Sax was singled out in the piece, which portrayed the logic behind single sex education as ‘pseudoscience.’
The piece is riddled with errors, says Sax:
On September 23 2011, the journal Science published an article entitled "The Pseudoscience of Single-Sex Schooling." I received the prepublication galleys the first week of September 2011. On Monday September 5 2011, I sent an urgent e-mail to the senior editors at Science, informing them that the article "is filled with astonishing errors of fact which invalidate the main thrust of the article -- astonishing in that errors of fact of this mangitude usually do not survive peer review." Many of those errors concerned me personally: the article misrepresents and distorts my view on many issues, and ignores most of my published writings and books which contradict the assertions made in the article.
And here’s a letter from Sax:
Friends, The article published last week in SCIENCE contains a section entitled "Negative Impacts of Highlighting Gender." This section has received extensive coverage in national and international media. It was the basis for the article last Friday in the Washington Post, "Single-sex education may do more harm than good"; "Children at single-sex schools more likely to be sexist" (The Daily Telegraph); "Single-sex schools make students sexist" (The Daily Beast) and so forth. But ALL the studies cited in the SCIENCE article regarding "negative impacts" were in fact studies involving a small number of PRE-SCHOOL students attending a COED pre-kindergarten. I am happy to provide the full text of the four articles on which the authors place such weight (my e-mail account won't allow me to attach such large files to an e-mail sent to more than 1 or 2 recipients at a time). The files are their references #24 (Martin & Halverson 1981.pdf), #25 (Hilliard & Liben 2010), #27 (Martin & Fabes 2001) and #28 (Fabes & Martin 1997). We're talking about that section of the article in Science under the heading, "Negative Impacts of Highlighting Gender." The authors begin this section by acknowledging that "some proponents of SS [single-sex] education claim it is well suited to countering sexism. . ." They then conjecture that "gender divisions are made even more salient in SS settings because the contrast between the segregated classroom and the mixed-sex structure of the surrounding world provides evidence to children that sex is a core human characteristic . . ." Presumably Halpern, Eliot, Bigler et al. believe that if we didn't have any single-sex classrooms, then children would have no evidence that "sex is a core human characteristic." More importantly, these authors provide no evidence for their substantive claim that "gender divisions are made even more salient in SS settings." In fact, this conjecture has been tested, and proven false, in multiple studies, none of which is cited by Halpern et al. One of the most compelling of these papers is the study published by Ursula Kessels and Bettina Hannover, in their paper entitled "When being a girl matters less: Accessibility of gender-related self-knowledge in single-sex and coeducational classes and its impact on students' physics-related self-concept of ability." The full text is attached as "2008 Kessels Hannover.pdf." These investigators randomly assigned 401 8th-grade girls from coed schools either to study science in an all-girls classroom or in a coed classroom. They found that girls in the all-girls classroom were less aware of "being a girl" and less aware of gender stereotypes regarding science, compared to girls who were randomly assigned to the coed classroom. This study directly refutes the conjecture of Halpern et al., yet they never mention it. (We will consider another similar study, by Carol Cronin Weisfeld, in a moment.) The authors then solemnly state, "Research has demonstrated that, when environments label individuals and segregate along some characteristic (e.g. gender, eye color, or randomly assigned t-shirt groups), children infer that the groups differ in important ways." Well, certainly. This is a robust finding which I discuss at length in my book Girls on the Edge, chapter 6, especially pages 174 - 180. If you don't have a copy of Girls on the Edge, let me know and I will send you a complimentary copy. I and others refer to this effect as "group contrast effects" although it goes by various names. People, including children, do tend to exaggerate the differences between groups. Halpern et al. then provide another conjecture (bottom of the 1st column, p. 1707): ". . .it is likely that these effects would be even more powerful when sex is used to divide children into entirely separate classrooms or schools rather than merely into separate lines to go to lunch." What evidence do the authors provide in support of this conjecture? Zero. Zilch. Nada. They got nothing. More troubling, they show no awareness that this conjecture has been tested and found to be false. For example, consider the classic