It’s been fascinating, not least from a sociological perspective, to read the comments threads under “The Faith and Fables of Thunderfoot” blog post and video I uploaded recently. (The video was mirrored at a number of other channels, so there are quite a few comments to browse in total. A big thank you to those who mirrored the video: Kristi Winters, Kevin Logan, Angry Basterds, and chrisiousity. And, of course, thanks also to all those who left intelligent, thoughtful comments. Much obliged.)
The majority of those commenting tend to echo the following sentiment (from a Reddit thread on the video/post):
It’s really quite astounding to me that Thunderf00t didn’t even attempt to provide a shred of empirical evidence to back up his “hypothesis” despite being presented with multiple opportunities.
There is, however, also a subset of comments from those who attempt to defend Mason’s stance on sexual dimorphism. These range from the clueless, willfully uninformed, and severely grammatically challenged  to a small number of rather more thoughtful and well-written replies. I deal with the latter in detail below but a few words on the former are also in order because, despite the vacuity of their responses, they provide further illuminating examples of the faith-based stance that was adopted in an attempt to support Mason’s evidence-free claims.
A number of those who have left comments in defence of Mason state specifically that they have not, and will not, read the blog posts that critique his arguments. This not only highlights a worrying aversion to reading — and it’s clear that quite a few of those who commented on the video did not read the detailed arguments in the associated blog posts — but is indicative of an inherently ‘tribal’, i.e. “in group” vs “out group”, attitude that really doesn’t care about evidence or reasoned argument. (We’re seeing similar gut-level responses in the EU referendum debate here.)
It was also amusing to find quite a few posting comments along the lines of “What Phil (Mason) is saying…/What Phil (Mason) means…/What Thunderfoot is pointing out…“, despite the fact that at no point during the exchange with Mason did he provide any type of (counter-)argument. I believe that the term Mason et al. would use under these circumstances is “white knighting“. (In addition, a number of particularly aggrieved commenters defending Mason’s honour claimed, in textbook ad hom style, that my core motivation was to simply get more YouTube views. Errmm, no. Some of us are motivated by factors other than YouTube view counts.)
Moreover, I very deliberately used “faith” in the title of the post and video; some of those commenting helped to strengthen that particular argument for me. This faith-based position was no better demonstrated than in this comment (and follow up). Note the absence of any attempt by “St. Thomas” to provide evidence to support their position. It’s just a gut-level, instinctive claim which is remarkable in its core certainty: Of course there’s lots of evidence for this.
That’s faith in action.
Another intriguing parallel with faith group thinking, and something I find remarkable for those who identify as atheists, is the persistent appeal to what’s best described as the “natural order of things”, i.e. women are just less suited to and/or less disposed to physics because of their (immutable) “nature” . Most of the time this is asserted with nothing more than the type of empty claim put forward by “St Thomas” above, but, on occasion, a more thoughtful analysis is given.
One of those who commented took the time to write a blog post (with the obligatory Sherlock Holmes reference, of course): Being Sherlock is edgy these days. That post makes the same core points as have been put to me (very) occasionally by the more literate/intelligent supporters of Mason and so is worth dissecting in detail. (I only wish Mason could have responded at this level). Let’s start.
“As I said back when we first clashed it is currently not necessarily easy to tease out what is innate and what is.”
There’s an unfortunate typo here but clearly what’s meant is the following: “It is currently not necessarily easy to tease out what is innate and what isn’t”. Indeed. This has been the core of my argument throughout. But “not necessarily easy to tease out” is a massive understatement. The balance of nature vs nurture is exceptionally difficult to determine in very many cases, and this is why there have been so many long-standing debates spanning decades. It’s worth reading the exchange in the comments section under this article to see just how bitter those nature-vs-nurture disputes can get, even among professionals in the field.
Arguably the most compelling recent evidence for the strong convolution of nature and nurture — as I outlined in the “When atheists ape creationists…” post — is the comprehensive (to put it mildly) meta-analysis carried out by Polderman et al., published last year: Nature Genetics 47,702–709(2015). (I’ve bypassed the paywall and am making made the .pdf of that paper available free of charge. It will remain available here unless Springer Nature, the publisher, decides to serve up a cease-and-desist order).
That meta-analysis is astounding in its scope. Quoting from the abstract of the paper, “We report a meta-analysis of twin correlations and reported variance components for 17,804 traits from 2,748 publications including 14,558,903 partly dependent twin pairs, virtually all published twin studies of complex traits.”
14.5 million pairs of twins!
Meta-analyses are not common in my research field of condensed matter physics/nanoscience — I struggle to think of a single example. They’ve been used in particle physics, however, for quite some time. Moreover, the concept of a meta-analysis appears to have been first introduced by astronomers and mathematicians in the 18th century. Meta-analyses are now a core part of the research firmament in a number of fields, including, of course, medicine.
