When we see alarming statistics about gender disparities in STEM disciplines, we quite naturally wonder how to fix this imbalance.
(This hope – by the way – isn’t simply a do-goody desire to sing “It’s a Small World After All.” If we believe that men and women can contribute equally to a scientific understanding of our world, then every girl discouraged is a contribution lost.
In other words: we ALL benefit if boys and girls contribute to science.)
So, how can we encourage girls to participate in science?
To answer this question, we might first answer a related question: what discourages girls in the first place.
If we can undo the discouragement, we are – indirectly but effectively – encouraging.
So, what discourages girls?
Is Science Education Itself the Problem?
Here’s a disturbing possibility.
When students learn about genetics, and specifically about the genetics of sex differences, they might infer that genders have a fixed, absolute quality. All boys (and no girls) are this way; all girls (and no boys) are that way.
It’s in the genes, see?
This set of beliefs, in turn, might reinforce a fixed mindset about gender and ability.
Through this causal chain, a particular science curriculum might itself discourage girls from pursuing science.
Yikes!
Researcher Brian Donovan and his team explored this question in a recent study. To do so, they asked students to read different lessons about genes and sexual dimorphism.
Some 8th – 10th graders learned about the genetics of human sexual difference.
Others learned about the genetics of plant sexual differences.
Others read a curriculum that explicitly contradicted the notion that genetic sex differences directly cause differences in intelligence and academic ability.
Did these curricular differences have an effect?
The Results Envelope Please
Unsurprisingly, students who learned that we can’t draw a straight line from genes to gender roles and abilities believed that lesson.
To make the same point in reverse: students who studied a seemingly “neutral” scientific curriculum – “we’re just talking about genes here” – drew unsupported conclusions about absolute differences between men and women.
Amazingly, this finding held true both for the students who studied the genetics of human sexual differences AND those who studied plant sexual differences.
WOW.
Perhaps surprisingly, students who learned that genetic sex differences don’t cause gendered ability differences also expressed a greater interest in science.
In particular, the girls who studied the “genetics only” lesson expressed meaningfully less interest in a science major than those who got the alternative lesson. (The two lessons neither encouraged nor discouraged the boys.)
But, Why?
Here’s the likely causal chain:
A science curriculum that focused “purely” on genetics seemed to suggest that men and women are utterly different beings.
Students who read this “pure” lesson inferred that some human abilities – like, say, scientific competence – might differ between genders.
This inference, in turn, made gender stereotypes (e.g., “men do better at science than women”) more plausible.
And so, the women who got that seemingly neutral science lesson, discouraged by the stereotype it reinforced, felt less inclined to pursue science.
By this roundabout route, a traditional science lesson might itself discourage students from learning science.
Alternative Explanations
Of course, the topic of gender differences – especially in the realms of math and science – can generate lots of energetic debate.
When I asked Donovan for alternative explanations for his findings, he was quick to emphasize that we need lots more research in this field. His is the first study done on this specific question. As always, teachers shouldn’t assume that any one study has found THE answer.
Some people do in fact argue that math and science ability (or interest) differ by gender because of genes. (Dr. Donovan explicitly rejects an explanation that moves directly from genes to gender differences.)
Here’s a recent book review by Lise Eliot, emphasizing that gender differences in brain regions
a) are often exaggerated and mis-reported, and
b) result from societies that emphasize gender differences.
Donovan’s research suggests that teachers can and should do more to be sure we’re not discouraging some students from particular academic interests and career paths.
For one set of practical suggestions, this interview with Sapna Cheryan outlines several ways we can promote “ambient belonging” in our classrooms.
You can understand why this study lit up my twitter feed recently. It makes a remarkable claim: women — but not men — experience working memory declines after a sleepless night.
Why We Care
We have at least two powerful reasons to care about this study.
First, it makes strong claims about gender differences. According to lead author Rangtell (and 8 colleagues), women’s performance on a working memory task gets worse after a sleepless night.
On the other hand, men’s working memory performance remains just as good as when they had a cozy 8-hour sleep.
(I’ve written about gender differences before. You may recall that I’m often skeptical of specific claims, but do think that there are some important differences at the population level.)
So, this study plays an important role in the ongoing debate.
Second, Rangtell’s study focuses on working memory. And, working memory is really important in school.
What is working memory?
