August 22, 2006

Leonard Sax on hearing

Leonard Sax's recent book, Why Gender Matters: What Parents and Teachers Need to Know about the Emerging Science of Sex Differences, presents an impassioned argument for single-sex education. His basic message is that "Girls and boys play differently. They learn differently. They fight differently. They see the world differently. They hear differently." He claims that these are not stereotypes, but scientific facts. And therefore, boys and girls should be educated differently -- and separately.

In an earlier post, we took a look at Sax's argument that girls and boys "see the world differently", and found that he builds his case by presenting data about sexual dimorphism in rat retinas as if it were data about human retinas, when the (easily available) comparable human data is not at all like the rat data. In this post, we're going to consider what Sax has to say about how girls and boys "hear differently". Part of the problem here is the usual rhetorical move of claiming that boys and girls (or men and women) are essentially different as groups, when in fact the difference in average values between the sexes is small relative to the within-sex variation. But in this case, bizarrely, one of the two pieces of research that Sax cites actually found a difference in means that's in the opposite direction from what he claims.

On the web site for Sax's book, he offers a page on sex differences in hearing, which expands on the content of the corresponding section of his book. Below, I'll quote some passages from that web page, in blue, with interspersed commentary, with some figures and quotations from one of the key sources that Sax cites.

The first systematic evaluation of the hearing of girls and boys was performed by Professor John Corso of Penn State University in the late 1950's and early 1960's. Dr. Corso simply used a soundproof booth, headphones, and a tone generator. He consistently found that the girls hear better than boys do, especially in the range of frequencies above 2 kHz. See John Corso, Age and sex differences in thresholds, Journal of the Acoustical Society of America, 31:489-507, 1959; also John Corso, Aging and auditory thresholds in men and women, Archives of Environmental Health, 6:350-356, 1963.

The title of Corso 1959 is actually "Age and sex differences in pure-tone thresholds", and its page numbers are actually 498-507, not 489-507. These bits of bibliographical sloppiness are trivialities, but they're an initial indication that Sax is not applying a very high standard of care here, despite the centrality of this work to his argument. There's a much more important problem about the way that Sax respresents the content of Corso's work. The difference that Corso found between average male and female hearing thresholds is generally about 1/4 to 1/2 of a standard deviation, depending on age and frequency. As a result, the distributions are so heavily overlapped that (as I put it in another post on the same topic), "[i]f you pick a man and a women (or a boy and a girl) at random, the chances are about 6 in 10 that the girl's hearing will be more sensitive -- but about 4 in 10 that the boy's hearing will be more sensitive." It's misleading, at best, to describe this by saying that as "[Corso] consistently found that girls hear better than boys do".

Now comes the bizarre part.

Pediatric audiologists Barbara Cone-Wesson, Glendy Ramirez, and Yvonne Sininger have done careful studies of the hearing of newborn babies. When any baby or child (or adult for that matter) hears a sound, there's an immediate reaction, called an acoustic brain response.

Actually, it's called the "auditory brainstem response". (As we'll see, this terminological sloppiness is followed by some extraordinarily careless -- i.e. backwards -- reading of the reported results.) It's also a bit misleading to call the ABR an "immediate reaction". The infants are not "reacting" in the sense of perking up and looking around -- the tests were done while they were sleeping. ABRs are event-related potentials, "very small electrical voltage potentials originating from the brain recorded from the scalp in response to an auditory stimulus", and representing the automatic passage of signals from the cochlea via the 8th cranial nerve through various brainstem structures. "Wave V", which is what Cone-Wesson et al. rely on, "is believed to originate from the vicinity of the inferior colliculus".

Cone-Wesson and her colleagues decided to measure the acoustic brain response of more than 60 newborn girls and boys.

For the record, they measured the "auditory brainstem response" of 72 neonates.