study by Carol Cronin Weisfeld and colleagues, who videotaped girls in Chicago playing dodgeball, first girls against girls, then girls and boys together. Then they did the same thing with Native American children on a Hopi reservation in New Mexico. When Chicago girls played against other Chicago girls, there was lots of variation in the style and quality of play. Some of the girls were really serious about the game: as soon as play began, those girls would adopt what coaches call the "athletic stance," knees bent, arms flexed, eyes focused, ready to jump for the ball. When play got under way, these girls were competitors: they would jump for the ball, grab it, sometimes even wrestling the ball away from another girl. The girls who were most engaged were, not surprisingly, the highest-skilled girls at playing the game. Other girls were not particularly excited about the game and certainly were not jumping and grabbing for the ball; again, not surprisingly, these girls were less skilled. Weisfeld and colleagues found the same variation in engagement and skill among the Native American girls in New Mexico. When boys were brought into the gym so that there were an equal number of girls and boys playing, the picture changed. New Mexico Hopi girls still participated in the game, but the high-skill girls no longer demonstrated their skill. They didn't want to fight the boys for the ball. When the boys were playing, the high-skill Hopi girls looked very much like the low-skill girls. Most of the high-skill Chicago girls didn't even hang around for the game when boys were playing. Instead, they left the playing area altogether and went off in little groups to dance with one another or to snack on potato chips. Most of these kids, girls and boys, were 12 years old. In this study, the average girl in Chicago and in New Mexico was bigger and taller than the average boy in Chicago and in New Mexico, respectively. Nevertheless, the high-skill girls seemed to lose much of their enthusiasm for the game when boys were playing. (This paragraph is largely borrowed from chapter 6 of my book Girls on the Edge.) This study, and others like it, demonstrate that group contrast effects are most pronounced when girls and boys are TOGETHER. When they are separated, girls are more free to explore different ways of being, and likewise boys. The best way to get boys to play the flute is to create an all-boys band (see the closing chapter of my book Why Gender Matters for an example from Calgary Alberta in this regard). When girls and boys are together, boys are more likely to say, "Flutes are for girls." Likewise for cooking. Likewise for dance. What's stunning is that the authors show no awareness of the extensive literature on this topic. On this topic, in this section, they cite only papers written by one of the eight co-authors of this commentary. They literally seem not to have read anything which they themselves did not publish. What evidence do the authors cite? In the next paragraph, they claim that "When teachers make children's sex salient, students choose to spend less time interacting with other-sex peers." Note first of all how they have changed the topic. The authors have slipped in the assumption that a single-sex classroom is equivalent to making children's sex salient. In fact, the empirical evidence - and my own first-hand observations at more than 300 schools - suggests just the opposite. At coed schools, there's always a class clown, and the class clown at coed schools is almost always a BOY. At girls' schools, there's always a class clown, and the class clown at a girls' school is always a GIRL. There is a great deal of evidence (e.g. Kessels and Hannover, cited above) that one's own biological sex is less salient when everybody in the classroom is a girl, then when girls and boys are together and everybody wonders whether Emily likes Justin. But let's proceed with the authors' assertion that "When teachers make children's sex salient, students choose to spend less time interacting with other-sex peers." What evidence do they provide in support of this assertion? A study by Lacy Hilliard and Lynn Liben (one of the paper's co-authors) of 57 four-year-olds! Hilliard and Liben found that when teachers emphasized that boys are boys and girls are girls, in this COED setting, that all sorts of bad things happened. Well, OK. And explain to me how this is relevant to an ALL-GIRLS classroom? It's not. The authors then assert that "boys who spend more time with other boys become increasingly aggressive." That's a very broad and general statement. What evidence do they provide in support of this extraordinary statement? Just one study, written by two of the paper's co-authors, Carol Martin and Richard Fabes (Martin & Fabes 2001). And who were the subjects in this study? The subjects were 61 four-year-olds, 28 boys and 33 girls. How large was this effect? The authors never state the size of the effect in the original study (see p. 439), except that it was "significant" at the p=0.05 level. Let's just state the syllogism here: 1. Four-year-old middle-class white boys who play with other four-year-old middle-class white boys in a coed preschool in Tempe Arizona are somewhat more likely, later in the school year, to be aggressive. 