When a student works on a word problem in math, she first has to select the key information from the sentences. Then she holds that information in mind. Third, she reorganizes all that information into the correct formula. And finally she combines pieces of that formula appropriately: for example, she combines “7x+8x” into “15x.”
Whenever students select, hold, reorganize, and combine information, they’re using working memory.
And, our students do that all the time. They use working memory to conjugate a new Spanish verb. And, when they apply new terminology (“protagonist”) to a specific book (“Sethe is the protagonist of Beloved.”) And, when they balance chemical equations.
Basically, schools are shrines we build to honor successful working memory functioning.
If there truly is a gender difference in working memory function, that’s a really big deal.
Sleeplessness Harms Women More Than Men?
This study is, conceptually, very straigtforward.
Ask some people to do a working memory task after a full night’s sleep. Then, ask them to do the same task after they’ve been up all night. Is there a difference in their working memory performance?
Rangtell and her colleagues say: for men, “no”; for women, “yes.”
However, this study includes a very serious problem. The task that they use to measure working memory DOESN’T MEASURE WORKING MEMORY.
(You read that right.)
The researchers asked these people to listen to a list of numbers, and then type those numbers into the computer in the same order.
That’s simply not a test of working memory. After all, the participants didn’t have to reorganize or combine anything.
Instead, that’s a test of short-term memory.
Now, short-term memory is related to working memory. But, “related to” isn’t good enough.
Imagine, for instance, I claimed that sleeplessness makes people shorter. The way I determine your height is by measuring the length of your arm.
Of course: arm length and height are related. But, they’re not the same thing. Tall people can have short-ish arms. I can’t measure one thing and then make a claim about a related but different thing.
So too, Rangtell can’t measure short-term memory and then make claims about working memory. She didn’t measure working memory.
Does sleeplessness harm women’s working memory more than men’s? We just don’t know.
(By the way: I’ve reached out to the lead researcher to inquire about the working memory/short-term memory discrepancy. I’ll update this post if I hear back.)
You doubtless know that Mindset Theory has been increasingly doubted–and increasingly defended–in the last two years or so.
(In this post, for example, the author updates his earlier criticism of Mindset Theory and largely ends up defending Dweck–or, at least, criticizing her critic. His back-n-forth on this question helpfully represents the nature of the current debate.)
Today’s News
A recently published study looks carefully at a specific set of claims often advanced in Mindset world:
First: that girls and women have a fixed mindset more often than boys and men, and
Second: the smarter the girls and women, the likelier they are to have fixed mindsets.
In other words, for Mindset
First: gender matters, and
Second: for girls and women, intelligence matters.
What Did The Researchers Find?
Nope, and nope.
In their research, which included not only college students but also adults in the population at large, Macnamara and Rupani found no consistent patterns in either direction.
That is: in their research, there was no consistent gender split on Mindset. And, for men as well as women, intelligence level didn’t consistently influence Mindset; nor did a Growth Mindset predict academic accomplishment.
In truth, as you’ll see if you look at the graphs, they got quite a complex muddle of results. It’s genuinely difficult to pick out meaningful patterns in all their data.
What Next?
In my experience, Dweck tends to be quite open and responsive to thoughtful critique. Unlike some researchers who refuse to recognize those who disagree with their work, she is remarkably comfortable acknowledging debate and rethinking her own research.
So: I’ll be curious to see if and how she responds to this study.
There is, by the way, a broader message here as well. Although Mindset Theory is quite well established in the field of education, it is still up for discussion in the field of psychology.
Those of us who shape our classrooms and our schools with such theories in mind should be sure to check back in and see if they are holding up over time.
Regular readers of this blog know that I’m a skeptic about gender differences in learning. Although they certainly do exist–I think particularly about differences in 3d mental rotation–I often think they’re overstated or overemphasized.
At the same time, my emphasis on this point might obscure the fact that at the population level, gender differences in learning do sometimes exist. Two articles are, I think, particularly helpful in understanding these ideas.
First, this weighty research review considers the number of women in STEM fields and reaches three broad conclusions:
“Males are more variable [than females] on most measures of quantitative and visuospatial ability, which necessarily results in more males at both high- and low-ability extremes; the reasons why males are often more variable remain elusive.”
