Talk:Variability hypothesis/Archive 1

Current text
The current wikipedia article states "In contrast, one meta-analysis found that males were slightly better at progressive matrices, while women actually had more variability" when discussing differences in male and female variability. In the referenced article, I can't find any mention of findings specific to differences in variability. Is anybody else able to see where this is discussed in the article? — Preceding unsigned comment added by 209.205.253.87 (talk) 16:24, 6 October 2015 (UTC)

No, I also can't locate the cited findings. --24.214.215.253 (talk) 06:50, 25 October 2018 (UTC)

Consistently biased - should be rewritten
There is a lot of current, peer-reviewed research showing that males do in fact show higher variability than females across many domains, not just in cognitive abilities. This is also true across species, and is consistent with predictions from evolutionary theory. The current text misquotes several sources that just reports on the differences. Instead it chooses to cite sources that have more of an ideological motivation. Koyos (talk) 16:37, 2 April 2017 (UTC)


 * Peer-reviewed research is not magically methodically perfect and unbiased. See Delusions of Gender: "However, later work by Fine published in the journal Neuroethics identified systematic issues in the way human neuroimaging investigations of sex differences tend to be investigated, contra the notion of a few 'bad apples'. [...] Many of the criticisms of such work made in Delusions of Gender were noted in the article "Perils and pitfalls of reporting sex differences" by Donna L. Maney, as part of a Philosophical Transactions of the Royal Society Theme Issue "Multifaceted origins of sex differences in the brain', compiled and edited by McCarthy in 2016."
 * In fact, in her book Delusions of Gender, Cordelia Fine has specifically attacked the variability hypothesis. This deserves mention in the article. --Florian Blaschke (talk) 23:24, 14 August 2018 (UTC)


 * Like most writers who object to the GMVH, Cordelia Fine isn't a scientist. Maybe criticism should be grouped in two sections, from within and without the relevant fields.Punslinger (talk) 08:01, 19 September 2018 (UTC)


 * "Peer-reviewed research is not magically methodically perfect and unbiased" No one said that it is. But, it's better to use a source that has been evaluated by experts in a field than it is to use an opinion piece that has not been evaluated.

This entire article reeks of ideological bias and comes off almost as an opinion piece written by someone that doesn't want to believe the hypothesis. For example, the largest section details criticisms, some of which seem irrelevant: " In an attempt to examine the validity of the variability hypothesis, while avoiding intervening social and cultural factors, Hollingworth gathered data on birth weight and length of 1,000 male and 1,000 female neonates. This research found virtually no difference in the variability of male and female infants.." This has nothing to do with some of the main conclusions of the original findings, and this comes across like a strawman argument and does nothing to address cognitive variability from other studies. So if there are no differences on one dimension, that negates differences on a completely unrelated dimension? "variability is not significant in and of itself, but rather depends on what the variability consists of, and" Whoever added this didn't think about the implications, and this is another strawman. "Significant" is subjective, unless you are referring to statistical significance (which can be influenced by variability). Variability refers to differences, or the spread, within or between groups. Wikipedia has become an unreliable (if it ever was reliable) source of information because of seemingly ideological and/or political differences currently inflaming certain sectors of western culture. — Preceding unsigned comment added by 2600:1702:1B0:AE90:BD0C:65F6:8452:E2C7 (talk) 21:27, 1 October 2018 (UTC)

I would also agree that this article consistently shows bias and cites misleading interpretations of the hypothesis at hand. The variability in question most often refers to a cognitive variability, so I'm not sure why exactly sources about physical variability are brought into focus and written out as if they speak to the verity of cognitive variability. Perhaps in rewriting, the physical, psychological, and cognitive variability should all be written out in their own sections. On that note, no research into psychological variability is even present, despite being referenced in the introduction. The inclusion of the line "...in which no objective evidence was found to support the idea of innate female inferiority." seems pretty sketchy as well considering no source cited suggests this. Male variability =/= male superiority, and to have an article written in a way that implies such undermines the legitimacy of the platform itself. --24.214.215.253 (talk) 06:50, 25 October 2018 (UTC)


