Listen to this PodcastMore 60-Second Science
-
Gravity's Engines
We’ve long understood black holes to be the points at which the universe as we know it comes to an end. Often billions of times more massive than the Sun, they...
Read More »
Why aren’t there more women physicists, and in senior positions? One factor may be unconscious biases that could keep women physicists from advancing—and may even prevent women from going into physics in the first place.
Amy Bug, a physicist at Swarthmore College, examined the bias question.* Her research team trained four actors—two men, two women—to give a 10-minute physics lecture. Real physics classes watched the lecturers. Then the 126 students were surveyed.
When it came to questions of physics ability—whether the lecturer had a good grasp of the material, and knew how to use the equipment—male lecturers got higher ratings by both male and female students.
But when asked how well the lecturer relates to the students, each gender preferred their own. And while female students gave a slight preference to female lecturers, male students overwhelmingly rated the male lecturers as being superior. The research appears in the journal Physics World. [http://bit.ly/b3ctOj]
Bug says the results may be evidence of inherent biases that could hold women back—along with economic inequalities, such as lower wages and smaller start-up grants. Which reduce career acceleration and thus the amount of force available to crack the glass ceiling.
—Cynthia Graber
*Correction (8/3/10): This sentence has been edited to correct an error conveyed by this podcast. Swarthmore College was originally identified as Swarthmore University.
[The above text is an exact transcript of this podcast.]
See what we're tweeting about
34 Comments
Add CommentIt is not "inherent biases" that cause students to prefer men speakers over women. Because of the high pitch and the softness of a woman's voice, it makes them very difficult to understand unless they are right up in your face, and the high pitch of a woman's voice makes their lecture seem more like a reprimand. A man's voice has more base in it and their words are clearer and easier to understand. It is not bias, it is the length of the sound wave.
Reply | Report Abuse | Link to thisI guess I have only 1 question... So?
Reply | Report Abuse | Link to thisWell, following your logic, then, the pitch of a female voice (which also generalizes all of the female biological sex as having high-pitched, soft voices and the reverse for males) would cause every student in every subject, not to mention every audience for every kind of public speaking situation, to always prefer hearing male voices. You are arguing, in essence, that no one prefers to hear a woman's voice in public speaking, ever.
Reply | Report Abuse | Link to thisThe conclusion assumes that the male presenters actually were not more competent or confident in their presentation - highly subjective, especially given the sample size of 4.
Reply | Report Abuse | Link to thisJamesDavis is just being a troll. Please don't feed the trolls. I do agree with Rhoark on this one. Not all actors are of the same caliber, and familiarity with the subject would make the job easier for any one of the actors to make a convincing performance. With only four actors, I don't think the given conclusions could be considered reliable.
Reply | Report Abuse | Link to thisDtraversscott, if I could hug you right now, I would.
Reply | Report Abuse | Link to thisI do not think that Cynthia reflects an attitude problem toward men in the slightest; she is merely bringing attention the fact that women in physics are not treated in the same ways, however vague they may be.
Rhoark - Good point.
Reply | Report Abuse | Link to thisIf none of the four actors were able to deliver a performance convincing enough to fool physics students, the student's preferences may have been based solely on physicist stereotypes. Did any happen to resemble Albert Einstein?
Seriously, if there is a general gender bias for male physicists, it may be related to the highly structured and regimented methods of physics, favoring highly established conservative approaches to all issues.
Regardless of any knowledge or abilities, physics instructor would be a difficult role for Carrot Top to pull off.
I had a female physics instructor in my senior year in high school. She was an idiot.
Reply | Report Abuse | Link to thisI also had a few female physics classmates in college, and they were some of the smartest people I've ever met.
As far as I care, it's about the performance, not the delivery.
As a female grad student in physics, I find this study fascinating, but I doubt that the listeners could adequately pinpoint why they liked one presenter over another. I think there are a lot of effects involved, from JamesDavis' apt observation that vocal pitch tends to affect listeners in different ways, to the fact that gender stereotypes and parental influences will affect how people react to presentations from people of each gender, which in turn likely affected how the actors presented the material in the first place. It also could have been a fluke that the males were better at presenting their material, as was pointed out by several people commenting about the sample size being so small.
Reply | Report Abuse | Link to thisI wouldn't say that this study *explains* the gender bias in physics at all, rather that it *reports* on some of the effects of society's general bias towards the perceived authoritativeness of males. The overall gender bias in the field of physics is due to gender bias in people's upbringing; it wasn't all that long ago that women were encouraged towards careers that were considered "more appropriate for females". The effects of this early gender-biased influence is still slowly trickling away, leaving a residual imbalance of female role models in the field. This imbalance is not CAUSED by any CURRENTLY perceived difference in ability, but it is still in the process of being rectified.
