly ESE)
their son: ILE
their daughter: LSI
her husband: SLI

father's sister 3: LSI
her long-term boyfriend: EIE

father's sister 4: ESI
her husband: ILI

After writing all this up, it occurs to me that only immediate family would be useful to you. Although I think that analyzing only parental relations makes more sense, statistically, and considering genetic factors.

I actually have calculated some statistics for my own typings--not just family, but all of them. The closest thing I have to a random sample is my high school grade's population (but it's not random, for a very good reason), and here it is:

I: 0.66
E: 0.34
N: 0.44
S: 0.54
T: 0.32
F: 0.67
J: 0.38
P: 0.6

INTJ: 0.066
ESFJ: 0.074
ISFP: 0.154
ENTP: 0.037
INFP: 0.096
ESTP: 0.037
ISTJ: 0.029
ENFJ: 0.037
INTP: 0.059
ESFP: 0.074
ISFJ: 0.088
ENTJ: 0.007
INFJ: 0.074
ESTJ: 0.007
ISTP: 0.074
ENFP: 0.066

Alpha: 0.331
Beta: 0.199
Gamma: 0.228
Delta: 0.221

N=136

Not really even. Very few ESTjs and ENTjs--which might be part of my typing bias. The statistics are noticeably different for the celebrities and fictional characters I've typed. The entire sample is much more even.

Why so many ISFps? I don't know. I go to a selective and competitive school, and I think it may have something to do with the ISFp's hidden agenda. Also, it's harder for me to differentiate between IF types (and ET types).

This comes from a Python program I wrote. I plan to incorporate Reinin dichotomies soon. The program is getting quite complicated; it extracts the typings from a text file according to a simple syntax. It can also do relationship stuff. I can send it to anyone who wants it; I think it makes it easier to handle large samples of typings.

Here is my total type distribution (people I personally know + celebrities + fictional characters):

I: 0.53
E: 0.46
N: 0.45
S: 0.53
T: 0.49
F: 0.49
J: 0.49
P: 0.49

INTJ: 0.105
ESFJ: 0.083
ISFP: 0.104
ENTP: 0.054
INFP: 0.051
ESTP: 0.061
ISTJ: 0.058
ENFJ: 0.044
INTP: 0.056
ESFP: 0.054
ISFJ: 0.056
ENTJ: 0.038
INFJ: 0.037
ESTJ: 0.059
ISTP: 0.052
ENFP: 0.058

Alpha: 0.346
Beta: 0.214
Gamma: 0.205
Delta: 0.206

N=781

Still a noticeable Alpha bias, mostly from celebrities. Probably because I enjoy/am better at spotting Alpha celebrities. Otherwise very even.

30. Originally Posted by pezzonovante

I am 100% sure about these typings. I have yet to type my sister accurately as I rarely get to see her.
I wonder why in hell the isfj estp marriage is so common?

31. thehotelambush: Quite a lot of work you've done there! What does your program do exactly, summary statistics for a formatted text file? Python is an interesting language to have picked up so early, did you use it in a course, or learn it on the side?

niffweed17: Still don't have your type! I think we could extend the frame to include anyone for whom interaction isn't strictly voluntary. Self-selection of teachers, professors, bosses, and so on would probably contribute around the same level of bias as any possible effects of genetics or upbringing. So feel free to give that info if you'd like, especially if you already have it available.

Everyone: Thank you to all those who've contributed so far. The overwhelming majority of the data points I've collected have been from ENTps and ENFps due to their prolific contributions and high visibility on the board. In this type of analysis, the chain is only as good as its weakest link (pun not intended)... if you belong to another type, especially the aforementioned six for which no information has been found, your contribution would be much appreciated! Please take the time to chip in

32. Originally Posted by FDG
I wonder why in hell the isfj estp marriage is so common?
Because the Moron is right ...?!

Originally Posted by Keirsey
Protectors and Promoters are apt to find their best match with each other. The outgoing, tough-minded Promoters (ESTPs) tend to have high periods during which they are in a whirlwind of euphoric activity, and the seclusive and friendly Protectors (ESFJs) enjoy preparing a quiet place for the high-rolling entrepreneur to crash. [...]

