Every semester, I emphasize that no one can fairly assess the debate between Scientific Creationists (or other creationists) and evolutionary theorists by relying on mass media coverage of that debate. As I'm sure you know, the quality of science reporting in the media varies enormously. During March-April, 2000, there was some press coverage of a 500-word report in Nature (one of the world's leading professional science journals) that helps to illustrate the latter point powerfully.
"Finger-length ratios and sexual orientation" TERRANCE J. WILLIAMS, MICHELLE E. PEPITONE, SCOTT E. CHRISTENSEN, BRADLEY M. COOKE, ANDREW D. HUBERMAN, NICHOLAS J. BREEDLOVE, TESSA J. BREEDLOVE, CYNTHIA L. JORDAN & S. MARC BREEDLOVE 30 March 2000 Nature 404, 455 - 456 (2000) © Macmillan Publishers Ltd.
ABSTRACT: Animal models have indicated that androgenic steroids acting before birth might influence the sexual orientation of adult humans. Here we examine the androgen-sensitive pattern of finger lengths, and find evidence that homosexual women are exposed to more prenatal androgen than heterosexual women are; also, men with more than one older brother, who are more likely than first-born males to be homosexual in adulthood, are exposed to more prenatal androgen than eldest sons. Prenatal androgens may therefore influence adult human sexual orientation in both sexes, and a mother's body appears to 'remember' previously carried sons, altering the fetal development of subsequent sons and increasing the likelihood of homosexuality in adulthood.
(Incorrect!) Description by editors of Nature:
"Is human sexual orientation influenced by the exposure to 'androgens' (male hormones) in the womb? In a Brief Communication this week, S. Marc Breedlove and colleagues, of the University of California, Berkeley, describe how, on the basis of their measurements of patterns in finger length, they believe that it could be. One difference between men and women is that women tend to have a smaller difference between the lengths of their second and fourth fingers - men have relatively shorter index fingers. This ratio, then, is a measure of fetal androgen exposure. From their finger-length patterns, lesbian women seem to have been exposed to more androgen in the womb. Oddly, the results for gay men are similar, suggesting that in general they too have been exposed to more fetal androgen than straight men. This pattern is reinforced if you subdivide the men by a well-established, but little known, factor predisposing men to homosexuality - having several older brothers. Broadly speaking, the more older brothers a man has, the more fetal androgens he was exposed to, and the more likely he is to be gay."
The San Francisco Chronicle MARCH 30, 2000
"Finger Length Points to Sexual Orientation; Anatomy quirk called possible biological clue"
by Carl T. Hall, Chronicle Science Writer
In a study sure to provoke a lot of self-examination, a University of California at Berkeley team has found that differences in the lengths of one's fingers may yield clues about sexual orientation.
Lesbians on average turned out to have more "masculine" hands than heterosexual women -- with the index finger significantly shorter than the ring finger. There was no such difference in the hands of gay and straight men, however.
A brief summary of the study, led by Berkeley psychologist S. Marc Breedlove and undergraduates Terrance J. Williams and Michelle E. Pepitone, appeared yesterday in the science journal Nature.
Although the results may seem a bit puzzling, experts said it's only the latest of several studies that suggest hormones in the womb have a powerful affect on sexual orientation and behavior throughout life. Previous research has found certain differences in brain structures as well as other anatomical distinctions between gays and straights.
The evidence is mixed, however, and nobody claimed the latest findings will settle the debate over how sexual orientation is shaped by biology versus environmental factors. Nor can finger lengths be used as a reliable guide to very much of anything, although it's long been known that men tend to have longer fingers than women. "The differences are subtle," said Raymond Blanchard, a pioneer in gender and sexuality studies at the University of Toronto's Center for Addiction and Mental Health. "There's no way anybody could use this to screen a date."
The latest study was based on the fact, which was already apparent to scientists, that males tend to have a different pattern of finger lengths than females. In men, the index finger tends to be a bit shorter than the ring finger. These same two fingers on women are typically about the same length. The male-female difference tends to be more pronounced on the right hand, for reasons experts cannot explain, regardless of whether the person is left- or right-handed. These gender differences can be picked up in very young children, strongly suggesting they arise because of prenatal influences.
Based on a host of animal studies and other research, it's almost certainly a product of androgen levels -- testosterone and other male hormones -- in the womb.
Scientists now suspect that finger lengths may be a marker of how much a fetus was exposed to these masculinizing sex hormones. The same prenatal influences may affect behavior and sexual identity throughout life.
Breedlove and his Berkeley colleagues appear to be the first academic researchers to bother checking whether there is any appreciable difference in finger-length ratios linked to sexual orientation in adults.
They set up booths last summer at a gay pride event in Oakland, the annual Solano Stroll in Berkeley and at the Castro Street Fair in San Francisco. Fairgoers were offered a free $1 lottery ticket in exchange for having their hands photocopied and filling out a detailed questionnaire. After examining the hands of 720 adults, Breedlove's team discovered that the average finger-length pattern for lesbians closely resembled that of males. That is, the index fingers on lesbians who were surveyed was judged to be significantly shorter than the ring finger. Researchers concluded that "at least some homosexual women were exposed to greater levels of fetal androgen than heterosexual women." Hand patterns between gay men and heterosexual men revealed no such difference.
But the surveys yielded other intriguing clues. It was already known that men with more than one older brother are slightly more apt to grow up to be homosexual than are firstborn males. The new study suggested that males with older brothers are subjected to higher levels of androgens in the womb than are eldest sons.
