Recent Posts

Pages: [1] 2 3 ... 10
My dog-eared copy of "Weird War" is radical leftist because it has people fighting Nazis?

Cool, thanks KiA!
Politics and Government / Re: Mr. Trump Goes to Washington
« Last post by dpareja on Today at 05:50:09 pm »

Read this... now do you really want Pence over Trump?
The Lounge / Re: Things That Annoy You
« Last post by RavynousHunter on Today at 12:14:25 pm »
Okay, is it just me, or is Facebook failing to load, like, half the images on anyone else's feeds?
Science and Technology / Re: Google facing multiple lawsuits
« Last post by Tolpuddle Martyr on Today at 02:10:55 am »

Then take a break from building strawmen. Nobody's forcing you to do it.

I'm building strawmen? Then surely you can show me where you answered Murdin's long list of points he raised in his post!

C'mon Lana, I'm the only one left in the thread asking you to answer the science points from the poster in the thread who knows his science at a level above "interested amatuer" and here you are wanting to trade insults.

Or...would you prefer to go back to shouting about Quinn and Antifa? Maybe speculating about your identity does it for you?

Because I saw a very detailed post challenging your reading of the science from someone who obviously knows their shit and here's you wanting to discuss anything but.

So, about Murdin's post:

(click to show/hide)

Askold already pointed out how the article resorts to lowbrow nitpicking to make its lame points. That's certainly true, but it won't be the focus of my criticism. There's worse, so much worse than that to be found in this Gish gallop, and this time our buddy will NOT be able to hide behind weak as shit reservations such as "there are a few things I'd have liked for it to cover in more depth, but overall, I think I agree with it". I wasn't lying when I said it was a very interesting read. There's a lot to learn here about the so-called "rationalist" mindset, and how it can become so utterly detached from the scientific enterprise.

And what better way to introduce this assassination of science, than with this audacious dismissal of the idea that correlations made by people the author agree with do not imply causation?

This is a case of what Garett Jones calls the Everest regression. He says that controlling for height, the atmospheric pressure there is not low. Or as I say, controlling for latitude, the Sahara desert has good weather.

The error here is that HDI and gender equality are substantially linked. Controlling for HDI or GDP is like controlling for gender equality. As a general case, all good things are correlated: technology, moral progress, GDP, country IQ, industrialisation tend to be coupled. We don’t need power to explain those differences.

Okay, let's follow his reasoning through. After controlling for height, the atmospheric pressure at the summit of the Everest is normal. This is obviously because altitude is the only factor (it actually isn't but w/e) that affects pressure at this position on the surface of Earth ; there's no need for an alternate explanation. Therefore, if after controlling for HDI the gender differences are normal for any set of gender-equality factors, it means... OH SH-

This would also be a questionable cause fallacy, of course. The entire point of Dr. Sadedin was to give an alternative interpretation of the same results (high HDI causes men to act more stereotypically masculine) that makes at least as much sense as the one this study was clearly designed around (gender equality causes people to act more stereotypically gendered), just to show that you can't easily conclude any causation from a mess of correlations that are also correlated with each other. The same mess of correlation that the blog's author actually mention while completely ignoring its actual implications. Science is fucking hard, guys.

The concept behind the "Everest regression" itself is a massive fallacy. Its implication that controlling for known external factors is fallacious is... beyond insane. You can't even argue that it denounces somehow "abusive" or "illegitimate" forms of control, either, because its textbook case IS logically sound and scientifically meaningful. We can evaluate the correlation between pressure and height (and maybe also temperature...) from other measurements. If the pressure at Mt Everest is NOT normal controlling for height, then there has to be an additional factor to explain this discrepancy.

Jones is an associate professor in economics at the Koch-funded George Mason University. From what I can guess, he invented his fallacy in order to defend IQ as some essential measurement of man, against trained psychologists who mostly see it as a tool which nicely correlates with many factors of social success. I'll let you make your own opinion of the man, his works, and whether his layman's stances on natural sciences are worth your consideration. Besides Googling his name, his Twitter is a good source of information ; Here's a reblog demonstrating his vision on what makes good science.

Incidentally, cursory knowledge of world geography would tell you the Sahara's weather is, in fact, particularly inhospitable even when accounting for latitude. Unless you define "good weather" as "sunny", in which case the Sahara has excellent weather regardless of latitude.

The paper says that initially, mental rotation differences were moderaly large, d=.59, for men primed male and women rimed female. (p=0.01).  For men and women both primed male, the effect was d=0.01. But what is the p-value or that? Well, p=0.94. Yes, 19 times larger than the standard 0.05 cutoff commonly accepted for statistical significance. For the whole set they report statistical significant results, but no effect size. We can also study statistical significance in the extreme case: female primed men and male primed women. If we plug in their data in a Welch’s t-test calculator, we get a p-value of 0.61. Again, not statistically significant.

(click to show/hide)

Like David Silverman in his interview by Bill O'Reilly, I... genuinely can't explain what I have in front of me. This is faux-scientific fetishism of the dumbest fucking kind, the end result of years of smug rationalist cargo cult enabled by the likes of Scott Alexander. That man's understanding of p-values is apparently limited to "low good, high bad".

