Total Posts:63|Showing Posts:1-30|Last Page
Jump to topic:

IQ and methodology

Hoppi
Posts: 1,655
Add as Friend
Challenge to a Debate
Send a Message
5/7/2016 10:56:43 AM
Posted: 7 months ago
I want to talk about BOT's favorite research paper which can be seen here.

https://www.udel.edu...

Rushton & Jensen claim to have 10 categories of evidence that show that the black-white IQ difference is genetic.

I've got a stressful, real life deadline and instead of facing it like a grown up, I thought I'd come here and go through the R&J evidence study by study.
Hoppi
Posts: 1,655
Add as Friend
Challenge to a Debate
Send a Message
5/7/2016 11:29:01 AM
Posted: 7 months ago
Okay, I'm skipping all that first philosophical framework bit for now. Let's start at the bottom of page 240 where they start to present evidence. They say that

Racial-group differences in IQ appear early. For example, the Black and the White 3-year-old children in the standardization sample of the Stanford"Binet IV show a 1 standard deviation mean difference after being matched on gender, birth order, and maternal education (Peoples, Fagan, & Drotar, 1995).

The samples were not matched on economic standing, though. From the peoples paper:

Roughly 50% of the African American children received welfare benefits (i.e., Aid to Dependent Children; ADC), whereas none of the European American children received ADC benefits. (p.72)

Poverty has been consistently shown to lower performance on everything, including IQ tests. Here are some links about it. Interestingly, one of them argues that poverty lowers IQ by up to 13 points which is greater than the average IQ difference between the samples in the Peoples et al study.

https://www.princeton.edu...
https://www.sciencedaily.com...
http://www.businessinsider.com.au...

In their discussion, Peoples et al point out that their findings are different from other studies, which did not find such a big IQ difference, and that those other studies matched the samples on economic status.

First, the racial groups were not matched for SES, in terms of earnings, which could contribute to the larger racial group differences found in this study as compared to those demonstrated in studies in which the groups were matched on SES. (p.81)

In summary, rich kids did better on a test than poor kids. I don't know why the researchers wanted to make that about race and trying to guess about their motives is depressing.
Hoppi
Posts: 1,655
Add as Friend
Challenge to a Debate
Send a Message
5/7/2016 11:44:20 AM
Posted: 7 months ago
They continue at the bottom of page 40:

Similarly, the Black and the White 21R60;2- to 6-year-old children in the U.S. standardization sample of the Differential Aptitude Scale have a 1 standard deviation mean difference. No data are available for East Asian children at the youngest ages. On the Differential Aptitude Battery, by age 6, however, the average IQ of East Asian children is 107, compared with 103 for White children and 89 for Black children (Lynn, 1996).

Lynn uses data from another study, he says that,

The DAS was standardized in the United States in 1986 on a representative sample of the population stratified by age, sex, race (black, Hispanic, white, Asian), geographical location, urban-rural areas, parental socio-economic status and educational preschool enrolment. The characteristics of the total population to which the standardization sample was matched were obtained from the Current Population Survey of 1986.

In 1986, black families were earning on average just over half of the white family average, according to this blog: http://www.epi.org...

So it's the same rich-poor comparison. Until that's accounted for, there's no justification for assuming that any differences are due to race.
Hoppi
Posts: 1,655
Add as Friend
Challenge to a Debate
Send a Message
5/7/2016 12:14:53 PM
Posted: 7 months ago
The next paragraph is about IQ scores in sub-Saharan Africa.

This paper is a really nice critique of that. Especially the part of Raven's matrices having a much lower loading on g in African samples. Which means, it's not a very good measure of intelligence across cultures.

http://www.iapsych.com...
dylancatlow
Posts: 12,244
Add as Friend
Challenge to a Debate
Send a Message
5/7/2016 2:28:47 PM
Posted: 7 months ago
They've done studies which control for economic factors and the findings are always the same: the black white IQ gap does not go away. In other words, if you look at a black family and a white family making the same income, the children from the white family still outperform the children from the black family.

Pretty much all of the studies purporting to show that poverty lowers IQ make the same elementary mistake: they interpret the correlation as meaning that poverty lowers IQ and not that low IQ causes poverty. Of course, if IQ tests are meanginful tests of intelligence to begin with, it pretty much has to be the case that low IQ and poverty will go together more often than pure chance would dictate.
Hoppi
Posts: 1,655
Add as Friend
Challenge to a Debate
Send a Message
5/7/2016 4:22:08 PM
Posted: 7 months ago
At 5/7/2016 2:28:47 PM, dylancatlow wrote:
They've done studies which control for economic factors and the findings are always the same: the black white IQ gap does not go away. In other words, if you look at a black family and a white family making the same income, the children from the white family still outperform the children from the black family.

This is a thread about evidence, not hearsay.

