Return to Socionics.us home  
 
Introduction to Socionics
Socionics Theory
Socionic Types
Intertype Relations
Socionics in Practice
Issues and Philosophy
Famous People's Types
Notes on Socionics Works and Articles
Interviews and Reports
Wikisocion
Socionics Blog
Socionics Tests
Socionics.us
Web



 
pop-up chart
socionic types
   

Socionic Type Distribution Statistics

You may choose to skip this discussion and go straight to the type statistics at bottom.


There are different opinions as to the distribution of socionic types in society. My own firmly held opinion is that they are distributed more or less evenly. Why should we care? I consider the issue of type distribution to be very important because it has the potential to help distinguish between good and bad typing, to avoid drifting too far from the foundations of socionics, and to clarify the differences between socionics and the MBTI and other Jung-based typologies. I have a number of reasons for believing that socionic types are distributed roughly evenly:

1. No consensus regarding the most common types or socionic traits
Among the socionists and socionics amateurs who claim directly or indirectly that the types are not at all evenly distributed, there is absolutely no consensus as to the nature of this unequal distribution, if it exists. This contrasts with Meyers-Briggs Typology, which officially recognizes the large statistical dominance of some types and traits over others. In socionics, however, some people are convinced that ethical intuitive extraverts and logical sensing introverts are most common (Gulenko's school), a few others claim that intuitive logical extraverts are most common ("Typologist" and others), and others have strong typing biases towards different quadras. This suggests that claims of uneven type distribution are more the result of individual socionists' typing quirks than an actual statistical imbalance.

Related to this is the data gathered in Oleg Khrulyev's Celebrity Benchmark List project (discussed in depth at this site). Here we see perfectly clearly that virtually every socionist favored certain types over others in typing celebrities, but the combined results showed no statistically significant dominance of certain types or socionic traits over others (this could also be due to the relatively small sample size) except for logic vs. ethics, which can be explained by the extreme prevalence of men in the list. Some socionists fail to see the importance of this benchmark list and continue to insist on the accuracy of their own quirky typing statistics, even if they show a certain typing lopsidedness that no one else happens to have noticed!

2. Real-life type samples
At bottom I compare four different sets of type samples: Aushra Augusta's random sample of 200 Lithuanians, the celebrity benchmark list, Bukalov and Karpenko's statistics of married couples, and my own sample from my corporate English classes (with information on my sample below the chart).

In gathering a sample to study type distribution, it is important to make sure you are not filtering out certain types of people. For example, the types of your friends or just "people you know" will not be representative of the population at large. You must type people who have not chosen you, nor chosen each other, and you must make sure you type each of them — not just the ones that are easy to type. This is not easy to manage, but it is the only way to get an objective picture of socionic type distribution.

My own attempts at typing people who come across my path more or less randomly suggest that the types are roughly evenly distributed. These are people I meet in my train compartment when I travel, people I meet in my English classes, and other random groups of people I encounter here and there. I make a point of paying particular attention to random situations and the kinds of people I encounter in them.

3. Avoiding typological drift
Augusta's writings suggest that she saw the types as being roughly evenly distributed (for example, see her random sample at bottom). It would seem strange to claim to follow her typology and yet consistently have completely different type distributions. That would imply that the socionist had basically created his own new typology. It would be logical that some of Augusta's followers would eventually come to type "better" than her, but it would be strange indeed to have little overlap with her typings. Strictly speaking, the only way to type "better" than the founder of socionics is to type people in such a way that relationships are explained better than if she had done the typing. That's the only real test of good typing that we have.

4. Logical arguments against irregular type distribution

Theoretically, we would expect type distribution to be roughly even for a number of reasons. Socionics as a typology was intended to explain relationships, and a highly uneven type distribution would mean an ineffecient typology of relationships. Each type would have radically different relationships with the rest of the world, which is something that has not be systematically observed by socionists. Some types would be in extreme demand, while others would be "throw-away" members of society.

Let's just think logically for a moment about what it would mean if 75% of people were, say, sensers...

