Thursday, October 9, 2014

H.J. Hornbeck should apologize to Sam Harris


Sam Harris has been accused of making some sexist statements about gender differences related to aggression and nurturing, and H.J. Hornbeck has opted to join that bandwagon of intellectual self-immolation with his own youthful exuberance (see comments).  At least he seems less twitchy than another of Harris' recent critics: The Dark Knight that Harris didn't deserve. The thread that ties criticisms of Harris' ideas together, whatever their source, is commonly a failure to thoughtfully and carefully consider what the man is actually saying. I fear that Hornbeck's zeal has clouded his judgement, too; while he manages to find sexism in just about every sentence that Harris makes, he has proven unable to demonstrate that they meet even his own definition of sexism.


I am, however, thankful for his acknowledgement that this conflict reduces to a question about empirical evidence which he, more than most, has been willing to discuss. We agree that his burden is to show that Harris’ original comments are, as all that upon which bigotry is ultimately founded, unwarranted or false. Today, I will again focus mainly, though not exclusively, on the seminal work of Janet S. Hyde, PhD, who coined the “gender similarities hypothesis” in her detailed meta-analysis of over 20 years of psychological studies that showed that the genders are much less different than we think. This is Hornbeck’s defiant position and he would have us believe that it is incompatible with Dr. Harris’ remarks.


But is that really the case?


There is a clear gender difference in (even non-physical) aggression


Hornbeck has been correct to point out that the gender difference related to physical aggression is greatest, but the problem for him is that it just doesn’t follow that there are no relevant differences in non-physical aggression.



Table 1 in Dr. Hyde’s study (reproduced in part above) provides information from 5 different meta-analyses of aggression between the genders.  Every single one of those studies found that males were more verbally (or otherwise non-physically) aggressive than females, with the magnitude of that difference overwhelmingly ranging from small to moderate (values in the right column from 0.11-0.65)*. Hyde analyzed 6 studies concluding that males were moderately more verbally aggressive. Eagly and Steffen analyzed 20 studies concluding that males were mildly more psychologically (as opposed to physically) aggressive. Knight analyzed over 50 studies concluding that males showed moderately more verbal aggression and aggression in a variety of emotional arousal contexts. Bettencourt & Miller analyzed over 50 studies concluding that males were moderately more aggressive under both neutral and provocative conditions. Archer analyzed almost 100 studies concluding that verbal aggression had a mild to moderate male predominance.


Consider this data from the perspective of evolutionary biology. Hornbeck does not challenge Hyde’s finding of large and moderate sized gender differences in physical prowess and physical aggression. But ask yourself, why would evolution build males this costly way and not provide them with psychological inclinations (that can be amplified by sociocultural influences) towards aggressive conflict? We simply should not be surprised that small to moderate differences between the genders in non-physical aggression shine through reams of data.


Hornbeck equivocates that because this gender difference isn't large, non-physical aggression "doesn't have much of a gender divide"[emphasis is mine], but as numerous investigators have shown over and over, the notion that males are, on average, mild-moderately more aggressive, including more aggressive in non-physical ways than females, seems unopen to dispute. And if you think that small-to-moderate effect sizes just can't be relevant in the real world, I encourage you to read Dr. Hyde's discussion on that topic in her paper^.


But there is more with which Hornbeck must contend ...


What about nurturing?


Ironically, Hornbeck claims that I’m “equivocating between aggression and not nurturing”, but this is a red herring. I wonder if what he really meant was that I’ve been conflating being aggressive with being less nurturing. Doing so is not as inappropriate as it may at first seem because these two personality variables lie in contrast to each other. Aggression is about attack, and hence, is offensive in posture, while nurturing is more about fostering and protecting, which is more defensive. Just as evidence that men are taller is relevant if one believes that women are shorter, evidence that males are more aggressive is relevant if one believes that females are more nurturing. And besides, only the most uncharitable reading could lead one to conclude that Harris doesn't weave both concepts into his comments.


But if Hornbeck specifically wants evidence that females are, on average, moderately more nurturing, well, that exists, too. In 2011, Yanna Weisberg, PhD, and colleagues examined personality traits between the genders in a variety of cultural traditions. It is noteworthy that this work was written after Hyde's meta-analysis, so its conclusions regarding both the general state of the evidence and the particular findings of that investigation regarding nurturing are more contemporary. In their introduction, they wrote:

“Gender differences in personality traits are often characterized in terms of which gender has higher scores on that trait, on average. For example, women are often found to be more agreeable than men (Feingold, 1994; Costa et al., 2001). This means that women, on average, are more nurturing, tender-minded, and altruistic more often and to a greater extent than men.”

Weisberg found few surprises relating to nurturing:

Replicating previous findings, there was a significant gender difference in Agreeableness [ie. being more nurturing, tender-minded, and altruistic] such that women tend to score higher than men, and this pattern was the same for the aspects, Compassion and Politeness, when measured in terms of raw scores or residualized scores. Compassion most clearly represents a tendency to invest in others emotionally and affiliate on an emotional level, encompassing traits such as warmth and empathy. Politeness describes the tendency to show respect to others and refrain from taking advantage of them, and is related to traits such as cooperation and compliance. Our findings that women score higher than men on both aspects are consistent with previous research showing women are more trusting and compliant than men (Costa et al., 2001)" - Weisberg et al 2011

But I found what came towards the end of Dr. Weisberg's paper most striking:

"We would caution against adopting such a dramatic interpretation of the pervasive gender differences in personality that we report in this study. All of the mean differences we found (and all of the differences that have been found in the past – e.g., Feingold, 1994; Costa et al., 2001) are small to moderate. This means that the distributions of traits for men and women are largely overlapping. To illustrate this fact, in Figure 10 [see graph below] we present the male and female distributions from our sample for the trait which showed the largest gender difference, Agreeableness. One can see that both men and women can be found across a similar range of Agreeableness scores, such that, despite the fact that women score higher than men on average, there are many men who are more agreeable than many women, and many women who are less agreeable than many men.” -Weisberg et al 2011
Figure 10. Weisberg et al.