Pretty much all of the studies purporting to show that poverty lowers IQ make the same elementary mistake: they interpret the correlation as meaning that poverty lowers IQ and not that low IQ causes poverty.

Lol at "pretty much all". I think that means you're guessing and you haven't read any except third hand.

Of course, if IQ tests are meanginful tests of intelligence to begin with, it pretty much has to be the case that low IQ and poverty will go together more often than pure chance would dictate.

Thanks for sharing your beliefs.
lamerde
Posts: 1,416
Add as Friend
Challenge to a Debate
Send a Message
5/7/2016 4:45:17 PM
Posted: 7 months ago
At 5/7/2016 4:22:08 PM, Hoppi wrote:

This is a thread about evidence, not hearsay.

Thanks for sharing your beliefs.

LOL
Why I ignore YYW:
http://www.debate.org...
http://www.debate.org...
Calling someone a bitch multiple times while claiming you're taking the high road is an art form, I suppose: http://www.debate.org...
dylancatlow
Posts: 12,244
Add as Friend
Challenge to a Debate
Send a Message
5/7/2016 5:54:34 PM
Posted: 7 months ago
At 5/7/2016 4:22:08 PM, Hoppi wrote:
At 5/7/2016 2:28:47 PM, dylancatlow wrote:
They've done studies which control for economic factors and the findings are always the same: the black white IQ gap does not go away. In other words, if you look at a black family and a white family making the same income, the children from the white family still outperform the children from the black family.

This is a thread about evidence, not hearsay.

Pretty much all of the studies purporting to show that poverty lowers IQ make the same elementary mistake: they interpret the correlation as meaning that poverty lowers IQ and not that low IQ causes poverty.

Lol at "pretty much all". I think that means you're guessing and you haven't read any except third hand.

Of course, if IQ tests are meanginful tests of intelligence to begin with, it pretty much has to be the case that low IQ and poverty will go together more often than pure chance would dictate.

Thanks for sharing your beliefs.

It's clear that you approach this issue from an objective standpoint lol
lamerde
Posts: 1,416
Add as Friend
Challenge to a Debate
Send a Message
5/7/2016 6:00:42 PM
Posted: 7 months ago
At 5/7/2016 5:54:34 PM, dylancatlow wrote:

It's clear that you approach this issue from an objective standpoint lol

He said the thread is about evidence, not hearsay, because you posted "They've done studies" without specifying who or posting links to said studies. How is that not objective? If you think what you said is rooted in evidence, post the links and keep it moving.
Why I ignore YYW:
http://www.debate.org...
http://www.debate.org...
Calling someone a bitch multiple times while claiming you're taking the high road is an art form, I suppose: http://www.debate.org...
dylancatlow
Posts: 12,244
Add as Friend
Challenge to a Debate
Send a Message
5/7/2016 6:33:27 PM
Posted: 7 months ago
At 5/7/2016 6:00:42 PM, lamerde wrote:
At 5/7/2016 5:54:34 PM, dylancatlow wrote:

It's clear that you approach this issue from an objective standpoint lol

He said the thread is about evidence, not hearsay, because you posted "They've done studies" without specifying who or posting links to said studies. How is that not objective? If you think what you said is rooted in evidence, post the links and keep it moving.

The reason I didn't cite evidence for the claim is because it's not even controversial, and should be well known to anyone who knows what they're talking about. Here's the data visualized:http://1.bp.blogspot.com...

College Board publishes sat scores of people based on race and income which also confirms the claim: http://www.jbhe.com...
someloser
Posts: 1,377
Add as Friend
Challenge to a Debate
Send a Message
5/7/2016 8:38:05 PM
Posted: 7 months ago
At 5/7/2016 11:29:01 AM, Hoppi wrote:
Okay, I'm skipping all that first philosophical framework bit for now. Let's start at the bottom of page 240 where they start to present evidence. They say that

Racial-group differences in IQ appear early. For example, the Black and the White 3-year-old children in the standardization sample of the Stanford"Binet IV show a 1 standard deviation mean difference after being matched on gender, birth order, and maternal education (Peoples, Fagan, & Drotar, 1995).

The samples were not matched on economic standing, though. From the peoples paper:

Roughly 50% of the African American children received welfare benefits (i.e., Aid to Dependent Children; ADC), whereas none of the European American children received ADC benefits. (p.72)

Not that it matters - IQ gaps are even larger at higher SES (even with children, see Jensen 1973, Gottfredson 2003, or Herrnstein and Murray 1994 p. 288). Data from pre-1994 SATs should confirm that it's g loaded as well.

Poverty has been consistently shown to lower performance on everything, including IQ tests. Here are some links about it. Interestingly, one of them argues that poverty lowers IQ by up to 13 points which is greater than the average IQ difference between the samples in the Peoples et al study.

https://www.princeton.edu...
https://www.sciencedaily.com...
http://www.businessinsider.com.au...

