One of my favorite twitter accounts deleted (On peer review and social “science”)

Recently, one of my favorite twitter accounts was deleted. The name of the account was @realpeerreview and the main focus of this account was to take excerpts from truly retarded social “science” research papers and show how stupidly our tax money is being wasted. You can see an archive of his work here and an archive of the archive in case that goes down. I would read through all those. It is both hilarious and infuriating at the same time. Why does the public have to pay for such stupidity?

Of course, other social “scientists” didn’t like public attention to their “research” and threatened to doxx the individual broadcasting how atrocious most social “science” research actually is. These cockroaches prefer to stay in dark under the rug and don’t like anyone lifting up a corner. @realpeerreview was, I guess, left no choice but to back off because his career was possibly on the line. Some one else quickly snatched up the account name after finding out about this typically leftist hiding of the truth to continue on the good work. I don’t know if she will do as good a job, but I hope she can give them hell.

Anyway, just one more example why academia, at least in the humanities, should not be trusted at all. I have written on stuff like this before on how stereotype threat is bunk and how standardized tests are rigged against boys, however, it was nice to see how widespread the sickness is. It spans hundreds of papers and hundreds of topics. Defunding the humanities would wipe out 100 times more cancerous tumor than healthy tissue. It is time for some emergency surgery.

Share Button

Wikipedia in Action on Race

I like to refer to Lewontin’s fallacy frequently when debating people who deny the biological basis of race. Wikipedia, while clearly not perfect, did have a reasonable article (at least for quick referral of lay-people) on the paper written by W.F. Edwards which coined “Lewontin’s fallacy.”(1) A brief overview is that in the 1970’s an academic social justice advocate published a paper(2) in which he claimed that there is more variation within individuals from one race than there is variation between different racial populations. So much that you can regularly find people of different races who are more similar to each other than they are to members of their own race. However, the first paper linked to above shows that the problem mainly stems from the fact that very few loci were studied by Lewontin. Allele frequencies differ between populations and with enough loci studied, the ability to distinguish between racial groups based purely on genetic information is quite high. Virtually 100%.

As is typical for pretty much all articles on Wikipedia, anything that isn’t politically correct can be expected to drift over time such that claims that are not PC are deleted, diluted, and placed next to a larger number of criticisms than is warranted such that it implies that the non-PC claims seem unsupported or only supported by very few outliers. Sometimes, like in this article, a paper which can be seen to support one conclusion actually supports the opposite on more careful inspection. All of this is the wikipedia version of death by 1000 cuts. I once tried editing the page on gender differences in intelligence and was basically run out and banned by marxist feminists. I assume this happens to anyone who objectively tries to include factual and balanced information into potentially politically incorrect articles. These same people got that article deleted or subsumed into gender differences in psychology for awhile, but it looks like it has been resurrected now. Honestly, the constant battle over these sorts of articles is just beyond all reason and I will never bother editing wikipedia again. Chances are your work is just going to get deleted and there are other platforms where that won’t happen.

Subjectively, it seems like this sort of thing has been happening to the Lewontin’s fallacy article, but I will let you be the judge:

Here is an old archived version of this article.

Here is an archived version of the current article.

Here is a direct link to the article. (It shouldn’t look different than the above link at the time of this post, but who knows what future changes will be made. In a year or two it could be interesting to compare these three versions)

The thing that is most obvious in my mind is that a paper discussed in an earlier version of the article which supported the concept of Lewontin’s fallacy has had any reference to it completely deleted. Here is the now deleted content:

Studies of human genetic clustering have shown that people can be accurately classified into racial groups using correlations between alleles from multiple loci. For instance, a 2001 paper by Wilson et al. reported that an analysis of 39 microsatellite loci divided their sample of 354 individuals into four natural clusters, which broadly correspond to four geographical areas (Western Eurasia, Sub-Saharan Africa, China, and New Guinea)

In addition, a paper which purports to undermine the concept that Lewontin’s thinking is fallacious is present at the end in both versions, but is quoted more (and very selectively) in the most recent version. In my opinion, the findings in both wikipedia versions are misrepresented.