Weisberg's last paragraph is written with many of the same cautionary words that Dr. Harris himself  employed to characterize his position:

“My work is often perceived (I believe unfairly) as unpleasantly critical, angry, divisive, etc. The work of other vocal atheists (male and female) has a similar reputation. I believe that in general, men are more attracted to this style of communication than women are. Which is not to say there aren’t millions of acerbic women out there, and many for whom Hitchens at his most cutting was a favorite source of entertainment. But just as we can say that men are generally taller than women, without denying that some women are taller than most men, there are psychological differences between men and women which, considered in the aggregate, might explain why “angry atheism” attracts more of the former ... How much is explained by normally distributed psychological differences between the sexes?” – Sam Harris

Harris went out of his way to ensure that his message would not be received as an endorsement of the idea that large, purely innate, psychological gender differences are the only explanation for the entire apparent gender imbalance in question, but this is what his critics would have you believe. To have done so would be to have made sexist remarks, but Harris' suggestion is clearly, and, as usual, more nuanced and sophisticated than the straw men from which his critics have been making much hay of late. He really isn't the sexist pig they're looking for.


The glaring irony here is that it is those who distort Harris' message that spread false and unwarranted claims. His remarks and later his explanation over at his blog convey notions supported by a significant body of literature that, while generally highlighting gender similarities, nevertheless points directly at psychological gender differences related to aggression and nurturing. Accordingly, it is entirely reasonable for Dr. Harris to consider that these real, albeit small to moderate, specific differences may play a relevant role in explaining the apparent gender imbalance in active, North American atheism. Note that the magnitude of the trait differences in studied populations don't have to mimic the magnitude of the apparent male predominance in active atheism. The former must only be large enough that certain contexts may tend to attract more males than females, helping to explain the latter.


It's time for Hornbeck to retract his charge of sexism and apologize. He either has to do that, or else show that none of this evidence could reasonably be believed, for only if that were the case would Hornbeck's own definition of sexism (that Harris' comments were false or unwarranted) run through. As you can see by looking at the evidence, that can't possibly be the case.


I hope that Hornbeck will use his voice on campus and in social media to help remove this ugly and penetrating stain on Dr. Harris' reputation, and encourage others to follow suit.


1. The Gender Similarities Hypothesis. American Psychologist 2005;60:581-92.
2. Gender differences in personality across the ten aspects of the big five. Frontiers in Psychology 2011; 2:178


*Archer found that females may have a greater propensity than men for indirect or relational aggression, which amounts to behaviour such as backbiting and gossiping to harm others by undermining their relationships, though the effect size crossed unity, indicating that in at least one study, males were more inclined towards this type of aggression.

^d values of 0.2-0.6, which seem to be the case for aggression, would be expected, based on Rosenthal's analogy, to result in absolute differences between men and women of 10-25%.

27 comments:

  1. I fear that Hornbeck's zeal has clouded his judgement, too; while he manages to find sexism in just about every sentence that Harris makes, he has proven unable to demonstrate that they meet even his own definition of sexism.

    And you complain I'm being hyperbolic? In the previous thread, I gave three specific examples of sexism in Harris' article. You have only made the barest attempt at refuting my arguments, and even then just argued by assertion; instead, you've run away and focused on a small point of Harris' rather than confront my definition head-on.

    For someone who claims to be good at philosophy, you're quite horrible at it. About the only thing you're worse at is sociology, which my next few comments should demonstrate.

    ReplyDelete
  2. And if you think that small-to-moderate effect sizes just can't be relevant in the real world, I encourage you to read Dr. Hyde's discussion on that topic in her paper.

    As do I, because as I pointed out last thread, her interpretation of her own work is at odds with yours (emphasis mine):

    Rosenthal (1991; Rosenthal & Rubin, 1982) has argued the other side of the case—namely, that seemingly small effect sizes can be important and make for impressive applied effects. As an example, he took a two-group experimental design in which one group is treated for cancer and the other group receives a placebo. ... an r of .32 between treatment and outcome, accounting for only 10% of the variance, translates into a survival rate of 34% in the placebo group and 66% in the treated group. Certainly, the effect is impressive.

    How does this apply to the study of gender differences? First, in terms of costs of errors in scientific decision-making, psychological gender differences are quite a different matter from curing cancer. So, interpretation of the magnitude of effects must be heavily conditioned by the costs of making Type I and Type II errors for the particular question under consideration. ... Second, Rosenthal used the r metric, and when this is translated into d, the effects look much less impressive.