Poor fems! Forever looking for a glimmer of difference that can be interpreted as bias. If gender is a bias then should we trust the tests and selections of Amy Bug. The reporting of Cynthia Graber?
Reply | Report Abuse | Link to thisWhere is the positive side of this issue? Certainly there must be more than four successful female physicists who could be identified and presented as a contradiction to any potential bias.
Has the researcher investigated other subjects? The said
Reply | Report Abuse | Link to this"bias" occurs in physics more significantly than in other subjects? I found it very hard to draw the author's conclusion from the evidence presented in the article.
Clearly this study does not effectively show the presence of bias in student preferences of physics lecturers. However, it at least attempts to highlight some of the remaining gender imbalances in higher education: namely that women in the sciences, mathematics and engineering are fewer in number and receive less pay and accolades than their male counterparts.
Reply | Report Abuse | Link to thisAs a non-feminist male, I think that it is quite relevant to consider whether unconscious (or conscious--i.e. when I was in high school I believed men were simply better than women at math) biases are reducing the opportunities that women have in these fields. Moreover, I think this sort of article is a refreshing change from the "men have more variation in science aptitude" explanation for these trends (an argument which seems nothing more than a misogynistic, non-scientific rationalization of what remains of the male-dominated status quo in the hard sciences). So let's have some real research done to determine whether the gender divide still needs to be rectified, and if so, what should be done to overcome it.
Physics was long held to be a highly professional and elite field, and was once practiced by aristocrats and nobles (ie, gentlemen). Ergo, for a long time, women were not even considered qualified to study it.
Reply | Report Abuse | Link to thisThere is a story of an incident at one of the old British universities (either Oxford or Cambridge) where a female graduate student in physics was one of two people to arrive for a lecture on a particular day (the other was a male). The standing policy at that university was that if only one student arrived for class, the class would be cancelled. The tale has it the professor saw his one male student and one female and stated, "As there is only one student present, I will not be conducting class today."
Interesting, I feel that men often mumble and the women's higher pitch is clear as a bell to me!
Reply | Report Abuse | Link to thisAmazing, no matter how much evidence is presented some people are still in denial about the unfair structure of this paternal society. I don't think you can get any less biased than training actors and giving them a script. Are you implying that even actresses are less confident and comptent than actors?
Reply | Report Abuse | Link to thisHistory has proven that people of lesser intelligence are usually the ones who have to resort to insults since they are lacking the ability to express themselves.
Reply | Report Abuse | Link to thisRoger that!
Reply | Report Abuse | Link to thisIt is really tough to remove 100's of years of gender bias; i.e., burning women at the stake for trying to learn and practicing science. It won't occur in 10 years or 50 years. In the same way as we have been battling racial biases for many decades. We can't even start attacking this problem until we agree that we have one. To me the discrepancy in gender representation and salaries is enough proof that we need to seriously look at this. Problems can always be explained away by picking apart little details until nothing is left over. This is a very well known strategy used in criminal cases. I believe one or two murderers got off the hook through desensitization. To all you critics, I hope that you will have a daughter or sister who helps you feel the pain that women physicists go through on a daily basis. Thanks for reading!
Reply | Report Abuse | Link to thisI agree, gender bias whether overt or unconscious in physics is both an issue, not always an issue, and a result of past discrimination that is in the process of fading.
Reply | Report Abuse | Link to thisBut a sample size of four, no matter the quality of the actors, is just so small!
Just to offer the counterpoint, since no one else has:
Reply | Report Abuse | Link to thisIt is foolish to suggest that males and females have identical capabilities. There are obvious hormonal differences that do affect thought as well as behavior. As I understand, there are also brain physiology distinctions between the sexes that produce statistically significant intellectual abilities. While statistically there is a large overlap between the two groups, outliers with exceptional abilities may produce greater abilities for one sex or the other in certain subjects.
While it has often been inferred that girls are not as good as boys in math and science, this conventional wisdom had become self-fulfilling prophecy, limiting opportunities for capable females. Sexual bias should not be installed as formal or implicit policy, but biological sexual distinctions cannot be completely ignored.
Hi, Amy here. I am trying to stay out of this exciting debate. But I had to jump in and say that I completely agree with you. 4 is small. I wish it were 40.
Reply | Report Abuse | Link to thisBut there is precedent for this kind of study in social psych. One tapes different actors engaged in the same activity. They rarely use more than a couple of actors - it is so tough and so costly to train them to perform in "identical" ways. (Truly identical is impossible of course - you and the actors do your best.) There is a some film on the PhysWorld website, so you can see how I did.
It also costs to recruit subjects, and this study was done on a shoestring budget. Bottom line, even though n=4 is small, and N=126 is small too, it was deemed fine by the social scientists we consulted before we began. Upon analysis, we found the data indicated a real effect. (p < 0.05 ... significant, by a popular definition in that field.)