In the old article this:

The paper claims that this masks a great deal of genetic similarity between individuals belonging to different clusters. Or in other words, two individuals from different clusters can be more similar to each other than to a member of their own cluster, while still both being more similar to the typical genotype of their own cluster than to the typical genotype of a different cluster. When differences between individual pairs of people are tested, Witherspoon et al. found that the answer to the question “How often is a pair of individuals from one population genetically more dissimilar than two individuals chosen from two different populations?” is not adequately addressed by multi locus clustering analyses. They found that even for just three population groups separated by large geographic ranges (European, African and East Asian) the inclusion of many thousands of loci is required before the answer can become “never”

On the other hand, the accurate classification of the global population must include more closely related and admixed populations, which will increase this above zero, so they state “In a similar vein, Romualdi et al. (2002) and Serre and Paabo (2004) have suggested that highly accurate classification of individuals from continuously sampled (and therefore closely related) populations may be impossible”. Witherspoon et al. conclude “The fact that, given enough genetic data, individuals can be correctly assigned to their populations of origin is compatible with the observation that most human genetic variation is found within populations, not between them. It is also compatible with our finding that, even when the most distinct populations are considered and hundreds of loci are used, individuals are frequently more similar to members of other populations than to members of their own population”

expanded into this:

In the 2007 paper “Genetic Similarities Within and Between Human Populations”,[20] Witherspoon et al. attempt to answer the question, “How often is a pair of individuals from one population genetically more dissimilar than two individuals chosen from two different populations?”. The answer depends on the number of polymorphisms used to define that dissimilarity, and the populations being compared. When they analysed three geographically distinct populations (European, African and East Asian) and measured genetic similarity over many thousands of loci, the answer to their question was “never”. However, measuring similarity using smaller numbers of loci yielded substantial overlap between these populations. Rates of between-population similarity also increased when geographically intermediate and admixed populations were included in the analysis

Witherspoon et al. conclude that, “Since an individual’s geographic ancestry can often be inferred from his or her genetic makeup, knowledge of one’s population of origin should allow some inferences about individual genotypes. To the extent that phenotypically important genetic variation resembles the variation studied here, we may extrapolate from genotypic to phenotypic patterns. […] However, the typical frequencies of alleles responsible for common complex diseases remain unknown. The fact that, given enough genetic data, individuals can be correctly assigned to their populations of origin is compatible with the observation that most human genetic variation is found within populations, not between them. It is also compatible with our finding that, even when the most distinct populations are considered and hundreds of loci are used, individuals are frequently more similar to members of other populations than to members of their own population. Thus, caution should be used when using geographic or genetic ancestry to make inferences about individual phenotypes”,[20] and warn that, “A final complication arises when racial classifications are used as proxies for geographic ancestry. Although many concepts of race are correlated with geographic ancestry, the two are not interchangeable, and relying on racial classifications will reduce predictive power still further.”

This paper… It had decent data and methodology actually. But as is almost always the case with these sorts of things, interpretations and framing of the results are key. It is clear that the people who wrote this are deliberately softballing their wording either to cover their ass (my guess) or to promote a more progressive narrative.

ω in the following quotes is defined as given a certain number of loci considered, the probability of individuals originating from two distinct geographical areas will be more similar to each other than to someone originating closer to them. I.E., the probability that two randomly selected individuals from different races will be more similar to each other than each is similar to a randomly selected member of their own race. Keep in mind that ω is not the same as determining what race a person is based on genetic data. Even with small numbers of loci and a high ω, there is very low probability of misclassifying the race of an individual person. From the very same paper used to undermine the Edwards’ paper:

[A relatively large ω is found with low numbers of loci] It breaks down, however, with data sets comprising thousands of loci genotyped in geographically distinct populations: In such cases, ω becomes zero.

With the large and diverse data sets now available, we have been able to evaluate these contrasts quantitatively. Even the pairwise relatedness measure, ω, can show clear distinctions between populations if enough polymorphic loci are used. Observations of high ω and low classification errors are the norm with intermediate numbers of loci (up to several hundred)

Thus the answer to the question “How often is a pair of individuals from one population genetically more dissimilar than two individuals chosen from two different populations?” depends on the number of polymorphisms used to define that dissimilarity and the populations being compared. The answer, ω, can be read from Figure 2. Given 10 loci, three distinct populations, and the full spectrum of polymorphisms (Figure 2E), the answer is ω ≅ 0.3, or nearly one-third of the time. With 100 loci, the answer is ∼20% of the time and even using 1000 loci, ω ≅ 10%. However, if genetic similarity is measured over many thousands of loci, the answer becomes “never” when individuals are sampled from geographically separated populations.

Molecular biologists and geneticists use a little bit different definition of polymorphism than some other branches in biology. In this case, they are referring to single nucleotide differences in the genome. This is equivalent to having one letter different in spelling a word. Prog and prig mean almost the same thing, but there is one letter difference which slightly changes the meaning. This is a reasonable analogy to the differences in the genetic code.