    If there's anything you can gather from her work over the last thirty years, it's that gender difference is so small as to be meaningless in most contexts (emphasis still mine):

    The main conclusion that can be reached from this analysis is that the gender differences in verbal ability, quantitative ability, visual-spatial ability, and field articulation reported by Maccoby and Jacklin (1974) are all small. Gender differences appear to account for no more than l%-5% of the population variance. The difference in means is only about one fourth to one half of a standard deviation ... . Generally, it seems that gender differences in verbal ability are smaller and gender differences in spatial ability are larger, but even in the latter case, gender differences account for less than 5% of the population variance.
    [Hyde, Janet S. “How large are cognitive gender differences? A meta-analysis using w² and d..” American Psychologist 36.8 (1981): 892.]

    ReplyDelete
  3. Doing so is not as inappropriate as it may at first seem because these two personality variables lie in contrast to each other. Aggression is about attack, and hence, is offensive in posture, while nurturing is more about fostering and protecting, which is more defensive.

    Except people can be both aggressive and nurturing at the same time; you could be nurturing to an infant in your care, while aggressive to other adults. Context makes a huge difference, and Hyde makes a similar point in her Gender Similarities paper:

    Gender researchers have emphasized the importance of context in creating, erasing, or even reversing psychological gender differences (Bussey & Bandura, 1999; Deaux & Major, 1987; Eagly & Wood, 1999). Context may exert influence at numerous levels, including the written instructions given for an exam, dyadic interactions between participants or between a participant and an experimenter, or the sociocultural level.

    In an important experiment, Lightdale and Prentice (1994) demonstrated the importance of gender roles and social context in creating or erasing the purportedly robust gender difference in aggression. ... The results indicated that in the individuated condition, men dropped significantly more bombs (M = 31.1) than women did (M = 26.8). In the deindividuated condition, however, there were no significant gender differences and, in fact, women dropped somewhat more bombs (M = 41.1) than men (M = 36.8). In short, the significant gender difference in aggression disappeared when gender norms were removed.


    So even if these studies show a difference, it's not clear that difference is due to social context or biological difference. Other researchers concur, and think the idea of a sex difference in aggression in meaningless:

    There are good reasons to doubt whether it is meaningful at all to debate whether one sex is more or less aggressive than another. One has to take into account the type of conflict. Firstly, is it a matter of aggression between groups, or of interpersonal aggression within a group, or within a family. Secondly, the sex of the opponent is of critical importance: male-male, female-female, and male-female encounters should clearly be distinguished from each other. With respect to interpersonal aggression,
    same-sex encounters are, for instance much more frequent than between-sex
    counters (Bjorkqvist and Niemeld, 1992; Burbank, 1987).
    Björkqvist, Kaj. “Sex differences in physical, verbal, and indirect aggression: A review of recent research.” Sex roles 30.3-4 (1994): 177-188.


    So both you and Harris are still at odds with the research.

    ReplyDelete
  4. Replicating previous findings, there was a significant gender difference in Agreeableness [ie. being more nurturing, tender-minded, and altruistic] such that women tend to score higher than men, and this pattern was the same for the aspects, Compassion and Politeness, when measured in terms of raw scores or residualized scores.

    Dr. Weisberg conflates statistical significance with personal significance. I can demonstrate with good statistical significance that a coin lands heads-up 53% of the time by tossing it a thousand times, but if you wanted a single yes-no decision, would you avoid using that coin? Of course not, for a single flip that difference has no personal significance.

    Take a look at his chart on Agreeableness. Notice how much overlap there is between men and women? I popped it into an image editor, and found 70% of the values overlapped. If we were gather a large number of men and women together, then asked the women to find a man with the same Agreeableness,78% would find a partner. If I were to pluck a random Agreeableness score out of all the ones found, and ask you to guess which gender it was associated with, on average you'd get it right 59% of the time, at best.

    On what basis does he argue men and women are different on Agreeableness? As far as I can tell, the similarity is far stronger than any difference, and if that's the case how can Weisberg argue for difference on any other metric, given that he admits every other one shows greater similarity? His finding of statistical significance does not translate into personal significance.

    ReplyDelete
  5. It's time for Hornbeck to retract his charge of sexism and apologize. He either has to do that, or else show that none of this evidence could reasonably be believed, for only if that were the case would Hornbeck's own definition of sexism (that Harris' comments were false or unwarranted) run through. As you can see by looking at the evidence, that can't possibly be the case.

    Actually, it totally is. Since you're coming at it from a background assumption of difference, all you see is difference. Consider the alternate hypothesis, and suddenly everything you've brought forward argues for similarity and pokes holes in Harrs' claim.

    This ignore the fact that Harris was almost certainly unaware of this research, and that not every statement of his that I declared to be sexist relies on this research.

    No apology is warranted, so I ain't givin' one.

    ReplyDelete
  6. Well I couldn't bring myself to give a full and proper reading to Harris' response post, but while it did rub me the wrong way, it wasn't horrid or awful, there where areas when he was clearly speaking outside of his expertise. So I'm just going to assume he didn't say anything full out sexist only that he (as he admitted himself in the post) doesn't have the back ground to talk about this stuff. The response post just came across as facile. This is probably because the sort of systematic oppression we are talking about isn't something one can learn in 2 hours. it's taken me probably over 150 hours to get a good grip on the subject material and I'm still learning all the time.

    It looks like HJ has defended himself well enough so really I'm in the position to say that in honest fact it doesn't matter nearly enough for an apology or heavy burden of blame be passed around. Harris probably said something causally sexist and a then the blog-o'sphere exploded. Though I will say this.