These are all citing the same study (Mani et al.), that Wicherts and Scholten 2013 (http://science.sciencemag.org...) picked apart at length. When Mani et al. responded to Wicherts and Scholten, it didn't exactly help their case very much ("it's standard procedure, our data was 'noisy'!")

Either way, the study's implications, assuming they were true, cannot explain why, between childhood SES and IQ, the latter is a much better predictor (Streze 2007, Colom and Flores-Medoza 2007). It should be noted that the hereditarian standpoint isn't exactly incompatible with their data's implications (whatever they may be at this point) either.

So, if you ignore dubious methodology, lack of an established causal relationship, and ignore all the data on the effects of IQ on later SES, it might prove what you think it does. But you shouldn't.

In summary, rich kids did better on a test than poor kids.
And is that why upper-class black kids do almost as badly as the poorest white kids (Leohlin Lindsey and Spuhler 1975, Black Journal of Higher Education)? The wealth explanation doesn't really pan out when you compare the IQs of, say, poor Chinese farmers with US blacks who earn ten times their income.
Ego sum qui sum. Deus lo vult.

"America is ungovernable; those who served the revolution have plowed the sea." - Simon Bolivar

"A healthy nation is as unconscious of its nationality as a healthy man of his bones. But if you break a nation's nationality it will think of nothing else but getting it set again." - George Bernard Shaw
Hoppi
Posts: 1,655
Add as Friend
Challenge to a Debate
Send a Message
5/8/2016 12:38:42 AM
Posted: 7 months ago
At 5/7/2016 6:33:27 PM, dylancatlow wrote:
At 5/7/2016 6:00:42 PM, lamerde wrote:
At 5/7/2016 5:54:34 PM, dylancatlow wrote:

It's clear that you approach this issue from an objective standpoint lol

He said the thread is about evidence, not hearsay, because you posted "They've done studies" without specifying who or posting links to said studies. How is that not objective? If you think what you said is rooted in evidence, post the links and keep it moving.

The reason I didn't cite evidence for the claim is because it's not even controversial, and should be well known to anyone who knows what they're talking about. Here's the data visualized:http://1.bp.blogspot.com...

College Board publishes sat scores of people based on race and income which also confirms the claim: http://www.jbhe.com...

You said children. The peoples et al study was looking a preschool children. They're not taking college entrance exams, and an unrefernced blog chart doesn't count as evidence. We need to see where the data comes from.
Hoppi
Posts: 1,655
Add as Friend
Challenge to a Debate
Send a Message
5/8/2016 12:55:33 AM
Posted: 7 months ago
At 5/7/2016 8:38:05 PM, someloser wrote:
At 5/7/2016 11:29:01 AM, Hoppi wrote:
Okay, I'm skipping all that first philosophical framework bit for now. Let's start at the bottom of page 240 where they start to present evidence. They say that

Racial-group differences in IQ appear early. For example, the Black and the White 3-year-old children in the standardization sample of the Stanford"Binet IV show a 1 standard deviation mean difference after being matched on gender, birth order, and maternal education (Peoples, Fagan, & Drotar, 1995).

The samples were not matched on economic standing, though. From the peoples paper:

Roughly 50% of the African American children received welfare benefits (i.e., Aid to Dependent Children; ADC), whereas none of the European American children received ADC benefits. (p.72)

Not that it matters - IQ gaps are even larger at higher SES (even with children, see Jensen 1973, Gottfredson 2003, or Herrnstein and Murray 1994 p. 288). Data from pre-1994 SATs should confirm that it's g loaded as well.

R&J were trying to argue that the IQ differences appear in preschoolers. What age-group do your studies look at? Because once children are part of the educational system, and taking tests in the classroom, the issues are different.

Poverty has been consistently shown to lower performance on everything, including IQ tests. Here are some links about it. Interestingly, one of them argues that poverty lowers IQ by up to 13 points which is greater than the average IQ difference between the samples in the Peoples et al study.

https://www.princeton.edu...
https://www.sciencedaily.com...
http://www.businessinsider.com.au...

These are all citing the same study (Mani et al.), that Wicherts and Scholten 2013 (http://science.sciencemag.org...) picked apart at length. When Mani et al. responded to Wicherts and Scholten, it didn't exactly help their case very much ("it's standard procedure, our data was 'noisy'!")

They criticized the first study. They didn't mention the second. Anyway, the point is that the study was presented as matched samples and it wasn't. In the paper itself, peoples et al referred to other studies with different findings, that did take SES into account. R&J didn't mention any of that, which shows that they are not presenting evidence fairly so far.

Either way, the study's implications, assuming they were true, cannot explain why, between childhood SES and IQ, the latter is a much better predictor (Streze 2007, Colom and Flores-Medoza 2007).

