Equal distribution assumption

[Phaedrus]

Even putting the fact that those are MBTT statistics aside, there's no way to know how accurate the information used in those statistics is (in other words, no way to know if those people had been accurately typed).

That you are unable to understand simple statistics is a remarkable phenomenon. Hardly what one would expect from an ENTj. But if you can't comprehend statistical facts, you can't comprehend statistica facts.

Of course men are going to generally appear (and test) less "feelery" than not. That doesn't make them logical types.

Men not only appear less "feelery" than women, they are also measured to be less "feelery". We know that a typical man's brain is different from a typical woman's brain, and one tested and indisputably existing difference is the difference that is captured in the T/F dichotomy, which is the same in both Socionics, MBTT, and the Five Factor Theory. It is totally impossible that men are not more logical (T) in general than women. It is an empirical fact that you just have to accept.

As far as whether there are more sensory types than intuitive types, it's quite possible.

It is not only possible, it is a proven fact. This is not an issue that is open for debate. You just have to accept the facts.

So anyways, based on what I've observed, the overall distribution of clubs is as follows: ST/SF > NF > NT.

People's personal observations about the distribution of the types are totally irrelevant and totally unreliable. We should pay no attention at all to people's experiences here. The only relevant thing is observations on a larger scale, and to do those we need test results. And we already have such test results.

Does it seem like there are more irrational types than rational types? There seem to be fewer EJ's and IJ's than EP's and IP's among the people I know. That could just be a perception based on my noting irrationality more than rationality though, similar to what I said about quadra values.

That's nothing but a totally irrelevant personal observation that has no value. Whatever you happen to observe in your own surroundings can be the result of pure chance.

Bleh. Anyways, it doesn't really matter to me how common types are. There's no way to KNOW, and even if there was it wouldn't really make a difference. (If there was a way to KNOW, I would be more curious though.)

You incompetent fool. Knowledge in these matters is very easily obtained -- at least for some people.

[Joy]

You should strongly assert that you're correct some more. And that your concept of what is accurate (and what is not) is infallible. Make sure you use a lot of absolutes.

[Logos]

You should strongly assert that you're correct some more. And that your concept of what is accurate (and what is not) is infallible. Make sure you use a lot of absolutes.

[mikemex]

Because I am loyal to the objective truth.

Seek for a therapist's help because you suffer from low self esteem. Don't believe me? Just observe how you've become an attention seeker. You find it difficult to stop compulsively answering people once you think you've got their attention. And even if you disguise your discussion as a "search for the truth", it becomes rather obvious that you don't have a clue of what objectivity means because you seem to believe that just because you read this or that it is automatically true.

Behind all your comments there is a hidden intention to prove yourself that you can hold an intellectual discussion. But since you focus more on appearing that you know what you're talking about (the impression you give to others) than in what you're actually saying, you undermine your chances of actually holding a high level conversations. It's something like the negative and positive loops of the enneagram, something like this:


http://mikemex.t35.com/socionics/esteem.gif

[Phaedrus]

Seek for a therapist's help because you suffer from low self esteem. Don't believe me?

No, I don't believe you.

Just observe how you've become an attention seeker.

Many people on this forum incorrectly assume that I am an attention seeker. And it doesn't seem to matter what I say to correct them -- they stick to that false belief anyway.

You find it difficult to stop compulsively answering people once you think you've got their attention.

I tend to correct people even when I haven't got their attention. It has nothing to do with the people involved. It has to do with the statements they make. I don't care much about people, but I care about the objective truth. I am kind of allergic to blatantly false statements.

And even if you disguise your discussion as a "search for the truth", it becomes rather obvious that you don't have a clue of what objectivity means because you seem to believe that just because you read this or that it is automatically true.

That is another of those false statements I am allergic to.

[Phaedrus]

Skeptical? I think it is utter crock and always am rolling my eyes behind my computer monitor when I see people make arguments based on type distribution 'statistics.' I mean, hell, look at how many people are questioning their type, other people's types on this one forum. Nope. I think to assert any claim about type distribution is nut-fuckery at best. I mean, it makes no sense whatsoever.

And the reason it doesn't make sense to you is because you don't understand probability theory and statistics. What I have said so far in this thread about the distribution of the types is necessarily true. If you don't understand it, you should study some mathematics.