    Harris probably did say something casually sexist out of his own ignorance, so if Harris wants to talk about this more in the future he should do the research. If not, then he should probably not talk about it in the future. "I honestly don't know enough to give a competent answer to that question." would probably work.

    However, those papers where very interesting to read, so I'll be critiquing more your data analysis then anything.

    However I would like to talk about the statistical significance of these to papers or how they are barely useful for drawing any sort of conclusions. It all comes to one simple fact about the data used, it's ordinal. For those read who are not in stats, ordinal data is data which is ordered by there is not set scale. So 1 could be 10 times smaller than 2, and three could simply be half as large again as two. For example.

    To be able to draw hard conclusions about a data set you need data which is scaled, and you know to a high degree how those values differ. (are the values logarithmic, exponential, or liner?)

    In both studies neither had this scaled data. Because of that while we can say that there probably are differences in those factors which showed that women and men differ we don't actually know to what extent those differences are. Even the small moderate and large labels they attach are arbitrary, and there is no guarantee that there is any relation between the number value as you move from one group to the next, or even one study to the next.

    Also nurturing really doesn't equate with a lack of aggression i.e submission. You can infact be aggressive in your attempts to nurture, and in how you nurture. Sure this generally isn't the case, but it also isn't an oxymoron. and again the data just isn't clear enough to make any strong claims about how this would affect how many women are in the atheist movement based on those differences. Yes I know his argument is somewhat more nuanced then that be regardless the data is just to vague.

    If I were to guess anything it would have more to do with how women are treated poorly in many atheist groups, how many atheist are openly hostile to feminism, and how there is already a majority of men which makes it harder for women to feel comfortable, or safe. While these differences may play a role I cannot make any claim on the validity of such a claim unless better info becomes viable, and it very likely that no one else can either. They might try, but it won't actually be back by anything.

    Though the data is pretty clear that the hypothesis that women and men are externally similar in mental and personality traits does hold well based on this information, since most traits show very little difference.
    That probably sounds a bit like cherry picking conclusions from the data, but data is more constantly seemingly insignificant, and if anything these tests probably overstate the differences rather then understate them. Given the very nature of the tests explained in the studies.

    ReplyDelete
  7. Thanks for your comments, HJ and hessianwithteeth. I am eager to respond, but other obligations will prevent me from doing so probably until next week.

    ReplyDelete
  8. Ok, HJ. You think that Hyde’s suggests that the up-to-moderate effect sizes found in her meta-analysis just can’t be relevant in the real world. Let’s discuss that.

    In her 1981 paper that you referenced, Hyde correctly points out that small differences (judged by d) account for small proportions of population variance. But one cannot look at Table 1 of her 2005 meta-analysis and not be struck at the overwhelming preponderance of consistent, moderate sized gender differences (d 0.35-0.65) in even non-verbal aggression. So the question isn’t whether small gender differences can account for a significant proportion of population variance, it’s whether up-to-moderate differences can. And that question can be further refined, given your accusation of sexism, to whether up-to-moderate psychological gender differences can account for enough of a proportion of the population variance for one to reasonably believe that those up-to-moderate differences might be playing a relevant role in explaining the phenomenon in question.

    Despite Hyde’s explicit position that similarities prevail in general, she specifically points out (in a section in her paper called “The Exceptions”) that the data on aggression & sexuality stand apart from the data on other gender differences and are repeatedly up-to-moderate with respect to non-physical aggression (d up to 0.65).

    Hyde’s comment about the different costs of making Type 1 and Type II errors in medicine and psychology can be ignored because Rosenthal’s choice of a medical example was arbitrary and only for the purpose of illustrating a population difference based on values of r (which appear to be linearly related to d). The illustration clearly applies to the values obtained in Hyde’s meta-analysis. Whether those differences satisfy oncologists is entirely another matter that is irrelevant to this discussion.

    The relationship between d and r seems perfectly linear, so the math is easy to do, and for small-moderate values of d, one would expect, according to Rosenthal’s BESD, real population differences of 10-25%, just as I indicated. This is not as high as the 30% difference between populations that Rosenthal’s example provided (where d was roughly 0.7), but not as low as the 5-10% population difference that Hyde calculated for small effect sizes like d=0.1-0.2.

    I’ve translated between values of d and r, and the results I calculated are clearly in a range where one could very reasonably expect them to be playing a relevant role in the phenomenon under question.

    Accordingly, Hyde’s work is not at all at odds with what I’ve written, and it supports you making an apology to Harris.

    ReplyDelete
  9. “So even if these studies show a difference, it's not clear that difference is due to social context or biological difference. Other researchers concur, and think the idea of a sex difference in aggression in meaningless …
So both you and Harris are still at odds with the research.” – HJ Hornbeck Oct 13

    No we aren’t, because for the Nth time, Harris’ defence doesn’t rest on the notion that the differences at play are or must be (even in part) biological. This is a red herring. How many times are you going to make me scold you for poor reading, HJ?

    ReplyDelete
  10. HJ, the 1994 paper by Bjorqkvist suggests that aggression between the genders might be the same but the type of aggression may differ with females more likely to engage in indirect aggression. This will do fine to warrant Harris’ original comments, which were aimed at addressing the apparent gender difference in direct aggression between believers and non-believers.

    (The indirect aggression you are taking against Harris is a reminder that these differences are not absolute.)