For a 1500 Hz tone played to the right ear, they found that the average girl baby had an acoustic brain response about 80% greater than the response of the average baby boy. Here are the references for those studies: Barbara Cone-Wesson and Glendy Ramirez. Hearing sensitivity in newborns estimated from ABRs to bone-conducted sounds, Journal of the American Academy of Audiology, 8:299-307, 1997.
Yvonne Sininger, Barbara Cone-Wesson, and Carolina Abdala. Gender distinctions and lateral asymmetry in the low-level auditory brainstem response of the human neonate. Hearing Research, 126:58-66, 1998.
The fact that girl babies have a more sensitive threshold for very quiet sounds does not by itself prove that girls hear better than boys when the sound is louder;
[emphasis added]

Here's where it gets weird. What these (eminent and deservedly respected) researchers found in their (exemplary) study was exactly the opposite of the way that Sax describes it in the emphasized phrase above. In their own words, Yvonne Sininger, Barbara Cone-Wesson, and Carolina Abdala, "Gender distinctions and lateral asymmetry in the low-level auditory brainstem response of the human neonate", Hearing Research, 126:58-66, 1998, wrote:

This study revealed three significant and surprising results. First, ABR thresholds are lower for male newborns than for females. Second, wave V amplitudes elicited with low-level stimuli are larger when the right ear is stimulated in male and female neonates. Third, wave V amplitudes elicited by low-level stimuli are larger in female than male neonates but only in the right ear.

In other words, the boy babies had a "more sensitive threshold" than the girl babies, on average, not the other way around! Repeating it another way, the infant boys showed a statistically significant ABR "wave V" response to softer stimuli, on average, than the girls did. Of course -- as usual -- the differences were small compared to the within-sex variation:

Fig. 1. Mean ABR thresholds for clicks and tone bursts plotted as a function of gender (A) and ear stimulated (B). Gender and ear effects are condensed across frequency in bar graphs at right. Only gender effects reach significance (P<0.05). Error bars indicate ±one standard deviation.

[For the geeks among you, SPL is "sound pressure level" in DB relative to a reference level (typically 20 µPa), and peSPL is the "peak equivalent" SPL, defined as the SPL of a 1000-Hz tone whose maximum amplitude is the same as the peak value of a transient signal like a click.]

At signal levels where responses were well established, girls' ABRs were (on average) somewhat greater in amplitude than the boys' ABRs were (in terms of voltages measured through the scalp, whatever that really means). Again, as usual, the between-sex differences were small relative to the within-sex variation:

Fig. 5. Mean wave V amplitude from neonates in response to low-level stimuli. Latency is plotted by gender (A) and by ear (B). Bar graph at right condenses data across frequency. Error bars indicate ±one standard deviation.

And the researchers note that the combination of lower thresholds and lower amplitudes for boys vs. girls is "paradoxical". So when Sax writes that "the average girl baby had an acoustic brain response about 80% greater than the response of the average baby boy", we move from carelessless to misrepresentation. Let's ignore the fact that he's got the terminology wrong. The cited researchers tested the baby's response to clicks and to tones at four different frequencies. He ignores (in the book) or inverts (on the web page) the threshold differences. It's only for one of the five stimulus conditions -- the 1500 Hz tone -- that there's much of a difference between the sexes in the response amplitude. The overall sex difference in Wave V amplitude (which is the only kind of amplitude data reported) is roughly .2 μV for the boys and .25 μV for the girls. I make that 25% greater, not 80% greater. The difference in means for the 1500 Hz tones might be 80%, but what about all the other stimulus conditions? (And look at the error bars in the 1500 Hz condition!)

[A couple of caveats are in order here. While I'm not an expert in this technology, it strikes me that there are a lot of things besides the brain in the electrical pathway between the brainstem and the electrodes measuring the voltage -- blood, bone, flesh, and skin among other things -- and it's not clear to me how we can be sure that the voltage differences didn't result from differences in these non-neurological components. There are some other questions to explore about this experiment as well -- thus "female infants were tested at an average of 127 h post partum whereas male infants were on average 78.3 h old", and "The slope of the regression line for the threshold hours post partum function is positive for both male (+0.032×hours) and female (+0.027×hours) infants." The study's authors test and reject this explanation on statistical grounds, but I'd feel better about it if I could explore their raw data a bit.]

