Bay 12 Games Forum

Please login or register.

Login with username, password and session length
Advanced search  
Pages: 1 ... 308 309 [310] 311 312 ... 339

Author Topic: SCIENCE, Gravitational waves, and the whole LIGO OST!  (Read 516502 times)

Starver

  • Bay Watcher
    • View Profile
Re: SCIENCE, Gravitational waves, and the whole LIGO OST!
« Reply #4635 on: January 01, 2017, 06:37:32 am »

I have a good friend who nonetheless subscribes to the whole IQ And The Wealth Of Nations malarkey, and all of its 'racialist' overtones. I try not to argue about this, any more.

(And I would class myself as good with IQ tests, but not a higher achiever because of that.)
Logged

Reelya

  • Bay Watcher
    • View Profile
Re: SCIENCE, Gravitational waves, and the whole LIGO OST!
« Reply #4636 on: January 01, 2017, 08:24:36 am »

Reelya, again, is there more than one guy who puts forward those views? Looking at the website, his style of talking about it, and his responses to comments, gives me the impression that's he's not exactly unbiased.

Well, that one guy has a Doctorate in Psychology and has published peer-reviewed papers. There are also many other articles, writers and experts that question the methodology

http://www.slate.com/articles/life/twins/2011/08/double_inanity.html
Quote
That identical twins do not, in fact, have identical DNA has been known for some time. The most well-studied difference between monozygotic twins derives from a genetic phenomenon known as copy number variations. Certain, lengthy strands of nucleotides appear more than once in the genome, and the frequency of these repetitions can vary from one twin to another. By some estimates, copy number variations compose nearly 30 percent of a person's genetic code.

On the other side of the fence isn't "the whole scientific establishment" however. There are just a handful of names that always crop up when defending twin studies from criticism, mainly Nancy Segal. She's always the one person they cite when the field is criticized. And ... she took funding from The Pioneer Fund to do her major research, which is often cited as the creme la creme of the twin studies. They're eugenicists who have been described as a "neo-Nazi organization closely integrated with the far right in American politics." And that's an actual quote from a mainstream newspaper. Yeah, sure, it's not proof that your research is bullshit, but hell, taking your funding from Neo-Nazis to do science that backs up their racist claims makes you and your claims 100% suspect.

And some of the "defenses" are bafflingly bullshit when you even think about them for a second. For example, one criticism of twin studies is that parents empirically treat identical twins more similarly than they treat non-identical twins (and each twin might treat the other more differently), so in identical vs fraternal twin studies, this is a confounding factor. After all, small differences in how kids are treated over time will amplify. e.g. for non-identical twins the parents are more likely to label each twin's abilities differently, e.g. "the smart one" vs "the athletic one", whereas for identical twins they might view them more as peas in a pod. So those are the criticisms.

So the leading pro-twin study researcher Nancy Segal (the one funded by the white supremacist eugenics lobby) did this "defense" which involved picking random "lookalikes" from the general public and showing that they aren't treated similar. Which makes no sense at all because it's got nothing to do with how parents treat siblings:
https://jasoncollins.org/2014/08/28/twin-studies-stand-up-to-the-critique-again/
http://www.sciencedirect.com/science/article/pii/S0191886912003698
How the fuck does picking random pairs of strangers who look similar, and showing that they're not similar in behavior, disprove the criticism of twin studies? it's a complete non-sequitur / red-herring. And this is from the #1 most cited "expert" on twin studies.

So in other words, they do feel under pressure to defend twin studies from that criticism, but they haven't got a coherent response that actually uses twin data. So they concocted the above stunt / clearly-bullshit-research effort. It has so little actual bearing on twin studies or the criticisms that it's clear they're grasping at straws to even think about doing it. It wouldn't pass muster if you were a first-year psychology student.