 * I won't weigh in on whether or not the article is biased. However, I found one instance where a source made the opposite claim than did the text that referenced it. I corrected that in an edit where I included a block quote in order to avoid confusion about the meaning (and also because the article behind a pay-wall). Bias or none, sources should support the claims made in the article. Three other sources were referenced in the same sentence, so I did my best to keep the previous editor's meaning in regards to the other sources. Here is the entire conclusion without ellipses:


 * Our analysis of international test score data shows a higher variance in boys' than girls' results on mathematics and reading tests in most OECD countries. How this translates into educational achievement is a matter open for discussion. Higher variability among boys is a salient feature of reading and mathematics test performance across the world. In almost all comparisons, the age 15 boy-girl variance difference in test scores is present. This difference in variance is higher in countries that have higher levels of test score performance.


 * Sex differences in means are easier to characterize: It is evident from the PISA data that boys do better in mathematics, and girls do better in reading. This has a compositional effect on the variance differences as well. The higher boy-girl variance ratio in mathematics comes about because of an increased prevalence of boys in the upper part of the distribution, but the higher variance in reading is due to a greater preponderance of boys in the bottom part of the test score distribution. Because literacy and numeracy skills have been shown to be important determinants of later success in life (for instance, in terms of earning higher wages or getting better jobs), these differing variances have important economic and social implications (22).


 * We therefore confirm that 15-year-old boys do show more variability than girls in educational performance, with specifics that differ according to whether mathematics or reading are being studied and tested. These results imply that gender differences in the variance of test scores are an international phenomenon and that they emerge in different institutional settings.