2. Therefore boys' elementary schools, boys' middle schools and boys' high schools are a bad idea for all boys. Huh? I think it would be more reasonable to conclude: 1. Four-year-old middle-class white boys who play with other four-year-old middle-class white boys in a coed preschool in Phoenix are somewhat more likely, later in the school year, to be aggressive. 2. Therefore you should be cautious about enrolling your son in a coed preschool in Tempe Arizona. This article has been widely reported in media such as The New York Times and the Washington Post as the authoritative and indisputable last word documenting the negative effects of single-sex education. But the bias in this article really makes Rush Limbaugh look like a model of impartiality and restraint. The authors of the article in Science continue this paragraph with the assertion that "girls who spend more time with other girls become more sex-typed" - and what's the sole citation provided in support of this claim? Once again, Martin & Fabes, 2001. So: if it's true of 33 "mostly Caucasian" girls from "middle- to upper-middle class families, with 70% of the children living in two-parent households", attending preschool, then it must be true of all girls everywhere, at all grade levels. No cautions, no disclaimers. Let it not be said that the authors of the article for Science ignore only the work of other authors. They ignore their own studies, if their own studies don't fit their preconceived ideas in this article. For example, Fabes, Martin, Hanish et al. 2003 - a paper written by three of the co-authors of last week's article for Science - documented some unexpected BENEFICIAL effects of same-sex play for both girls and boys (Fabes Martin Hanish 2003.pdf, attached). But the authors of last week's paper for Science never mention the paper, even though they wrote it. Or at least three of them did. I have posted this comment, with links to all the articles cited, at my Psychology Today blog, www.psychologytoday.com/node/75335. I will post this comment and more to come at www.leonardsax.com/Science.htm where you will also find the full text of the article for Science and the pre-publication galleys, demonstrating the remarkable changes made to the manuscript between the time of my e-mail to Science on September 5 and actual publication of the article on September 22. At the moment I am busy preparing for our 7th International NASSPE Conference, in Orlando, less than two weeks away: Saturday and Sunday October 8 and 9. More information about the conference is available at www.singlesexschools.org/conference.html. The full conference program is now posted online at www.singlesexschools.org/2011.doc, so you can see the full list of 91 presenters and descriptions of the 58 presentations. Leonard
Leonard Sax MD PhD
We’re talking about that section of the article in Science under the heading, “Negative Impacts of Highlighting Gender.” The authors begin this section by acknowledging that “some proponents of SS [single-sex] education claim it is well suited to countering sexism. . .” They then conjecture that “gender divisions are made even more salient in SS settings because the contrast between the segregated classroom and the mixed-sex structure of the surrounding world provides evidence to children that sex is a core human characteristic . . .”
Presumably Halpern, Eliot, Bigler et al. believe that if we didn’t have any single-sex classrooms, then children would have no evidence that “sex is a core human characteristic.” More importantly, these authors provide no evidence for their substantive claim that “gender divisions are made even more salient in SS settings.” In fact, this conjecture has been tested, and proven false, in multiple studies, none of which is cited by Halpern et al. One of the most compelling of these papers is the study published by Ursula Kessels and Bettina Hannover, in their paper entitled “When being a girl matters less: Accessibility of gender-related self-knowledge in single-sex and coeducational classes and its impact on students’ physics-related self-concept of ability.” The full text is attached as “2008 Kessels Hannover.pdf.” These investigators randomly assigned 401 8th-grade girls from coed schools either to study science in an all-girls classroom or in a coed classroom. They found that girls in the all-girls classroom were less aware of “being a girl” and less aware of gender stereotypes regarding science, compared to girls who were randomly assigned to the coed classroom. This study directly refutes the conjecture of Halpern et al., yet they never mention it. (We will consider another similar study, by Carol Cronin Weisfeld, in a moment.)