I won't claim any more. I dislike scientists who over-claim just as much anyone. We set up an experiment, did it, gathered data, and ANOVA said that there were statistically significant effects. Moreover, the effects were consistent with a theory of implicit bias.
Did we "prove" implicit bias was operating? Hardly! Not even a part of our experimental design. Descartes himself pointed out that there is an inevitable gap between data and theory. But we did what we did and we got what we got, and it is surely plausible that implicit bias theory might be relevant.
Lastly: It isn't good news to me that gender was a main effect in evaluation, and that male and female students had statistically different responses to the same videotape. Gender matters to how students evaluate their professor in a field as objective as physics? That's bad. If my gender matters in evaluation, does it also affect how much my students can learn from me? That's even worse.
But should I suppress data or misrepresent them even if I don't like them, even if I know I'm going to be flamed? (I knew.) That isn't right. Can't do it.
Thanks to you for commenting, writelhd, and anyone else reading this.
If I recall my self taught statistics correctly, the 4 actors posing as lecturers represent the sampled variable's potential values. The number of samples would be the number of students surveyed. Statistical procedures would return indication of statistical significance depending on the number of students surveyed.
Reply | Report Abuse | Link to thisThe failure of this test revolves around the selection criteria. There are far too many uncontrolled variables contributing to instructor preference other than gender. Some students may have selected an instructor, subconsciously, on the basis of wardrobe, others by language accent, others by hair color, etc., etc. As described, the test apparently does not ensure that selection is made solely on the basis of the instructor's gender. As a result, the conclusions are invalid.
BooRadley - Regarding a glass ceiling limiting career opportunities for females, while there may be some gender bias in student evaluations, perhaps other factors are considered in promotional selections. Certainly the government mandated positive bias for women and minority races offers a significant advantage in initial employment, at least (I'm not that familiar with personnel policies, etc., & its been 30 years since I applied for a job). There are many well qualified males that have not been able to find employment in their field of expertise.
Reply | Report Abuse | Link to thisRegarding gender bias in general, unless a gender neutral biological unit becomes available, the fundamental distinctions between males and females may produce same gender communications benefits that cannot be eliminated by training or positive bias.
However, many foreign language speakers who do not as clearly communicate with native speakers can become very effective professionals. That would seem to indicate that natural biases can be overcome in a professional work environment.
Hi jtdwyer,
Reply | Report Abuse | Link to thisLike you, I am also self-taught in statistics. I defer to my collaborator who is a social psychologist and has this stuff under control. But I wonder if we are just claiming different things.
Every experiment has uncontrolled variables. That is a given. It would be wrong, as you say, to claim that "selection is made solely on the basis of the instructor's gender". Indeed, "gender" is a fuzzy category involving many of the other characteristics that have been mentioned (clothing? pitch of voice? mannerisms?)
When folks do an experiment like ours, they create categories (man or woman , 1st tier vs. 2nd tier university, honors vs. nonhonors student) and see if they affect the mean ... a "main effect". This is a time-honored method of social psychology. The null hypothesis is that a variable has no effect on the mean result. You then prove or disprove within confidence limits. One of our categories was lecturer identity. We only had 1 of each identity. That's even worse than having 2 men and 2 women! Doesn't matter, you can still do ANOVA. It yielded an effect. The better ranked individual was significantly better than the least ranked individual. Why was Greg ranked more highly than Denise? Great question. No great answer.
How shall we interpret our results? We were motivated to do this study by other studies looking at evaluation of women and men. We can only say that our data *supports* results of certain other studies, and *supports* certain theories. No proof. Not possible. Can't prove a theory. (Not even with dozens of lecturers ... as was done in meta-analysis of real world science teaching.)
One can question the whole paradigm of experiments like this. You want to. I respect that. I can only reiterate that we did something quite standard in social sciences, and it yielded a violation of the null hypothesis . To be honest, that even just 2 people saying the same words and writing the same equations would get statistically distinguishable scores on questions like "He/she knows the material" ... that is the simple fact that astounds me about our experimental results.
It would be amazing if someone had the time and funding to repeat it with dozens of lecturers, all trained to recite the same 10 minute physics monologue, word for word. I suspect though, that even if the null hypothesis is violated in that case, it would remain a very controversial result.
Thanks for your comment!!!
BooRadley - Hi, Amy. You're correct (LOL) - I am generally skeptical of the standard methods of social science, and some other sciences as well. I'm sure I won't be changing any of that - good luck in your chosen field.
Reply | Report Abuse | Link to thisIt does seem pretty scary that there is so much gender alignment in intellectual pursuits. It took me many years to realize that others did not necessarily have the same perspective that I do, and decades longer to realize that the methods of reason employed by women were even more different than other people. Having a daughter and granddaughters helped me with that. Keep up the good work!