 * Also, one of those sources is dated to 1994, so whether it constitutes current research might be up for discussion.KristinaLu (talk) 01:02, 28 April 2020 (UTC)
 * KristinaLu, like I stated with this edit where I reverted you (followup comment here), your "some researchers" wording was WP:Synthesis. Also see WP:Weasel words. If the "In general" wording is not supported by any of the sources, that should be changed as well. We should stick to what the sources state. We should also summarize. Including the blockquote gives WP:Undue weight to that source. And like I asked in my edit summary, why are you focusing on a 2008 source, given the research since? Also, per WP:SCHOLARSHIP, secondary sources are preferred to primary sources. Flyer22 Frozen (talk) 18:51, 28 April 2020 (UTC)
 * "Some researchers" is something the sources back, "most researchers" can't be backed by references to researchers that do. --mfb (talk) 01:15, 29 April 2020 (UTC)
 * What sources explicitly state "some researchers"? That's my point with regard to WP:Synthesis. And like I stated, if the "In general"/"most researchers" wording is not supported by any of the sources, that should be changed as well. We can find some other way to word the matter. For example, using "Multiple studies indicate [...]" or "While multiple studies indicate [...]." Even "some research" is more appropriate than "some researchers" if "some research" aligns with WP:Due. This is because "some research" is less likely to be tagged with Template:Which than "some researchers" or "most researchers" is to be tagged with Template:Who. But even for "Multiple studies indicate [...]" or "While multiple studies indicate [...]", it's best if we are reporting on what a WP:Secondary or tertiary source is stating. Google Books has secondary sources on the topic of the variability hypothesis. And looking on Google Books, I see sources such as this 2015 "History of Psychology: Ideas and Context" source, from Routledge, page 308, commenting, "Briefly stated, the variability hypothesis held that men are physically and psychologically more variable than women. According to the hypothesis, women tended toward averages in all things, whereas men showed greater variety; men were were thought to be more courageous but more cowardly, more virtuous but more corrupt, more intelligent but more stupid, and more violent but more peaceful. Superficial support for the variability hypothesis was found in the achievements of men compared to women and in the finding that more boys compared to girls were diagnosed as intellectually deficient." The source goes on to state more about the hypothesis. Notice that it states "held", as though the hypothesis is simply a past matter or outdated? It also refers to the hypothesis as "so-called." Flyer22 Frozen (talk) 17:39, 29 April 2020 (UTC) Updated post. Flyer22 Frozen (talk) 17:49, 29 April 2020 (UTC)
 * As I said my goal was to correct what was the most glaring problem with the passage: the source strongly contradicted the claim made in the article. (Note: My edit focused on this particular source in order to correct the error. This was communicated in both my edit-summary and talk-page contribution). I felt (and still feel) that my edit was a significant (and I felt obvious) improvement upon what came before. "In general most researchers have failed" may have been a poorly executed summary, a conflict that arose between edits by different editors or something else; in any case it was (and now is) a dramatic misrepresentation of at least one of the sources and needs to be corrected.
 * On the other hand (as was pointed out by mfb) "some researchers" is clearly not WP:Synthesis; if clarity is needed please see the guidelines for the resolution of summation conflicts. It is clearly stated that editors may/should (ie that no secondary source is needed to do this) summarize several sources or point out that sources conflict with each other. Perhaps, in order to better compare/constrast of the conflicts in the sources, someone can expand "some researchers" by using WP:INTEXT as I did for the article in Science.
 * As for WP:PSTS, I'm glad this was brought up. The utility of tertiary sources will be quite limited however with regards to current research trends (although they would be great for the History section etc). This point is illustrated by the above quote mentioned by Flyer22 Frozen, where "Superficial support for the variability hypothesis was found in the achievements of men compared to women and in the finding that more boys compared to girls were diagnosed as intellectually deficient" is refering to the research of "the late nineteenth and early twentieth centuries." (Attention should be paid to context when interpreting the context of such claims!) Secondary sources (eg reviews published in academic journals especially when authored by a recognized authority in the field) are more appropriate for documenting recent research trends and any consensus that has been reached. These sources should be used in the Modern studies section rather than less specialized sources like encyclopedias.KristinaLu (talk) 22:18, 4 May 2020 (UTC)
 * These are not original research is not a Wikipedia policy or guideline. WP:Weasel wording is a guideline. With this edit, you were still giving WP:Undue weight to "a 2008 analysis of test scores" by putting it up against those sources. Again, why are you focusing on a 2008 source as a contrast point? It should instead be placed in chronological order in that section, and that is how I presented it with this edit. Your "although" wording that puts the 2008 source up against other research clearly falls under WP:EDITORIALIZING, which states, "More subtly, editorializing can produce implications that are not supported by the sources. Words used to link two statements such as 'but', 'despite', 'however', and 'although' may imply a relationship where none exists, possibly unduly calling the validity of the first statement into question while giving undue weight to the credibility of the second." When summarizing what sources state, editors should still not tie sources together in that way and they should adhere to WP:STICKTOTHESOURCE. I removed "some" per what I've stated. Unless you can show that it's explicitly supported by one of the sources, it's best not to add it back. And per WP:EDITORIALIZING, I removed your "although" wording. As for WP:PSTS, our personal opinions on it do not trump what it states; it's clear that secondary and tertiary sources are preferred to primary sources. As for "attention should be paid to context when interpreting the context of such claims", I paid attention to the source, but I perhaps should have given more context to what I was citing (even though editors can clearly click on the source and read for themselves, like you did). But then again, it is not like the source is only looking at Hollingworth from a "late nineteenth and early twentieth century" perspective. For instance, it states, ""An example of research that attacked and helped change public and scientific attitudes is visible in the work of Leta Stetter Hollingworth." It then goes on to explain, which is why it delves into the historical matter. You stated, "Secondary sources (eg reviews published in academic journals especially when authored by a recognized authority in the field) are more appropriate for documenting recent research trends and any consensus that has been reached." Whether it's better to use a secondary source over a tertiary source is a WP:CONTEXTMATTERS thing. You stated, "These sources should be used in the Modern studies section rather than less specialized sources like encyclopedias." That's an opinion. And, again, WP:CONTEXTMATTERS.