    ReplyDelete
  11. In complete opposition to what you ‘ve written,HJ, Weisberg (whom you’ve erroneously assumed is a *man* - how ironic is *that*?), went out of her way to avoid conflating statistical significance with practical significance, and I even quoted her doing that.

    It simply doesn’t follow from the fact that similarity exceeds difference, that difference doesn’t exist or has no practical significance. The 30% non-overlap in agreeableness (ie. nurturing) you found with your image editor certainly seems sufficient to warrant Harris’ hypothesis and to therefore diffuse your charge of sexism.

    ReplyDelete
  12. “As far as I can tell, the similarity is far stronger than any difference, and if that's the case how can Weisberg argue for difference” HJ Hornbeck Oct 13

    This is precisely where you go wrong, and again, it’s a non-sequitur. It just doesn’t follow that because similarities exceed differences, that there is no difference, or that the difference isn’t relevant.

    Remember those side-by-side picture puzzles you looked at when you were a kid and you had to find 3 or 4 subtle differences between them? Those 2 pictures were 95% similar, but that doesn’t mean that somebody arguing for a difference between them is wrong in doing so. How can you not see these things, HJ?

    I’m not coming at this from a background assumption of difference. I’ve done nothing but concede that similarities exceed differences just as the data shows. But you repeatedly ignore the up-to-moderate differences in precisely the features that warrant Harris’ hypothesis. That’s completely self-serving, just as is your completely unwarranted accusation that he has made no effort to familiarize himself with the literature before writing a blog post about it.

    ReplyDelete
  13. Hessianwithteeth,

    Thanks for your comments about the data in question being ordinal. Your point about the magnitude of the difference being unclear is well taken. Importantly, ordinal data is what we are going to have to deal with to some extent, for how does one quantify aggression and nurturing? Despite your criticism, there is no dispute that ordinal data can certainly identify differences when they exist, and while the magnitude of the difference between ordinal categories could be overestimated (as you point out) it could also be *underestimated*.

    There are multiple studies – so many that Hyde could do a meta-analysis of meta-analyses – and her conclusion is that there is evidence of gender differences related to physical features and ability, aggression, and sexuality. Weisberg also finds consistency with respect to nurturing (agreeableness) in the literature and in her work.

    Let’s agree that the magnitude of the difference remains unclear, but that there is plenty of evidence of difference. I’m afraid that that cannot be disputed. And please note that from the fact that there is evidence of similarity in a much larger variety of other features, it doesn’t follow that the differences identified aren’t real or significant.

    ReplyDelete
  14. Hey Yorgo,

    Yes and I am not saying that the differences don't matter. What I'm saying is there has not been any solid link between these differences. So to make any claims about how they affect a women's like position in a given organization is folly. (I'm pretty sure this is a point HJ would make as well). Besides I think we do have to stop talking as though Sam Harris was talking from a point of authority which I think is clear he was not. Defending/Criticizing his hypothesis is simple not a good use of time simplely because he admits he wasn't talking from expetise and there was quite a bit of a blow up.

    From some review and because of some distant I think it safe to say that our responses towards Harris as a community and the medias portrayal of him are unfair. Perhaps an apology is in order perhaps not. Honestly I really doubt what we've said had any further effect on him. compared to what else has gone on.

    If Sam Harris has any problem it is that his most quotable moments tend to be unflattering, and easy to take out of context. However, i have not followed him long enough, and I do not have anything like the time to go back and look at his public peaking record or read all his many books. So I apologize for what ever extent I bought into all this (I was always some what skeptical, but now I'm even more skeptical).

    I think Sam Harris has been painted into a bad guy by numerous people, including myself. I also think he's said some very silly and problematic things. I even think more recently he unintentionally said some sexist things out of ignorance of the topic, and lacking the time to think out his answer in greater detail. Do the gender differences we discussed have anything to do with all of this. Kind of? Maybe? Who the hell knows? I know I don't. I do know I can't find any real links other then the ones we wrote in, so meh. This has all be rather unfortunate, and the forces and players involved here are way more complex then and single person or problem. (and I'm quite certain that gender difference are not the primary cause,,if cause at all)

    I don't mean to dismiss any of it, but I honestly never should have gotten involved as this rabbit hole goes deeper then I'm willing to venture, and to do it justice is not something I'm going to do. Finally apologies should only be given when earnest, so if HJ doesn't think he should apologize then he shouldn't. Perhaps he can be convinced, perhaps not, but really the issues surrounding Harris are looking like a real enigma and would take some major digging to start sorting them out.

    Tyler

    ReplyDelete
  15. Thanks for your comments, Tyler. Does this mean that you are hessianwithteeth?

    "I am not saying that the differences don't matter. What I'm saying is there has not been any solid link between these differences. So to make any claims about how they affect a women's like position in a given organization is folly. " - Tyler

    Can you explain that to me better. I don't understand. Plus, Harris was never trying to explain the disposition of any single woman in particular. So you’ll have to explain how your comment might influence very large groups.

    "[Sam Harris] admits he wasn't talking from expetise" - Tyler

    Can you tell show where he admitted that?

    "Do the gender differences we discussed have anything to do with all of this. Kind of? Maybe? Who the hell knows? I know I don't. " - Tyler

    It isn't an area with clear answers, but that's how it is with so many areas of interest and concern. Reasonable people can and should be free to recognize what they consider to be reasonable and/or tentative conclusions. HJ should be free to do so, but he should recognize that people like Harris should be free to do so also, provided that they aren't talking out of their asses, and as long as they aren’t, they shouldn’t have to fear accusations of sexism at every turn.