What this paper says (and it should be said with less tip-toeing) is that if you only consider a small number of these single nucleotide polymorphisms, there is a high degree of error and you can often erroneously conclude that two people from different races are more similar to each other than they are to individuals of their own race. The key word here is erroneously. This is a statistical problem, not biological fact. If you consider thousands of SNPS at once, then you have virtually no chance of encountering this problem. The authors of this paper found that Edwards was right and Lewontin was wrong. Individuals from two different races are never more similarly related than people from the same race, and the genetics supports this when you consider enough loci. It is pretty unambiguous. The quotes in the Wikipedia article and in the paper don’t really represent what the researchers actually found. The researchers had to dress this language up the way they did because of progressive influence in academia. Chances are they wouldn’t have gotten published if they were straight forward about what they found, and even if they could have published political heresy they may have had their careers ruined by SJWs in academia. See what happens when you don’t toe the line with the progressive narrative by reading what happened to a University of Texas researcher who didn’t find the “right” conclusions with regards to gay couples raising children. Though there is a huge problem with how Wikipedia articles are written and “maintained,” they wouldn’t have been able to misconstrue these results so badly if it weren’t from the same sorts of SJWs in academia malevolently influencing researchers. Though it shouldn’t be understated that the wikipedia editors did in fact selectively quote from this already bludgeoned paper. Two layers of SJW influence changed the findings of this paper to mean the exact opposite of what it actually found. Unbelievable. It is truly amazing that this sort of shenanigans is allowed to go on.

You might object that “thousands” is a huge number and that this demonstration of statistical problems convincingly shows that races don’t differ if it takes that many to reduce error to zero. However, the human genome is about 3 billion base pairs long. If you were to use 3000 base pair SNPs, which is consistent with the minimum in the paper, then you need to utilize only .0001% of the whole genome to reduce this error to zero. Or, if you want to consider SNPs only, there are about 10 million SNPs in the human genome. A sample of 3000 SNPs is only .003% of the total number of SNPs that could be used. This is a conservative estimate because their figure 2 indicates it only takes about 1000 SNPS to minimize this error. In other words, it only takes a vanishingly small fraction of the genome to relieve you of this statistical error that can find that humans from two different races are more similar to each other than either is to their own race.

Yet this paper, which so conclusively shows that human races are different from each other on the genetic level, is used to debunk the original Edwards’ paper. The author’s of the paper attempt to debunk themselves or at least pretend like they found the opposite of what they actually did. This paper is absolutely one of the worst instances of doublethink I have ever come across. It literally blows my mind. As a society, we seem to have a real hatred for truth when it comes to biological realities and the uninformed are clearly being purposefully told lies.

Sidenote: I know there was another article on cathedral entryism on Wikipedia in the alt-right in the last year or so, but for the life of me I can’t find it. If anyone can provide a link I would appreciate it. Edit: Found it.

(1) Bioessays. 2003 Aug;25(8):798-801. Human genetic diversity: Lewontin’s fallacy. Edwards

(2) The Apportionment of Human Diversity. R. C. Lewontin. 1972

(3) Genetics. 2007 May; 176(1): 351–359. doi:  10.1534/genetics.106.067355 PMCID: PMC1893020 Genetic Similarities Within and Between Human Populations J. Witherspoon, S. Wooding, A. R. Rogers, E. E. Marchani, W. S. Watkins, M. A. Batzer, and L. B. Jorde

Share Button

Stereotype threat and pseudo-scientists

The politically acceptable explanation for gender differences in intelligence studies and tests is that discrimination accounts for all current disparities between men and women in intellectual Fields, starting first and foremost with the standardized tests themselves. The question is: does the data support this? The most fashionable (possibly faddish) explanation for test differences between gender (and race) come from social psychology and is termed ‘stereotype threat’. Stereotype threat is a type of implicit discrimination that is supposed to result from traditional stereotypes based on the gender of test takers. It is supposedly all pervasive throughout society, and constantly present everywhere you are. This obscure and unfalsifiable ether is supposed to depress the scores of females in tests. Society has historically held that women were not as intelligent as men. Stereotype threat proponents argue that knowledge of this history combined with the manner in which questions are designed and phrased leads to lower test scores for women in a sort of self fulfilling prophecy.