As a result, the study says, men with more than one older brother have a significantly more masculine right hand than men without older brothers. That is, the right index finger is more often shorter than the ring finger on the same hand. There was no such effect in men with older sisters.
Breedlove said it's a complete mystery as to how a mother's body could "remember" how many male children she had borne, where this signal was kept and how it could influence hormone levels of a later-born child. He also cautioned against making too much of the differences in the hands of gay and straight women.
© 2000 The Chronicle Publishing Co.
In correspondence, Professor Breedlove described his (relatively) positive experience being interviewed by Diane Sawyer for "Good Morning America":
I just got home from being "interviewed" for Good Morning America and, as you might expect, I must immediately circulate a correction to [my fellow researchers]!
First let me assure you that Ms. Sawyer did most of the talking. My professor's disease made me, despite my intense nervousness, want to talk more. So I kept trying to say things when I thought she'd stopped talking, only to find she had more to say. She's not easy to interrupt!
Before I was "on" a producer's voice in my ear asked me if it was fair to say that these most recent findings were part of a growing body of evidence that biological factors influence sexual orientation. I said, "well, specifically prenatal events, but yes, that's fair".
Then seconds later Ms. Sawyer says, on air, that this is the latest in a body of evidence showing that only biological factors determine sexual orientation (I haven't seen the show yet, so I don't recall the exact wording). I just want to assure you that Ms. Sawyer's statement does not represent some [radical change in view] for me. I am still [confident that environment in the womb and after birth are signficant factors in development of orientation], and I've tried to tell everyone I've spoken to about this that I think our findings indicate there is some effect of early androgens, not that they are the only factor or even a major factor.
She then proceeds to ask me questions where each question is far longer than the answer I can give, so I had no chance to correct.
It was quite an experience to be ready to say all sorts of things, to be very nervous and then find I have almost no opportunity to say anything! Why be nervous if you aren't actually going to say anything?
She pointed out that, there in the studio, they'd looked at their hands and it didn't work. <sigh> Hmm...., I think, "shall I give the lecture about sample size and statistical power??" but of course it gave me a chance to say that you can't classify people (read children or fetuses) accurately with this method.
<stuff deleted>
S. Marc Breedlove, Psychology Department, University of California Berkeley, CA 94720-1650 http://ls.berkeley.edu/dept/psychology/breedbio.htm
NPR ran a brief (3 mins. 39 secs.) report on the reported research. Here's how their report was described on the relevant web page:
http://search.npr.org/cf/cmn/cmnpd01fm.cfm?PrgDate=03%2F30%2F2000&PrgID=3
Finger-Length Ratios? ... -- NPR's Richard Harris reports that psychology professor Marc Breedlove at UC Berkeley xeroxed the hands of 720 people in order to compare the length difference between the index and the ring fingers. On most women, there is little difference, but on men, the ring finger is longer. Breedlove's survey found that lesbians have more of a male finger-length ratio on their left hands. His findings also suggest that gay men with one older brother were more likely to have shorter index fingers. Breedlove's study isn't being received warmly. (3:39) [RealPlayer file at [http://www.npr.org/ramfiles/me/20000330.me.12.rmm]
You can listen to the report using RealPlayer by downloading from the URL just given. (It'll probably work better over a 'fast' connection, e.g., on campus or via cable or DSL connection, though it should work even over a 28.8Kpbs modem connection. I'm not sure for how long the file will be available there.)
For the reasons he gives below, Professor Breedlove found the NPR report seriously flawed. In giving me permission to quote the material below, Professor Breedlove asked me to say that he later regretted not waiting to cool down a bit before writing to NPR. Given that the NPR report mixes readily avoidable inaccuracies and irrelevant personal slurs, his reaction seems, however, to be quite reasonable.
From: [Professor S. Marc Breedlove]
Date: Fri, 31 Mar 2000 13:00:40 -0800
To: [Richard Harris et al, National Public Radio]
Subject: NPR's hatchet [job]
Dear Mr. Harris,
I did not catch the broadcast of your story about our paper in Nature this week, but when two different colleagues, one in Chicago and one in Texas, independently described it to me as a "hatchet job", I decided I'd better check it out, so I just listened to it on the NPR web site:
http://search.npr.org/cf/cmn/cmnpd01fm.cfm?PrgDate=03%2F30%2F2000&PrgID=3
I am appalled at such sloppy science journalism. You substitute rumor, opinion and insinuation for getting the story straight.
Firstly, your "segment description" text on the web page has several facts wrong, as either reading the paper Nature gave you or attending to my phone conversation should have told you:
1) the difference between lesbians and heterosexual women was in the right hand, not the left.
2) the birth order effects on men were seen across all men, not just in the gay men. The broadcast itself has additional errors of fact, but is also overlaid with innuendo about the motivations of people (me) and institutions (Nature) without a shred of evidence to support the idea.
First your many errors of fact in the broadcast:
1) You say that I chose "arbitrarily" to measure fingers from the crease at the palm to the tip. The choice was not at all arbitrary. I felt we must first replicate John Manning's findings on finger length and this is how his published reports described making the measurements. In lovely replication of his results, we found a sex difference in exactly the direction he reported, and we replicated his finding that the sex difference is more robust on the right hand than the left. (We also replicated both these findings in a pilot study we did with undergrads before going to street fairs.) It would have been "arbitrary" to begin measuring fingers anywhere other than where Manning measured them.