[Phaedrus]

I barely read through the stuff on this thread but I was simply replying to Salawa's comment since it had to do with my thoughts on this thing for a long time now (it started when I went from MBTI where there a lot of stats to socionics and now I don't believe in any exact numerical statistics but equal distribution makes sense on a theoretical level.)

Salawa made a stupid comment, and the argument that an equal distribution of the types makes sense on a theoretical level is even more stupid. The big mistake you and others (like Rick) are making is to ignore reality and stick to your subjective understanding of the model. There is empirical data to not only support but to prove that the types are unequally distributed. The only support for the false equal-distribution-argument is your own prejudices.

And I am horrible at math so not a chance there.

Then don't have an opinion. Try to find the truth instead of forming beliefs that have no empirical support.

If you have a belief about it, fine, but I choose to think it's something that cannot essentially be proven because of the subjectivity of personality theory in the first place.

I don't have a belief, I have knowledge. I know that the types are unequally distributed, because it is a proven fact. You must be dumb as a bat if you don't understand what I say here. And you admit that you don't understand math, so try to learn that first. It is essential if you are gonna understand how it can be a proven fact that the types are unequally distributed.

We are making vague assumptions about the matter at best, from my understanding, as flawed as my understanding probably is, but at least I am being honest with myself.

You damn idiot. You understanding of this is definitely flawed, and yet you defend your stinking attitude by saying that you are honest with yourself!? That's simply disgusting. You should be more concerned with the objective truth than with your personal feelings here.

[Joy]

I have an idea. We should have 10,000 people roll Bionicgoat's type dye to determine their types. Once we've recorder the results, anyone who understands probability theory and statistics will know how types are distributed (and anyone who doesn't should study some mathematics). There's absolutely, positively no way the results could be disputed.

[Joy]

If MBTI testing correlates with MBTI descriptions, and there is correlation when socionists and the general public compare sets of MBTI and socionics descriptions (as has been demonstrated on Lytov's site)

I'm rather skeptical about the accuracy of the typings, tbh.

It's interesting that you see a lot of LIE's. I have wondered if my identicals may stand out to me the least of all types. I can only think of a handful that I've known though.

[Joy]

http://translate.google.com/translate?u=http%3A%2F%2Fwww.famo.ru%2Flitovi.php& sl=ru&tl=en&hl=en&ie=UTF-8
http://64.233.179.104/translate_c?hl...GALsZzrMMN6gSA

[Joy]

I don't know about the distributions... my point was more related to the convergence between type lists of professional Socionists.

btw, Loytov's test showed that I was clearly ILE.

[Joy]

Okay... so the difference here is that I put more weight on quadra values and you put more weight on the E/I, S/N, T/F, P/J dichotomies...

[Phaedrus]

Okay... so the difference here is that I put more weight on quadra values and you put more weight on the E/I, S/N, T/F, P/J dichotomies...

Lytov's test, as well as most other socionic tests, tries to determine the four dichotomies, because that is the most reliable method to use if we want to find a person's correct type. Going by quadra is a much less reliable typing method, and that has also been confirmed by evaluations of various typing methods.

[Joy]

Lytov's test, as well as most other socionic tests, tries to determine the four dichotomies, because that is the most reliable method to use if we want to find a person's correct type. Going by quadra is a much less reliable typing method, and that has also been confirmed by evaluations of various typing methods.

So I'm ILE again? Sweet. Look out Alpha, here I come!

[Phaedrus]

So I'm ILE again? Sweet. Look out Alpha, here I come!

You are obviously an idiot. Either you are an ILE or you have got an incorrect result on Lytov's test. Is that too simple for you to understand?

[Joy]

You relative to I, that is possible. The dichotomies relative to each other -- I would say I approach these ones about equally, with variation depending upon the availible evidence and application. I discuss quadra value dichotomies heavily in relation to subtype theory, for example.

Even if all quadra values are obvious, one still must choose between a set of 4. Temperament and club can largely explain how one seperates functional usage within the quadra. If a test at least gives you 4 or 5 of the aforementioned dichotomies correctly -- the test gives you a lot of valuable information. At the very least, you would have narrowed down the type and function descriptions you will most want to read in narrowing the typing itself down further.

also bear in mind that we were discussing T/F at length earlier. whether or not the foundational dichotomies are most important for in-practice typing, test result with respect to the foundational 4 dichotomies is still relevant data for what we were discussing.