The mistake is to think that psychometrics (which includes personality-test and stats-based stuff like IQ and twin studies) is "science" in the sense of biology, chemistry and physics. It really isn't.

~~~

EDIT: Here's another peer-review research paper. When you add in pre-natal (maternal) effects, which previous studies assume is negligible, it accounts for 20% of the similarity between twins and 5% of similarity between siblings, cutting right into the "genes did it" explanation.
http://www.nature.com/nature/journal/v388/n6641/full/388468a0.html
Quote
most previous models have assumed different degrees of similarity induced by environments specific to twins, to non-twin siblings (henceforth siblings), and to parents and offspring. We now evaluate an alternative model that replaces these three environments by two maternal womb environments, one for twins and another for siblings, along with a common home environment. Meta-analysis of 212 previous studies shows that our ‘maternal-effects’ model fits the data better than the ‘family-environments’ model. Maternal effects, often assumed to be negligible, account for 20% of covariance between twins and 5% between siblings, and the effects of genes are correspondingly reduced

That's basic proof of the unscientific nature of twin studies claims. It's a "god of the gaps" argument in other words. "Genes did it" is so vague, it can fill any gap we haven't explained yet in other more concrete terms. So this new model chipped right into the gap, and thus we reduce the amount that's attributable to "genes". But if it's so easy to find other ignored factors that cut huge chunks into the claimed heritability of IQ, then clearly more detailed and specific research on those lines would carve further big chunks out of it.

That's the big problem here, they're taking a component which is "unexplained" then making the claim that "genes" fits the gap in explained causes. Then if part of that chunk is explained by another factor, the remaining gap is still "genes". But that's negative proof, not positive proof. The onus is on the claimant to prove their factor caused the thing, not on critics to prove it didn't.
« Last Edit: January 01, 2017, 10:39:50 am by Reelya »
Logged

Rolepgeek

  • Bay Watcher
  • They see me rollin' they savin'~
    • View Profile
Re: SCIENCE, Gravitational waves, and the whole LIGO OST!
« Reply #4637 on: January 01, 2017, 03:34:05 pm »

Psychometrics is science in the same way economics is. Really damn hard to figure out.

It's very apparent that twin studies have flaws, and are an imperfect means of testing for genetic components. I'm not sure of any other way to test for complex trait effects (when there's less than a dozen or so I think genetic sequencing and comparisons become fairly useful? Not certain on that though) being influenced by genes, however, and using imperfect methods is still what science does, until a better method comes along. I'll admit that it's probable that it's very possible that IQ &related is less genetic than these studies purport, but genes are essentially Occam's Razor for development. The basic building blocks of our development and makeup having an impact on the development and makeup of our brains really isn't a stretch. Whether a parent considers their child the athletic or the academic child having an impact is a bit more of a stretch. Though I suppose that's subjective.

My apologies: I meant that from what I can tell it's generally accepted that twin studies are a useful way of gaining insight; most of the field uses, cites, and accepts the use of twin studies, not that the field as a whole defends them from criticism. As you've pointed out (many of these I was, in fact, unaware of), twin studies do have issues. So do most other methods of study, however, such as animal models, and when the field continues to use them...yes, I'm aware that the logic there is close to being circular "the field uses it so it's legitimate so the field should continue to use it". Again, if there's a better method, great.

I think my conclusion here, if nothing else, is that science once again proves to be really hard and figuring out what is and isn't true turns out to be really hard, once again. >.<

Though, considering that from what I've read, shared environment is very rarely found to influence outcomes, or that when it does it does so minimally, though usually it only considers family environment rather than the extremely important stage of development that is the nine months before birth, so I'm curious about that now.
« Last Edit: January 01, 2017, 03:39:10 pm by Rolepgeek »
Logged
Sincerely, Role P. Geek

Optimism is Painful.
Optimize anyway.

Reelya

  • Bay Watcher
    • View Profile
Re: SCIENCE, Gravitational waves, and the whole LIGO OST!
« Reply #4638 on: January 01, 2017, 08:56:27 pm »