2) You say the report has not been received "warmly". To date, I know of only one journalist and one scientist who has judged the report frivolous, uninteresting or statistically unsound, and that's you and your single scientific consultant. Feel free to correct me if you know of others. Certainly no one described the paper as Jerry Springer-like, cocktail party chatter, "titillating" or that it should not be taken seriously except you.
3) You say of the birth order effect that a similar trend in heterosexual men was ignored, but the paper makes the point that we see this effect in all men, and when I spoke to you the day before we taped, I emphasized to you that I feel sure the birth order effect in men applies to both orientations. You accuse me of ignoring an issue I discuss in the text of our paper and display on one of the graphs.
4) You (or your consultant Steven Austad of the University of Idaho) say that there is only one other study of finger lengths. In fact the paper references several and, had you actually contacted the scientist who first described the sex difference, John Manning, he would have told you about several more in press. Why you consulted a biologist in Idaho who has never studied or published about any of the relevant topics rather than the world's expert is beyond me, but certainly doesn't speak well of your competence as a science reporter.
5) You (or your consultant) say there's very little consistency in the measures between labs, focusing on the different absolute ratios seen. Of course, Dr. Manning and I may measure fingers differently (I've never met or spoken to the man, just read his papers), so I don't think anyone would be surprised if our absolute values differ slightly despite my attempts to replicate his methods. But what makes this statement untrue is that you ignore our replication of Manning's reports on: the presence of a sex difference, the direction of that sex difference, and the stronger sex difference on the right hand than the left.
6) The importance of our findings of a birth order effect on finger ratios is that the same birth order effects were found (by Dr. Ray Blanchard and colleagues) on male sexual orientation, as our paper clearly indicates in black and white. You suggest that we "factored" out birth order as a desperate attempt to find differences, but anyone who knows the literature of human sexual orientation would know that this is a well-established factor influencing orientation in men, but not women. By failing to provide this background, you falsely indicate we were "data trolling", an idea you reinforce by saying that I "measured and measured and measured". How very droll and sophisticated of you.
7) You suggest that "measurement error" could easily "skew" a "faint" difference. Fortunately, scientists discovered earlier in this century a method that makes sure that we can judge whether even a "faint" difference is likely to be real or erroneous.
We call it statistics.
My lab used and reported these statistical analyses and neither you nor your expert challenged whether we used the proper statistical tests or whether we performed them properly. As indeed you should not, since they were correctly applied and correctly performed. So no, Mr. Harris and Dr. Austad, measurement error cannot "skew" results
Now for your errors of judgment as a science reporter:
1) For scientists to comment on a study of finger ratio differences between people of different sexual orientations and the possible implication of hormones, I would expect a journalist to choose either a) someone who studies fingers, b) someone who studies sexual behavior, preferably in humans, or c) someone who studies the influence of hormones on behavior. But you choose Dr. Steven Austad of the University of Idaho. I'm reasonably well acquainted with the field of Hormones and Behavior, as I'm on the editorial board of that journal and a co-author of one of the two textbooks on that subject, so I knew he didn't work in that field. I keep up with the field of sexuality research and I'd never heard of him. I've read the entire finger length literature (that's easy because it's quite new), but hadn't run into his name. Consultation of Medline indicates that Dr. Austad's specialty is aging research, which has no relationship to the methods, issues or disciplines of our report. I suppose I should be glad you didn't choose someone who studies rocket science.
2) The voice over before you come on indicates that our report is "outlandish". This is hardly a neutral, dispassionate position for a journalist. In whose opinion is it outlandish? yours and Dr. Austad's alone? Are these informed opinions? And you imply that Nature published it because it was outlandish and they found it "irresistible", but never provide any evidence apart from Dr. Austad's opinion, to support the idea.
3) Austad describes the results as only useful for "cocktail party chatter", an ad hominem, wholly subjective judgment (from someone unfamiliar with any of the relevant fields of study).
4) Your expert asserts, without any evidence, so this is his personal opinion alone, that Nature accepted the paper because of its "Jerry Springer" like appeal and the "titillation" of the paper. Again, you present to a national audience insinuations that our paper was not legitimate, and that Nature was not meeting its responsibilities, on the basis of one man's personal opinion. And again, he doesn't do research in any of the relevant fields.
5) You say that the paper was peer reviewed but, "of course Nature hand-picked the reviewers and was free to ignore the advice it solicited". All right, Mr. Harris, who did you want to pick the reviewers if not Nature? me? you? Dr. Austad? And have you any evidence that they ignored the reviewer's advice? Did you try to gather any evidence? I have the reviews and they both are very positive, one explicitly saying it should be published and the other offering only the reservation that it would be difficult to shorten to the size of a Brief Communication. Both reviewers suggested quite explicit revisions, all of which Nature asked me to make, and all of which I did make. So you now know that Nature did not ignore their reviewers, right? You could have learned that before broadcasting rumors by asking me about it. You could have seen the reviews for all I cared. You didn't ask, you just smeared my reputation and Nature's without even bothering to investigate. Gosh, reporting is a lot easier than science.
6) You smugly close your report by saying that "Breedlove defends his report. He's likely to be doing a lot of that". Again, yours is the only report I've seen that says anything I disagree with, and that includes many statements of caution from scientists who actually know what they're talking about. This email is the only defense I've had to make and it's clear that I'm having to defend myself from you because you made numerous errors and behaved unprofessionally.
Mr. Harris, you and NPR owe me a public apology.