“Females tend to excel in verbal abilities, with large differences between females and males found when assessments include writing samples. “
“We conclude that early experience, biological factors, educational policy, and cultural context affect the number of women and men who pursue advanced study in science and math and that these effects add and interact in complex ways. There are no single or simple answers to the complex questions about sex differences in science and mathematics.”
The article stands out to me not only for its thoroughness, but for its all-star list of authors. Janet Shibley Hyde, for example, is well known for her skepticism about gender differences; in fact, she authored a widely-cited article called The Gender Similarities Hypothesis. If a known skeptic is on board with these conclusions, then I’m comfortable being there too.
(Another author, Diana Halpern, by the way, is a former president of the American Psychological Association.)
Second, Hyde has published an exploration of the first argument above: that men show greater variability in quantitative and visual abilities. This hypothesis suggests that–although large populations of men and women will have the same average math scores–we would expect to see more men who are very good at math (say, the top 5%) and also who are very bad at math (say, the bottom 5%).
Hyde’s article shows the complexity of this hypothesis. In particular, given that these variations differ from country to country, and can change over time, we have to recognize the social and historical context of any data set.
Two articles jumped out at me today because of the illustrative way they clash with each other.
Writing on Twitter, and providing helpful links to several sources, Adam Grant argues that “Differences between Men and Women are Vastly Exaggerated.”
Whereas Neuroscience News published a summary of a recent research study (by Daniel Amen) with the headline “Women Have More Active Brains Than Men.”
So, which is it? Are differences between the sexes exaggerated? Or do male and female brains operate very differently?
Let’s use three lenses to look at that question.
The First Lens: Discipline
Oversimplifying for the sake of clarity, we can say that neuroscience studies brains–that is, physical objects. It looks at neurons and blood flow and neurotransmitters and electrical energy. Things.
Psychology studies the behavior of brains–that is, what people do with those physical objects. It looks at a student’s ability to remember, or an athlete’s ability to concentrate, or an adult’s ability to learn a new language. Behaviors.
Obviously, both neuroscience and psychology are fascinating. But, which discipline is more useful?
Of course, the answer to that question depends on your definition of “useful.”
I myself think that teachers benefit from learning about the behavior of brains (that is, psychology) more than we do from learning about brains as objects (that is, neuroscience).
For example, if I tell you how brains change physically when long-term memories form, that information is interesting. (In fact, I often share this information when I talk with teachers.)
But, if I tell you what kind of teaching behavior makes long-term memory formation more likely, that information is really useful.
For this reason, I think Grant’s summary–which focuses on psychology–is likely to be more useful than the Amen study–which focuses on neuroscience.
For example: Grant’s summary looks at anti-stereotype-threat strategies that combat gender differences in college majors or professions. Teachers can do something with this information.
The Amen study, on the other hand, tells us about different levels of brain activity as measured by Single Photon Emission Computed Tomography (SPECT). I don’t know exactly what SPECT is, and I certainly don’t know how I would teach differently given this information.
So for me, again, neuroscience is fascinating, and psychology is useful.
(To be clear, I have several colleagues–whose judgment I highly respect–who disagree with me strongly on this point; that is, they think the neuroscience is just as important for teachers as the psychology. So, if you think I’m wrong, you’re not the only one.)
The Second Lens: The Population Being Studied
Whenever you use brain research to help your teaching, you should focus on the participants in the study. The more the participants resemble your own students, the likelier it is that the research findings will benefit your students.
So, if you find a study that says three repetitions of a practice exercise benefits long-term memory, that study might be very helpful. But: if the participants in the study were college students at an elite university, and you teach 1st graders who are already struggling with formal education, the study might not mean much to you.
After all, your students differ from those in the study so substantially that there’s no way to be sure the conclusions apply to your teaching context.
Grant’s research summary chooses several very large analyses. When he looks at (very small) gender differences in math scores, for example, his source draws on almost 4,000 studies. It seems likely that such broadly supported research will apply to my students too.
Amen’s study looks at a very large population–almost 27,000 people. However, and this is a big however, all but 119 of those people were suffering from “a variety of psychiatric conditions such as brain trauma, bipolar disorders, mood disorders, schizophrenia/psychotic disorders, and attention deficit hyperactivity disorder (ADHD).”
For obvious reasons, it’s hard to draw conclusions about neurotypical brains by studying aneurotypical brains.