It is important to consider who should be affected by stereotype threat and other types of discrimination and what kind of pattern in the distribution of scores should result if it was having a widespread impact. Discrimination of this type is assumed to be universal and omnipresent. It is supposed to be present in both the society at large as well as in the test that all students are taking. If it is universally present then it should have a universal effect. It should affect women equally at all levels of test taking proficiency and should result in a uniform downward shift of the score distribution compared to men. In the graph below, hypothetical male and female test score distributions are super-imposed and the expected influence of stereotype threat is shown. The male distribution does not have as high a peak and extends out farther in either direction to reflect the greater male variability in scores found in virtually all IQ tests. I am not the best artist so you will have to forgive me if it isn’t as pretty as it could be.

Hypothetical male and female IQ distributions before and after stereotype threat

Stereotype threat voodoo action from the ether

If it does not fit this pattern, explanations for why women at different test taking levels react differently to stereotype threat must be invented. The more disagreements from this trend, the more explanations that must contrived, and the more parsimony is lost (which generally means a theory is weaker). The difference between average IQ at least suggests that this might be happening, but the 3-5 iq points generally reported is a relatively small difference and mainly suggests that whatever universal influences do exist, their importance must be relatively minor. (However, there is reason to suspect that male IQ advantage is severely underestimated)

The discrimination theory is not the only possible explanation for this overall shift. The difference could just as easily, in fact probably much more easily, be explained by biological differences in brain development. Especially the fact that males in general grow to be larger, which translates into larger brain sizes on average, which for reasons that should be obvious correlates with higher IQ.

The discrimination and biological differences above have one thing in common: they are universal and thus are not good explanations for greater male variance which is the root source of most male/female disparity in the highest levels of achievement. A consistent universal factor should have a consistent universal effect for all levels of ability as shown in the figure above, and the only consistent universal difference in mean IQ scores between gender are small. Assuming stereotype threat is real, which is doubtful, it is not impossible that of the small difference that does exist, stereotype threat only makes a small contribution in addition to other factors like biological development. In such a case, the individual contribution of stereotype threat would be vanishingly small and would approximate complete irrelevance.

In the case of gender, stereotype threat is pretty much ruled out for the above reason. However, racial gaps do take a form that would be consistent with the idea of Stereotype threat. However, there are other reasons why it is also doubtful in the case of race. For more exploration on why conclusions drawn from stereotype threat studies are doubtful for methodological reasons, I recommend this paper by law professor Amy Wax: Stereotype Threat: A case of overclaim syndrome? (I have a special love for this paper because insisting on using it in the sex and intelligence wikipedia page years ago is what brought down a flock of feminist harpies who eventually got not only the paper, but also my user account banned from wikipedia.) Seems like Wax really likes sticking her neck out and fighting the good fight.

At best, stereotype threat is something that exists and has only a very small effect and at worst it is an example of publication bias amongst journals where positive results that support politically progressive ideas (like discrimination against women) are overwhelmingly published relative to studies that don’t confirm progressive beliefs or which might positively refute progressive beliefs.

Diederick Stapel was previously a highly regarded and influential Dutch social psychologist who did a lot of work on stereotype threat, among other things, until it came to light that he “routinely falsified data and made up entire experiments.” Another example of his politically biased work was a “scientific” article which sanctimoniously claimed to find that meat eaters were more selfish and less agreeable than vegans. Unfortunately, it is impossible to be surprised by outspoken priggishness from vegans. Thanks to this media attention, Stapel is now the most notorious charlatan in the field of social psychology, which is saying a lot for what appears to be a regularly fraudulent and pseudo-scientific discipline. Social Psychologists as a group do not make the data they collect available for outside review 2/3rds of the time. This stinginess with data is actually against the ethical rules established by Social psychologists themselves and suggests that there are likely many more Stapels out their who simply haven’t been caught. A survey by the Harvard business school found that 70% of social psychologists admitted to cutting corners in reporting data, 30% reporting unexpected findings as if they were expected from the start, and 1% admitted to falsifying data. Another meta-analysis of papers published in high-tier psychology journals found that 50% of papers surveyed contained at least one statistical error and 15% contained an error so severe that the conclusion drawn would have had to have been reversed. Yet another meta-analysis which looked at whether or not positive results from stereotype threat studies could be replicated found that almost half could not, and that a further 25% were confounded by methodological issues. A substantial majority of the findings were unreliable.1,2,3

