Absolutely not. But I must admit that this is the first time I’ve seen the error go in this direction!
Consider the below (with the obvious translation to causal/correlation jargon):
A = y*noise - x*B + C
B = z*noise + C
I expect rhetoric and reason to be positively correlated, largely through intelligence
That seems to be a safe assumption.
Do you expect that, holding intelligence constant, there is a negative correlation?
Not particularly. An expectation on that relationship isn’t implied by my reasoning and here and there are too many other likely important variables at play for me to trust anything I came up with. I take it as implied, by the way, that such things as as education and personality type were to be held constant. By virtue of describing education as crystallised intelligence or otherwise. If those aren’t controlled for obviously the correlation would be positive.
It is possible to have a negative influence on something while also being positively correlated with it. For example, tax-paid-this-financial-year is positively correlated with net worth despite being a negative contribution. (At least, it is sure to be once we control for legal expenses!)
The equations just point roughly to how (with the right x) a set could be produced where B tends to have a negative causal influence on A despite being positively correlated. I’m not sure how useful they are in understanding the flaw in the statement “surely causality is correlation”. But then, I also would have thought “um… no?” to be more than sufficient. The only interpretation I can make of that claim that doesn’t seem completely ridiculous is if he means ‘an overriding dominant causal influence will guarantee correlation’. Even then, it’s not what he said and it doesn’t remotely fit the context.
The original sense would be that intelligence makes one good at rhetoric and logic, but practising rhetoric then makes one worse at logic. My personal experience (weakly) confirms this.
Absolutely not. But I must admit that this is the first time I’ve seen the error go in this direction!
Consider the below (with the obvious translation to causal/correlation jargon):
That seems to be a safe assumption.
Not particularly. An expectation on that relationship isn’t implied by my reasoning and here and there are too many other likely important variables at play for me to trust anything I came up with. I take it as implied, by the way, that such things as as education and personality type were to be held constant. By virtue of describing education as crystallised intelligence or otherwise. If those aren’t controlled for obviously the correlation would be positive.
Could you explain this, please? I don’t think many people understand.
It is possible to have a negative influence on something while also being positively correlated with it. For example, tax-paid-this-financial-year is positively correlated with net worth despite being a negative contribution. (At least, it is sure to be once we control for legal expenses!)
The equations just point roughly to how (with the right x) a set could be produced where B tends to have a negative causal influence on A despite being positively correlated. I’m not sure how useful they are in understanding the flaw in the statement “surely causality is correlation”. But then, I also would have thought “um… no?” to be more than sufficient. The only interpretation I can make of that claim that doesn’t seem completely ridiculous is if he means ‘an overriding dominant causal influence will guarantee correlation’. Even then, it’s not what he said and it doesn’t remotely fit the context.
Ahhh, excellent, thankyou.
The original sense would be that intelligence makes one good at rhetoric and logic, but practising rhetoric then makes one worse at logic. My personal experience (weakly) confirms this.