When attempting to determine the genetic/biological vs societal underpinnings of particular aptitudes, it is important not to rely on individual, isolated studies. For one thing, and as highlighted by Poldermann and co-workers, the balance of nature vs nurture is generally close to 50:50.This means that the effective ‘signal-to-noise ratio’ for sexual dimorphic effects can be weak and thus the associated statistical analysis of the data needs to be exceptionally robust. Meta-analysis can help to provide this powerful statistical basis.
Back to that blog post…
What we know is that in highly talented samples there is an extreme ability difference. Data comes from several sources. First when it comes to mechanical reasoning, a category highly relevant to physical science, there is a large gap in mean ability, 3/4 of a standard deviation between men and women. Source.
The problem with the (single) cited source here is that the study does not attempt to normalise out environmental/societal influences. Moreover, the suggestion in the blog post (and the cited paper) would appear to be that the differences are “hard-wired” and immutable. (I’ll get back to this). As the — anonymous, of course — writer of the post stated from the outset, teasing out just what is innate and what isn’t is not easy…
Reading up on some of the papers that cite the study linked to above (i.e. Lemos et al.), we find (i) a meta-analysis that highlights the importance of the relationship between vocational interests and cognitive abilities [this]; (ii) a study that investigates the link between socioeconomic level and cognitive ability (this), finding that, as stated in the abstract, “socioeconomic level had more influence than sex on most of the cognitive tests“; and (iii) a distinct warning against using comparisons of g scores across gender.
I cite these papers not to suggest that any of them is the definitive last word on the subject. In fact, I cite them precisely because they’re clearly not the definitive last word on the subject. It is exceptionally important not to cherry-pick individual studies and consider their findings in isolation. This is true in the physical sciences, but it is orders of magnitude more true outside the neatly controlled confines of experimental physics where there are so many confounding, and confounded, variables that too often cannot be adequately taken into account. This is one reason Internet Guy here doesn’t appreciate that the abstracts he’s cited (after a quick search for keywords with Google Scholar) may not be quite as “damning” as he thinks…
Moreover, when a huge percentage of research in a particular field is irreproducible, meta-analyses, rather than single studies, become critically important.
Such a difference in mean has, when assuming a normal distribution (which is not a bad approximation, see here )of ability massive differences at the tail of the distribution. For example if physcists need +3SD of ability to succeed this would mean that the cutoff for the female distribution is 3 +3/8 SD above their mean while for the males it is 3 -3/8 over their mean, leading to a ratio of male to female of 11.6:1.
First, I have no bone to pick with regard to the normality/’Gaussianity’ of intelligence levels (although I have many bones to pick with the concept of the pseudo-quantitative estimation of intelligence that is the IQ level. IQ tests demonstrate one’s ability to…do IQ tests). The central limit theorem tells us that a Gaussian is the natural result of the convolution of different probability distribution functions so, given the complexity of the nature-nurture process as described above, I’m happy to accept normality. 🙂
But where does the metric of “+3 standard deviations to succeed”arise? Where is the evidence for this claim? Or was it chosen simply to fudge the figures so as to get a preferred male:female ratio? I note that the author of the blog post doesn’t provide a citation.
Further we have several pieces of evidence that at the tails (not at the mean) there actually are significant differences in mathematical ability. For one at the higher end of SAT-M scores (700-800) the ratio of boys to girls is 1.6http://media.collegeboard.com/digitalServices/pdf/research/TotalGroup-2012.pdf
Yet again, this takes no account of environmental/societal factors. (I’ll reiterate that the author of the post herself/himself pointed out that separating out “innate” and “non-innate” differences is problematic.) In any case, the question of ‘gendered’ ability in mathematics (where the gender balance is close to 50:50 in the US, and currently stands at ~ 40:60 (F:M) in the UK) has been studied in considerably more detail than for physics. For example, in a well-cited paper (based on the findings of a number of meta-analyses), Hyde and Mertz show that “girls in the US have reached parity with boys in mathematics performance” and that “greater male variability with respect to mathematics is not ubiquitous. Rather, its presence correlates with several measures of gender inequality. Thus, it is largely an artifact of changeable, sociocultural factors, not immutable, innate biological differences”.
The lack of immutability is key because if cognitive differences between males and females really were “hard-wired” and entirely dominated over societal influences, we would not expect to see significant differences in uptake/aptitude for various subjects over short periods of time (i.e. decades). This very important point is made very well by ObjectiveReality a number of times in the comments thread for “Faith and Fables…” .
