Debate user BoxerShorts47 on "strawmans" and logical fallacies, definitions of ephebophilia, how to MAGA, religion, Sailor Moon and more

BoxerShorts47 said:
HEXACO looks like it still has the same fundamental problems as Big 5. 1. Not MECE or even close. 2, Vague "what is agreeableness exactly." 3. Interpretation because on the one hand you don't want to be that obnoxious contrarian all the time but at the same time you don't want to be a yesman so I find a factor like agreeableness to be completely useless.

Also extroversion and introversion should NEVER be sampled on their own. They should always be sampled together with Perception & Judgement.
Extroversion + Judgement = Choleric.
Extroversion + Perception = Sanguine.
Sampling I or E by itself is bad because it combines the 4 Temperaments from antiquity into 2 factors.

I personally use my own modified version of socionics. I like enneagram, it works well for a person's motivations or drives but it's not as deep or rigorous as socionics. I really hate Big 5. I think it's standard social sciences trash.
Click to expand...
Huh. So the fact that Big 5 replicates, while socionics and enneagram don't, doesn't bother you at all?
 
  • Thunk-Provoking
Reactions: Dead Mime and Troon Draugur
BoxerShorts47 said:
TITS or GTFO!
You know the rules.
Click to expand...

Okay
Man tits.jpg
 
  • Informative
  • Like
Reactions: Captain Chorizo, Stardust, HIVidaBoheme and 3 others
Zonga said:
Huh. So the fact that Big 5 replicates, while socionics and enneagram don't, doesn't bother you at all?
Click to expand...
Replicating junk doesn't matter. Do you think Google or FaceBook use Big 5 when they do targeting? I've never heard it used for any serious analysis outside of the humanities.
 
  • Dumb
  • Mad at the Internet
  • Deviant
Reactions: Balanced Boomer, Μusk, HIVidaBoheme and 7 others
Zonga said:
...huh? You're kidding, right?
Click to expand...
No. I've never heard it used outside the humanities for serious business.

From what I understand. Carl Jung developed a psychological types model. Myers and Briggs used his research to create a questionnaire. Their questionnaire (MBTI) wasn't that good. But people really liked introversion and extroversion so these concepts were reused (without the original context) in Big 5. Big 5 is basically a Frankenstein model.
 
  • Mad at the Internet
  • Dumb
  • Autistic
Reactions: Balanced Boomer, Be patient I have lurkism, Μusk and 8 others
BoxerShorts47 said:
No. I've never heard it used outside the humanities for serious business.

From what I understand. Carl Jung developed a psychological types model. Myers and Briggs used his research to create a questionnaire. Their questionnaire (MBTI) wasn't that good. But people really liked introversion and extroversion so these concepts were reused (without the original context) in Big 5. Big 5 is basically a Frankenstein model.
Click to expand...

That's... Not what happened, at all. What happened is, computers got good enough that a bunch of psychologists who were into different personality metric systems got together and said "you know what? Let's just build the biggest questionaire we can, throw all sorts of questions into it, run some regression analysis and see what clusters pop out. Then let's build a second questionnaire based on those clusters, try again and see if it replicates." And... Well, it did. Repeatedly.

And it just so happened that when they looked at the clusters, there were five of them, and one of them was almost identical to Jung's "Extroversion/Introversion" scale. But they couldn't line the others up with Judging/Sensing etc., so they had to come up with new terms for them. The ones they picked were probably not the best words to describe what they're measuring, but now they're "baked in". But the actual metrics are real - although it turns out there's a distortion, due to the original researchers wanting the test to be "morally neutral", when real world personality actually has a moral component (hence HEXACO).
 
  • Informative
Reactions: Dead Mime, Troon Draugur, Pope Fucker and 1 other person
Zonga said:
That's... Not what happened, at all. What happened is, computers got good enough that a bunch of psychologists who were into different personality metric systems got together and said "you know what? Let's just build the biggest questionaire we can, throw all sorts of questions into it, run some regression analysis and see what clusters pop out. Then let's build a second questionnaire based on those clusters, try again and see if it replicates." And... Well, it did. Repeatedly.