I challenge you to devote more than 3 minutes to letting me tell my side of the story in response to your slimey innuendo.
Perhaps the greatest irony is that you accuse me of engaging in Jerry Springer-like titillation, when that is exactly the best description of your conduct. Was it fun to smear someone unfairly?
Marc Breedlove
I'm sending a copy of this e-mail to Mr. Harris and colleagues at NPR to see if they wish to make any comments. I'll forward to you anything relevant that I receive.
Given all of the above, it would, of course, be pretty silly to engage in frenzied finger-measuring (or sibling census-taking, etc.); it'd even make for a very dull party-game.
By the way, if you are interested in current research on the origin and nature of sexual orientation and gender, I'd recommend the book by Stein and the books by Fausto-Sterling for starters, all of which you should find quite accessible. The first is a pellucid gem of "applied philosophy." The two by biologist Anne Fausto-Sterling offer very helpful analyses some of the ways in which social prejudices have influenced biologically-based research on orientation and gender. All three books have extensive bibliographies.
Anne Fausto-Sterling, Sexing the Body: Gender Politics and the Construction of Sexuality (Basic Books, 2000)
Anne Fausto-Sterling, Myths of Gender: Biological Theories about Women and Men 2nd ed.(BasicBooks, 1992)
In Chapter 1, you were acquainted with Professor Breedlove's misgivings about Mr. Richard Harris's National Public Radio report concerning research published in Nature.
Mr. Harris kindly forwarded his reply to Professor Breedlove's e-letter, and it appears immediately below. Following it is a copy of Professor Breedlove's rebuttal, also kindly forwarded to me. I hope that you will find both items interesting. They do illustrate an important aspect of the connections between scientific research and science journalism and should encourage a healthy skepticism.
Mr. Harris's note to me and the appended reply to Professor Breedlove:
Dear Dr. Austin,
Since Marc Breedlove sent you his critique, I thought I'd share with you my reply. I think it could generate an interesting discussion in your class about the methodology of science -- and the media. In case you are wondering, I do not have an axe to grind about this topic. In fact, I reported favorably on a similar study concerning sexuality and otoacoustic emissions, by Dennis McFadden.
Sincerely,
Richard Harris, Science Correspondent, National Public Radio, 635 Massachusetts Ave., NW, Washington DC 20001; voice: 202 414-2786; efax NOTE AREA CODE: 707 516-0368
Dear Dr. Breedlove,
This is in response to the email you sent me and others about my story last week regarding your work. As a result of my reporting, I concluded that your study was fatally flawed and that Nature should have caught that flaw with a competent peer review. I stand by those conclusions.
Let me address the particular issues you raise.
First, I do not write the material that ends up on our website. I have asked the web editor to correct the minor errors you brought to my attention.
As for the items in my story:
1) You made an arbitrary choice to select a line on the skin as a surrogate for finger length. A person's fingers do not start at a wrinkle line. You did not present any data showing that is a valid substitute for bone length or other meaningful measure of a finger. Likewise, you chose a questionable surrogate for the end of a person's finger -- when flesh lifts off a flat surface. According to the scientists I interviewed, this is a critical flaw in your paper and invalidates your results. More about that in a minute.
2) The paper was extremely poorly received by the two scientists I showed it to. In addition to Dr. Austad, I sent a copy for review to Dr. John Bailar, at the University of Chicago. For many years he was the New England Journal of Medicine's lead reviewer for biostatistics. He told me the statements you made in the paper were contradicted by the data you presented. Few other journalists who covered you paper bothered to seek careful comment. The Philadelphia Inquirer was one of the few papers do so. That report quotes Dr. John Manning saying that you should have corrected for ethnicity, since he has found major differences in finger length based on ethnic background. The Inquirer also quotes Dr. Robert Trivers saying the evidence for "hyper masculinity" is "very weak."
3) Your data on gay versus straight men showed no significant differences. Both Austad and Bailar found your statements about this comparison in the paper to be misleading.
4) Austad and I found only one other paper in the literature relevant to finger length and your hypothesis -- the Manning paper you cited. I didn't see a reference in your paper to a similar study. The reason Austad raised that point is because your results are markedly inconsistent with Manning's. I called Steven Austad because he is a field biologist who has studied sex ratios in animals (Austad & Sunquist. Sex ratio manipulation in the common opossum. Nature 324:58-60, 1986). I have spoken to him over the years and found him to be a thoughtful and credible scientist. He took the care to read Dr. Manning's previous work on this subject and to experiment with his own hand on the issue of measurement error. Austad raised a serious scientific issue about your paper. I pressed you on that subject during our interview and you could not rebut his fundamental criticism.
5) The issue of replication is important because it shows that there is a very large error in the way fingers are measured. In Dr. Austad's view, this overshadows the consistency you report with other findings (findings that have similar methodological problems).
6) As stated above, several researchers have found your analysis of the men to be weak. My statement about "measuring and measuring and measuring" was not intended to be derisive, and I doubt our listeners interpreted it that way. The point was simply that you gathered a lot of data.
7) This gets to the fatal flaw of your paper. As we discussed during the interview, there are two kinds of errors in scientific research. First is measurement error, that is uncertainty about each data point in the research. The second is statistical error, which deals with whether the effect you have measured is simply due to chance. The scientists I have consulted for this story say that if your measurement error is the same size as your effect, you can't correct for that with simple statistics. They say you need to use a "total error model" which factors in the measurement error in your data points. Your effect is a 1 percent difference in finger ratios, which translates to a fraction of a mm in finger length. I asked you twice what your measurement error was and you were unable to tell me. In order for your results to be valid, the surrogate measure you used -- hand crease to tip lift-off needs to be accurate to well under 1 percent -- or otherwise accounted for in your analysis. I spoke with a top statistician at the National Academy of Sciences today, and he agrees that your results aren't valid if you don't measure and correct for this source of error.