Predictor of what?

It should be noted that the hereditarian standpoint isn't exactly incompatible with their data's implications (whatever they may be at this point) either.

If you can't articulate what the implications of the data are, then how could any standpoint be incompatible?

So, if you ignore dubious methodology, lack of an established causal relationship, and ignore all the data on the effects of IQ on later SES, it might prove what you think it does. But you shouldn't.

You're missing the point. R&J argued that there's a difference in IQ by race in preschoolers, but the samples being compared were not matched, and so the evidence isn't sufficient to support their view.

In summary, rich kids did better on a test than poor kids.
And is that why upper-class black kids do almost as badly as the poorest white kids (Leohlin Lindsey and Spuhler 1975, Black Journal of Higher Education)? The wealth explanation doesn't really pan out when you compare the IQs of, say, poor Chinese farmers with US blacks who earn ten times their income.

This is data from other age groups, i assume. R&J have 10 categories of evidence, so maybe that will come up later in the paper.
Hoppi
Posts: 1,655
Add as Friend
Challenge to a Debate
Send a Message
5/8/2016 1:13:43 AM
Posted: 7 months ago
At the bottom of page 241, they talk about South Africa. The first couple of studies are uncontrolled (no comparison group).

Page 242, Rushton & Skue gave the Raven's matrices to psychology students. The average score was 83. They then got training in how to do those sorts of problems, and the score rose by 13 points to 96! That's more than 1 standard deviation!

So their own studies show clear training and practice effects, of a magnitude equal to the racial differences they're carrying on about.
someloser
Posts: 1,377
Add as Friend
Challenge to a Debate
Send a Message
5/8/2016 1:41:17 AM
Posted: 7 months ago
At 5/8/2016 12:55:33 AM, Hoppi wrote:
At 5/7/2016 8:38:05 PM, someloser wrote:
At 5/7/2016 11:29:01 AM, Hoppi wrote:
Okay, I'm skipping all that first philosophical framework bit for now. Let's start at the bottom of page 240 where they start to present evidence. They say that

Racial-group differences in IQ appear early. For example, the Black and the White 3-year-old children in the standardization sample of the Stanford"Binet IV show a 1 standard deviation mean difference after being matched on gender, birth order, and maternal education (Peoples, Fagan, & Drotar, 1995).

The samples were not matched on economic standing, though. From the peoples paper:

Roughly 50% of the African American children received welfare benefits (i.e., Aid to Dependent Children; ADC), whereas none of the European American children received ADC benefits. (p.72)

Not that it matters - IQ gaps are even larger at higher SES (even with children, see Jensen 1973, Gottfredson 2003, or Herrnstein and Murray 1994 p. 288). Data from pre-1994 SATs should confirm that it's g loaded as well.

R&J were trying to argue that the IQ differences appear in preschoolers. What age-group do your studies look at? Because once children are part of the educational system, and taking tests in the classroom, the issues are different.
My mistake. I think only Gottfredson looked at children specifically. Here's some data on the gap in preschoolers: http://humanvarieties.org...

Poverty has been consistently shown to lower performance on everything, including IQ tests. Here are some links about it. Interestingly, one of them argues that poverty lowers IQ by up to 13 points which is greater than the average IQ difference between the samples in the Peoples et al study.

https://www.princeton.edu...
https://www.sciencedaily.com...
http://www.businessinsider.com.au...

These are all citing the same study (Mani et al.), that Wicherts and Scholten 2013 (http://science.sciencemag.org...) picked apart at length. When Mani et al. responded to Wicherts and Scholten, it didn't exactly help their case very much ("it's standard procedure, our data was 'noisy'!")

They criticized the first study. They didn't mention the second.
Missed that one. Tucker-Drob and Gates 2015? Much better paper, though it didn't find IQ depression, just slightly lowered heritability (a' = .074) in impoverished groups in the US. It might have more of an effect on children, though, given that heritability is lower for them across the board. The effects were negligible, when not opposite, outside the US.

Anyway, the point is that the study was presented as matched samples and it wasn't. In the paper itself, peoples et al referred to other studies with different findings, that did take SES into account. R&J didn't mention any of that, which shows that they are not presenting evidence fairly so far.
Fair enough

Either way, the study's implications, assuming they were true, cannot explain why, between childhood SES and IQ, the latter is a much better predictor (Streze 2007, Colom and Flores-Medoza 2007).

Predictor of what?
Future individual SES, can't believe I didn't notice that got left out.

It should be noted that the hereditarian standpoint isn't exactly incompatible with their data's implications (whatever they may be at this point) either.

If you can't articulate what the implications of the data are, then how could any standpoint be incompatible?
Because the implications are pretty uninformative. Population x does better than population y, when sorted on the basis of differences in economic status.
Ego sum qui sum. Deus lo vult.