So, again, because the Grant summary includes students like mine, and the Amen study doesn’t, I’m likelier to benefit from Grant’s conclusions.
(By the way, it’s entirely possible that your students seem more like Amen’s participants than those included in Gran’s summary–in which case, you may be more swayed by Amen’s findings.)
The Third Lens: Biases
In the world of science, “bias” isn’t necessarily a bad thing. All analysis–including yours, including mine–includes bias. Our goal should not be to eliminate bias (we can’t), but to recognize it in ourselves and others, and to do the best we can to look for countervailing biases.
So, let me be up front with you: my bias is, I’m usually skeptical about strong claims of gender difference in education. This skepticism has many sources–but, no matter how good those sources are, you should know that I’m not an impartial author delivering truth from on high.
I am, instead, someone who rarely finds evidence of gender difference in education persuasive…and (surprise!) my post has twice concluded that the “gender makes little difference in education” article is more useful and persuasive than the “there are big gender differences in brains” article.
Now that you know my bias, you should a) look for people with the opposite bias, and see if you find their arguments more persuasive than these, and b) recognize your own biases, and do your best to counterbalance them.
After all, one thing is certainly true about male and female brains: we’re all faster to believe ideas that support our own prior conclusions.
——————————————
Two final notes:
First, my thanks to Stephanie Sasse (prior editor of this blog) and Maya Bialik (former writer for this blog) for their idea of “lenses” as a way to analyse brain research.
Second, brain research generally hasn’t come to grips with people who fall outside a male/female gender dichotomy. Our understanding of gender and learning will be stronger and more useful when it does.
The short version: rates of diagnosis continue to increase.
The longer version: depending how you analyze the categories, you get very different results. For children younger than 5, the rates are — in fact — falling. For adults over 65, however, the rate rose 348% from 2008-9 to 2012-13.
(That is not a typo: 348%).
One important point as you review these data: percentages are interesting, but so too are the absolute numbers. Diagnoses among those over 65 can increase so much as a percentage because the absolute numbers are relatively low.
By the way: analysis by gender shows an interesting pattern. Among adults, both diagnosis and medication are increasing faster for men than women. Among children, however, that pattern is reversed.
How can we encourage young women to pursue STEM fields?
In the German state of Baden-Württemberg, school leaders tried a substantial reform: they increased the math requirement during the final two years of high school. Instead of taking math three days a week, all students had to take math four days a week.
What were the results of increasing the math requirement by 1/3 for 2 years? (That sentence sounds like a word problem, no?)
A mixed bag.
The good news: this reform reduced the gap between male and female achievement scores in math. On the surface, in other words, it seems young women learned more.
This result should be very exciting. However…
The so-so news: this additional math work did very little to increase women’s participation in STEM fields in college. Instead, it increased the STEM interest of male college students–the enrollment gap remained about the same.
And, the bad news: although the women learned more math, they felt worse about their own math abilities.
The reason for this last result isn’t clear — the author’s hypothesis honestly sounds a little convoluted to me.
But, given the size of the data pool behind this study, the conclusion seems clear: requiring more math may boost math learning, but — for women — it’s not sufficient to boost math confidence and interest in STEM fields.
At a minimum, the study suggests that we should think not only about how much math students learn, but how they learn it.
A further point: I don’t know how the math curriculum in a typical Baden-Württemberg high school compares to that of a school in the US. Before we try this intervention, we should (again) think not only about how much math students learn, but what math they learn.
It has long been true that men are diagnosed with dyslexia more often than women. This article (rather technical, by the way) offers one potential explanation: processing speed.
What is processing speed? It’s an unusually straightforward concept in psychology.
Imagine that I show you a piece of paper with several rows of different shapes. There might be a square, and then a star, and then a rectangle, and then a circle. And so forth.
To test your processing speed, I simply ask you to name all those shapes as quickly and accurately as you can. Or, I might ask you to say the colors of the shapes: the first one is green, the second is purple, and the third orange.
If you accomplish these tasks relatively quickly, you have a high processing speed.
Overall, women have slightly higher processing speed than men–especially in verbal tasks. The authors of this new study find that this difference in processing speed gives women an edge in reading fluency–and reduces the likelihood that they will be diagnosed with dyslexia.
There are no immediate teaching implications of this finding; however, anything that helps us understand how learning differences come to be…and, come to be diagnosed…might help us improve reading and learning in the future.