As for my science reporting skills, I have already addressed the issues you raise, such as Dr. Austad's credentials. Again, I believe he was well qualified to comment on your experimental design.
Nature is well known for publishing flashy results that get good play in the media, but that don't always stand up to scientific scrutiny. According to my reporting, a serious review of this paper would have caught the errors that were evident to the first two scientists I faxed it to. A spokeswoman for Nature admitted to me that it was reviewed by a scientist who, in my judgment, had a strong motivation to endorse the findings. In discussing your paper, that spokeswoman also volunteered that Nature is free to ignore critical reviews. Perhaps I read too much into that remark, but it was a reasonable conclusion to reach under the circumstances.
If you wish to pursue this matter further, please feel free to contact NPR's Ombudsman, Jeffrey Dvorkin. He can be reached at 202 414-3245.
Sincerely,
Richard Harris
Professor Breedlove's rejoinder:
Date: Mon, 03 Apr 2000 15:07:57 -0700
To: Richard Harris <RHarris@NPR.ORG>
Cc: Jeffrey Dvorkin <JDvorkin@npr-01-msg.npr.org>, Kevin Klose <KKlose@NPR.ORG>, jtmann@liverpool.ac.uk, mcfadden@psyvax.psy.utexas.edu, austad@uidaho.edu, r.cotter@nature.com, nature@nature.com, David_Austin@ncsu.edu, trivers@rci.rutgers.edu, jcbailar@midway.uchicago.edu, jpalca@NPR.ORG
Subject: Re: My reply
Dear Mr. Harris,
I note that your reply to my complaint ignores most of the points I made in my message. Hence I submit these to the NPR Ombudsperson. Dr. Dvorkin, please inform me of NPR's procedures for demanding an on-air retraction or a rebuttal. I'm sure a review will convince you that Mr. Harris overstepped the bounds of fair journalism by presenting as wide-spread belief the ad hominem, derisive comments of a single scientist from another field.
At 04:18 PM 4/3/00 -0400, Richard Harris wrote: First, I do not write the material that ends up on our website. I have asked the web editor to correct the minor errors you brought to my attention.
OK, this mistake is that of someone else at NPR, not you. But it has to work both ways, Mr. Harris--I do not write the material that ends up in Nature press releases. I don't even get to see them, much less review them.
"1) You made an arbitrary choice to select a line on the skin as a surrogate for finger length. A person's fingers do not start at a wrinkle line. You did not present any data showing that is a valid substitute for bone length or other meaningful measure of a finger."
It is only your opinion that this is arbitrary, and your welcome to it but it is ludicrous. But my choice was not arbitrary as I told you before--the reported sex difference in two year old children used the wrinkle line as the base. Here you are clearly misleading the public by pretending that I made this measure up when I was replicating the literature. Had I used any starting point other than the one already used in the literature, that would have been arbitrary.
"Likewise, you chose a questionable surrogate for the end of a person's finger -- when flesh lifts off a flat surface. According to the scientists I interviewed, this is a critical flaw in your paper and invalidates your results."
You are free to question the use of xeroxes, although the other Manning paper we referenced (ref4) which you admit (below) you and your advisor did not consult, used exactly this method of using xeroxes. Yet despite your opinion and that of your consultants, we replicated a) Manning's report (ref1) of sex differences, b) Manning's report of the direction of sex differences, c) Manning's report of the sex difference being stronger on the right than the left. So far our results are not invalidated, but validated because we replicated the literature. That xerox would affect the absolute ratios is entirely possible, yet the detection of early androgen (as revealed by the sex differences) came through just fine.
"2) The paper was extremely poorly received by the two scientists I showed it to."
Neither of whom study hormones or human sexuality. You would have soundly criticized Nature for picking inappropriate reviewers if they had sent it out to these two you picked.
"In addition to Dr. Austad, I sent a copy for review to Dr. John Bailar, at the University of Chicago. For many years he was the New England Journal of Medicine's lead reviewer for biostatistics. He told me the statements you made in the paper were contradicted by the data you presented.
What statements in the paper? Here you resort to nebulous, blanket statements to avoid defending your statements. It is already clear that Nature sent out a press review, which I have still not seen, that contradicts both what I believe and what the paper says. You and your reviewers, in making derogatory comments about my work, should have been reviewing my work, not that of the Nature press corp. You are probably confusing those, but we can't tell when you won't be specific. It was only the Nature press release that, apparently, said the birth order effect was present only in gay men. As I said before, and you conveniently ignore, I present a graph making the point that the birthorder effect is strongest when all the men are considered.
"Few other journalists who covered you paper bothered to seek careful comment. The Philadelphia Inquirer was one of the few papers do so. That report quotes Dr. John Manning saying that you should have corrected for ethnicity, since he has found major differences in finger length based on ethnic background."
Yet because we found differences, despite not correcting for ethnicity (a correlate that has not yet been published, so I do not yet accept as demonstrated), simply indicates that the effects could be seen despite additional sources of variance. By the way, I liked the Philadelphia Inquirer piece quite a lot. In fact, I'm not unhappy with any of the other dozens of reports about our research, only yours.
