Hacker News Reader

Comment by hackitup7 | Hacker News Reader

I'm not making the claim that MBTI is useful, but anecdotally people who are familiar with it seem to be able to guess close acquaintances' types with above chance accuracy. Leads me to suspect that it's getting at something, although perhaps in a coarse or low utility way. I've never seen someone guess horoscope signs above chance.

View on HN

Mine told me I had the same personality type as Optimus Prime and nobody's taking that away from me.

That's why it's stuck around so long—it tells EVERYONE that they have the same personality type as Optimus Prime, so nobody wants to dump it.

Not everybody, just your ENFJ types. :-)

Honestly, I was at a company who got everybody to do the personality tests and then had a marriage counselor sit down with every team to discuss the results. The goal was to help people try to relate to each other more effectively.

It was honestly a decent exercise. I learned that I was one of 3 extroverted programmers at the company (of 150 devs). I don't know that it had much of a lasting impact, but it was solid effort on the part of that company.

The only reason I know the Optimus Prime thing is that they asked us to find a famous figure with a matching personality type to include on our team's internal wiki page.

Not a statistician, but just because the results of the test are predictable, I’m not sure the test is useful. Meaningful tests/models have predictive power. Is Meyers Briggs capable of predicting anything? (Repeatedly and reliably)

Eh, if my boss listens to the description of my type and approaches dealing with me with that in mind I’m pretty much guaranteed to be happier.

If it's stable over time, surely that allows some predictions to be made.

It's not even an issue of statistics - yes the traits might exist, and people might fit into it, but that doesn't mean it's predictive of any outcome related metric.

I did it three times and ended up with three different types. My "personality" as measured by the test is more of product of environment, who is around me and how I adjust to it rather then something inherent to me.

Also, the questions (at least on last test) thinking/feeling scale seemed to measure values rather then actual personality.

If your “personality” is a product of environment then it’s not really personality, is it? The notion of “personality” that personality tests are supposed to get at is not dependent on environment, or at least, that is the way that personality is defined in this context.

Yep. That is my point. It is not measuring something that would be inherent to me.

Personality varying with context is a well accepted concept. Most of these kinds of tests acknowledge it.

Well yeah, it's not difficult to know if someone is introverted or extraverted is it? Would you be impressed with a test that determined your gender, retirement status and race?

The test might ask such insightful questions as "are you over 65?" and "are both of your parents white?". Then it would use complex algorithms to determine that you are a white retiree.

> perhaps in a coarse or low utility way.

I think this is right, and a lot of the "this is astrology" people seem to be conflating the fact that something can be useful (which, to be clear, I don't believe astrology is) with the fact that something can be misused.

> I've never seen someone guess horoscope signs above chance.

I've met one person who was David-Blaine-party-trick-level good at this. Out of ~20 people she never met (including me) I think she missed one person's water/earth/fire/air sign, and was probably >80% on guessing their month sign, it was very entertaining.

From my understanding of it and having to take it a couple times at different jobs, it seems like a reasonably coherent categorization system for certain aspects of personality.

Where it falls apart is making predictions about how individuals will react to a certain situation or how people will interact with each other. In that I've never seen it do better than just what a socially competent person would accomplish by going with their judgement and intuition, and often worse.

http://www.socionics.com/rel/rel.htm

This is a link that describes how people will relate with each other, and I’ve found it to be accurate. Note the type encoding system is slightly different from MBTI, for introverts the last letter has to be flipped.

I want to point out that MBTI is basically a brochure version of an underlying theory. The value comes from understanding the individual cognitive functions themselves. Once you understand that, you can know the _process_ someone uses to think and make decisions. That gives you a better idea of why they think the way they do, especially if their cfs are different / confusing to you. It’s like being able to boot up someone else’s mental… OS in a VM (taxing and slow, but if you really want to run something, necessary.)

I think it’s not actually testing personality, but some kind of way that people choose to present themselves. The difference is that personality is supposed to be more stable over time and across different circumstances.

Anecdotally people who believe horoscopes say they can guess above average. Now, if you can do it in a controlled setting, that's more meaningful.

There could be something to the horoscope thing. Someone born in winter will have slightly different initial conditions than someone born in summer. Seasonal variations could affect early years.

That's not even getting into how month-of-birth could affect how a child fits into the school system. Some months will be on average older than their classmates, some younger.

I think the second effect is probably stronger in the modern (air-conditioned, artificially-lit, state-schooled) era.

Well.. there could be. But there doesn't seem to be. People can't hit the horoscope sign even roughly. They can't do it better than change for all of the year.

Funnily enough, looking into it, there are apparently some measurable differences between winter and summer births.

From the abstract: "Winter/spring birth, while associated with an increased risk of schizophrenia, is generally associated with superior outcomes with respect to physical and cognitive development."

Source: Schizophrenia Research 81 (2006) 91–100. "Season of birth is associated with anthropometric and neurocognitive outcomes during infancy and childhood in a general population birth cohort". https://dash.harvard.edu/bitstream/handle/1/3743679/sdarticl...

Is there even such a thing as "your type" if as the article states 50% of people get a different result the second time they take the test?

A suggestion is people project themselves as what they want, and not what they are - and answer that way. Self reporting & self contextualization of meaning of the questions could be considered low accuracy.

Reply on news.ycombinator.com