This would mean that only 50% of people would have any real chances of having activators and duals as close friends and partners (all of the intuiters and one-third of the sensers). The other 50% would be "doomed" to psychological imbalances and constant emotional discomfort because of not having anyone to take care of their super-id functions on a regular basis. All of these people would have problems of a similar nature — a lack of intuitive assistance in life. The intuiters, on the other hand, would be in great demand. There would be on average three duals and three activators per intuiter, and all of them would be in need of that intuiter's assistance. These intuiters would be the "lucky guys" of society, prized by so many people. And what would happen? They would reproduce more compared to the sensers, which would lead to the proportion of intuiters and sensers gradually evening out...

The socion's self-regulating mechanisms
This brings us to the topic of self-stabilizing mechanisms in the socion and in society at large. If we look at the ratio of men to women in the population, we find very minor differences of just a few percentage points from country to country. I am convinced that the same is true of socionic types. The presence of too many traits of a certain kind opens the door for opposing traits to enter and quickly gain a reproductive advantage (I discussed this in my essay "Socionics and Evolution"). In socionics, when the proportion of people of one type in a location becomes too high, the relative worth of each of them drops because the cumulative demand for their type's kind of behavior and information product remains constant despite the growing availability of the type (think supply and demand). The presence of so many people of one type also creates more and more demand for their duals, activators, and other types who can provide complementary "information products" and services to these people. This means that the likelihood is very high that 1) some of the people of the overly common type will leave the group and find a more favorable location, or 2) people of other types will come into the scene to fill the information vacuum and reap the rewards of being unusually useful.

A similar thing happens at the quadra level. Each quadra also represents a kind of specialized type of adaptation to the environment. If there are too many people of one quadra in a location, or part of society, that opens the door for other quadras to come in and reap great rewards by taking advantage of the areas left unnoticed by the dominant quadra.

These self-regulating mechanisms can easily be observed in small and medium-sized groups. For example, I personally have never seen a group consisting only of people of the same type larger than two people! (assuming they were not forced to be together). The repulsion effect and desire to find a more favorable environment for self-expression is simply too great, and, given a choice, people will never stay in such a three or foursome. Such groups have no cohesiveness and can't find anything to hold them together.

In my experience with informal groups (where people choose to participate of their own free will), there always seems to be a nearly equal number of introverts and extraverts. The other dichotomies seem to have a greater degree of flexibility, but they also tend towards a 50-50 proportion unless the group is focused on some specialized task (but then it probably is not an informal group).

Informal groups can be found with as much as a 75%-25% ratio of some socionic trait, but such groups feel imbalanced to their members and hence don't tend to live long. Subjectively it feels like there is "something wrong" with the group. Hence, members begin to look elsewhere until they find new people to add to the old group or leave to join a new group themselves. On the group level there is also a sort of natural selection; only the fittest groups survive.

How applicable are MBTI statistics to socionic types?
The Myers-Briggs Type Indicator shows a prevalence of extraverts (>60%) and sensing types (>70%) in the U.S. population. Unfortunately, I don't have references, but this is common knowledge. Seeing that Myers-Briggs type-watchers generally have quite different ideas about intertype compatibility than socionists, it would seem that the types themselves conceptually do not coincide. Indeed, a number of studies by socionists have shown significant differences in type descriptions. Here is an article attempting to reconcile the MBTI type statistics by Aleksandr Bukalov from 1996 (in fair English).

My conclusions
There has to be a well-balanced type distribution for any group to be cohesive, stable, and competitive. I see no reason for this not to be true at the societal level as well. I would expect self-regulating mechanisms to keep a stable equilibrium of socionic traits, and would be very surprised to see more than 55% of people in a society (any country of the world) share any one socionics dichotomy trait, or for any type or quadra to be more than twice as prevalent as any other.