Your report is far and away the outlier, Mr. Harris. Not just critical, but insulting.
"The Inquirer also quotes Dr. Robert Trivers saying the evidence for "hyper masculinity" is "very weak."
Again, it was the Nature press release that said, apparently, that gay men were hypermasculinized. What I said, in the paper and to you, was that all laterborn men were hypermasculinized and that I believed this hypermasculinization could contribute to the increased chance of laterborn boys growing up gay. Again, you're scouring other reports posthoc to find anything you can construe as negative--you haven't spoken to either Trivers or Manning, have you? They would have been appropriate commentators. Why don't you ask them now their opinions of your report if you stand by it?
"3) Your data on gay versus straight men showed no significant differences. Both Austad and Bailar found your statements about this comparison in the paper to be misleading."
By misleading, you mean in every way technically correct, right? I agree entirely that gay and straight men appeared the same, as I say in the paper and said to you. Again, apparently the Nature press release says they differed, but I never did. It's true that the birthorder effect in straight men did not reach statistical significance, but this is probably due to the smaller sample size of straight men.
"4) Austad and I found only one other paper in the literature relevant to finger length and your hypothesis -- the Manning paper you cited. I didn't see a reference in your paper to a similar study."
Then you now admit that you and Austad missed it--see ref 4--Trivers is another co-author. This is also the one that uses the method Dr. Austad "found" invalid--xeroxing hands.
"The reason Austad raised that point is because your results are markedly inconsistent with Manning's.
Only in absolute ratios which, as you quote Dr. Manning as saying to you, could be due to different ethnicities in his sample and mine. Yet we replicated the sex difference, the direction of the sex difference and the laterality of strength of the sex difference. Only a dullard would characterize this as "markedly inconsistent". As you know, Manning was one of the reviewers, and if he considered this inconsistent with his own results, he didn't say so. If you asked him (and I'll bet you were careful not to), I'll bet he would have said our results are quite consistent with his and that the differences in absolute values are to be expected. Hmm... you could ask him that now, but I'll bet you won't.
"I called Steven Austad because he is a field biologist who has studied sex ratios in animals (Austad & Sunquist. Sex ratio manipulation in the common opossum. Nature 324:58-60, 1986). I have spoken to him over the years and found him to be a thoughtful and credible scientist. He took the care to read Dr. Manning's previous work on this subject and to experiment with his own hand on the issue of measurement error."
None of which you mentioned to me or in your report. If Dr. Austad wants to offer actual data instead of off-the-cuff judgments, he's welcome to actually do a blind test and submit it for publication. I bet he won't. Why bother actually conducting experiments and publishing papers?
"Austad raised a serious scientific issue about your paper. I pressed you on that subject during our interview and you could not rebut his fundamental criticism."
You have yet to say what the criticism is. Is it that placing your hands in different ways will result in different ratios? Of course. That's why we had everyone place their hands the same way as much as we could. Is Dr. Austad actually suggesting that men with older brothers consistently place their hands down differently from men with no older brothers, and that's why we got a difference? To my mind that would be a far more remarkable statement than any I have made. You see here the problem of presenting Dr. Austad's "experiment" without any scientific review.
And the irony is that it is you and Dr. Austad who are corrupting the scientific process here (a crime you accuse Nature of committing). Failure or success to replicate happen all the time, but requires every party to submit their assertions to anonymous peer review. But Dr. Austad, because you have found him a "serious and credible" scientist, gets to offer his hip-shot opinions to millions of people without any scientist reviewing his data (if indeed he did anything as serious as gathering data). You criticize me for publishing data in Nature with anonymous peer review, then you and Dr. Austad broadcast his data to millions more people without any review at all.
"5) The issue of replication is important because it shows that there is a very large error in the way fingers are measured. In Dr. Austad's view, this overshadows the consistency you report with other findings (findings that have similar methodological problems)."
Again, Dr. Austad is welcome to his guesses, but I won't be impressed with his view alone--real scientists require data. Yes, replication is important and that's what Dr. Austad should have attempted. He is free to submit a paper of his own. But he and you couldn't be bothered that much before offering your "views" as established fact.
Does this mean, Mr. Harris and Dr. Austad, that if other labs replicate our results, you'll go on Morning Edition and admit you were wrong? I don't think either of you are big enough people to do that. Feel free to prove me wrong.
"6) As stated above, several researchers"
you mean by several, two, right? And these are researchers in other fields, right?
"have found your analysis of the men to be weak."
Again, you resort to nebulous term analysis without specifying what you mean. If you were specific about what you mean about analysis of men, I think you'd see you and your experts are confusing the Nature press release, for which I have no responsibility, with my paper.
"My statement about "measuring and measuring and measuring" was not intended to be derisive,"
I don't believe you and you know you're being hypocritical here. You might feel that unfair of me, but you made judgments about what you thought my motives were (and those of the Nature editors) and broadcast them to millions of people. None of which you ever brought up with me. I, at least, share my opinions of your motives with you. So clearly my distrust of your reported intentions is much more fair than yours.
How about "Jerry Springer-like", "cocktail party chatter", "titillating" and "not to be taken seriously"--we're those intended to be derisive Mr. Harris? Are you going to hide behind Dr. Austad when it was you who chose to broadcast his remarks? How about your sneering closer, "Dr. Breedlove is likely to be doing a lot of that [defending his paper]"? Was that intended to be derisive?
"and I doubt our listeners interpreted it that way."