"America is ungovernable; those who served the revolution have plowed the sea." - Simon Bolivar

"A healthy nation is as unconscious of its nationality as a healthy man of his bones. But if you break a nation's nationality it will think of nothing else but getting it set again." - George Bernard Shaw
someloser
Posts: 1,377
Add as Friend
Challenge to a Debate
Send a Message
5/8/2016 2:40:07 AM
Posted: 7 months ago
At 5/8/2016 1:13:43 AM, Hoppi wrote:
So their own studies show clear training and practice effects, of a magnitude equal to the racial differences they're carrying on about.
Sure, if you're specifically training to beat an IQ test. Is there any evidence to suggest whites across the socioeconomic spectrum, en-masse, are engaging in any such activity?
Ego sum qui sum. Deus lo vult.

"America is ungovernable; those who served the revolution have plowed the sea." - Simon Bolivar

"A healthy nation is as unconscious of its nationality as a healthy man of his bones. But if you break a nation's nationality it will think of nothing else but getting it set again." - George Bernard Shaw
someloser
Posts: 1,377
Add as Friend
Challenge to a Debate
Send a Message
5/8/2016 2:51:26 AM
Posted: 7 months ago
Additionally, if you train to beat an IQ test, then the extra points wouldn't be g loaded. So you wouldn't see any of the real-world gains associated with a higher IQ that would be expected from a "real" IQ gain.
Ego sum qui sum. Deus lo vult.

"America is ungovernable; those who served the revolution have plowed the sea." - Simon Bolivar

"A healthy nation is as unconscious of its nationality as a healthy man of his bones. But if you break a nation's nationality it will think of nothing else but getting it set again." - George Bernard Shaw
Hoppi
Posts: 1,655
Add as Friend
Challenge to a Debate
Send a Message
5/8/2016 5:04:10 AM
Posted: 7 months ago
At 5/8/2016 1:41:17 AM, someloser wrote:
At 5/8/2016 12:55:33 AM, Hoppi wrote:
At 5/7/2016 8:38:05 PM, someloser wrote:
At 5/7/2016 11:29:01 AM, Hoppi wrote:
Okay, I'm skipping all that first philosophical framework bit for now. Let's start at the bottom of page 240 where they start to present evidence. They say that

Racial-group differences in IQ appear early. For example, the Black and the White 3-year-old children in the standardization sample of the Stanford"Binet IV show a 1 standard deviation mean difference after being matched on gender, birth order, and maternal education (Peoples, Fagan, & Drotar, 1995).

The samples were not matched on economic standing, though. From the peoples paper:

Roughly 50% of the African American children received welfare benefits (i.e., Aid to Dependent Children; ADC), whereas none of the European American children received ADC benefits. (p.72)

Not that it matters - IQ gaps are even larger at higher SES (even with children, see Jensen 1973, Gottfredson 2003, or Herrnstein and Murray 1994 p. 288). Data from pre-1994 SATs should confirm that it's g loaded as well.

R&J were trying to argue that the IQ differences appear in preschoolers. What age-group do your studies look at? Because once children are part of the educational system, and taking tests in the classroom, the issues are different.
My mistake. I think only Gottfredson looked at children specifically. Here's some data on the gap in preschoolers: http://humanvarieties.org...

That link uses the R&J paper as a reference! It's like a tertiary source. Do you have a particular reason for linking to that?

Poverty has been consistently shown to lower performance on everything, including IQ tests. Here are some links about it. Interestingly, one of them argues that poverty lowers IQ by up to 13 points which is greater than the average IQ difference between the samples in the Peoples et al study.

https://www.princeton.edu...
https://www.sciencedaily.com...
http://www.businessinsider.com.au...

These are all citing the same study (Mani et al.), that Wicherts and Scholten 2013 (http://science.sciencemag.org...) picked apart at length. When Mani et al. responded to Wicherts and Scholten, it didn't exactly help their case very much ("it's standard procedure, our data was 'noisy'!")

They criticized the first study. They didn't mention the second.
Missed that one. Tucker-Drob and Gates 2015? Much better paper, though it didn't find IQ depression, just slightly lowered heritability (a' = .074) in impoverished groups in the US. It might have more of an effect on children, though, given that heritability is lower for them across the board. The effects were negligible, when not opposite, outside the US.

Anyway, the point is that the study was presented as matched samples and it wasn't. In the paper itself, peoples et al referred to other studies with different findings, that did take SES into account. R&J didn't mention any of that, which shows that they are not presenting evidence fairly so far.
Fair enough

Either way, the study's implications, assuming they were true, cannot explain why, between childhood SES and IQ, the latter is a much better predictor (Streze 2007, Colom and Flores-Medoza 2007).