Socionic Type Samples

Aushra Augusta:

Random sample of 100 Lithuanian men and 100 women.
Source: The Dual Nature of Man

Type breakdown:
Alpha Quadra - 61 Beta Quadra - 47
ILE - 15
M - 11
F - 4
ESE - 13
M - 3
F - 10
EIE - 14
M - 5
F - 9
SLE - 12
M - 9
F - 3
SEI - 22
M - 7
F - 15
LII - 11
M - 6
F - 5
LSI - 9
M - 6
F - 3
IEI - 12
M - 4
F - 8
Gamma Quadra - 36 Delta Quadra - 56
SEE - 13
M - 4
F - 9
LIE - 6
M - 4
F - 2
LSE - 7
M - 5
F - 2
IEE - 16
M - 10
F - 6
ILI - 10
M - 8
F - 2
ESI - 7
M - 3
F - 4
EII - 13
M - 3
F - 10
SLI - 20
M - 12
F - 8

Comments:
Of these fairly evenly distributed numbers what stands out is the high percentage (60%) of irrational types. Also, the difference between the Alpha (30.5%) and Gamma (18%) quadras is quite large.

In addition, Augusta adds that "among 50 randomly selected married couples 17 (34%) turned out to be optimal" (i.e. dual).
  Statistics:
 
Total
Men
Women
no. % no. % no. %
     dichotomies:
rational 80 40% 35 35% 45 45%
irrational 120 60% 65 65% 55 55%
extraverts 96 48% 51 51% 45 45%
introverts 104 52% 49 49% 55 55%
intuitive 97 48.5% 51 51% 46 46%
sensing 103 51.5% 49 49% 54 54%
logical 90 45% 61 61% 29 29%
ethical 110 55% 39 39% 71 71%
     quadras:
Alpha 61 30.5% 27 27% 34 34%
Beta 47 23.5% 24 24% 23 23%
Gamma 36 18% 19 19% 17 17%
Delta 56 28% 30 30% 26 26%
     clubs:
int. + log. 42 21% 29 29% 13 13%
int. + eth. 55 27.5% 22 22% 33 33%
sens. + log. 48 24% 32 32% 16 16%
sens. + eth. 55 27.5% 17 17% 38 38%

Bukalov and Karpenko:

Random sample of 119 married couples from Russia and Ukraine. In addition to type distribution statistics, this report provide statistics of various intertype relations between spouses (socionists did not know who was married to whom until afterwards).
Source: International Institute of Socionics site (Russian only)

Type breakdown:
Alpha Quadra - 54 Beta Quadra - 69
ILE - 17
M - 11
F - 6
ESE - 8
M - 1
F - 7
EIE - 14
M - 3
F - 11
SLE - 23
M - 13
F - 10
SEI - 16
M - 7
F - 9
LII - 13
M - 10
F - 3
LSI - 16
M - 12
F - 4
IEI - 16
M - 8
F - 8
Gamma Quadra - 63 Delta Quadra - 52
SEE - 15
M - 6
F - 9
LIE - 15
M - 13
F - 2
LSE - 11
M - 8
F - 3
IEE - 11
M - 4
F - 7
ILI - 15
M - 11
F - 4
ESI - 18
M - 2
F - 16
EII - 17
M - 3
F - 14
SLI - 13
M - 7
F - 6

Comments:
Very even distribution of socionic traits with the exception of logic (71.4% of men) and ethics (68.1% of women).

In addition, this survey furnished the following intertype relations statistics:
within the same quadra - 64%
duality - 45%
activation - 8%
identity - 6%
mirror - 5%
request - 10%
supervision - 5%
others - 13%
with the opposing quadra - 8%
  Statistics:
 