    Harris has an excellent track record of not talking out of his ass. People frequently think that he’s done so, but he is open to challenges, publishes interactions with his most formidable opponents at his website, and always reveals being much more well informed than anybody imagined. Have you seen the man? He’s calm as a cucumber and always has been. He speaks slowly, carefully, and thoughtfully. He tries hard not to mischaracterize the views of others. His writing is remarkably well referenced. He’s just too smart to make comments about such a heated topic as gender differences without having any idea what he’s talking about. He’s also inclined to never lie, and when asked that question by Boorstein in DC, I think he just gave his honest answer, and one that he would not have offered had it not been at least somewhat informed. Let’s face it. One needn’t look that hard to find very reasonable evidence of relevant psychological gender differences. I sure didn’t have to, and I didn’t have a guy like Steven Pinker at my disposal. He and Harris are friends.

    Speaking of Pinker, you might enjoy this debate on a very similar question from 2005, when the question was at a new peak of interest: http://edge.org/events/the-science-of-gender-and-sciencepinker-vs-spelkea-debate

    "Finally apologies should only be given when earnest, so if HJ doesn't think he should apologize then he shouldn't" - Tyler

    That's a good point, which I like.

    Thanks again, Tyler.

    ReplyDelete
    Replies
    1. I am in fact the Withteeth of HessianWithteeth (it's a two person blog)

      Well looking back it seem I entirely miss read the point where he said he hasn't spent 5 minute thinking about it. He was talking about how he hasn't thought about changing his tone. That was my bad.

      Rather then answering all the points you brought up which are fair enough. I think the discourse may be improved if I make some further clarifications, as I don't think we have gotten to the root of our disagreement. I'm not worried so much that there exist some gender differences. You saying wait, but that's totally what you said before. Yes and no, what I'm really trying to get at is yes there are some obvious gender differences, but the problem is: People conflate gender and biological sex.

      This should end up answering at least answer one of your questions above. It isn't that I'm opposed to saying there are *some* differences between the genders, but given the scientific observations I've been exposed too. We are in no position to separate nature vs nurture. We can't make claims the gender women are such and such a way, because they have the primary and secondary sex characteristics of a women. Because not all women have all those traits, and that's before your start looking at trans* or gender non-conforming people, who tend to show more variety of traits.

      To restate my point again, I feel the root of many confusions i
      here is the notion that both HJ and I are pushing back against the false notion that gender = biological sex which lead people to assume that physiological difference have to have some physiological basis. Even if gender = biological sex that wouldn't necessarily b the case, but sine they don't even line up very we in real life we can be very sure that things are not likely to be that simple.

      This really heavily gets int gender politics and culture stuff, but this is why I want everyone to be really hesitant to jump into drawing conclusions about physiological connections to gender differences, largely because it fall too neatly into the narrative instilled into us by our culture, but also excludes an unquantified, but not insignificant part of the population, part of the population which do not fit nicely into those gender expectations. As well as giving undue credence to numerous sexist ideas, which do not seem to be accurate, or have been shown to be largely false, but are till passed around as fact. Such as women being worse are science, or men are worse at writing.

      Tyler

      Delete
  16. Hi HJ,

    I’m going to try to summarize our interaction here. You claimed that three of Sam Harris’ ideas are sexist. Not probably sexist, or sexist if certain conditions are true, but clearly, flat-out sexist. We agree that this amounts to charging that the following 3 claims are clearly, flat-out false or unwarranted:

    1. Harris claimed that he tends to respect women more than men.
    2. Harris claimed to be confident that psychological gender differences play a role in the Fortune 500 phenomenon
    3. Harris claimed that psychological gender differences might explain a greater male fondness for a perceived style of religion bashing. Given the context, the proposed gender differences would make sense if they were in the areas of aggression and nurturing.

    To have succeeded at #1, you had to know why Harris respects women more than men and argue that that reason represents insufficient warrant for doing so. But you can’t read Harris’ mind and you haven’t asked him. So you had to argue that there *can be no such warrant*.

    To have succeeded at #2, you had to show that the only reasonable position is to believe that no psychological gender difference could possibly play any role in the Fortune 500 gender phenomenon.

    To have succeeded at #3, you had to show that the only reasonable position is to believe that psychological gender differences related to aggression or nurturing couldn’t possibly play a relevant role in the active atheism gender phenomenon.

    That’s three very, very tall orders. Please pause and consider for even just a moment what’s actually required to support these three charges of sexism as opposed to how easy it is for you to toss the accusation around.

    I had a *much* smaller burden than you did. In suggesting that you owe Harris an apology, all I had to do was show that your accusations were unwarranted. That is, all I had to do was point out your burden and see if you met it or not. So when you wrote that my refutation assumes Harris knows more about the literature than is in evidence, and that therefore, your charges of sexism stand, you were making the mistake of confusing my burden of proof with yours. You’re the one claiming that Harris’s statements are sexist. I don’t have to prove that Harris was familiar with the evidence of psychological gender differences. You have to prove that he wasn’t.

    You have, on all 3 counts, failed to meet your burden (see below).

    ReplyDelete
  17. Continued from above.