I don't disagree with you. I guess I just tend to shy away from typing by dichotomy because I've witnessed the results... People who are obviously Fe/Ti insisting that they're ILI, for example.

I also don't think testing is a good way to determine someone's type. It can provide a point from which to start evaluating types, but too often people become convinced by test results that are wholly inaccurate. My first MBTI test result, for example, was INFJ. Test results can only be as accurate as a person's understanding of themselves.

Also, test results tend to focus on common manifestations of functions (instead of on information elements themselves), and the same behavior can be a manifestation of different functions in different people. Environmental factors can also play a large role. Tests that do focus on the information elements themselves tend to be too abstract for most people and aren't any more useful.

The afforementioned problems with tests also apply to most type descriptions.

The bottom line is that study of the theory is the first step to typing oneself or others. Once someone understands quadra values conceptually and witnesses their own and other people's interactions and behavior in a setting in which they're not trying to determine their type, then they'll have a working understanding of quadra values. Once someone has a working understanding of quadra values, then they'll be able type others with relative accuracy. However, even the people who are best at typing will not have perfect accuracy rates, and their accuracy rates are sure to be even lower when they're typing people based on tests and interviews instead on observation of someone in his/her day to day life. This is especially true for individuals whose behavior changes over time as the result of environmental factors or their level of wellbeing. (Studying biographies and television/magazine interviews to type a person is also problematic for all of the above reasons.)

Dichotomies are useful when someone is obviously intuitive, obviously ethical, etc., but most of the time not all of the dichotomies are obvious.

Using myself as an example, the only dichotomy that's entirely unmistakable is N/S. T/F is more apparent when evaluating my behavior in day to day life and the types of problems that I've had over the years related to strengths/weaknesses, but E/I is not entirely apparent. Also, according to pop MBTT descriptions/tests I could easily be seen as a P type. It's only by observing me when I'm at my healthiest/happiest and evaluating dual pairs that my temperament becomes apparent.

The bottom line is that I don't believe it's possible for anyone or any method to type people quickly and accurately enough to gather information that would be valid for creating reliable statistics.

[Ezra]

What ifmd95 said (the bit Joy quoted at the top of page 7) about Lytov's site is wrong. Firstly, it's not the general public or their views that are listed in that table; it's a circle of socionists. Secondly, even if it was the general public of whom Lytov tallied the opinions concerning socionics/MBTI correlations, why should we take their views into consideration? Why the fuck should I trust a typical Russian member of the public over my own generally well-researched ideas and opinions?

[Expat]

I agree with Joy's post overall, especially with this:

I guess I just tend to shy away from typing by dichotomy because I've witnessed the results... People who are obviously Fe/Ti insisting that they're ILI, for example.

[Joy]

You are obviously an idiot. Either you are an ILE or you have got an incorrect result on Lytov's test. Is that too simple for you to understand?

I obviously got an incorrect result. I'm sure a lot of people do. That's my point.

[Phaedrus]

I obviously got an incorrect result. I'm sure a lot of people do. That's my point.

And if you got incorrect test results, you are not likely to type yourself correctly by other methods. You don't know which type you are until you get the correct test results, and you don't understand the types if you get incorrect test results.

[Logos]

There seems to be quite a number of those people.

[Ezra]

And if you got incorrect test results, you are not likely to type yourself correctly by other methods. You don't know which type you are until you get the correct test results, and you don't understand the types if you get incorrect test results.

Why do you always assume the test-maker to be infallible, and people's own perceptions to be nearly always flawed?

Firstly, there are several tables. (Expat has the "circle of socionists" comparisons archived here: http://www.the16types.info/vbulletin...d.php?p=210440) Secondly, Lytov hasn't been my only source for MBTI correlations. I listed him as one example.

Yeah, I agree with that table. The only people who identified with SLE were xNTJs. I think this is right if you read about NTJs in MBTT.

The point in question was the varacity of self-response tests. Identifying oneself in a description is likewise a form of self-response, and a more detailed form as stated earlier. To see convergence in self-response from an increasing number of angles, the self-respondant must be doing something right.

