Accuracy in typing

[ncassidy]

How accurate do you all think the collective Socionics community is at determining the types of the individuals we study? What factors contribute to our accuracy, and how accurate are we both individually and as a group?

Are our individual determinations clouded by a desire to follow leaders in the forum, agree with the reasoning of logical members, or some other undesirable influence? Are we more inclined as a group to attribute a certain temperament, quadra, or even type to any given individual?

[Expat]

I don't think there is a general answer to that question.

In the cases where detailed descriptions of an individuals behavior, and of the relationships of that individual, are available - and they all fit well - I think we can say that the typing is very accurate.

In other cases, it varies. Who can say for sure?

I will say this: those here who say stuff like "I see the types, but I don't buy the intertype relationships" are clearly doing something wrong.

[Phaedrus]

I will say this: those here who say stuff like "I see the types, but I don't buy the intertype relationships" are clearly doing something wrong.

I agree, but the opposite is also true -- see my latest post in this thread:

http://the16types.info/forums/viewto...r=asc&start=30

If you can't see the types you cannot even get started.

[ncassidy]

If we assume Socionics itself is accurate and flawless, how close do we as individuals and forums come to perfect typing? How important is bias? I've tried to come up with a statistical method to assess it, but it's complicated beyond belief as a type is a vector, and a binary one.

[FDG]

If we assume Socionics itself is accurate and flawless, how close do we as individuals and forums come to perfect typing? How important is bias? I've tried to come up with a statistical method to assess it, but it's complicated beyond belief as a type is a vector, and a binary one.

What are you trying to evaluate, the efficacy in typing of the forum as whole? Unfortunately, I doubt its possibility - there is no way to be sure that a person is of a given type, therefore it's impossible to conduct experiments on accuracy. In any case - I think that, as a whole, if we avoid considering the cases of individual with utmostly skewed perception, I can say that the accuracy is very good.

[tcaudilllg]

If we assume Socionics itself is accurate and flawless, how close do we as individuals and forums come to perfect typing? How important is bias? I've tried to come up with a statistical method to assess it, but it's complicated beyond belief as a type is a vector, and a binary one.

flawless...? Are we talking about Augustan socionics here...?

[snegledmaca]

I remember a psychological study that showed the unskilled workers hugely overestimate their abilities. Being that they are unskilled they also lack the skills to make an assessment. This is unbelievably present on typology forums. Everybody thinks they can type, while the truth is that they have no ability at all and are like that door man at the gates of the corporation thinking he plays a vital role.

[ncassidy]

FDG - We can at the very least least determine the effects of sampling bias and groupthink on our agreement over a type. It's entirely possible, I would even say probable, that the group comes to a different consensus based on the order of postings, or the impact conferred by the reputation or social power of their authors.

tcaudilllg - That's the last assertion I would make, I'm just staying within those bounds to avoid opening up a whole new can of worms.

snegledmaca - I agree completely.

We may be working with a misspecified model... see labcoat's "Hidden temperaments" thread. I don't claim to fully understand what he's come up with there, but the argument that more types may be possible could be important.

[Raver]

A lot of people in this forum type terribly mainly because their beliefs, views and opinions cloud their judgement. However, if you get rid of these evil demons, type accuracy would jump tenfold.

[ncassidy]

Even the other ENFPs!?

[Raver]

Even the other ENFPs!?

ENFPs are awesome in general, so not really.

[ncassidy]

Yea I was about to say.