In other words, you don't know. Shall we take a survey and see if your listeners found your report "derisive"? Please do that, it might be good data for legal action later.
"The point was simply that you gathered a lot of data."
Indeed, and if I had not, you would have used that to criticize the paper. At least that criticism would have been valid.
"7) This gets to the fatal flaw of your paper. As we discussed during the interview, there are two kinds of errors in scientific research. First is measurement error, that is uncertainty about each data point in the research."
Yes, and one traditionally addresses that by taking the same measurement twice (blindly, and with time in between each measure) and asking whether one gets the same measurements. We did this and got r values ranging from +0.95 to +0.99, as reported in our paper. How did your statistical experts explain this high correlation if our measurement error was so high? Are you sure they read that part of the paper?
"The second is statistical error, which deals with whether the effect you have measured is simply due to chance. The scientists I have consulted for this story say that if your measurement error is the same size as your effect, you can't correct for that with simple statistics. They say you need to use a 'total error model' which factors in the measurement error in your data points. Your effect is a 1 percent difference in finger ratios, which translates to a fraction of a mm in finger length. I asked you twice what your measurement error was and you were unable to tell me."
I thought you wanted to know how much resolution I brought and I told you: to the nearest half millimeter. I didn't know you wanted to know about repeatability because you never said that, plus it's described in black and white in the paper.
"In order for your results to be valid, the surrogate measure you used -- hand crease to tip lift-off needs to be accurate to well under 1 percent -- or otherwise accounted for in your analysis. I spoke with a top statistician at the National Academy of Sciences today,"
Does he have a name? Or is he unwilling to actually commit himself to the statements you attribute to him?
"and he agrees that your results aren't valid if you don't measure and correct for this source of error."
Good. Luckily, we did. Again you are so nebulous that I can't be sure what you mean here. But let me assure you, Mr. Harris, even if the mean sex difference in height is only 10 inches, you could with sampling detect that difference by measuring people to the nearest foot.
"As for my science reporting skills, I have already addressed the issues you raise,"
Not at all, the most serious reporting mistake was broadcasting one person's ad hominem remarks about my motives and those of Nature, without any corroborating evidence and apparently relying on your mind-reading abilities alone. Also, significantly, you presented that one person's remarks as if they were representative of the entire field, without a single third-party presenting an opposing view. More about that at the end of my remarks.
"such as Dr. Austad's credentials. Again, I believe he was well qualified to comment on your experimental design."
Again, he has never studied hormones, does not study human sexual behavior, but published a paper about sex differences over a decade ago. It would be dead easy for any journalist to find a more qualified source to comment on our paper. Tell me more about your history of turning to Dr. Austad--how did you get to know him exactly?
"Nature is well known"
By "well known" you mean, "I have no evidence but would still like to assert that"
"for publishing flashy results that get good play in the media, but that don't always stand up to scientific scrutiny. According to my reporting, a serious review of this paper would have caught the errors that were evident to the first two scientists I faxed it to."
If they are correct and the reviewers are not, an untested assumption of yours.
"A spokeswoman for Nature admitted to me that it was reviewed by a scientist who, in my judgment, had a strong motivation to endorse the findings."
I take this to mean Manning, whom I have never met but who asked not to remain anonymous. (By the way, since we know Manning is one of the two "hand-picked" (your words) reviewers, are you really criticizing Nature for sending a paper claiming to replicate and extend a scientist's publication to that scientist?) I emailed him after he identified himself and learned that he was himself gathering finger ratios related to sexual orientation. That he gave the paper a thumbs-up, despite our scooping him, could only be seen as a conflict of interest by a reporter trying to justify his precipitate behavior. Any outsider would see his behavior as commendable, not self-serving.
"In discussing your paper, that spokeswoman also volunteered that Nature is free to ignore critical reviews."
As are all journals.
"Perhaps I read too much into that remark, but it was a reasonable conclusion to reach under the circumstances."
And so you felt free to "perhaps...read too much into the remark" and slander me to millions of listeners. In other words, you admit that you have absolutely no corroborating evidence that the Nature review process was tainted, right? Funny, that was not the message you broadcast.
Furthermore, you avoid my main complaint about your journalistic method. Find a SINGLE person somewhere who's willing to say derogatory, ad hominem things about a paper ("Jerry Springer-like", "cocktail party chatter", "titillating", "not to be taken seriously") and put him on the air with no one who has heard his complaints allowed to answer, then give listeners the impression that all scientists feel this way. Even the most rudimentary sense of fairness would have put on a scientist who approved of the report, or at least one that had heard Dr. Austad's concerns. You didn't even tell me about his concerns and allow me to answer them. But you had no interest in being fair, did you? That's the complaint about your journalistic sensibilities, Mr. Harris. To my mind, your behavior resembles that of Jerry Springer more than does my own. But again, I have no need to broadcast this opinion of mine as fact to millions of people, as you did.
This is why, Mr. Dvorkin, NPR owes me an on-air rebuttal.
As for the final test, you are quite right that the issue of replication will decide whether your dismissive, abusive and unreviewed report was correct or whether my serious, peer-reviewed paper was correct. Are you willing to wager, Mr. Harris?