Predictor of what?
Future individual SES, can't believe I didn't notice that got left out.

I haven't read those papers. Did they really use a longitudinal design or did they find a higher correlation between adult IQ and adult SES than between retrospectively reported childhood SES and adult SES?

It should be noted that the hereditarian standpoint isn't exactly incompatible with their data's implications (whatever they may be at this point) either.

If you can't articulate what the implications of the data are, then how could any standpoint be incompatible?
Because the implications are pretty uninformative. Population x does better than population y, when sorted on the basis of differences in economic status.

If the implications are uninformative, you can believe what you like...
someloser
Posts: 1,377
Add as Friend
Challenge to a Debate
Send a Message
5/8/2016 5:09:23 AM
Posted: 7 months ago
At 5/8/2016 5:04:10 AM, Hoppi wrote:
That link uses the R&J paper as a reference!
Yeah... and it's far from their only reference -- the post's analysis (i.e. the reason I linked it in the first place) controlled for SES.
Ego sum qui sum. Deus lo vult.

"America is ungovernable; those who served the revolution have plowed the sea." - Simon Bolivar

"A healthy nation is as unconscious of its nationality as a healthy man of his bones. But if you break a nation's nationality it will think of nothing else but getting it set again." - George Bernard Shaw
Hoppi
Posts: 1,655
Add as Friend
Challenge to a Debate
Send a Message
5/8/2016 5:11:27 AM
Posted: 7 months ago
At 5/8/2016 2:40:07 AM, someloser wrote:
At 5/8/2016 1:13:43 AM, Hoppi wrote:
So their own studies show clear training and practice effects, of a magnitude equal to the racial differences they're carrying on about.
Sure, if you're specifically training to beat an IQ test. Is there any evidence to suggest whites across the socioeconomic spectrum, en-masse, are engaging in any such activity?

You're confusing where the burden of proof lies in this discussion. R&J set out to prove that there is a genetic racial difference in IQ and blade_of_bigotry has used this paper as justification for his claim that white people are superior to black people. To support this claim, there needs to be evidence.

If IQ tests are sensitive to training and practice effects - and they are - then one obvious explanation for any differences between groups is a difference in practice and training experience. The burden is on R&J to rule this out. Maybe they will later on in the paper.

But yes, the education system provides plenty of practice experience on IQ-like tasks. Here's an example of a pattern recognition task that would set children up to do tests like Raven's matrices.

http://www.k5learning.com...
someloser
Posts: 1,377
Add as Friend
Challenge to a Debate
Send a Message
5/8/2016 5:15:41 AM
Posted: 7 months ago
At 5/8/2016 5:04:10 AM, Hoppi wrote:
I haven't read those papers. Did they really use a longitudinal design or did they find a higher correlation between adult IQ and adult SES than between retrospectively reported childhood SES and adult SES?
The former (which is Strenze 2007, not "Streze", that was a typo on my end) did. It is actually a meta-analysis of other longitudinal studies. Not sure about the latter
Ego sum qui sum. Deus lo vult.

"America is ungovernable; those who served the revolution have plowed the sea." - Simon Bolivar

"A healthy nation is as unconscious of its nationality as a healthy man of his bones. But if you break a nation's nationality it will think of nothing else but getting it set again." - George Bernard Shaw
Hoppi
Posts: 1,655
Add as Friend
Challenge to a Debate
Send a Message
5/8/2016 5:20:20 AM
Posted: 7 months ago
At 5/8/2016 5:09:23 AM, someloser wrote:
At 5/8/2016 5:04:10 AM, Hoppi wrote:
That link uses the R&J paper as a reference!
Yeah... and it's far from their only reference -- the post's analysis (i.e. the reason I linked it in the first place) controlled for SES.

The only study in there that seems to be controlled for SES is the headstart one, but I can't see what analysis he used. when I followed the link, there's a dataset, but is it even coded for race?
Hoppi
Posts: 1,655
Add as Friend
Challenge to a Debate
Send a Message
5/8/2016 5:21:35 AM
Posted: 7 months ago
At 5/8/2016 5:15:41 AM, someloser wrote:
At 5/8/2016 5:04:10 AM, Hoppi wrote:
I haven't read those papers. Did they really use a longitudinal design or did they find a higher correlation between adult IQ and adult SES than between retrospectively reported childhood SES and adult SES?
The former (which is Strenze 2007, not "Streze", that was a typo on my end) did. It is actually a meta-analysis of other longitudinal studies. Not sure about the latter

If it's a meta-analysis, then it's not a longitudinal study. If you don't know what the evidence is, then it's not worth discussing here.
someloser
Posts: 1,377
Add as Friend
Challenge to a Debate
Send a Message
5/8/2016 5:24:53 AM
Posted: 7 months ago
At 5/8/2016 5:11:27 AM, Hoppi wrote:
If IQ tests are sensitive to training and practice effects - and they are - then one obvious explanation for any differences between groups is a difference in practice and training experience. The burden is on R&J to rule this out. Maybe they will later on in the paper.
It's obvious, if you assume that one group is collectively being trained to beat IQ tests throughout the entire socioeconomic spectrum.