 * For future reference, I prefer not to be WP:Pinged to talk pages I'm watching. Since I'm watching this one, all editors need to do is simply comment; I'll see it. Flyer22 Frozen (talk) 23:38, 4 May 2020 (UTC) Updated post. Flyer22 Frozen (talk) 23:53, 4 May 2020 (UTC)

Ted Hill's work
It was published. It passed peer review, which is the important step we require here. Your original revert reason is simply not applicable. Your second edit comment is original research from you. A high quality secondary source would be better but in the absence of that a primary source is the best we have. It is clearly research relevant to the topic, we should mention it. --mfb (talk) 05:45, 17 September 2018 (UTC)
 * An article that was briefly published and then retracted is not "published"; it is "retracted" - the journal is not currently saying that it meets the content and review standards of that journal. (The lack of a formal retraction here does not lessen the point that the editors of the journal have effectively retracted it. Passing peer review is not the decision of the reviewers, it is the decision of the editors.) If my second revert reason is OR then so is the summary that was in the article, as it is just an interpretation of what Hill was saying. He himself is pretty clear in the paper that his model is a very simplified mathematical model (for that reason I think it is in fact not clear that it is relevant to the topic) and much further work is necessary to see if his assumptions are satisfied in real systems (see section 8 of ) so the summary that I removed oversold even what Hill claimed to have done. But that is why we shouldn't be putting raw primary research straight into Wikipedia articles -- to let a secondary source do the interpretation for us, so we're not doing original research in interpreting it. If we are going to discuss this article, we need to do it through secondary sources. CapitalSasha ~ talk 06:55, 17 September 2018 (UTC)
 * A retraction would need an announcement that it was retracted. It was deleted without further comment - and not for scientific reasons (otherwise we would have an explanation why). Hill's research is not our OR, it is research that did pass peer review: The editors said it is okay. We have many primary sources, several of them are references here. As long as we present them with "X said Y" not "Y is" that is not a big deal. --mfb (talk) 19:45, 17 September 2018 (UTC)
 * Your opinion on the correct retraction process, and your assumption that the retraction was not for scientific reasons, sound a lot like OR to me. The editors of the journal initially decided to publish it and then decided not to publish it, for reasons we don't know for sure. Moreover, it had been submitted to math journals, which I would not normally call high-quality sources for an article about biology. A paragraph in this article about an arxiv posting about a very simplified mathematical model seems like undue weight to me. CapitalSasha ~ talk 20:58, 19 September 2018 (UTC)
 * Can we let the decision if it is very simplified be made by a journal please? There is no need to repeat your personal opinion every time. We know there is no reason given for deleting the article. A proper retraction is done with a comment why it was retracted, this didn't happen here. The paper is about a mathematical model for a biological phenomenon, what is wrong with mathematical journals? You'll find mathematics used in physics in mathematical journals as well. --mfb (talk) 21:31, 19 September 2018 (UTC)
 * 1. Perhaps the retraction was not "proper", but that does not mean it was not for scientific reasons -- that is an inference. 2. The fact that the model is very simplified is not my personal opinion -- it is stated in the conclusion of the preprint:

The goal here has been neither to challenge nor to confirm Darwin’s and other researchers’ observations of greater male variability for any given species or any given trait, but rather to propose an elementary mathematical theory  based  on  biological/evolutionary  mechanisms  that  might  serve  as  a  starting  point to help explain how one gender of a species might tend to evolve with greater variability than the other gender. The precise formal definitions and assumptions made here are clearly not applicable in real-life scenarios, and thus the contribution here is also merely a general theory intended to open the discussion to further mathematical modeling and analysis.
 * But I think what has become clear from this discussion is that it is impossible to determine the relevance of the paper to this article without interpreting a math paper's relevance to the biological sciences (since this article is about the biological sciences, not about mathematics), and that requires a secondary source. CapitalSasha ~ talk 05:34, 20 September 2018 (UTC)
 * Clearly Hill's arXiv paper 1703.04184 should be cited! It was accepted for publication, it then turned into a controversy, causing it to become unpublished. The controversy clearly proves that the paper is relevant. Neither its acceptance, nor its unpublication, proves if the paper is "true" or "false". But it is not up to us to decide what is "true". If published evidence demonstrates its falsehood, then that can be cited along with Hill's paper. But either way, the work is relevant and we must cite it. If we can not agree on this we should ask other wikipedia editors to make a decision for us. MvH (talk) 19:31, 6 October 2018 (UTC)
 * Why on earth does a controversy prove that the work is relevant to the scientific issue of the variability hypothesis? CapitalSasha ~ talk 03:59, 10 October 2018 (UTC)
 * Papers that have no relevance to controversial topics are not treated this way. (Edited:) In the acknowledgments there is a colleague who worked on computer simulations who wants to remain anonymous. The message is very clear: If you write a paper relevant to this topic (even a math paper!) but the paper supports the "wrong" side, you had better be in a point in your career where you have nothing left to lose (i.e. retired). People who are considering working on this subject will certainly want to know this. In any case, the efforts to stop this paper are highly counterproductive, they make this paper much more widely known. MvH (talk) 19:45, 23 October 2018 (UTC)
 * See WP:EXCEPTIONAL. Papers that show shocking and unexpected results that go against established science require stronger evidence; in particular, WP:RSMED warns against relying on individual papers with shocking or controversial results that were not clearly borne out by later research.  Our job as an encyclopedia is to reflect the scientific consensus, not what you personally feel the scientific consensus ought to be; that means that controversial one-off papers aren't good sources. --Aquillion (talk) 23:50, 21 August 2019 (UTC)
 * A paper that is allegedly too simplified to be relevant but at the same time important enough to get several warriors trying to make it disappear where ever they can (that's mainly directed at journals). That's a curious combination. If the paper would be as irrelevant as people claim it would simply be ignored, just like a large share of other publications is. Deletion without a proper retraction is an exceptionally unusual process, I have never heard of that happening before. --mfb (talk) 09:26, 22 August 2019 (UTC)

Bell curve image
I removed the bell curve image because the image does not include a source for the data in its description, so it is a presentation of unsourced data. If we want to include this image we definitely need a source. Pinging who added the image. CapitalSasha ~ talk 21:40, 26 September 2018 (UTC)
 * But just because you know the mean and standard deviation of a distribution doesn't mean you know that it's normal! CapitalSasha ~ talk 06:47, 1 October 2018 (UTC)
 * I can't understand when you argue the meaning "that it's normal". The Stanford-Binet (SB) score or Wechsler score (16SD and 15SD with mean 100 for all populations) are the most popular tests to get general IQ but there exist a lot of "normal" comparable tests that psychologists uses as the ASVAB-gc to get he general intellectual quotient. Even, all the test scores can be homologous with a rarity chart and bell curves comparison. Espíritu nocturno (talk) 17:32, 1 October 2018 (UTC)
 * Sorry, by "normal" I mean normally distributed (Gaussian). The way the chart is presented makes it look like it is a plot of the real distribution of scores. That's not the case - it's a plot of two normal distributions with the same means and standard deviation as the scores looked at in that study. CapitalSasha ~ talk 18:52, 1 October 2018 (UTC)
 * To add to this, it is rather standard in statistics to assume that things like test scores should be approximately normally distributed. But that does not mean that you should plot a normal distribution when you plot the data. CapitalSasha ~ talk 18:54, 1 October 2018 (UTC)
 * Both curves are the normal distribution for each gender instead of the classical normal for all groups (SB or Weschler). Even, lot of psychologist and mathematical scientists know lot of scores where the results tends to be platikurtic for male curves and leptokurtic for female. If someone ploted numbers is not an original research because is well know that SD and mean always indicate a Gaussian curve. At least is a secondary source, and secondary sources is allowed in encyclopedias.Espíritu nocturno (talk) 23:27, 1 October 2018 (UTC)
 * OK, now I have no idea what you are talking about. The mean and standard deviation can be computed for any distribution, not just a Gaussian one. What is being plotted are Gaussian distributions with the same mean and standard deviations as the true datasets -- not the data sets themselves. I don't think we have a source that justifies doing this. Even if we have a source saying that it's a good approximation (I don't doubt at all that it is), I question the wisdom of plotting the approximation rather than true data. CapitalSasha ~ talk 05:03, 2 October 2018 (UTC)
 * Yeah, is right – the curves constitute WP:Original research for the reason that the shapes of the curves are not known simply because the means and standard deviations are (for instance: the data set of the first five positive integers has a mean equal to 3 and a standard deviation equal to the square root of 2.5, but is not normally distributed). However, we don't need a real example to illustrate the concept – an image of two distributions with equal means but different variabilities would suffice. Something like this with a caption along the lines of "Two distribution curves with identical means but different variabilities. Note that the curve with the greater variability yields higher values in both the lowest and highest ends of the range." would do the trick. TompaDompa (talk) 11:12, 4 December 2018 (UTC)
 * I would have no problem with an abstract picture like that. CapitalSasha ~ talk 17:30, 4 December 2018 (UTC)
 * Very well, . TompaDompa (talk) 20:12, 4 December 2018 (UTC)