Total
Men
Women
no. % no. % no. %
     dichotomies:
rational 112 47.1% 52 43.7% 60 50.4%
irrational 116 52.9% 67 56.3% 59 49.6%
extraverts 114 47.9% 59 49.6% 55 46.2%
introverts 124 52.1% 60 50.4% 64 53.8%
intuitive 118 49.6% 63 52.9% 55 46.2%
sensing 120 50.4% 56 47.1% 64 53.8%
logical 123 51.7% 85 71.4% 38 31.9%
ethical 115 48.3% 34 28.6% 81 68.1%
     quadras:
Alpha 54 22.7% 29 24.4% 25 21%
Beta 69 29% 36 30.3% 33 27.7%
Gamma 63 26.5% 32 26.9% 31 26.1%
Delta 52 21.8% 22 18.5% 10 25.2%
     clubs:
int. + log. 60 25.2% 45 37.8% 15 12.6%
int. + eth. 58 24.4% 18 15.1% 40 33.6%
sens. + log. 63 26.5% 40 33.6% 23 19.3%
sens. + eth. 57 23.9% 16 13.4% 41 34.5%

Celebrity benchmark list:

Typings of famous people (mostly from the former Soviet Union) with a high level of convergence among different socionists. Not to be considered a random sample in any sense!
Source: Celebrity benchmark list

Type breakdown:
Alpha Quadra - 44 Beta Quadra - 45
ILE - 10
M - 8
F - 2
ESE - 7
M - 7
F - 0
EIE - 12
M - 12
F - 0
SLE - 14
M - 13
F - 1
SEI - 11
M - 11
F - 0
LII - 16
M - 16
F - 0
LSI - 11
M - 11
F - 0
IEI - 8
M - 8
F - 0
Gamma Quadra - 34 Delta Quadra - 37
SEE - 5
M - 3
F - 2
LIE - 6
M - 6
F - 0
LSE - 12
M - 10
F - 2
IEE - 6
M - 6
F - 0
ILI - 15
M - 12
F - 3
ESI - 8
M - 6
F - 2
EII - 6
M - 4
F - 2
SLI - 13
M - 7
F - 6

Comments:
See here.

  Statistics:
 
Total
Men
Women
no. % no. % no. %
     dichotomies:
rational 78 48.8% 72 51.4% 6 30%
irrational 82 51.3% 68 48.6% 14 70%
extraverts 72 45% 65 46.4% 7 35%
introverts 88 55% 75 53.6% 13 65%
intuitive 79 49.4% 72 51.4% 7 35%
sensing 81 50.6% 68 48.6% 13 65%
logical 97 60.6% 83 59.3% 14 70%
ethical 63 39.4% 57 40.7% 6 30%
     quadras:
Alpha 44 27.5% 42 30% 2 10%
Beta 45 28.1% 44 31.4% 1 5%
Gamma 34 21.3% 27 19.3% 7 35%
Delta 37 23.1% 27 19.3% 10 50%
     clubs:
int. + log. 47 29.4% 42 30% 5 25%
int. + eth. 32 20% 30 21.4% 2 10%
sens. + log. 50 31.3% 41 29.3% 9 45%
sens. + eth. 31 19.4% 27 19.3% 4 20%

My English students:

All my students from corporate English classes in Kiev, Ukraine (this chart will be updated as I gain new students and modify some typings).
Source: my personal records.

Type breakdown:
Alpha Quadra - 16 Beta Quadra - 14
ILE - 2
M - 1
F - 1
ESE - 5
M - 2
F - 3
EIE - 5
M - 2
F - 3
SLE - 3
M - 2
F - 1
SEI - 5
M - 5
F - 0
LII - 4
M - 2
F - 2
LSI - 4
M - 2
F - 2
IEI - 2
M - 1
F - 1
Gamma Quadra - 25 Delta Quadra - 11
SEE - 6
M - 1
F - 5
LIE - 7
M - 5
F - 2
LSE - 1
M - 1
F - 0
IEE - 2
M - 2
F - 0
ILI - 6
M - 4
F - 2
ESI - 6
M - 2
F - 4
EII - 2
M - 1
F - 1
SLI - 6
M - 2
F - 4

Comments:
The sample (66 people) is still too small to draw any serious conclusions, but so far the dichotomies are quite evenly distributed in both men and women, but with a clearly predominant Gamma quadra (38.3% of students), possibly due to the fact that the vast majority of my students are from corporate classes.

