3 ms·
> Education level, as a variable, has predictive power ... Yes, it certainly does. For example, here's a list of billionaire college dropouts: http://en.wikip
by lutusp 14y ago
> Education level, as a variable, has predictive power ...
Yes, it certainly does. For example, here's a list of billionaire college dropouts:
http://en.wikipedia.org/wiki/List_of_college_dropout_billionaires http://en.wikipedia.org/wiki/List_of_college_dropout_billion...
> ... as shown in a couple of studies that I've come across, one of which is cited there.
It has precisely no predictive power in the cause-effect sense one expects to see in a scientific study. Studies that indicate otherwise are simply wrong -- are drawing conclusions not based on evidence.
There are any number of "studies" that describe without trying to explain (very common in the "social sciences"). Those that try to explain often do so by ignoring cause-effect possibilities other than the preferred one. Most such studies are funded by institutions of higher learning, institutions with a vested interest in a particular outcome.
> That's beside the point though.
No, it is the point. You're trying to show a correlation between political achievement and education, but you're using schooling as a convenient but very misleading indicator of education.
Today's important lesson: correlation is not causation.
http://en.wikipedia.org/wiki/Correlation_does_not_imply_causation http://en.wikipedia.org/wiki/Correlation_does_not_imply_caus...
- mas_namerif 14y agoI can't do much with your point about the trustworthiness of research in the social sciences given that it's anecdotal and I have nothing to compare it to (anecdotally), nor do I see how referencing a list of 31 college dropout billionaires has much to do with the argument about predictive power. I'm aware of logical implication and correlation vs causation and so that seems like a whopping non-sequitur. I don't want to get into a pissing match, but I genuinely am interested in knowing how you would suggest codifying education?
- lutusp 14y ago> ... nor do I see how referencing a list of 31 college dropout billionaires has much to do with the argument about predictive power. You mean, apart from falsifying the notion that there's any meaningful correlation between education and worldly success? > I'm aware of logical implication and correlation vs causation and so that seems like a whopping non-sequitur. Knowing that correlation is not causation doesn't constitute a license to ignore its implications. > I genuinely am interested in knowing how you would suggest codifying education? You can't. Education is not biology, it is not neuroscience, it's psychology, and psychology is not a science. The reasons: http://arachnoid.com/trouble_with_psychology http://arachnoid.com/trouble_with_psychology It's important to remember that there are things not open to meaningful scientific investigation. This is one of them.
- mas_namerif 14y agoNowhere do I claim that. What I said was that there is research that suggests that the educational level of political leaders is a significant positive predictor of economic growth, based on a sample of 2000 or so leaders over a period of 129 years. That is quite different than showing "correlation between education and worldly success" as you paraphrased it into. The final two points and link are appreciated. I'll read your article with my morning coffee tomorrow.
- lutusp 14y ago> Nowhere do I claim that. No, you just provide graphs comparing educational achievement and worldly achievement that strongly imply that they're meaningfully correlated, i.e. show a cause-effect relationship. > What I said was that there is research that suggests that the educational level of political leaders is a significant positive predictor of economic growth Your use of the term "research" implies a scientific basis for the result. But there is no scientific basis for the claim --- it's a raw correlation with no real meaning, and beyond this, it pretends that there is a correlation between education and schooling. In other words, it relies on two meaningless correlations in series. Like this: Worldly achievement <|> Years in school <|> Education My advice is not to support such unscientific speculations by dressing them up in Web pages that try to confer unearned status to guesses. Do you know what distinguishes what you're doing from science? In science, an explanation is offered, one that can be tested and potentially falsified. Your page only describes, although it pretends to explain. But it cannot do that, for lack of evidence.
- mas_namerif 14y agoI really lack the motivation to go on with this, as appreciative as I am of your passion for science and, apparently, experimental rigor. If you're arguing against the agnosticism of a statistical regression model, then there's nothing I can say to that. This isn't scientific investigation, nor is it masquerading as one. I'm not basing a dissertation on this nor am I using it to develop formal models for Curiosity. It's visualization, it's design, an aesthetic transformation - driven by data, which as I've already pointed out, the analysis of which is published and peer reviewed and available for your and others' scrutiny. If you do not wish to accept it because of your distrust of or dislike for the social sciences for whatever anecdotal reason, then that's fine, but it's not an argument that I can do much with. The standard you're using in this context is misplaced, the zeal unnecessarily antagonistic and a subset of arguments borderline fallacious.