Bias is rampant in the humanities, but especially in social psychology, both among individual researchers and among the journals publishing papers. Beyond the objective critical evaluation of papers, the field itself is essentially an ideological and political echo-chamber that is considerably more left-wing politically than the general population. 80% of social psychologists identify as liberal, while only 3 out of 1000 identify as conservative. Contrast this with the general population which is 40% conservative and only 20% liberal. Were these sorts of numbers occurring with a protected class, these same people wouldn’t hesitate to use it as incontrovertible proof of discrimination. Considering what is now known about the biological origins of cognition and intelligence, it is generally difficult to take claims of discrimination seriously when groups also display a relatively lower intelligence profile. However, in this case there is no reason to think that conservatives as a group have an intellectual profile below the general population. Social conservatives tend to be a little lower in intelligence relative to liberals, but free-market conservatives (libertarians) tend to be smarter. Being very partisan, either liberal or conservative, tended to be associated with high IQ as well. Increased income levels, which are a proxy for IQ, also moves people right ideologically. In other words, there is nothing that differentials in biologically determined intelligence can do to explain the lack of conservatives, and even moderates, in the humanities.4,5,6,7 Presumably academia wasn’t always so partisan, and thus its current state is a classic case of successful entryism.

In a survey of social psychologists, it was found that conservative responders feared negative consequences from revealing their political affiliation and that they were right to do so as liberal responders expressed willingness to discriminate against conservatives in approving papers, grant proposals, and hiring decisions. The more liberal a social psychologist is or the more consequential the decision would be for the conservative, the more willing liberal social psychologists are to discriminate. That willingness to discriminate against (or for) articles and proposals for ideological reasons has been empirically confirmed in several instances. In one study, reviewers were sent a manuscript which purported to show the mental health of a group of leftist political activists compared to a control group. Reviewers who were sent a version which showed that the activists had better mental health consistently felt that the paper was more publishable and even felt that the statistics were more adequate than reviewers sent a version that showed the activists had lower mental health. In another case, a research proposal which either wanted to study discrimination or reverse-discrimination was sent to 150 review boards. The proposal on discrimination was approved twice as often as the proposal on reverse-discrimination. In college admissions, it was found that reviewers would attach greater value to the criteria (grades vs. test scores) which would allow them to pick the candidates with similar partisan politics. Lastly, controlling for research productivity and academic achievement, another study found that conservative researchers were working at lower quality institutions relative to equivalent liberal colleagues than would be expected. The irony that a group which commonly publishes on the asserted negative consequences of discrimination would prove to itself be extraordinarily discriminatory is stunning.8,9,10,11

The pattern of ideologically driven academics significantly undermines the ability of an objective outsider to trust the conclusions coming out of certain fields, especially when it is related to such a politically charged subject as gender (and race) differences in test scores. It is quite clear that the overwhelming majority of researchers working on this topic possess a politically desired outcome of these studies. The great potential for this systemic Lysenkoism to motivate the production of inaccurate results which are contrary to reality can’t be overestimated. The objectivity of the field concluding stereotype threat is a real and large effect phenomenon is highly questionable. Calling this cynical skepticism “anti-intellectual,” a common criticism of conservative thinkers, is only so in the sense that these “scientists” have mis-defined the word “intellectual” to describe their political ideology and therefore themselves. Like most things on the right the “anti-science” feeling exposed by some is just a reaction to leftist entryism in academia and the dominance of pseudo-scientific articles surrounding politically partisan topics.

  1. Can stereotype threat explain the gender gap in mathematics performance and achievement? Stoet, Gijsbert; Geary, David C. Review of General Psychology, Vol 16(1), Mar 2012, 93-102. http://dx.doi.org/10.1037/a0026617
  2. The (mis)reporting of statistical results in psychology journals. Marjan Bakker and Jelte Wicherts. 2011. Behavior Research methods.
  3. Psychology rife with inaccurate research findings. Psychology today. 2011
  4. Jonathon Haidt’s post-partisan psychology page.
  5. Social scientist sees bias within. New York Times. 2011
  6. Is there a relationship between political orientation and cognitive ability? A test of three hypotheses in two studies. Markus Kemmelmeier. 2008
  7. Income and Ideology: How personality traits, cognitive abilities, and education shape political attitudes. Rebecca Morton. Jean-Robet Tyran. Erik Wengstrom.
  8. Political Diversity in social and personality psychology. Yoel Inbar, Joris Lammers. 2012
  9. Publish or politic: Referee bias in Manuscript review. Stephen Abramowitz, Beverly Gomes, Christine Abramowitz. 1975
  10. Human subjects review, personal values, and the regulation of social science research. Ceci, Peters, Plotkin 1985.
  11. Political Partisan Prejeduce: Selective Distortion and weighting of Evaluative categories in college admission applications. Munro, Lasane, Leary.
Share Button