And it just so happened that when they looked at the clusters, there were five of them, and one of them was almost identical to Jung's "Extroversion/Introversion" scale. But they couldn't line the others up with Judging/Sensing etc., so they had to come up with new terms for them. The ones they picked were probably not the best words to describe what they're measuring, but now they're "baked in". But the actual metrics are real - although it turns out there's a distortion, due to the original researchers wanting the test to be "morally neutral", when real world personality actually has a moral component (hence HEXACO).
Click to expand...
So many problems.

1. Jung's model was incomplete( and socionics really improves it).

2. Jung's model was deduction, based on axioms like introversion and extroversion. But surveys are inherently inductive. You're not supposed to create questionnaires (inductive prompts) for a deductive model. Huge no-no. Induction and deduction are completely different approaches.

3. I've made questionnaires and done principle component analysis and the problem that emerges is "what are you seeing," I gave you 1 example which is agreeableness, what does this even mean? Is that good? or bad? I'm not saying you cannot find patterns in the data but I'm saying Big 5 traits are worthless. Go deeper, with agreeableness, are you just measuring a proxy for a man's test levels?
 
  • Dumb
  • Autistic
  • Mad at the Internet
Reactions: Balanced Boomer, Be patient I have lurkism, Μusk and 7 others
Stardust said:
@BoxerShorts47, you are soooo mad you cannot refute what I've said, that you resort to old memes as a coping mechanism. Is this 2003? Y/N.

Negrate me again, crybaby, as validation you're more butt hurt than when @SIGSEGV has fucked you raw with his thundercock.
Click to expand...
I need the tits. I wanna see those huge udders. If you want me to answer your question, you gotta make it worth my while.
 
  • Dumb
  • Islamic Content
  • Deviant
Reactions: Balanced Boomer, sick_hypnotic, Be patient I have lurkism and 12 others
BoxerShorts47 said:
So many problems.

1. Jung's model was incomplete( and socionics really improves it).

2. Jung's model was deduction, based on axioms like introversion and extroversion. But surveys are inherently inductive. You're not supposed to create questionnaires (inductive prompts) for a deductive model. Huge no-no. Induction and deduction are completely different approaches.

3. I've made questionnaires and done principle component analysis and the problem that emerges is "what are you seeing," I gave you 1 example which is agreeableness, what does this even mean? Is that good? or bad? I'm not saying you cannot find patterns in the data but I'm saying Big 5 traits are worthless. Go deeper, with agreeableness, are you just measuring a proxy for a man's test levels?
Click to expand...
Well, Agreeableness happens to be the least cohesive one of the Big Five, so it makes sense that you'd pick that one to pick apart. But there's a lot of evidence for Extraversion, Stability, and Conscientiousness, and a reasonable amount of evidence for Openness-independent-of-IQ, but I grant that Agreeableness gets fuzzy. (it also happens to be the one that "comes apart" when you add moral questions).
 
  • Informative
Reactions: Dead Mime, Troon Draugur and Stardust
Zonga said:
Well, Agreeableness happens to be the least cohesive one of the Big Five, so it makes sense that you'd pick that one to pick apart. But there's a lot of evidence for Extraversion, Stability, and Conscientiousness, and a reasonable amount of evidence for Openness-independent-of-IQ, but I grant that Agreeableness gets fuzzy. (it also happens to be the one that "comes apart" when you add moral questions).
Click to expand...

1. With extroversion what are you actually measuring? Like I said, by itself E/I is kinda worthless. You're better off just sampling the 4 Temperaments since E/I in effect consolidates Choleric & Sanguine = Extrovert and Melancholic & Phlegmatic = Introvert. You're losing data using E/I by themselves with P/J.

2. Stability could mean many different things, including overlapping with basically every type that isn't Sanguine. Do you have data that shows the 5 traits are independent and not correlated. I'm pretty sure the traits are correlated which means the model isn't mutually exclusive. https://www.researchgate.net/figure/Correlation-among-the-Big-Five-personality-traits_tbl1_284349515
 
  • Autistic
  • Dumb
  • Mad at the Internet
Reactions: Balanced Boomer, Be patient I have lurkism, Μusk and 7 others
BoxerShorts47 said:
1. With extroversion what are you actually measuring? Like I said, by itself E/I is kinda worthless. You're better off just sampling the 4 Temperaments since E/I in effect consolidates Choleric & Sanguine = Extrovert and Melancholic & Phlegmatic = Introvert. You're losing data using E/I by themselves with P/J.