    Regarding #1, you have left that immense burden of proof completely unmet. You didn’t even try to argue that Harris could have no warrant for respecting women more than men. You just said that that was “obvious”. Nevertheless, I provided four examples of warranted reasons for Harris to do so. You tried to challenge one of them by repeatedly demeaning the challenges that millions of everyday women rise to daily – no hourly - in a male-dominated world. If you want to make the case that the majority of hard working women are “queen bees” or actually trying to maintain the male-dominated status quo because they are influenced more by SJT than ego or group motives, then you have to do more than just link to descriptions of what those terms mean. You have another unmet burden of proof here, in addition to having to argue that Harris could not possibly be warranted in respecting women more than men, which is not at all “obvious”.

    See below.

    ReplyDelete
  18. Continued from above.

    Regarding #2, You first argued that there was a logical contradiction in what Harris said about gender differences contributing to the Fortune 500 phenomenon, though there wasn’t one. You also repeatedly asked how Harris could be confident that biological reasons explain the Fortune 500 phenomenon, even though Harris never indicated that biology had to be playing any role whatsoever. As for the evidence that gender differences exist and could play a role, the very scientists that you quote (like Janet Hyde, PhD, perhaps the foremost researcher in the field) recognize that relevant gender differences, including moderate to large ones, do seem to exist, even if they are outweighed by other similarities. You kept referencing that similarities outweigh differences, but from that fact (with which I readily agree), it does not follow that differences capable of playing a role in the Fortune 500 phenomenon do not exist. The evidence shows that some relevant differences exist, and that they are sizeable enough to play a role.


    Since I showed that evidential warrant for Harris’ comment exists, you began emphasizing that Harris’ comment couldn’t have been warranted by that evidence because he couldn’t have known about it. Once again, your accusation requires mind reading. Never mind that Harris could have just had a single conversation with his friend, Dr. Steven Pinker, who debated this topic at the height of its controversy at Harvard in 2005, to be aware of the copious evidence of gender differences in some relevant areas. Never mind that Harris has an excellent track record of being very well informed about matters that he writes about. Never mind that he is a scientist and that he has advocated strongly for an evidence-based epistemology from his first book, The End of Faith. Never mind that a simple Google search readily brings up Hyde’s research. Never mind that he uses the very same language to describe his position as researchers in the field use to report their findings. Apparently, you still think that Harris just happened, by accident, to be right about the evidence. Make no mistake: you have come nowhere near establishing that Harris was oblivious to the evidence before writing a blog entry on the subject despite asserting that the idea “is untrue”. This line of argumentation is ad hoc (self-serving) and it is by far your least convincing.

    See below.

    ReplyDelete
  19. Continued from above.

    For #3 to succeed, you had to do the same thing required for charge #2 to succeed, specifically related to psychological gender differences in aggression and nurturing. The same problems with your defense of charge #2 apply here, as well. In addition, you tried to suggest that small differences in aggression and nurturing are too small to be relevant, while ignoring the fact that the differences identified are consistently moderate or up-to-moderate.


    You have failed to meet your burden on all 3 accusations of sexism, HJ. You should either meet the massive burden of proof that comes along with all three smears, or you should regret those accusations, and begin moving forward by humbly apologizing.

    ReplyDelete
  20. Whew, I have some catching up to do.

    In her 1981 paper that you referenced, Hyde correctly points out that small differences (judged by d) account for small proportions of population variance. But one cannot look at Table 1 of her 2005 meta-analysis and not be struck at the overwhelming preponderance of consistent, moderate sized gender differences (d 0.35-0.65) in even non-verbal aggression.

    You didn't read very closely earlier. Try again:

    The difference in means is only about one fourth to one half of a standard deviation ... . Generally, it seems that gender differences in verbal ability are smaller and gender differences in spatial ability are larger, but even in the latter case, gender differences account for less than 5% of the population variance.

    Don't see it? A half a standard deviation difference between the means is a d of 0.5. In other words, those "moderate-sized" gender differences account for about 5% of the population deviance. 80% of men will be able pair up with a woman with the same aggression level, and presented with a random aggression score you'll guess the gender correctly 60% of the time, at best.

    And yet all you see is difference.

    The relationship between d and r seems perfectly linear, so the math is easy to do, and for small-moderate values of d, one would expect, according to Rosenthal’s BESD, real population differences of 10-25%, just as I indicated.

    So 62% of the treatment group gets better compared to 38% of the control? The BESD is an inappropriate metric here, as there's no "treatment" involved.

    ReplyDelete
  21. HJ, the 1994 paper by Bjorqkvist suggests that aggression between the genders might be the same but the type of aggression may differ with females more likely to engage in indirect aggression.

    She also points out that researchers used to think there was a huge difference in overall aggression. They moved the goalposts when they found out it wasn't, however.

    Given how badly they were wrong the first time around, and how many former "obvious" differences have turned out to be negligible, we've got good reason to be skeptical of the new differences they see. And that's before we even consider what proportion of those differences are due to socialization vs. biology.

    ReplyDelete
  22. To have succeeded at #1, you had to know why Harris respects women more than men and argue that that reason represents insufficient warrant for doing so.

    No, have you forgotten my definition? That's what made it a sexist act: it was an action contingent on at least one false or unwarranted assumption related to sex or gender, where "action" includes both speech and inaction when action was possible and sufficient. I can satisfy that definition by showing Harris was false to claim that women are inherently more worthy of respect, easily: when equality is achieved, are women still worthy of more respect? No? Then there is nothing intrinsic with woman-hood that grants respect. Harris thus made a false claim, and that statement is sexist.