The authors then solemnly state, “Research has demonstrated that, when environments label individuals and segregate along some characteristic (e.g. gender, eye color, or randomly assigned t-shirt groups), children infer that the groups differ in important ways.” Well, certainly. This is a robust finding which I discuss at length in my book Girls on the Edge, chapter 6, especially pages 174 - 180. If you don’t have a copy of Girls on the Edge, let me know and I will send you a complimentary copy. I and others refer to this effect as “group contrast effects” although it goes by various names. People, including children, do tend to exaggerate the differences between groups. Halpern et al. then provide another conjecture (bottom of the 1st column, p. 1707): ". . .it is likely that these effects would be even more powerful when sex is used to divide children into entirely separate classrooms or schools rather than merely into separate lines to go to lunch.” What evidence do the authors provide in support of this conjecture? Zero. Zilch. Nada. They got nothing. More troubling, they show no awareness that this conjecture has been tested and found to be false. For example, consider the classic study by Carol Cronin Weisfeld and colleagues, who videotaped girls in Chicago playing dodgeball, first girls against girls, then girls and boys together. Then they did the same thing with Native American children on a Hopi reservation in New Mexico. When Chicago girls played against other Chicago girls, there was lots of variation in the style and quality of play. Some of the girls were really serious about the game: as soon as play began, those girls would adopt what coaches call the “athletic stance,” knees bent, arms flexed, eyes focused, ready to jump for the ball. When play got under way, these girls were competitors: they would jump for the ball, grab it, sometimes even wrestling the ball away from another girl. The girls who were most engaged were, not surprisingly, the highest-skilled girls at playing the game. Other girls were not particularly excited about the game and certainly were not jumping and grabbing for the ball; again, not surprisingly, these girls were less skilled. Weisfeld and colleagues found the same variation in engagement and skill among the Native American girls in New Mexico. When boys were brought into the gym so that there were an equal number of girls and boys playing, the picture changed. New Mexico Hopi girls still participated in the game, but the high-skill girls no longer demonstrated their skill. They didn’t want to fight the boys for the ball. When the boys were playing, the high-skill Hopi girls looked very much like the low-skill girls. Most of the high-skill Chicago girls didn’t even hang around for the game when boys were playing. Instead, they left the playing area altogether and went off in little groups to dance with one another or to snack on potato chips. Most of these kids, girls and boys, were 12 years old. In this study, the average girl in Chicago and in New Mexico was bigger and taller than the average boy in Chicago and in New Mexico, respectively. Nevertheless, the high-skill girls seemed to lose much of their enthusiasm for the game when boys were playing. (This paragraph is largely borrowed from chapter 6 of my book Girls on the Edge.)
This study, and others like it, demonstrate that group contrast effects are most pronounced when girls and boys are TOGETHER. When they are separated, girls are more free to explore different ways of being, and likewise boys. The best way to get boys to play the flute is to create an all-boys band (see the closing chapter of my book Why Gender Matters for an example from Calgary Alberta in this regard). When girls and boys are together, boys are more likely to say, “Flutes are for girls.” Likewise for cooking. Likewise for dance.
What’s stunning is that the authors show no awareness of the extensive literature on this topic. On this topic, in this section, they cite only papers written by one of the eight co-authors of this commentary. They literally seem not to have read anything which they themselves did not publish.