That being said, there is evidence that IQ "practice" (like the type you provided) has some effect on the gaps. Note, for example, that the size of the gaps increases the more g-loaded a test is (te Nijenhuis, 2012). This would imply that the IQ training is artificially shrinking the gaps.
Ego sum qui sum. Deus lo vult.

"America is ungovernable; those who served the revolution have plowed the sea." - Simon Bolivar

"A healthy nation is as unconscious of its nationality as a healthy man of his bones. But if you break a nation's nationality it will think of nothing else but getting it set again." - George Bernard Shaw
Hoppi
Posts: 1,655
Add as Friend
Challenge to a Debate
Send a Message
5/8/2016 5:27:28 AM
Posted: 7 months ago
At 5/8/2016 5:24:53 AM, someloser wrote:
At 5/8/2016 5:11:27 AM, Hoppi wrote:
If IQ tests are sensitive to training and practice effects - and they are - then one obvious explanation for any differences between groups is a difference in practice and training experience. The burden is on R&J to rule this out. Maybe they will later on in the paper.
It's obvious, if you assume that one group is collectively being trained to beat IQ tests throughout the entire socioeconomic spectrum.

That being said, there is evidence that IQ "practice" (like the type you provided) has some effect on the gaps. Note, for example, that the size of the gaps increases the more g-loaded a test is (te Nijenhuis, 2012). This would imply that the IQ training is artificially shrinking the gaps.

If you're going to reference papers, please link to them.
Hoppi
Posts: 1,655
Add as Friend
Challenge to a Debate
Send a Message
5/8/2016 5:29:19 AM
Posted: 7 months ago
At 5/8/2016 2:51:26 AM, someloser wrote:
Additionally, if you train to beat an IQ test, then the extra points wouldn't be g loaded. So you wouldn't see any of the real-world gains associated with a higher IQ that would be expected from a "real" IQ gain.

I disagree. IQ is a measure of potential for learning in a traditional classroom environment. Obviously, it's going to be affected by literacy and familiarity with test-taking. That's part of the construct.
someloser
Posts: 1,377
Add as Friend
Challenge to a Debate
Send a Message
5/8/2016 5:46:24 AM
Posted: 7 months ago
At 5/8/2016 5:20:20 AM, Hoppi wrote:
At 5/8/2016 5:09:23 AM, someloser wrote:
At 5/8/2016 5:04:10 AM, Hoppi wrote:
That link uses the R&J paper as a reference!
Yeah... and it's far from their only reference -- the post's analysis (i.e. the reason I linked it in the first place) controlled for SES.

The only study in there that seems to be controlled for SES is the headstart one, but I can't see what analysis he used. when I followed the link, there's a dataset, but is it even coded for race?
Looks like it. I'm linking for Malloy's analysis more than the cited section either way.

At 5/8/2016 5:21:35 AM, Hoppi wrote:
If it's a meta-analysis, then it's not a longitudinal study. If you don't know what the evidence is, then it's not worth discussing here.
It's a meta-analysis of longitudinal studies on the exact same subject. The evidence is right there.

At 5/8/2016 5:27:28 AM, Hoppi wrote:
If you're going to reference papers, please link to them.

Sure. Here's te Nijenhius 2012: https://lesacreduprintemps19.files.wordpress.com...

And here's Strenzen 2007: http://www.emilkirkegaard.dk...

At 5/8/2016 5:29:19 AM, Hoppi wrote:
At 5/8/2016 2:51:26 AM, someloser wrote:
Additionally, if you train to beat an IQ test, then the extra points wouldn't be g loaded. So you wouldn't see any of the real-world gains associated with a higher IQ that would be expected from a "real" IQ gain.

I disagree. IQ is a measure of potential for learning in a traditional classroom environment. Obviously, it's going to be affected by literacy and familiarity with test-taking. That's part of the construct.
IQ, the metric, will. Cognitive g (the thing that actually matters) won't. IQ != g.

Sure, if you familiarize kids with taking tests, they'll do better on tests. Your cognitive capacity - what IQ is supposed to measure and gives it good predictive value - won't actually improve.
Ego sum qui sum. Deus lo vult.

