Anyone want to help make socionics scientific?

[Lao Tzunami]

@ApeironStella Like I said, I want to extend this to all levels of socionics, but for simplicity's sake, let me show you what I have for reinin dichotomy system. This is the equation that sums all levels of correlations.


Here is a program that generates all the correct correlation combinations:
https://repl.it/@ajsindri/Reinin-Dic...tion-Genorator

I have this correlation structure working in my excel program, but for this to work in real life, you would at least need to weight and normalize each dichotomy input, and maybe more steps. I'm still working on this. This is an analysis of how I rate myself on the reinin dichotomies. You can see definite banding of each dichotomy. If it was a random input, it would look like tv static.


I ran 100,000 random dichotomy inputs and got this curve. You can see it is close to a normal curve, but skewed slightly the right. The total possible range the harmony score can be is -35 to +91, and the alpha less than or equal to 5% point is about 1.4 (this value changes a bit depending on the number of likert options and items per dichotomies. This was with 5 likert options and 1 question per trait). That means if you tested a bunch of people, you could use this to determine if the dichotomies were random. This has the same problem as ANOVA, where you can detect and effect, but you don't know exactly what factors are causing it.