What evidence do the authors cite? In the next paragraph, they claim that “When teachers make children’s sex salient, students choose to spend less time interacting with other-sex peers.” Note first of all how they have changed the topic. The authors have slipped in the assumption that a single-sex classroom is equivalent to making children’s sex salient. In fact, the empirical evidence - and my own first-hand observations at more than 300 schools - suggests just the opposite. At coed schools, there’s always a class clown, and the class clown at coed schools is almost always a BOY. At girls’ schools, there’s always a class clown, and the class clown at a girls’ school is always a GIRL. There is a great deal of evidence (e.g. Kessels and Hannover, cited above) that one’s own biological sex is less salient when everybody in the classroom is a girl, then when girls and boys are together and everybody wonders whether Emily likes Justin.
But let’s proceed with the authors’ assertion that “When teachers make children’s sex salient, students choose to spend less time interacting with other-sex peers.” What evidence do they provide in support of this assertion? A study by Lacy Hilliard and Lynn Liben (one of the paper’s co-authors) of 57 four-year-olds! Hilliard and Liben found that when teachers emphasized that boys are boys and girls are girls, in this COED setting, that all sorts of bad things happened. Well, OK. And explain to me how this is relevant to an ALL-GIRLS classroom? It’s not.
The authors then assert that “boys who spend more time with other boys become increasingly aggressive.” That’s a very broad and general statement. What evidence do they provide in support of this extraordinary statement? Just one study, written by two of the paper’s co-authors, Carol Martin and Richard Fabes (Martin & Fabes 2001). And who were the subjects in this study? The subjects were 61 four-year-olds, 28 boys and 33 girls. How large was this effect? The authors never state the size of the effect in the original study (see p. 439), except that it was “significant” at the p=0.05 level.
Let’s just state the syllogism here:
1. Four-year-old middle-class white boys who play with other four-year-old middle-class white boys in a coed preschool in Tempe Arizona are somewhat more likely, later in the school year, to be aggressive.
2. Therefore boys’ elementary schools, boys’ middle schools and boys’ high schools are a bad idea for all boys.
Huh? I think it would be more reasonable to conclude:
1. Four-year-old middle-class white boys who play with other four-year-old middle-class white boys in a coed preschool in Phoenix are somewhat more likely, later in the school year, to be aggressive.
2. Therefore you should be cautious about enrolling your son in a coed preschool in Tempe Arizona.
This article has been widely reported in media such as The New York Times and the Washington Post as the authoritative and indisputable last word documenting the negative effects of single-sex education. But the bias in this article really makes Rush Limbaugh look like a model of impartiality and restraint.
The authors of the article in Science continue this paragraph with the assertion that “girls who spend more time with other girls become more sex-typed” - and what’s the sole citation provided in support of this claim? Once again, Martin & Fabes, 2001. So: if it’s true of 33 “mostly Caucasian” girls from “middle- to upper-middle class families, with 70% of the children living in two-parent households”, attending preschool, then it must be true of all girls everywhere, at all grade levels. No cautions, no disclaimers.
Let it not be said that the authors of the article for Science ignore only the work of other authors. They ignore their own studies, if their own studies don’t fit their preconceived ideas in this article. For example, Fabes, Martin, Hanish et al. 2003 - a paper written by three of the co-authors of last week’s article for Science - documented some unexpected BENEFICIAL effects of same-sex play for both girls and boys (Fabes Martin Hanish 2003.pdf, attached). But the authors of last week’s paper for Science never mention the paper, even though they wrote it. Or at least three of them did.
I have posted this comment, with links to all the articles cited, at my Psychology Today blog, www.psychologytoday.com/node/75335. I will post this comment and more to come at www.leonardsax.com/Science.htm where you will also find the full text of the article for Science and the pre-publication galleys, demonstrating the remarkable changes made to the manuscript between the time of my e-mail to Science on September 5 and actual publication of the article on September 22.
At the moment I am busy preparing for our 7th International NASSPE Conference, in Orlando, less than two weeks away: Saturday and Sunday October 8 and 9. More information about the conference is available at www.singlesexschools.org/conference.html. The full conference program is now posted online at www.singlesexschools.org/2011.doc, so you can see the full list of 91 presenters and descriptions of the 58 presentations.
Leonard
Leonard Sax MD PhD