Okay, fair point.

[Phaedrus]

Why do you always assume the test-maker to be infallible, and people's own perceptions to be nearly always flawed?

Most socionic tests are based on the four dichotomies. In order to understand the types, you need to understand the four dichotomies -- and you need to understand them correctly. If you understand the four dichotomies correctly, you will get the correct test result on every test you take. And what this demonstrates is that almost every single test is a more accurate and reliable indication of your correct type than anything you can come up with yourself as far as typing methods are concerned.

A person who doesn't understand the four dichotomies will mistype him- or herself -- inevitably. And a person who does understand the four dichotomies will always know for sure which type he or she is -- and that is the same thing as getting the correct result on a typing test.

[Joy]

And if you got incorrect test results, you are not likely to type yourself correctly by other methods.

I didn't, at first. I didn't understand Socionics or myself all that well, so how could I?

I didn't figure out my type until I had tried a number of other types on for size. It wasn't until I started looking at quadra values (duality in particular) that my type became apparent. Since then I've considered a few other types when they were brought up (since I see it as extremely foolish to believe that there's no possibility that you could be wrong), but none has made as much since as LIE.

You don't know which type you are until you get the correct test results

?????

and you don't understand the types if you get incorrect test results.

This statement is very telling. You're right, someone who understands the types is able to get any test result they desire. That has nothing to do with what type they actually are though.

In order to get a correct test result, two things are necessary. First of all, the test has to be good. Secondly, the person has to know him/herself.

[Joy]

Most socionic tests are based on the four dichotomies. In order to understand the types, you need to understand the four dichotomies -- and you need to understand them correctly. If you understand the four dichotomies correctly, you will get the correct test result on every test you take. And what this demonstrates is that almost every single test is a more accurate and reliable indication of your correct type than anything you can come up with yourself as far as typing methods are concerned.

A person who doesn't understand the four dichotomies will mistype him- or herself -- inevitably. And a person who does understand the four dichotomies will always know for sure which type he or she is -- and that is the same thing as getting the correct result on a typing test.

The order of business you're describing here... let me see if I've got this straight...

1.) A person understands the dichotomies.
2.) A person knows their type, based on their understanding the dichotomies.
3.) A person confirms their type by knowing how to answer test questions so that their result is the type that he/she already knows him/herself to be???

[Phaedrus]

You're right, someone who understands the types is able to get any test result they desire. That has nothing to do with what type they actually are though.

You are wrong. A person who understands the types and the four dichotomies will have no desire to get a result that does not reflect his correct type. So it has everything to do with the type they actually are.

In order to get a correct test result, two things are necessary. First of all, the test has to be good. Secondly, the person has to know him/herself.

Yes, but that's trivial. If you don't know yourself, you will not type yourself correctly. But if you do know yourself, you will type yourself correctly -- if you understand the four dichotomies. And 9 out of 10 tests are sufficiently good for typing purposes. Almost every test is accurate enough to give you your correct type, if it is based on the four dichotomies.

[Joy]

Phaedrus applies personality theory in general inadequately. I don't think he is a typical counterexample. Perhaps you aren't, either. You did go through what seems like an above-average number of different self-types, but you aren't the first person to consider test results.



These are problems with any condensed "snapshot" of interaction, whether one condenses it to a test question, a description, or a narrative – yes. I have argued before that intertype relations are one of the strongest empirical basises for Socionics.

However there are personality theory applications where it is difficult to gather sufficient data in-depth and hands-on. These include distributional statistics, or even to an extent assigning any external characteristic like VI. It's possible these are fruitless pursuits worth conceding. But it's also possible to make up for loss of quality with quantity. This if there isn't some systematic bias. But we can check against biases by considering multiple detached sources and how they compare. I think what I said to Ezra about convergence and the materials applies here too.



I think I covered this earlier: "Even if all quadra values are obvious, one still must choose between a set of 4. Temperament and club can largely explain how one seperates functional usage within the quadra." To understand how a quadra together achieves its values, one still must identify interactions related to the foundational dichotomies.



My bottom line is that while I agree methods besides understanding intertype interaction have limited use in individual typings, there may still enough information available to make unbiased inferences about groups and the Socion itself, where accuracy less depends upon the precision of individual typings.