President Barack Obama greets 2010 Fermi Award recipient Dr. Mildred S. Dresselhaus, in the Oval Office, May 7, 2012. (Official White House Photo by Pete Souza)
If you watched the Oscars this past weekend, or simply had lucky t.v. timing over the past few weeks, you may have caught GE’s newest commercial featuring MIT scientist Millie Dresselhaus. The ad aims to promote GE’s upcoming diversity endeavor: 20,000 women in science, technology, engineering, and math (STEM) jobs by 2020. It’s a lofty goal, and I’m rooting for ‘em.
This initiative comes in response to only 18% of GE’s technical workforce being female. Although worrisome for both equity and economic reasons, this statistic is not unusual in the STEM student or professional world. We may be wondering: how has GE, and numerous other similar companies, achieved such low female employment and retention? Which is also to ask: what does it mean for women to persist in the STEM world, and what kind of internal oomph does it take? Luckily, researchers have begun to tackle both questions.
(It’s Not Just GE)
Fewer girls and young women engage with STEM at the advanced placement, college, and career levels than do males. A report published by the National Science Foundation (NSF) found that women represent only around 35% of college students enrolled in physics, mathematics, and computer science courses, and less than 10% of those studying physics and engineering at the graduate level. [1]
Also of concern is the high rate of attrition seen in those who complete undergraduate study and enter the workforce. This turnover leaves women holding only 22% of the math and science jobs available in the U.S.
A Couple of Questions Out of the Way…
Researchers have approached this gender disparity from different angles in hopes of better understanding what is happening.
Some have asked: is it perhaps so that males are just more able mathematicians than females? This conclusion seems unlikely, given that gender differences in math performance barely exist early in development, and tend to emerge at the high-school and college levels. [2]
Others hypothesized: maybe boys are just more interested in STEM than girls? This also seems a stretch, with research showing that throughout at least the elementary school years, a high percentage of both boys and girls (68 and 66 percent, respectively) report liking science. [3]
(See also my fellow-blogger’s post about raising girls’ levels of math participation in the US and India.)
A different approach, then. Several studies have gone beyond theories of disproportionate aptitude and interest and begun to question if social pressures and expectations affect girls’ pursuit of STEM. If yes, then how?
How?
In one such study, researchers at Brown University and Williams College studied the interaction of stereotype threat and mathematics performance in female students. [4]
Stereotype threat is a social occurrence whereby the targets of intellectual inferiority stereotypes, such as women or racial or ethnic minorities, perform more poorly at a task when in an environment that reminds them of this stereotype, such as in the presence of males or the racial or ethnic majority. [5]
In the experiment, college students completed a challenging math or verbal task in groups of three. Each trio included the study participant plus two others: two people of the same gender as the participant (the same-sex condition) or two people of the opposite gender of the participant (the minority condition).
The researchers found that women in the minority condition performed more poorly on the math test than did women in the same-sex condition (males did well on the test despite condition).
The female participants’ performance was also found to be proportional to the number of males in their group, such that women in a mixed-sex majority condition (i.e. two women and one male) still experienced performance deficits as compared to women in the same-sex condition.
Given their findings, the researchers suggested that women in STEM courses or jobs, where their colleagues are predominantly male, may experience stereotype threat. As a result, they are at-risk of performing below their ability, and thus at-risk for attrition.
In another study, students at two universities completed assessments of working memory and mathematics, as well as a self-report anxiety scale regarding their feelings about math. [6] The results statistically demonstrated a chain effect: female students had higher anxiety about math, which in turn affected the working memory resources they needed to complete the math tasks, which in turn lead to lower performance.
The researchers discussed this chain as lending support for Processing Efficiency Theory, which suggests that anxiety negatively affects the central executive component of working memory. The central executive is responsible for processing information stored in verbal and visuospatial working memory, both of which come in pretty handy when completing mathematics tasks.