This study goes a long way toward establishing the conclusion that its authors came to see in its data. They began with the assumption that the well-known sex differences in hearing were due to nurture, not nature:

Surveys of auditory sensitivity have consistently shown adult females to have superior hearing (lower thresholds) above 1000 Hz (Corso, 1963; Glorig and Roberts, 1965; Axelsson and Lindgren, 1981) whereas low frequency sensitivity (500 Hz and below) seems to be better in males (Jerger et al., 1993). Hearing decrement with age is also more rapid in males than females (Pearson et al., 1995; Kryter, 1983). Sex-related differences in auditory sensitivity are often attributed to greater occupational and recreational noise exposure in males (Kryter, 1983) because no significant differences in hearing related to sex were found when studying adult tribal populations in non-industrialized areas free of high levels of noise (Rosen et al., 1962; Goycoolea et al., 1986). This led to the conclusion that no inherent differences existed in sensitivity between males and females and left the assumption that gender-related differences in hearing could be attributed to differences in susceptibility to noise-induced hearing loss alone.

As a result, they did not really plan their study for the purpose of studying sex differences:

It is important to note that the number of subjects in this study is relatively small. The original purpose of this study was to characterize ABR thresholds in newborns. Whereas gender and ear randomizations were included in the original design, it was not anticipated that distinctions in ABR threshold based on these factors would be found.

However, they did find sex differences in their data, and they make a convincing argument that the differences are probably due to "nature" (i.e. hormonal or sex-linked gene expression differences) rather than "nurture" (i.e. different life experiences). BUT, crucially, the differences between the sexes are so small relative to the within-sex variation that no possible conclusions can be drawn about sex-related educational policy -- and, as it happens, the differences in thresholds are in the opposite direction from Sax's description.

Because the between-groups differences are so small, and the situation with newborns is apparently so different from the situation even a short time later in life, these results have no logical bearing at all on matters of educational policy. However, Sax's mistake in describing the direction of the threshold effect does matter, at least to his rhetoric if not to his logic. The point of all this, for Sax, is not just that boys and girls are different, but that they're different in particular ways that happen to coincide with a common set of sexual stereotypes about sensitivity vs. insensitivity (and other metaphorically-related oppositions). Sax writes (Why Gender Matters, p. 18):

The difference in how girls and boys hear also has major implications for how you should talk to your children. I can't count the number of times a father has told me, "My daughter says I yell at her. I've never yelled at her. I just speak to her in a normal tone of voice and she says I'm yelling." If a forty-three-year-old man speaks in what he thinks is a "normal tone of voice" to a seventeen-year-old girl, that girl is going to experience his voice as being about ten times louder than what the man is hearing. [...]

The gender difference in hearing also suggests different strategies for the classroom. More than thirty years ago, psychologist Colin Elliot demonstrated that eleven-year-old girls are distracted by noise levels about ten times softer than noise levels that boys find distracting. ... One reason for that difference, of course, it that eleven-year-old girls hear better. If you're teaching girls, don't raise your voice .... [but] the rules are different when you're teaching boys.

But one of the major pieces of research that Sax cites actually finds that newborn boys have somewhat more sensitive hearing, on average, than newborn girls do! I haven't tracked down the references that Sax gives to support his striking statements about girls perceiving voices and noises as "ten times louder" than boys do, but I'm willing to guess what I'll find if I look into it. [Update: and, alas, I guessed right. See "Girls and boys and classroom noise", 9/9/2006).]

My conclusion from all this is that Leonard Sax has no serious interest in the science of sex differences. He's a politician, making a political argument. For all I know, his political goal -- single-sex education -- might be a good thing. But he should stop pretending that he's got science on his side, or else he should start paying some minimal attention to what the science actually says.

[A list of Language Log posts on Leonard Sax and Why Gender Matters:

"David Brooks, cognitive neuroscientist" (6/12/2006)
"Are men emotional children?" (6/24/2006)
"Of rats and (wo)men" (8/19/2006)
"Leonard Sax on hearing" (8/22/2006)
"More on rats and men and women" (8/22/2006)
"The emerging science of gendered yelling" (9/5/2006)
"Girls and boys and classroom noise" (9/9/2006)

[Update 5/29/2008 -- see Dr. Leonard Sax's response here, and some further discussion by me at

"Sax Q & A", 5/17/2008
"Liberman on Sax on Liberman on Sax on hearing", 5/19/2008


Posted by Mark Liberman at August 22, 2006 06:18 AM