Back on IQ, the real contention is the claim that they're measuring some "general" and innate "thing" called "g", which is a single slider that all "intelligent" people have. Even if there's massive variation, the idea is that there is some innate single "smartness" factor that at least partially correlates with test scores.

https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4557354/

Quote
Typically, in constructing a test, cognitive problems or items thought to engage aspects of intelligence are devised for presentation to testees in trials. Those items on which differences in performance agree with differences in the criterion are put together to make up an intelligence test. There are many other technical aspects of test construction, but this remains the essential rationale. Thus, nearly all contemporary tests, such as the Stanford-Binet or the Woodcock-Johnson tests, rely on correlations of scores with those from other IQ or achievement tests as evidence of validity.

So the problem with using correlations between IQ tests to prove that they're working is that it's circular logic. A lot of effort goes into making tests that correlate well with the Stanford Binet. Which implies that if you just throw an IQ test together there's little correlation between that and "mainstream" tests. Which sort of undermines the idea of "g" when you think about it. If g is so universal it shouldn't need massive statistical trickery to make test results match it.

Quote
It is widely accepted that test scores predict school achievement moderately well, with correlations of around 0.5 (Mackintosh, 2011). The problem lies in the possible self-fulfilment of this prediction because the measures are not independent. Rather they are merely different versions of the same test. Since the first test designers such as Binet, Terman, and others, test items have been devised, either with an eye on the kinds of knowledge and reasoning taught to, and required from, children in schools, or from an attempt to match an impression of the cognitive processes required in schools. This matching is an intuitively-, rather than a theoretically-guided, process, even with nonverbal items such as those in the Raven's Matrices. As Carpenter, Just, and Shell (1990) explained after examining John Raven's personal notes, “ … the description of the abilities that Raven intended to measure are primarily characteristics of the problems, not specifications of the requisite cognitive processes” (p. 408).

In other words, a correlation between IQ and school achievement may emerge because the test items demand the very kinds of (learned) linguistic and cognitive structures that are also the currency of schooling (Olson, 2005). As Thorndike and Hagen (1969) explained, “From the very way in which the tests were assembled [such correlation] could hardly be otherwise” (p. 325). Evidence for this is that correlations between IQ and school achievement tests tend to increase with age (Sternberg, Grigorenko, & Bundy, 2001). And this is why parental drive and encouragement with their children's school learning improves the children's IQ, as numerous results confirm (Nisbett, 2009; Nisbett et al., 2012).

They go on to explain in detail why large meta-analyses that claim a high correlation between IQ and work performance are dubious. Partly because many of the studies included weren't even IQ tests, they merely took any and all "work performance" reviews where there was at least one "test" done of some type, and made the assumption that "test scores" on literally anything were equally valid measures of IQ. Which is clearly some faulty reasoning right there, because it's posited on the conclusion "g factor makes you good on tests. All tests." which doesn't actually follow from how IQ tests are deliberately fine-tuned to correlate with each other, thus the correlation itself is a deliberate statistical artifact.

Quote
In a study of salespersons, Vinchur, Schippmann, Switzer, and Roth (1998) found that “general cognitive ability” showed a correlation of .40 with supervisor ratings but only .04 with objective sales.

So the test scores match subjective interpersonal ratings better than they match objective results? Interesting. But that could also be because of:

Quote
Another problem is the difficulty investigators have experienced in establishing reliabilities for supervisor ratings. Accurate reliabilities are needed, of course, in order to achieve the corrections to correlations. But they tend to be available for only a minority of the studies incorporated in the commonly cited meta-analyses. The strategy of Schmidt and Hunter and other meta-analysts has been to simply extrapolate from the average of those actually available. That strategy, of course, involves many assumptions about representativeness, randomness, uniformity across disparate samples, and so on. Using such a strategy, Hunter and Hunter (1984) assumed a reliability of 0.6 for their corrections, which some investigators have considered to be too low (Hartigan & Wigdor, 1989). Bertua et al. (2005) used the same figure for their meta-analysis of British studies. Moreover, that estimate was based on inter-rater reliability. Murphy and DeShon (2000) pointed out that differences between raters should not be considered error to be corrected because different raters may be looking for different things in a worker. Instead, intra-rater reliabilities should be used. However these tend to be much higher: 0.86 rather than 0.6. according to the meta-analysis carried out by Viswesvaran, Ones, and Schmidt (1996). The lower the value adopted, of course, the bigger the inflation to raw correlations. Using the reliability of 0.6, for example, inflates the correlations by 29%.