At this point I'm wondering what the value of distribution statistics even is?

I mean, it seems like a lot of trouble for something that isn't even completely reliable. Even if it was completely reliable, it still seems like a lot of trouble, given the amount of effort necessary to correctly type even one person.

If there was a fast and reliable way to determine people's types, such as a blood test, I would say that we should get the largest samplings possible in as many areas of the world as possible and compile statistics on type distribution in different areas, industries, among married couples and families, within various races, etc. (Even then we'd have to make sure we got a random sampling, as it is possible that some types are more apt to have the blood test done than others.)

Since there is not, however, I'm not sure that our "good enough" statistics that require a lot of effort due to the amount of time/effort required to type even one person are of enough value to merit the compilation and study of them.

[Phaedrus]

The order of business you're describing here... let me see if I've got this straight...

1.) A person understands the dichotomies.
2.) A person knows their type, based on their understanding the dichotomies.
3.) A person confirms their type by knowing how to answer test questions so that their result is the type that he/she already knows him/herself to be???

If you understand the four dichotomies, you don't need to take a test, because then you already know how to type yourself -- just by reading about the four dichotomies and determine your four letter combination. There is no more reliable typing method available than the four dichotomies if you are a beginner, taking your first steps in typology. And you can never ever find out that you are some other type than the type that is in accordance with your understanding of the four dichotomies, because the four dichotomies ALWAYS coincide with your correct type.

[Joy]

You are wrong. A person who understands the types and the four dichotomies will have no desire to get a result that does not reflect his correct type. So it has everything to do with the type they actually are.

I can think of a few examples of someone wanting to be a certain type. Often it's because they either want to see themselves as strong in a certain area or they want to explain their weaknesses.

Yes, but that's trivial. If you don't know yourself, you will not type yourself correctly. But if you do know yourself, you will type yourself correctly -- if you understand the four dichotomies. And 9 out of 10 tests are sufficiently good for typing purposes. Almost every test is accurate enough to give you your correct type, if it is based on the four dichotomies.

Many people do not know themselves all that well. That is the primary reason why test results are unreliable.

[Joy]

If you understand the four dichotomies, you don't need to take a test, because then you already know how to type yourself -- just by reading about the four dichotomies and determine your four letter combination. There is no more reliable typing method available than the four dichotomies if you are a beginner, taking your first steps in typology. And you can never ever find out that you are some other type than the type that is in accordance with your understanding of the four dichotomies, because the four dichotomies ALWAYS coincide with your correct type.

I don't believe you. You should use more absolutes, and then maybe I will.


Seriously though, there are far too many bad descriptions of the dichotomies to make this valid. And there's still also the matter people's incorrect self-perceptions.

[Phaedrus]

I can think of a few examples of someone wanting to be a certain type. Often it's because they either want to see themselves as strong in a certain area or they want to explain their weaknesses.

Those persons are not objective, and they don't understand the types. The given premise is that you do understand the four dichotomies and want to find your correct type. If you don't have that desire, you will live a life in delusion.

Many people do not know themselves all that well. That is the primary reason why test results are unreliable.

Many people don't know themselves very well, and that is a major source of mistypings, especially on this forum. Too many people are simply idiots in that respect. But you are totally wrong about test results being unreliable. Test results are much more reliable than people's self reports, and when you have such a large sample of test results (several millions), the individual's inability to be objective becomes irrelevant. The overall pattern is very accurate and objective, and it reflects the distribution of the types extremely well, even though a single test results can be wrong quite often.

[Joy]

Those persons are not objective, and they don't understand the types. The given premise is that you do understand the four dichotomies and want to find your correct type. If you don't have that desire, you will live a life in delusion.

*ahem*

The problem is that these people think they are being objective.

Many people don't know themselves very well, and that is a major source of mistypings, especially on this forum. Too many people are simply idiots in that respect. But you are totally wrong about test results being unreliable. Test results are much more reliable than people's self reports, and when you have such a large sample of test results (several millions), the individual's inability to be objective becomes irrelevant. The overall pattern is very accurate and objective, and it reflects the distribution of the types extremely well, even though a single test results can be wrong quite often.

How can people's test results be reliable if their test answers are not an accurate portrayal of what they're actually like?

[Phaedrus]

The problem is that these people think they are being objective.