There’s Probably More to Persist Through than we Think
Studies such as these suggest that our social environment likely has a large impact on how women navigate, among other things, the STEM world. So we ask: what engenders women’s worried feelings around mathematics? How are messages of inferiority transmitted? After all, most girls grow up with some combination of family members, teachers, and/or other role models reciting the message that girls can excel in STEM just as boys can. Yet we still see that, by age six, girls are more likely to group boys into the category of “really, really smart” than they are to categorize their fellow ladies as such. [7]
First, let’s not underestimate the messages that waft in the background of girls’ daily lives. For example, picture a middle school student sitting down at her kitchen table to work on science homework. She can faintly hear CNN from the living room t.v. as she works on a diagram of Newton’s laws of motion. And what CNN happens to be covering that evening is Nobel scientist Tim Hunt’s rationale for promoting gender-segregated workplaces, which is that women in science laboratories are too at-risk of falling in love with their male colleagues.
Collective groan…but so what? Surely CNN will move to another story, the diagram will be skillfully completed, and the student will clear her books from the table so that she can eat her dinner. But not so surely will that story’s message be erased from her subconscious. And that’s a big ‘what’.
Second, as adult women and educators, we should try to get in the habit of taking a look at our own emotional navigation of STEM. Again, let’s not underestimate: one study recently found that heightened math anxiety in female teachers at the beginning of the school year is associated with lower math performance over that school year for their female students. [8] And that anxiety is communicated much more subtly than seeing a math problem and making a run for it.
In other words: better understanding our own subconscious relationship with, and reactions to, STEM disciplines can help us better understand the implicit messages that we transmit to young girls.
Third, let’s talk about it. Now, I would be remiss not to include the caveat that we cannot fully encourage girls to pursue, and persist in, STEM without also considering the importance of encouraging boys and young men to pursue female-dominated fields, such as nursing and early childhood education. Nonetheless, researchers have suggested that efforts to mitigate gender differences in math-related fields are inadequate unless they target specific factors, such as worry about math, in girls and women. [9]
So let’s talk about that worry. And, given what we know about social psychological phenomena (e.g., the prevalence of stereotype threat), the positive effects of such conversations may be maximized within all-girls STEM classes and extracurriculars. A quick Google search can lead us to organizations, such as Girls Excelling in Math and Science (GEMS), the Laurel School’s Center for Research on Girls, and NSF’s National Girls Collaborative Project, that are eager to provide guidance and resources for exactly this purpose.
Because oddly enough, the best way to empower girls to brush off gendered nonsense like Tim Hunt’s argument for workplace segregation, may just be to separate boys and girls for a bit.
National Science Foundation (1996). Women, minorities, and persons with disabilities in science and engineering: NSF Publication No. 96–311. Arlington, VA: Author. [link]
Lindberg, S. M., Hyde, J. S., Petersen, J. L., & Linn, M. C. (2010). New trends in gender and mathematics performance: A meta-analysis. Psychological Bulletin, 136, 1123–1135. [link]
National Science Foundation (2007). Back to school: Five myths about girls and science. NSF Press Release No. 07-108. Arlington, VA: Author. [link]
Inzlicht, M. & Ben-Zeev, T. (2000). A threatening intellectual environment: Why females are susceptible to experiencing problem-solving deficits in the presence of males. Psychological Science, 11, 365-371. [link]
Aronson, J., Lustina, M.J., Good, C., Keough, K., Steele, C.M., & Brown, J. (1999). When white men can’t do math: Necessary and sufficient factors in stereotype threat. Journal of Experimental Social Psychology, 35, 29–46. [link]
Ganley, C. M., & Vasilyeva, M. (2014). The role of anxiety and working memory in gender differences in mathematics.Journal of Educational Psychology, 106 (1), 105-120. [link]
Bian, L., Leslie, S.J., Cimpian, A. (2017). Gender stereotypes about intellectual ability emerge early and influence children’s interests. Science, 355(6323), 389-391. [link]
Beilock, S. L., Gunderson, E. A., Ramirez, G., & Levine, S. C. (2010). Female teachers’ math anxiety affects girls’ math achievement. Proceedings of the National Academy of Sciences, USA, 107, 1860–1863. [link]
Ganley, C. M., & Vasilyeva, M. (2014). The role of anxiety and working memory in gender differences in mathematics.Journal of Educational Psychology, 106 (1), 105-120. [link]
According to new research, a key difference might be the choice of opponent. Whereas men typically prefer to compete against others, women often choose to compete against themselves.
(As always: be careful about oversimplifcation of gender roles. I myself am much likelier to compete against myself than others. As Todd Rose notes, averages often give us useful information about groups, but never about individuals.)
