Socionics Research Plan

[Jonathan]

I've been thinking about this for a long time, but just never started a thread on how to actually make progress in Socionics (or at least not recently). Most of us, I think, go around in circles a bit, feeling that Socionics provides valid insights in our lives, but having doubts that can never be answered through introspection or casual observations. There are a few ways we can use Socionics without any scientific basis (e.g., "It works for me," or "I just use it as an inspiration in art/stories"). But to move to the next level, there must be some valid research. This thread is to talk about how we could actually put that into practice. The existing research that I've read about doesn't seem to answer fundamental questions. A typical study involves a small focus group where (for example) they divide people into Socionics-related groups, and then with a facilitator, people write down all the ways they're similar or different. This provides interesting hypotheses, but in no way does it help establish fundamental assumptions. There is the matter that people may influence each other, affecting the results, but more importantly, the research designs aren't really aimed at the key questions a skeptic of Socionics would ask.

There are hypotheses that could be scientifically tested...for example:

Hypothesis: There exists a standardizable method of typing people whereby the quality and nature of the interaction between two people who never met before can be predicted to a statistically significant degree.

Sub-hypothesis: This method provides superior predictive value compared to randomly chosen psychological traits that are not correlated with it.

Now, to do a test on this, someone needs to have the means, time, money, etc., to actually do it. Perhaps some of the people here are involved in a university program that could help underwrite such a study; maybe there's grant money available. Ultimately, though, there needs to be a plan to get it done. This thread is to discuss how to organize such an endeavor.

[ragnar]

to move to the next level, there must be some valid research. This thread is to talk about how we could actually put that into practice. ... establish fundamental assumptions. ...

There are hypotheses that could be scientifically tested...

Hypo 1: The set of all healthy American adults is partitioned into 16 subsets as per socionics theory.

Test1:
Select M "independent" socionics experts and N random people from off the street or whatever. Have each "expert" independently submit an opinion of each person's type, then afterwards you test the resulting MxN matrix for statistically significant correlation.

Find new randoms and repeat the experiment with various cues made inaccessible to the experts, in order to prevent partitioning based on visual or vocal or other non-personality traits.

Done properly, the test will show the 16 types exist.

[aestrivex]

it is hypothesized that socionics types of identical or adjacent quadras (as assessed by as effective an assessment method as you can implement efficiently; multiple different assessment methods may be employed) at a close psychological distance will get along better, by various self-report measures of interpersonal compatibility, than types of opposite quadras. randomly paired college freshman roommates constitute an ideal testing situation of close psychological distance.

[aestrivex]

to move to the next level, there must be some valid research. This thread is to talk about how we could actually put that into practice. ... establish fundamental assumptions. ...

There are hypotheses that could be scientifically tested...

Hypo 1: The set of all healthy American adults is partitioned into 16 subsets as per socionics theory.

Test1:
Select M "independent" socionics experts and N random people from off the street or whatever. Have each "expert" independently form an opinion of each person's type, then test the MxN matrix for statistically significant correlation.

Find new randoms and repeat the experiment with various cues made inaccessible to the experts, in order to prevent partitioning based on visual or vocal or other non-personality traits.

Done properly, the test will show the 16 types exist.

this is a terrible idea because of multiple comparisons and because your results depend upon the expertise of individuals without demonstrating that socionics has any real-world validity or empirically measurable effect; any findings using your suggestion would be of absolutely no clinical significance, trivial to dismiss as unverifiable in the absence of subjective expert opinion, and completely unpublishable.

[Jonathan]

it is hypothesized that socionics types of identical or adjacent quadras (as assessed by as effective an assessment method as you can implement efficiently; multiple different assessment methods may be employed) at a close psychological distance will get along better, by various self-report measures of interpersonal compatibility, than types of opposite quadras. randomly paired college freshman roommates constitute an ideal testing situation of close psychological distance.

Yeah, something like that would work. I think it's best if people's types are assessed before they meet though (or at least before they spend any significant time). If people live together and get along based on non-Socionic reasons, a skeptic could say that will influence their type results (e.g., they may shift towards similar quadra values).

Actually, one could simplify it further: Assess only two dichotomies: judicious/decisive and merry(subjectivist)/serious(objectivist). If both match, then the quadra is the same, and if one matches, the quadras are adjacent. However, I don't think the Reinin descriptions would be the best way to assess these, because I don't think that they're correct. They could, however, be one assessment method that could be employed, among others.

Are you currently involved in any sort of psych research?