X Chromosome Inactivation
https://en.wikipedia.org/wiki/X-inactivation Offers a plausible explanation for the phenomenon described here. Males (biological/chromosomal males, "biological" is hereafter omitted but implied) have a single X chromosome whereas women have two.

In women, during the early embryological life, in every cell, one of the two X chromosome is disabled at random, for life (except in germ cells IIRC).

If a male inherits an allele on the X chromosome that codes for an extreme trait, that's all he's got. If a woman does, she may inherit of a different allele on the other X that can counterbalance that (about half of a woman's cells will have one allele active, the other half having the other allele).

As a consequence, you'd expect a greater amount of individuals with extreme traits among males than among females (for all traits coded by genes on the X chromosome). That makes women better rounded, on average, and men more prone to being very good on some axes yet very bad on others.

You can also have alleles of a gene that interact in peculiar ways, giving rise, in sub-populations, to more variability in women than in men (think tetrachromacy in women that inherit only one allele for protanomaly or deuteranomaly).

This also explains why for some traits (like birth weight) the variance is identical: those traits are not influenced by genes on the X.

Assuming this is true, we'd expect the opposite patterns in birds and other species with W and Z chromosomes (where the males are ZZ and females WZ). Edit: They do: https://onlinelibrary.wiley.com/doi/full/10.1111/evo.12224

I'm surprised this isn't discussed in the article. 2A01:E35:8B10:A450:A524:905E:1ACF:2AB2 (talk) 23:29, 17 January 2019 (UTC)

I actually looked up this article expecting a discussion of the chromosome hypothesis. I'm not surprised that it's not included, though. It both supports and explains the hypothesis. Therefore I do not expect it to be in a Wikipedia article. This is a place of social justice warfare, not science; this article is a fine example. Still I am glad I learned something from this talk section, which hopefully will not be censored.212.72.180.15 (talk) 17:43, 28 December 2019 (UTC)

Leta Hollingworth's studies
This article places far too much emphasis on Leta Hollingworth's studies, given the number of studies with better methodology / better sampling / much greater sample size that have been conducted since then. — Preceding unsigned comment added by Strallus (talk • contribs) 00:40, 7 June 2019 (UTC)

I agree it is another very biased part of the article. Stating that Hollingworth's position "allowed her the opportunity to refute the variability hypothesis" is quite the claim. Also the mentioning of her doctoral dissertation adds nothing to the article, as it clearly addresses a completely different subject. It apparently has been added to discredit the variability hypothesis as having a misogynistic motivation (by trying to demonstrate that Hollingworth also "refuted" other "misogynistic claims"). 212.72.180.15 (talk) 17:36, 28 December 2019 (UTC)
 * I removed some off-topic stuff. --mfb (talk) 02:09, 30 December 2019 (UTC)
 * Surely Hollingworth's contributions are old enough now (all before 1940) that this can be a subsection within History. This research to having its own section goes against WP:UNDUE.KristinaLu (talk) 20:56, 4 May 2020 (UTC)
 * Given the substantial influence and commentary on Hollingworth's research, it's not undue for it to have its own section. Flyer22 Frozen (talk) 23:40, 4 May 2020 (UTC)

Pseudoscience claim
The latest edit claims that this is pseudoscience without attribution. What's the justification? 2A00:23C5:F88:7900:FD06:588E:DFA3:8CDA (talk) 23:33, 2 January 2022 (UTC)