    It's sad that I find myself explaining the obvious to you.

    To have succeeded at #2, you had to show that the only reasonable position is to believe that no psychological gender difference could possibly play any role in the Fortune 500 gender phenomenon.

    No, I merely have to show Harris' confidence is unwarranted. I did so: we have substantial evidence for gender difference due to sociological factors, and at best very little evidence for biological difference, plus the differences are within the range of what can be explained by social factors. Ergo we are not warranted to claim biology must play a role, and thus Harris is unwarranted to say so, making that a sexist statement.

    To have succeeded at #3, you had to show that the only reasonable position is to believe that psychological gender differences related to aggression or nurturing couldn’t possibly play a relevant role in the active atheism gender phenomenon.

    No, as you yourself point out Harris claimed that psychological gender differences might explain a greater male fondness for a perceived style of religion bashing. And yet neither of you have shown this is a greater male fondness. Even the much-vaunted gender difference in belief only shows there's a difference in belief, you simply assume that difference must be due to a difference in tone. Here's an alternate theory: churches provide better day care and support networks, and as women are disproportionately saddled with child care they have material reasons for staying within the faith.

    Now we have two explanations for the same thing. And given the historic failures in demonstrating gender differences, no matter the source, the priors side with my theory as the most likely. Again, Harris is proposing an explanation without warrant, and again he is being sexist.

    Never mind that Harris could have just had a single conversation with his friend, Dr. Steven Pinker

    And yet you have the best evidence for difference in front of you, and it instead shows the genders are more similar than different. So what evidence can convince Pinker otherwise, assuming he's aware of it, and assuming he and Harris discussed it at some point.

    Never mind that Harris has an excellent track record of being very well informed about matters that he writes about.

    He wasn't writing about gender differences, he made an offhand remark during an interview. You have no evidence Harris knows of the research. There is no reason to think Harris is aware of the research. Harris demonstrates no understanding of the research, otherwise he would have invoked it during his defense against sexism.

    All you have are assumptions and blind stubbornness at this point. And you consider yourself a skeptic?

    ReplyDelete
  23. Thanks for your comments, HJ. I don’t see anything new in your comments from Oct 29 to which I haven’t already responded. It’s been eye-opening for me, so thanks again.

    ReplyDelete
  24. I've been following this from afar and laying low, but I'd like to add my 2 cents after considering both sides carefully.

    HJ, you're all mixed up and tripping over yourself to stubbornly avoid admitting that the evidence clearly shows psychological gender differences that might explain Sam’s position. You do this by repeatedly claiming that since the similarities are larger than the differences, it's inappropriate to point out the differences that exist. The similarities don’t make the differences go away. You also keep coming back to arguing against the claim that the differences have to be biological despite Yorgo repeatedly pointing out that Harris never even said that. If your accusation of sexism is because you think that Harris said that biological, innate differences must be at play, you're just making a straw-man and Yorgo is right to blow you off.

    The BESD isn't inappropriate at all. The treatment is just an analogy for a certain exposure. Consider the exposure the treatment in the medical example, or exposure to a particular gender in the other example.

    Lastly, whether you realize it or not, you've admitted that Sam's greater respect for women can be justified by their PRESENT inequality, which is what Yorgo has been arguing all along.

    This has been a good exercise for me in weighing evidence and carefully considering arguments. Keep it up, Yorgo. Enjoy your blog.

    ReplyDelete
  25. Anonymous,

    Thanks for your interest in my blog and your kind words.

    I agree with everything you wrote and I'd just like to add one thing to your last point. Harris never said that women are "inherently" or "intrinsically" worthy of more respect than men are. He simply said that he tends to respect women more than men, and we seem to agree that that statement can be justified by the extra mile to which they have to go to succeed in a male dominated world. As long as there is but one context in which Harris’ comment can be warranted (and I provided four), HJ’s accusation of sexism dissolves. You see, HJ doesn’t have a clue how Harris warrants that statement, so to say that it’s sexist, HJ has to show that it cannot possibly be warranted, and he hasn’t even tried to do that. Most of Harris’ critics distort what he’s actually said into what they think he’s said. This all boils down to how easy it is to toss accusations of sexism around, especially when one is frivolous with one’s interpretations of other people’s ideas.


    HJ would probably respond that we don’t know how Harris warrants his statement either, so we aren’t in a position to say that his statement isn’t sexist, but this would be to mistake who has the burden of proof. HJ’s the one accusing Harris of sexism. He has the burden to show that the statement can’t be warranted. In saying that HJ should retract the accusation and apologize, all we have to do is show that HJ’s failed to meet his burden, which we have done.

    The same goes for HJ’s condescending insistence that I am assuming that Harris wasn’t familiar with the evidence on psychological gender differences in general or in aggression and nurturing in particular. HJ can puff that up all that he wants, but he’s again guilty of shifting the burden of proof. He’s the one making the claim, and he has to back it up, or retract it. He’s the one who has to show that Harris wasn’t aware of the evidence, and he hasn’t even come close to doing that, either.

    Thanks again.

    ReplyDelete
  26. I don’t see anything new in your comments from Oct 29 to which I haven’t already responded.

    I agree, we're not speaking to each other here. You keep insisting I've done nothing to show Harris is sexist, while ignoring my arguments to the contrary, and dismiss my arguments for similarity without response. There's nothing new for me to respond to, hence I don't think further response is warranted.

    ReplyDelete