Hi jtdwyer,
Reply | Report Abuse | Link to thisThank you so much for your reply! I'm having the same experience - as I get older I'm getting to realize that there are lots of thoughtful people with valid perspectives out there - my view of the world is just one of many. I wish all the best to you, and to your daughter and granddaughters. (I have sons.) -Amy
Just another example of hustling data to meet our own paradigms or conception of what the world should be. Truth is that since no two people are alike then all people are naturally biased. Good, bad or indifferent this is a basic truth that begins at born and develops as experience grows. The data itself probably has little to do with the instructors themselves. Whether it is vocal inflection or male or female, from birth children react differently to their fathers than their mothers. This trait is carried throughout their adult life. Exploiting this characteristic based on biased empirical evidence to support cheap political gain is an affront to true science and a full understanding of the world around us and how that world operates.
Reply | Report Abuse | Link to this"Please grant me the strength to change the things I can change, the serenity to accept the things I cannot change and the wisdom to know the difference." The wise move for this 'study' is either to expand it to include additional parameters or accept the basic fact that people are biased and how to best work with that fact for the betterment of all.
I think it's important to distinguish between "this study would be more convincing if it had been done a different way" comments and "there is no gender bias" comments. The former are worth considering. Yes, I agree more actors would increase credibility. Yes, I agree controlling for pitch of voice would increase credibility. As an alto who has had male physics students tell me they were upset because they couldn't "get to me" during my biochemsitry class, I agree that there is a problem that this study is addressing. Whether it is addressed successfully is a separate issue.
Reply | Report Abuse | Link to thisLaura Hoopes
Female Biochemistry Professor - Yeah, I got an axe to grind with physics, too - especially astrophysics!
Reply | Report Abuse | Link to thisRoark is wrong. The are 126 observations, which is plenty (the law of large numbers kicks in at 30). By that logic, observing the sun every day of the year would yield -- one observation :). The result doesn't surprise me; it is the flip side of the same stereotype that inhibits new parent/couples from hiring guys to babysit infants. We are making progress, since the guys as new parents perform about 2/5 of all childcare (see my Monthly Labor Review piece, Oct 2009), but Physics is a tougher nut to crack. I hope we get to equality, and this study helps us understand the continuing barriers.
Reply | Report Abuse | Link to thisThere are not 4, but instead 126 observations (by that logic, observing the sun every day for a year would yield -- one observation :). The result doesn't surprise me; it is the flipside of young, het couples not hiring male sitters for their infants. But things are changing; guys in those couples perform 2/5 of all childcare (see my October 2009, Monthly Labor Review article), and presumably things are equal, on average, among gay and lesbian couples :). Physics is a tough nut to crack, and the study helps to explain why. Thanks!
Reply | Report Abuse | Link to thisdrago - You passed the stats quiz (see my comment posted at 08:39 PM on 08/04/10). However, its unlikely that the two genders (and all sub-variations) will ever be treated identically, since they are fundamentally different. While that may seem unfair it would be unreasonable to ignore those distinctions. While we all vary individually, there are definite biological distinctions between gender populations that include brain chemistry.
Reply | Report Abuse | Link to thisThese distinctions can be clearly demonstrated by the high percentage of females who emit an 'ahhh' at the sight of a baby and the large percentage of mature males who become somewhat aroused at sight of a woman's breast. If the biometrics of each group were also monitored it would be demonstrated that these responses are biological, not just not social adaptations.
That being said, there's no reason that social conditions not be applied equally and fairly to all population subgroups. I can testify as a single working father having independently raised my youngest son that there's a lot of gender bias in the social support for 'working mothers' (automatically excluding males in identical circumstances). There's no basis for such social distinctions.
On the other hand, there is a reasonable economic basis for distinguishing between male and female employees, since females have a much higher incidence of pregnancy, This distinction could be eliminated if all employees were required to take an extended paid leave when a child is born.
Unfortunately, all this concern for equal treatment by society become moot if society collapses due to overpopulation increasing resource consumption and global warming disrupting resource production and established infrastructure. Sorry for being so pessimistic, but I think even social sciences have ignored this potentiality for far too long.
B y the way, having received directions to an unknown location from an intelligent lady yesterday, I'm reminded of this article. She, as do most females, attempted to relate instructions by following an instructional path of landmarks and turns, while I, like many males, attempted to envision spatial map coordinates relative to some known location. IMO, the genders are fundamentally different in specific intellectual capabilities, not specific academic disciplines.
Reply | Report Abuse | Link to thisI'd consider this study a lot more significant if they got these results with a group of 40 or 400 actors impersonating Physicists, but not 4. With so few, there is a legitimate possibility that the male lecturers just happened to be better, or maybe the women were having off-days.
Reply | Report Abuse | Link to thisTry it again, with more people this time.