Oh, right, they assumed that human error got into the supervisor's ratings so they added a correction factor which boosted correlation by 0.29. Basically, most of the claimed correlation of the meta-analysis was contained in their correction factor which boosted correlation. And this is for the #1 cited meta-analysis of the entire IQ industry. That's even before we add in all the other methodological problems, such as most of the tests not even being IQ tests etc.

Quote
Schmidt and Hunter's approach (1977) insists that correcting for measurement error provide an estimate of the “true” correlation between the underlying constructs. Borsboom and Mellenbergh (2002), on the basis of classical test theory, have vehemently disagreed with this because it also assumes what it is trying to prove, namely the validity of that construct being revealed through the test-criterion correlation.

Ah right, the #1 cited study aimed at proving the validity of IQ testing needed to use circular logic that any data which didn't match it's conclusion was suspect and needed to have an adjustment factor added which made it match their input data better. In fact, the various adjustments and assumptions seem to add up to about 90% of the correlation they claim in the study's conclusion. Here's the final nail in the coffin of the meta-analysis which claims a 0.5 correlation between IQ testing an job performance

Quote
When Hartigan and Wigdor corrected the newer 264 studies for only sampling error (because they were suspicious of the empirical justification for other corrections) the correlations were very low (0.06–0.07) and virtually identical across job families.

Without all the statistically trickery and using larger better-designed sample sets, the true correlation drops by 90%, and they're no different for jobs which are supposed to need IQ as ones that aren't supposed to. e.g. the correlation between IQ test scores and job performance doesn't explain variance in performance for brain surgeons, mathematicians or physicists, any better than it does for mailmen, garbage collectors or McDonald's employees. This was actually the final nail in the coffin for IQ testing. IQ test variation can't even predict your real-world success at doing even things that are included in the test questions. But, of course the pro-IQ people still cite the old uncorrected results as proof.
« Last Edit: January 01, 2017, 10:40:08 pm by Reelya »
Logged

Loud Whispers

  • Bay Watcher
  • They said we have to aim higher, so we dug deeper.
    • View Profile
    • I APPLAUD YOU SIRRAH
Re: SCIENCE, Gravitational waves, and the whole LIGO OST!
« Reply #4639 on: January 03, 2017, 05:46:46 pm »

I just expect to see it all go to hell at some point when some other "clever" scientists find a link between race and intelligence, at which point the already established institutions will be quickly used to erase the "lesser" races from existence.
Like it matters, the established institutions are full of people who don't have kids, they can't do shit lmao

Btw this is retarded logic, I'm pretty certain we can give infertile women kids without becoming Hitler. K thanks for trying come again

Starver

  • Bay Watcher
    • View Profile
Re: SCIENCE, Gravitational waves, and the whole LIGO OST!
« Reply #4640 on: January 03, 2017, 06:07:57 pm »

I really have no idea where on earth you're going there, except for maybe just Something Generically Edgy™ and a chance to use your favourite ETLA. Care to be constructive, ever?

Edit1: Yes, I see the irony of saying that.
Edit2: Above written before seeing your "Internet habits which annoy you" post...  <sef>
« Last Edit: January 03, 2017, 06:12:51 pm by Starver »
Logged

Loud Whispers

  • Bay Watcher
  • They said we have to aim higher, so we dug deeper.
    • View Profile
    • I APPLAUD YOU SIRRAH
Re: SCIENCE, Gravitational waves, and the whole LIGO OST!
« Reply #4641 on: January 03, 2017, 06:17:23 pm »

I really have no idea where on earth you're going there, except for maybe just Something Generically Edgy™ and a chance to use your favourite ETLA. Care to be constructive, ever?

Edit1: Yes, I see the irony of saying that.
Edit2: Above written before seeing your "Internet habits which annoy you" post...  <sef>