No, they usually don't think that they are objective. They don't even think in such terms. And it is no problem -- except from the annoying fact that they are idiots.

How can people's test results be reliable if their test answers are not an accurate portrayal of what they're actually like?

People's test results taken together are reliable, but a single test result is not. That is no more strange than the fact that we know for a fact that there are general differences in both personality and skills between men and women, and we know their exact nature too. But from that general knowledge we cannot conclude anything for sure about a single man or a single woman.

If 9 out of 10 people are totally lost and don't understand anyhing about a certain subject and are simply guessing or making things up when they are asked about their opinion, but one person out of 10 is close to an expert and will answer correctly, you will inevitably get a very clear overall pattern from which the correct answer can be deduced. It doesn't matter that 90 % of the participants are wrong, because their results will be random, while the 10 % correct results will have a significant and easily observable influence on the general pattern.

[Joy]

The thing that you both seem to be assuming is that inaccuracies will balance themselves out. This would mean that if 30% of people test as logical types when they're really ethical types, then it's just as likely that an equal number of logical types will test as ethical types.

[Joy]

I'm not sure what you mean by "balance themselves out", but i did discuss bias. Although there may be persistant bias in the samples such as types portraying other types, you can try to "outsmart" it by analyzing from an increasing number of points of view.

If 20% (for example) of ethical men believe themselves to be logical when they are actually ethical, then it doesn't matter how many men or groups of men in the sample. 20% is 20% whether it's 20% of 10 or 20% of 100,000.

i expected gender to have an unequal distribution (not only because of MBTI, but also Jung's writing on gendered archetypes.)

I don't understand how this addresses the aforementioned bias. It sounds like you were assuming that the MBTI and Jung results were correct.

Type descriptions can be quite extensive though, whether it be physiological tendencies, speech construction, or dual-seeking behaviors. If test results can be linked to those, I would think the test results have some more veracity than just the skin-deep appearances

I agree, but that's easier said than done.

While each of these methods are flawed, it is unlikely they are all flawed in the same way.

?

[Phaedrus]

The thing that you both seem to be assuming is that inaccuracies will balance themselves out.

They always do -- at least in situations like the one we are talking about here. The four dichotomies are reflected more or less correctly in all the tests we are talking about, and therefore we will inevitably end up with a correct overall (rough) distribution of the socionic types based on millions of MBTI test results. We would get the exact same general pattern if we used socionic tests instead, because almost all socionic tests are based on the same principles and the same four dichotomies as the MBTI tests. It is simply statistically out of the question and totally impossible that the types are equally distributed. It can never happen.

This would mean that if 30% of people test as logical types when they're really ethical types, then it's just as likely that an equal number of logical types will test as ethical types.

Yes, because otherwise the theory which the test is based on (if it is Socionics or MBTT or both doesn't matter here) is false, and the real types would be different from the way they are described to be. In that hypothetical case we would have to reconstruct the whole theory or at least the four dichotomies.

[Joy]

They always do -- at least in situations like the one we are talking about here. The four dichotomies are reflected more or less correctly in all the tests we are talking about, and therefore we will inevitably end up with a correct overall (rough) distribution of the socionic types based on millions of MBTI test results. We would get the exact same general pattern if we used socionic tests instead, because almost all socionic tests are based on the same principles and the same four dichotomies as the MBTI tests.

We disagree on the premise that testing is a good way to determine types.

It is simply statistically out of the question and totally impossible that the types are equally distributed. It can never happen.

I never suggested that types are equally distributed. That's not the issue at hand.

Yes, because otherwise the theory which the test is based on (if it is Socionics or MBTT or both doesn't matter here) is false, and the real types would be different from the way they are described to be. In that hypothetical case we would have to reconstruct the whole theory or at least the four dichotomies.

Again, the issue is whether testing is a valid means to determine people's types.

[Joy]

Okay Phaedrus, let's take this one step at a time.

1.) Do you believe that individuals will always know themselves well enough to answer test questions accurately?

[Joy]

But... why would a larger sampling improve it if 20% is 20% whether it's 20% of 10 or 20% of 100,000.

[Joy]

My point is that men a larger percentage of men is always going to test/appear logical than the percentage that actually is logical, no matter the size of the sampling or the number of samples.