Indeed, when we look at mathematics, we find that the gender balance in ability is certainly not locked in place (as stated clearly in “When atheists ape creationists…“):
“…it does not seem that biology is limiting intelligence in any way because biology alone cannot explain the vast improvement of female performance on certain measures such as the increasing numbers of females scoring at the highest end of the SAT math test (Blackburn, 2004).”
My correspondant should note the “at the highest end of the SAT test” qualifier in the quote above before they make assertions re. means vs tails of distributions. Hyde and Mertz also addressed this distinction (at length) in their paper. Moreover, they cite work by Penner (Am J Sociology 114:S138 – 170) which reaches the following conclusion: “The common assumption that males have greater variance in mathematics achievement is not universally true“.
It’s also worth reading some — or, indeed, time permitting, all (!) — of the papers that cite Hyde and Mertz’s work. These include “Do the maths: An analysis of the gender gap in mathematics in Africa” by Dickerson and co-workers. (I have a particular interest in education in Africa, having visited Ethiopia recently). Once again, the authors conclude that there is a substantial socioeconomic/societal component underpinning performance:
There is a significant difference in maths test scores in favour of boys, similar to that previously observed in developed countries. This difference cannot be explained by gender differences in school quality, home environment, or within-school gender discrimination in access to schooling inputs. However, the gender gap varies widely with characteristics of the regions in which the pupils live, and these regional characteristics are more predictive of the gender gap than parental education and school characteristics, including teacher gender.
I should stress yet again that I am not suggesting that Dickerson et al.’s paper is the last word on gender differences in maths in Africa (or elsewhere in the world). I cite it simply to show that, as one might expect from that pioneering meta-analysis of Poldermann et al discussed above, nature and nurture are inherently convolved. It is entirely unscientific to state that the nature (i.e. genetic/biological) component dominates aptitude/preference for physics when there is no evidence to support that conclusion.
Back to my correspondant’s blog post…
So to summarize my first and most important point: The proximate cause of gender differences in accomplishment in physical and mathematical science is likely differences in the number of highly talented individuals.
That’s a remarkable claim on the basis of just a handful of cited papers, particularly when the literature has addressed, and rebutted, those claims at length, as discussed above. (c.f. Internet Guy). Note, in particular, Penner’s paper referred to above (American J. Sociology 114:S138 – S170 ), a substantial piece of work, and the section entitled “Do females exist who possess profound mathematical talent” (and references therein) in Hyde and Mertz’s paper.
It’s worth quoting from the introduction to Penner’s paper as he explains the key point of his work,
“Given the inextricable link between the biological and social, I show that one way to proceed is to examine these differences internationally…If gender differences vary across countries (and they do) then social factors are important”.
This “geographical” variation complements the temporal variation discussed previously.
Back to the dissection of the blog post…
Proximate social causes like discrimination in universities are bad candidates as explanation since they a, ignore ability differences, and b do not explain ability differences that are allready present in 12 year olds. Whatever the reason for the difference it starts early.
This point is bizarre. My entire argument (and that of many of the papers I’ve cited above) is that we have to consider both nature and nurture components. But the societal effects obviously don’t just kick in at university — they’re present throughout life, from early stage (primary/elementary school) learning, and before. To argue that the “difference(s) start early” does not provide any type of evidence that we should discount societal/environmental effects in favour of a genetic/biological dominance.
In any case, when it comes to mathematics, the claim that the differences are already present “early” has been contested. For example, it’s been argued that boys and girls in preschool grasp number concepts at the same rate (see Spelke, Amer. Psychology 60, 50 (2005)).
Practice makes perfect?
I’m going to close this lengthy post with a discussion of the flawed concept that aptitudes for STEM subjects — or any subject — are immutable, with a particular focus on the topic of spatial reasoning. This is of keen interest to me because, although I’m now a physicist (and have loved science and physics from an early age), when I did an aptitude test in the early years of secondary school my spatial reasoning scores were rather lower than I would have hoped, and certainly made me (momentarily) question whether I was cut out for physics.
There’s a lot of spatial reasoning in physics. This is particularly the case in my area of research — condensed matter physics/surface physics/nanophysics (call it what you will) — where we have to consider crystal structures, symmetry groups and operations, different arrangements of atoms on various low- and high order crystal planes, etc…
What made a huge difference to my ability to consider and analyse structures in both real and reciprocal space was… practice.
And what’s made a huge difference in my ability to do physics of any type? Practice.
That’s one reason I found this particular article so fascinating. Questions just like the “Rotate This” poser in that article formed part of the aptitude test on spatial reasoning I did years ago. 34 years on from doing that aptitude test, it’s second nature to solve that puzzle. As a teenager, however, I clearly must have struggled. My experience mirrors that of Sheryl Sorby, described in the post:
As Sorby took more engineering courses, she got better at spatial cognition tasks, until eventually she found herself teaching engineering graphics, the very course that almost derailed her as an undergrad. “The brain is pretty plastic when it comes to spatial skills,” Sorby says. “I have improved my spatial skills vastly as an adult.”