More about my sample and where it comes from below.

  Statistics:
 
Total
Men
Women
no. % no. % no. %
     dichotomies:
rational 34 51.5% 18 51.4% 17 58.6%
irrational 32 48.5% 17 48.6% 12 41.4%
extraverts 31 47.0% 16 45.7% 15 48.4%
introverts 35 53.0% 19 54.3% 16 51.6%
intuitive 30 45.5% 18 51.4% 12 38.7%
sensing 36 54.4% 17 48.6% 19 61.3%
logical 33 50.0% 19 54.3% 14 45.2%
ethical 33 50.0% 16 45.7% 17 54.8%
     quadras:
Alpha 16 24.2% 10 28.6% 6 19.4%
Beta 14 21.2% 7 20.0% 7 22.6%
Gamma 25 37.9% 12 34.3% 13 41.9%
Delta 11 16.7% 6 17.1% 5 16.1%
     clubs:
int. + log. 19 28.8% 12 34.3% 7 22.6%
int. + eth. 11 16.7% 6 17.1% 5 16.1%
sens. + log. 14 21.2% 7 20.0% 7 22.6%
sens. + eth. 22 33.3% 10 28.6% 12 38.7%

About my type sample above
I am a part-time English teacher. I teach group English classes at corporations that are recruited by a language consulting firm. There is a high turnover of students as people come and go, joining or leaving existing groups, and as old clients discontinue classes and new ones start them. The type statistics above represent nearly all my students (all that I can remember), regardless of how many classes they attended. Some came to just two or three classes, while others have been attending regularly for months. This way I have come close to achieving a random sample for typing purposes. However, there might still be some type selection taking place. Different types might not be equally likely to work at the medium-sized and large corporations that I visit, and different types might be more or less likely to want to sign up for English lessons of this format in the first place. I believe this second factor is less important, since many people who don't like group lessons like mine are still curious enough to try at least a lesson or two and are given the same weight in my statistics as everyone else. The first factor I intuitively feel might indeed produce a quadra bias (so far I am showing a Gamma quadra bias). We shall see. A few people's types I am not 100% sure of and will continue to revise, but my practice shows that my revisions have little effect on the overall picture of type distribution.

Conclusions

Based on the type samples presented here I can hypothesize that men have an ever-so-slightly higher percentage of extraverted, intuitive, and logical types, and that women have an ever-so-slightly higher percentage of introverted, sensing, and ethical types. The disbalance among logical and ethical types is probably somewhat more pronounced.




08/15/2006 econdude
Rick, I wondered if you have ever considered that half of the population is -S-j (Guardians in Keirsey's lingo) with the rest of the population split evenly between the remaining three groups. This is not as outrageous as it sounds, since a medical doctor here in the United States specializing in the brain writes that about half of the population has a GABA dominant brain, which is associated with the SJ type.
08/29/2006 Author
I am acquainted with Keirsey's type distribution statistics, but with the conceptual differences, typing differences, and distribution differences it is best to think of the two typologies as separate, but distantly related. This page shows that Keirsey types famous people quite differently than socionists, assuming that the types used are somehow "equivalent." There is nothing wrong with Keirsey's typology, as it captures what it is meant to capture, but it has a different basis than socionics. Strictly speaking, it is not even the "same" typology as Meyers-Briggs. According to Keirsey's definitions, it may be true that half of the population are Guardians. However, that doesn't imply that half of the population are rational sensers according to socionics.

10/17/2006 Fortunato
In regard to Keirsey's linked distribution scores, also it should be noted it's not a random sample but applies only to the segment of the population that (1) has online access and (2) is interested and willing to set aside time to seriously fill out the test form.

An accurate set of statistics used to examine "humankind" in general, or a particular population, would have to more rigorously create a true random sample.

I did expect the NF Idealist to be overly represented (which it seems to be, when compared to what was anticipated by Keirsey and when compared to NT).

I suppose the real question is, how under-represented is the SP Artisan type in Keirsey's online survey?