2. Stability could mean many different things, including overlapping with basically every type that isn't Sanguine. Do you have data that shows the 5 traits are independent and not correlated. I'm pretty sure the traits are correlated which means the model isn't mutually exclusive. https://www.researchgate.net/figure/Correlation-among-the-Big-Five-personality-traits_tbl1_284349515
Click to expand...

In Big 5, at least, the cluster labeled "extraversion" is essentially measuring "sensitivity to external rewards", which essentially is a proxy for the sensitivity of the dopamine reward feedback loop. Likewise, the cluster labeled "stability" seems to measure something like "sensitivity to external threat", and conscientiousness measures something like "sensitivity to immediacy vs delay of rewards or threats". So the Stability/Neuroticism trait and the Conscientious/Spontaneous dimensions actually roughly encode the P/J and S/N traits, but with the eigenvectors rotated about 45 degrees, and the remaining data captures details about patience/need for immediacy, which it turns out clusters tightly with the need for order you see in J-types.
 
  • Informative
Reactions: Dead Mime, Troon Draugur and Stardust
With big 5, you have 5 traits and they can be high or low, 2^5=32 combinations. So if the model is strong, you should be able to explain these 32 psychological types. I've never even that done with Big 5. I think it's impossible because the traits are correlated with each other. I would argue that the overall is weaker than MBTI system but I can see why social scientists would use Big 5 since it can be repeated but I would counter that begin able to repeat junk is still junk.

> actually roughly encode the P/J and S/N trait

Not not even close to what P/J and S/N measure.
 
  • Autistic
  • Dumb
  • Mad at the Internet
Reactions: Balanced Boomer, Μusk, HIVidaBoheme and 7 others
BoxerShorts47 said:
With big 5, you have 5 traits and they can be high or low, 2^5=32 combinations. So if the model is strong, you should be able to explain these 32 psychological types. I've never even that done with Big 5. I think it's impossible because the traits are correlated with each other. I would argue that the overall is weaker than MBTI system but I can see why social scientists would use Big 5 since it can be repeated but I would counter that begin able to repeat junk is still junk.
Click to expand...
They aren't low/high though. These aren't binary divisions the way the Jungian aspects are. There are going to be clusters, and within those clusters certain traits will correlate more strongly, but all such clusters will be fuzzy rather than clean and binary. Hence there won't be 32 combinations. The "big 5" describes a 5 dimensional landscape (or six dimensional if you use HEXACO), and within that landscape you'll probably get some big 'blobs' of correlation that represent particular "archetypes" - but those blobs won't be neatly arranged at the exact endpoints of each of the eigenvectors, because why would they be?

I'd suggest that if you need things to be that "neat" in order to get any work done, the problem likely isn't with the system you're using.
 
  • Informative
Reactions: Dead Mime, Troon Draugur and Stardust
Zonga said:
They aren't low/high though. These aren't binary divisions the way the Jungian aspects are. There are going to be clusters, and within those clusters certain traits will correlate more strongly, but all such clusters will be fuzzy rather than clean and binary. Hence there won't be 32 combinations. The "big 5" describes a 5 dimensional landscape (or six dimensional if you use HEXACO), and within that landscape you'll probably get some big 'blobs' of correlation that represent particular "archetypes" - but those blobs won't be neatly arranged at the exact endpoints of each of the eigenvectors, because why would they be?

I'd suggest that if you need things to be that "neat" in order to get any work done, the problem likely isn't with the system you're using.
Click to expand...
It seems to me like your IQ is low. 1. If you build a model, it needs to have solid prediction powers. 2. clusters will emerge on the spectrum of above avg and below avg aka high and low. 3. If factors are correlated then you don't really have 5 factors. It means your factors are poor.
 
  • Dumb
  • Mad at the Internet
  • Deviant
Reactions: Balanced Boomer, Be patient I have lurkism, Μusk and 8 others
You must log in or register to reply here.
Back