I recommend you read the entire post but I’m going to quote at length from it in any case because it flags up (for the n-to-the-nth time) how it is nigh-on-impossible to credibly or definitively separate nature from nurture in so many cases.
“We don’t know what’s cause, and what’s effect,” Cashdan says. What is clear is that cultural biases have an effect.Consciously or unconsciously, girls are nudged away from activities that would help them develop spatial skills almost as soon as they’re born. As they grow, parents respond to their kids’ interests, quickly compounding what may start out as very slight biases.
“Parents are very invested in gender differences, and any differences between a son and a daughter tend to be attributed to sex,” says Lise Eliot, a neuroscientist at the Chicago Medical School of Rosalind Franklin University of Medicine and Science, author of Pink Brain, Blue Brain: How Small Differences Grow into Troublesome Gaps, and What We Can Do About It.
Over time, “boy” toys reinforce skills that are proven brain boosters. Playing with Legos and blocks, taking a shop class in high school and time spent playing 3-D computer games have all been shown to boost scores on mental rotation tests.
Ultimately, separating nature from nurture may be impossible. But Sorby and others who study gender differences say it may not matter. Nora S. Newcombe, a cognitive and developmental psychologist at Temple University, who has researched gender differences in spatial cognition, bristles at the concept that the dearth of women in science is due to hard-wired deficiencies. “I think there might be a biological mechanism, but it doesn’t seem that important in terms of human potential,” she says. “It seems like an excuse.” An excuse not to do the hard work necessary to improve in places we might be lacking.
Old dogs, different drums
Finally, I’m also interested in the nature vs nurture issue from the perspective of education in general (as distinct from, and in addition to, gender balance issues). Malcolm Gladwell’s book Outliers introduced the “10,000 hours” concept, i.e. it apparently takes 10,000 hours of practice to achieve mastery of a sport, or a game such as chess, or a musical instrument etc… This ‘meme’ has spread across the internet like wildfire since Outliers was published back in 2008. Gladwell’s arguments have been thoroughly critiqued since then with many making the rather obvious point that it’s not just any old practice regime that’s important: it has to be targeted and focused. Gladwell has always stressed, you guessed it, the importance of the nurture component of the nature-nurture question.
The targeted practice idea resonates with me because over the last year or so I’ve been spending an hour a day learning to drum (specifically, double bass drumming) with the wonder of Aerodrums. As discussed in the video below, my practice regime has been very focused. (Not easy for me). I also mention Gary Marcus’ Guitar Zero in the video — a fantastic book which challenges the age-old, and clearly flawed, adage that you can’t teach an old dog new tricks.
The brain, even in middle- to late-adulthood, is much more plastic than was previously thought. What’s also intriguing is that strong connections between physical activity and brain plasticity have been found. Erickson and co-workers have reviewed the research literature in this area, concluding that “physical activity is a promising intervention that can influence the endogenous pharmacology of the brain to enhance cognitive and emotional function in late adulthood.“. Drumming, of the Aero or traditional variety, would therefore seem to be an especially powerful method of enhancing cognitive function, combining physicality with learning an instrument.
And it finally gives the lie to all those “stupid drummer” jokes. (You know the ones… “What do you call a drummer with half a brain? Gifted”;”How do you tell if the stage is level? The drummer is drooling from both sides of his mouth.”)
 Some who seek to defend Mason claim that my pointing out the deficiencies in his writing is somehow an ad hominem fallacy. This shows a distinct lack of understanding of that particular fallacy. I did not attack Mason’s lack of communication skills in lieu of countering his groundless claims re. sexual dimorphism. Instead, I presented a detailed rebuttal of his claims and, in parallel, highlighted the deficiencies in his written communication. Indeed, in the video I introduced my criticism of Mason’s communication skills by referring to it as a “peripheral point”.
In addition, I found it amusing and illuminating to be chastised for writing “pretty language“. Although I took that chastisement very much as a compliment, it again flags up the increasing inability/unwillingness of many to read and digest even moderately sophisticated arguments. This is something that has concerned me for a while, particularly as I may well be contributing to the problem. See this post (or, for those who prefer not to read, this video).
 One of the most frustrating aspects of the inability/unwillingness in some quarters to read anything more complex than a grammatically garbled YouTube slur is that I end up having to repeat myself. Repeatedly.
 There is a tendency among Mason et al.’s followers to irrationally dismiss results published by social scientists solely on the basis of the discipline. For the reasons I discuss in “When atheists ape creationists…” this is an appallingly weak position to adopt.