"America is ungovernable; those who served the revolution have plowed the sea." - Simon Bolivar

"A healthy nation is as unconscious of its nationality as a healthy man of his bones. But if you break a nation's nationality it will think of nothing else but getting it set again." - George Bernard Shaw
Hoppi
Posts: 1,655
Add as Friend
Challenge to a Debate
Send a Message
5/8/2016 8:23:10 AM
Posted: 7 months ago
At 5/8/2016 5:24:53 AM, someloser wrote:
At 5/8/2016 5:11:27 AM, Hoppi wrote:
If IQ tests are sensitive to training and practice effects - and they are - then one obvious explanation for any differences between groups is a difference in practice and training experience. The burden is on R&J to rule this out. Maybe they will later on in the paper.
It's obvious, if you assume that one group is collectively being trained to beat IQ tests throughout the entire socioeconomic spectrum.

That being said, there is evidence that IQ "practice" (like the type you provided) has some effect on the gaps. Note, for example, that the size of the gaps increases the more g-loaded a test is (te Nijenhuis, 2012). This would imply that the IQ training is artificially shrinking the gaps.

What gaps? The paper you linked doesn't compare groups.
someloser
Posts: 1,377
Add as Friend
Challenge to a Debate
Send a Message
5/8/2016 8:35:36 AM
Posted: 7 months ago
At 5/8/2016 8:23:10 AM, Hoppi wrote:
At 5/8/2016 5:24:53 AM, someloser wrote:
At 5/8/2016 5:11:27 AM, Hoppi wrote:
If IQ tests are sensitive to training and practice effects - and they are - then one obvious explanation for any differences between groups is a difference in practice and training experience. The burden is on R&J to rule this out. Maybe they will later on in the paper.
It's obvious, if you assume that one group is collectively being trained to beat IQ tests throughout the entire socioeconomic spectrum.

That being said, there is evidence that IQ "practice" (like the type you provided) has some effect on the gaps. Note, for example, that the size of the gaps increases the more g-loaded a test is (te Nijenhuis, 2012). This would imply that the IQ training is artificially shrinking the gaps.

What gaps? The paper you linked doesn't compare groups.
te Nijenhuis? Yes it did... it's even in title ("The Flynn effect, group differences, and g loadings") and the abstract - not to mention the paper itself.

From the conclusion:

The research literature overwhelmingly showed that group differences and g loadings are strongly correlated
Ego sum qui sum. Deus lo vult.

"America is ungovernable; those who served the revolution have plowed the sea." - Simon Bolivar

"A healthy nation is as unconscious of its nationality as a healthy man of his bones. But if you break a nation's nationality it will think of nothing else but getting it set again." - George Bernard Shaw
Hoppi
Posts: 1,655
Add as Friend
Challenge to a Debate
Send a Message
5/8/2016 8:35:47 AM
Posted: 7 months ago
At 5/8/2016 5:46:24 AM, someloser wrote:
At 5/8/2016 5:20:20 AM, Hoppi wrote:
At 5/8/2016 5:09:23 AM, someloser wrote:
At 5/8/2016 5:04:10 AM, Hoppi wrote:
That link uses the R&J paper as a reference!
Yeah... and it's far from their only reference -- the post's analysis (i.e. the reason I linked it in the first place) controlled for SES.

The only study in there that seems to be controlled for SES is the headstart one, but I can't see what analysis he used. when I followed the link, there's a dataset, but is it even coded for race?
Looks like it. I'm linking for Malloy's analysis more than the cited section either way.

He's using data that doesn't take SES into account when matching samples.

At 5/8/2016 5:21:35 AM, Hoppi wrote:
If it's a meta-analysis, then it's not a longitudinal study. If you don't know what the evidence is, then it's not worth discussing here.
It's a meta-analysis of longitudinal studies on the exact same subject. The evidence is right there.

At 5/8/2016 5:27:28 AM, Hoppi wrote:
If you're going to reference papers, please link to them.

Sure. Here's te Nijenhius 2012: https://lesacreduprintemps19.files.wordpress.com...

And here's Strenzen 2007: http://www.emilkirkegaard.dk...

At 5/8/2016 5:29:19 AM, Hoppi wrote:
At 5/8/2016 2:51:26 AM, someloser wrote:
Additionally, if you train to beat an IQ test, then the extra points wouldn't be g loaded. So you wouldn't see any of the real-world gains associated with a higher IQ that would be expected from a "real" IQ gain.

I disagree. IQ is a measure of potential for learning in a traditional classroom environment. Obviously, it's going to be affected by literacy and familiarity with test-taking. That's part of the construct.
IQ, the metric, will. Cognitive g (the thing that actually matters) won't. IQ != g.

Sure, if you familiarize kids with taking tests, they'll do better on tests. Your cognitive capacity - what IQ is supposed to measure and gives it good predictive value - won't actually improve.

That idea about cognitive capacity is a belief not based on evidence. An IQ test is a set of tasks. Like any other task, an individual's performance will be affected by motivational factors, environmental factors, mood, previous experience, etc.

You can make up stories about the meaning of the test, but why should anyone believe those stories?