I made a constructive post but the response was "yeah, but what if nazis?"

So yeah nah

Sergarr

  • Bay Watcher
  • (9) airheaded baka (9)
    • View Profile
Re: SCIENCE, Gravitational waves, and the whole LIGO OST!
« Reply #4642 on: January 05, 2017, 11:34:06 am »

So Google has decided to test their next iteration of AlphaGo online against the best human players in existence, in secret. Result? It beat all of them, with a more than 50-0 win streak.

Looks like Lee Sedol's first win against AlphaGo may also be human's last victory in Go against top-line AI. One step closer to the AI revolution!
Logged
._.

forsaken1111

  • Bay Watcher
    • View Profile
    • TTB Twitch
Re: SCIENCE, Gravitational waves, and the whole LIGO OST!
« Reply #4643 on: January 05, 2017, 12:09:53 pm »

So Google should be banned for botting?
Logged

Sergarr

  • Bay Watcher
  • (9) airheaded baka (9)
    • View Profile
Re: SCIENCE, Gravitational waves, and the whole LIGO OST!
« Reply #4644 on: January 05, 2017, 10:08:44 pm »

So Google should be banned for botting?
That's probably not in the rules. Besides, banning Google is like banning yourself from the Internet.

Meanwhile, some Indians have managed to install a carbon capture technology on their coal plant. They bake soda with it - and they also didn't use any subsidies to do it, so it should be financially viable. One step closer to the stable environment!
Logged
._.

wierd

  • Bay Watcher
  • I like to eat small children.
    • View Profile
Re: SCIENCE, Gravitational waves, and the whole LIGO OST!
« Reply #4645 on: January 05, 2017, 10:16:16 pm »

Widespread adoption will drive bicarbonate prices way down, due to supply saturation. (See why current oil prices are depressed)

Without a good industrial use for bicarbonate that does not involve future liberation of the co2, that drives up demand for the end product, this is only a niche solution.

If you ask me, synthetic asphalt is the big one.

People won't stop driving just because we are axing fossil fuels. Road surfaces are made from oil refinery waste, and demand for surfacing material won't go away. With supply side contraction caused by reduced demand for fuel oil, there will be less asphalt on the market and prices will rise, causing economic havok for growing civic centers. Artificial asphalt would provide replacement supply for the surfacing needs of society, and give a place for carbon to be sequestered for at least several decades at a time.

Logged

Rolepgeek

  • Bay Watcher
  • They see me rollin' they savin'~
    • View Profile
Re: SCIENCE, Gravitational waves, and the whole LIGO OST!
« Reply #4646 on: January 05, 2017, 11:30:05 pm »

I figure we just use the carbon for specialized greenhouses. Or find a way to resynthesize it into plastics.
Logged
Sincerely, Role P. Geek

Optimism is Painful.
Optimize anyway.

Reelya

  • Bay Watcher
    • View Profile
Re: SCIENCE, Gravitational waves, and the whole LIGO OST!
« Reply #4647 on: January 06, 2017, 12:04:25 am »

Look into the processes for making sodium carbonate right now, they're pretty polluting, and need their own specialized plants. So this new stuff will drive some of the existing bicarb out of the market.

https://en.wikipedia.org/wiki/Solvay_process

^ If sodium carbonate as a byproduct comes in they'll be doing less of this. The Solvay process makes about 30 million tons of sodium carbonate per year, which is 75% of global production. The carbon plant in India produces 66000 tons. So it would need 500 similar plants to completely replace what's made by the Solvay Process. But that's a good thing:

For every ton of soda ash the Solvay Process produces, it creates two tons of crud that needs to be landfilled, and that is polluting water systems. So this new process that creates it as a byproduct of another process is only good news.
« Last Edit: January 06, 2017, 12:10:03 am by Reelya »
Logged

alway

  • Bay Watcher
  • 🏳️‍⚧️
    • View Profile
Re: SCIENCE, Gravitational waves, and the whole LIGO OST!
« Reply #4648 on: January 06, 2017, 01:06:31 am »

Er, the article which was linked to is literally using that same method. It's just a slightly modified solvay process that happens to be sourcing its CO2 from a coal plant. And it's really not ever going to be a significant way to deal with CO2 unless you can find a use for billions of tons of baking soda per year. Clean coal is a scam. As is coal power in general these days, considering how wind, solar, and natural gas prices have dropped through the floor. Hence why all the coal related companies are going bankrupt.
Logged

Egan_BW

  • Bay Watcher
    • View Profile
Re: SCIENCE, Gravitational waves, and the whole LIGO OST!
« Reply #4649 on: January 06, 2017, 01:07:46 am »

Quick, pretend that steam engines are back in style!
Logged
I would starve tomorrow if I could eat the world today.
Pages: 1 ... 308 309 [310] 311 312 ... 339