The difference in test results between men and women both primed male is tiny. This results in a very high p-value. A valid interpretation of this p-value is that this kind of result would be very likely to be found if there was no difference between the two studied groups with regard to the studied characteristic. Or, to use a handy "Everest regression": when controlling for male priming, there's no observed difference in tests results between men and women. This tends to corroborate Dr. Sadedin's hypothesis that gender priming, not biological sex, is to blame for the widely measured disparity between men and women on spatial reasoning skills.

Meanwhile, I'm calculating a p-value around 0.014 for women primed female vs women primed male. I can't get the exact value without the group sizes, but my other calculations fit quite well with the blog author's numbers. In any case, that's actually quite significant. Obviously, this is also good for Dr. Sadedin's claim that gendered priming has an influence on test results.

I'm aware this study cannot be the be-all end-all on the subject. In fact, I'm almost certain more data will come out or already exist, that directly contradicts these results. It doesn't matter. Even if the scores behind this study were found to be completely forged, it does not excuse or justify the blog author's hatchet job in any way whatsoever.

There is a significant overlap, yes. But if we look at the tails, as I’ve been stressing over and over, one can still see massive differences.

The defilement of science is less eye-gouging than in the two previous exhibits, but there's a lot of different wrongs in this single point.
  • The linked article was written by an economist. Incidentally, the same economist with no background in natural sciences that pulled the Everest regression out of his ass.
  • It is, in fact, a libertarian political tract poorly disguised as a scientific study. Which is admittedly par for the course for an economist.
  • The blog author was trying to address the differences in software engineering skill between men and women. The relevant part of the article is about IQ instead.
  • Said part is based on a survey from Scotland, made in... 1932. That's right, 85 year old data from a fairly small and culturally homogeneous population.
  • On 11 year old kids.
  • The "massive differences" touted by the blog author... simply aren't that massive. Even at the very tail end of the chart, we have 277 boys for 203 girls, which is a bit over four boys for every three girls.
  • Inflated claims and abusive use of IQ as a measure for skill notwithstanding, this number does not even come close to explaining the truly massive gender disparities in software engineering.
From a more personal perspective, as a software engineer myself, I'm highly skeptical of the underlying claim that doing my job competently actually requires such extraordinary mental prowess.

As for the rest of the article past this point... it becomes pretty boring, to be quite honest. The author keeps talking past Dr. Sadedin's points, often rephrasing what she just said in a marginally more favorable way and then calling it a win. For a while, he just quotes relevant studies from actual scientists, wisely abstaining himself from adding his own commentary or conclusions. Then both the original response and the blog post drift into politics and I can finally be excused for not giving a shit. There isn't much to say about the author's self-congratulating conclusion, either.

(click to show/hide)

Obviously not a direct question this time, buddy, but... why do you hate science so much? Why do you keep using it as a blunt weapon against your rhetorical opponents, without showing any respect to its most fundamental principles? Why do you spew self-righteous bullshit like "I guess science is sexist now" or "rather than approaching this ideologically, let's look at it scientifically", only to effectively disown it by including such enormities in your narrative?

I mean... is it really worth it? What are you even trying to achieve here, and I actually do mean, here? You've already claimed Religion and Philosophy, Politics and Government, Society and History, was that not enough for you? Did you really have to bring your usual drivel to Science&Tech, incidentally the only place where I would give a fuck about it in in the first place, and then gloat openly over Queen taking the bait? And then take a blatant bait yourself, deliberately or not? Did you think you were the only one "clever" enough for that kind of dirty trick, or did you just decide to go along with the ride?

Because, unless making people exhaust themselves was somehow actually part of your goal, I'm pretty sure you haven't achieved anything here.

I'd love to hear your opinion on all this, please do elucidate us!
Science and Technology / Re: Google facing multiple lawsuits
« Last post by ironbite on October 16, 2017, 08:04:10 pm »
Dynamic Paragon?  Ultimate Dragon?

Ironbite-Anal Reverse?
Science and Technology / Re: Google facing multiple lawsuits
« Last post by Skybison on October 16, 2017, 07:56:01 pm »
I said "long", not "perfect".

It's a very convenient thing to be forgetting though.  Reminds me of another person who was here before who always made such mistakes.  Like when he made a "typo" that his brother was trans that he "forgot" to correct.  Gee what was his name again?
Science and Technology / Re: Google facing multiple lawsuits
« Last post by niam2023 on October 16, 2017, 07:33:52 pm »
Lana, why don't you just accept that nobody here is liable to be "converted" to believing the google dudebro did nothing wrong.
Politics and Government / Re: Mr. Trump Goes to Washington
« Last post by niam2023 on October 16, 2017, 06:26:19 pm »
I am reminded of stories of Hitler making fun of Himmler for being a weird nerd with esoteric beliefs.
Politics and Government / Re: Mr. Trump Goes to Washington
« Last post by dpareja on October 16, 2017, 06:23:15 pm »
Apparently Trump loves to mock Mike Pence for his Fundie views, ridiculing him and comparing him to the Taliban for "wanting to hang gays" and making fun of his plans to overturn roe vs wade since "blue states will just make it legal on their own."

Stopped clocks and all.

More like clocks that lose a second every year.
Religion and Philosophy / Disqus was hacked
« Last post by Sharon_at_home on October 16, 2017, 06:13:50 pm »
Has anyone else seen this?

 Security Alert: User Info Breach
Posted by Jason Yan on October 06, 2017

Disqus was hacked into so you should go check it out if you think it may have affected you.
Pages: [1] 2 3 ... 10