Promoters and Protectors are likely to get along famously. It is great fun for the tough, outgoing Promoters (ESTPs), born gamblers and wheeler-dealers, to see if they can win over the reserved and friendly Protectors (ISFJs) and persuade them to let go of their caution and concern for a little while, and take a few risks.

33. Originally Posted by Phaedrus
Originally Posted by FDG
I wonder why in hell the isfj estp marriage is so common?
Because the Moron is right ...?!

Originally Posted by Keirsey
Protectors and Promoters are apt to find their best match with each other. The outgoing, tough-minded Promoters (ESTPs) tend to have high periods during which they are in a whirlwind of euphoric activity, and the seclusive and friendly Protectors (ESFJs) enjoy preparing a quiet place for the high-rolling entrepreneur to crash. [...]

Promoters and Protectors are likely to get along famously. It is great fun for the tough, outgoing Promoters (ESTPs), born gamblers and wheeler-dealers, to see if they can win over the reserved and friendly Protectors (ISFJs) and persuade them to let go of their caution and concern for a little while, and take a few risks.
No thanks. I don't want to marry an ENTp.

34. Keirsey wrote:
Protectors and Promoters are apt to find their best match with each other. The outgoing, tough-minded Promoters (ESTPs) tend to have high periods during which they are in a whirlwind of euphoric activity, and the seclusive and friendly Protectors (ESFJs) enjoy preparing a quiet place for the high-rolling entrepreneur to crash. [...]

Promoters and Protectors are likely to get along famously. It is great fun for the tough, outgoing Promoters (ESTPs), born gamblers and wheeler-dealers, to see if they can win over the reserved and friendly Protectors (ISFJs) and persuade them to let go of their caution and concern for a little while, and take a few risks.
This sounds a bit more like LIE + ESI to me...

35. Originally Posted by Rick
Keirsey wrote:
Protectors and Promoters are apt to find their best match with each other. The outgoing, tough-minded Promoters (ESTPs) tend to have high periods during which they are in a whirlwind of euphoric activity, and the seclusive and friendly Protectors (ESFJs) enjoy preparing a quiet place for the high-rolling entrepreneur to crash. [...]

Promoters and Protectors are likely to get along famously. It is great fun for the tough, outgoing Promoters (ESTPs), born gamblers and wheeler-dealers, to see if they can win over the reserved and friendly Protectors (ISFJs) and persuade them to let go of their caution and concern for a little while, and take a few risks.
This sounds a bit more like LIE + ESI to me...
Still, he seems to have the types right, even if the description points towards LIE (maybe)

Yeah anyway, I don't know Phaedrus. I know myself that it's far easier for me to attract ISFjs than INFps, irl. Or rather...INFps are harder to keep.

36. Originally Posted by eunice
No thanks. I don't want to marry an ENTp.

37. My family

Parents:
ESFj, ISTp.

Three kids, (male
ESFp, INFp, INTp.

38. There are an awful lot of mothers who are xSFj around here. I have a feeling that mothers are the most mistyped species on earth...

39. Originally Posted by Kim
There are an awful lot of mothers who are xSFj around here. I have a feeling that mothers are the most mistyped species on earth...
eheh, i agree

40. Originally Posted by ncassidy
i included myself as being an ILI in my immediate family, although not completely sure of this even still

I think we could extend the frame to include anyone for whom interaction isn't strictly voluntary. Self-selection of teachers, professors, bosses, and so on would probably contribute around the same level of bias as any possible effects of genetics or upbringing. So feel free to give that info if you'd like, especially if you already have it available.
well, i would say that there might be a slight bias in that i might tend to pick math and science classes, and so the non-required classes that i take might be slightly more biased towards NT types.

nonetheless, here goes a long list of past teachers, again bolded those that i am sure of, which will make up a greater percent of these.

SEE, LIE, LIE, SEI, LSI, EII

ILI, ILI, LII, SEE

LSI, SEE, LII, LIE, EIE

EII, ILE, LII, ESI, LSE

ESE, IEE, ESI, SLI, SEI, ILE, LSE

Page 1 of 2 12 Last

#### Posting Permissions

• You may not post new threads
• You may not post replies
• You may not post attachments
• You may not edit your posts
•