Marc Breedlove
More from Breedlove:
By the way, Mr. Harris and Dr. Austad, as long as you enjoyed the Philadelphia Inquirer article, I thought I'd point out one of my favorite passages:
"Robert Trivers, an evolutionary biologist at Rutgers University who also has studied finger-length ratios, said that a graduate student at Rutgers got similar results in a study of lesbian and heterosexual women. The average ratios for the two groups were different, he said, but there was considerable overlap between lesbians and straight women." A LINK OF FINGER SIZE, SEXUAL ORIENTATION WHAT OUR HANDS MAY TELL US ABOUT OUR HEARTS. By Stacey Burling PHILADELPHIA INQUIRER Thursday, March 30, 2000 Page: A01.
In other words, there's been a replication of our main finding already!
How curious, given that we were doing something "Jerry Springer-like", for "titillation" good only for "cocktail hour chatter" that "should not be taken seriously" and was so deeply, statistically flawed (indeed, had a "fatal flaw" to quote Mr. Harris) that two eminent scientists, chosen by Mr. Harris, saw them clearly even when the two peer reviewers had been too incompetent to notice.
I look forward to Dr. Trivers' future papers on this topic, and I'm sure we'll both enjoy many happy comparisons of differences and similarities in our findings. We will probably manage to do that without the ad hominem comments Dr. Austad prefers.
Notice also that this reporter, unlike you, got it right that I never reported a difference between gay and straight men in finger ratios. But then, that's what you'd expect of a professional. The only difference between the two groups of men we reported was in proportion of brothers among older siblings. (Curiously, given what crummy, derisory scientists we have been declared by NPR across the nation, that finding replicated previous work from other labs, too).
Feel free to apologize any time.
Cheers, Marc Breedlove
From: breedsm@socrates.berkeley.edu
Date: Thu, 06 Apr 2000 12:18:33 -0700
Subject: Re: Digit ratio & androgenization
At 01:38 PM 4/6/00 -0400, Ray Blanchard wrote:
"I was thinking this morning about whether a negative correlation of fraternal birth order with 2D:4D ratio would necessarily imply that later-born males are exposed to higher levels of prenatal androgen. I can think of another, albeit perhaps far-fetched, possibility. In Blanchard and Klassen (1997, p. 375), I cited the following study: 'Another possibility is that H-Y antibodies interfere with the customary organizational role of some sex hormone, perhaps by producing an exaggerated version of that hormone's usual effect, with paradoxical consequences. Evidence that relatively specific immunohormonal interaction might occur comes from a study by Flaherty et al. (1979). These authors found that cell surface changes involving H-Y and components of the major histocompatibility complex are induced in the presence of H-Y antibody (Flaherty et al., 1979). This may be physiologically relevant because testosterone but not estradiol induces the same changes. Any changes involving the major histocompatibility complex at the cell surface would have profound effects on embryonic development.' Blanchard, R., & Klassen, P. H-Y antigen and homosexuality in men. Journal of Theoretical Biology, 185, 373-378, 1997. Flaherty, L., Zimmerman, D. & Wachtel, S. S. (1979). H-Y antigen: Cell surface mapping and testosterone-induced supramolecular repatterning. J. Exp. Med. 150, 1020-1027."
This doesn't sound so far fetched to me. Firstly, we presently have no means of distinguishing a birthorder effect on androgen exposure from a birthorder effect on androgen responsiveness. Among possible ways birthorder might affect androgen responsiveness, an immune system connection is a strong possibility. Likewise, it's possible that the maternal immune system is responsible for the underlying sex difference in finger ratios by a mechanism independent of androgen. One of the jobs of the immune system is to remember. In that case, the lesbian/heterosexual women difference would presumably be non-androgen-mediated, too, but it's harder to see how maternal immune attacks affecting the fingers would favor any masculinization of the female brain....
Cheers, Marc
S. Marc Breedlove, Psychology Department
Robinson, SJ; Manning, JT (jtmann@liverpool.ac.uk), "The ratio of 2nd to 4th digit length and male homosexuality" Evolution and Human Behavior, 2000, Vol 21, Iss 5, pp 333-345
ABSTRACT: Sexual orientation may be influenced by prenatal levels of testosterone and oestrogen. There is evidence that the ratio of the length of 2nd and 4th digits (2D:4D) is negatively related to prenatal testosterone and positively to oestrogen. We report that (a) 2D:4D was lower in a sample of 88 homosexual men than in 88 sex- and age-matched controls recruited without regard to sexual orientation, (b) within the homosexual sample, there was a significant positive relationship between mean 2D:4D ratio and exclusive homosexuality, (c) overall, there was a decrease in 2D:4D from controls to homosexual men to bisexual men and (d) fraternal birth order, a positive predictor of male homosexuality, was not associated with 2D:4D in a sample of 240 Caucasian men recruited without regard to sexual orientation and 45 homosexual men. Further work is needed to confirm the relationships between 2D:4D and sexual orientation. However, these and other recent data tend to support an association between male homosexuality and high fetal testosterone. Very high testosterone levels may be associated with a sexual preference for both men and women. © 2000 Elsevier Science Inc.
A correction from author JT Manning:
"There is a mistake on page 341 middle paragraph. The dependent variable was number of older brothers NOT 2D:4D. The independent variables were 2D:4D and group. As reported there was no relationship between 2D:4D and number of older brothers (t=0.64, p=0.52) but there was between group and number of older brothers (t=3.57, p=0.0004). This meant that the homosexual sample had more older brothers than the heterosexuals. ... we then confirmed this with a t test. We then performed three more multiple regression tests with older sisters, younger brothers and younger sisters as dependent variables and 2D:4D and group as independent variables. There were no significant relationships between group and number of sibs in any of these."