It’s all about self-citation, says the Economist

I received a link to a piece in the Economist from a colleague this week. The lamentable lack of female professors discusses a recent piece of research by Barbara F. Walters which demonstrates that in the International Relations literature, there are a couple of clear gender differences: men cite themselves more than women, and men also cite other men more than women.

These facts can and should be interpreted in light of a few facts: men and women are differently represented at different types of institutions, and child-related career breaks have a disproportionate effect on the publication record of women.  But the authors have controlled for a wide range of variables – including career stage, type and quality of publication, etc.  They also point to the curious fact that women are four times more likely to collaborate with men, than men are with women.

If you are aware of the literature on bias, including the effect of your own stereotypes on your own behaviour, then this research is a useful demonstration of one of the mechanisms by which bias impacts negatively on the careers of women. However, the Economist’s simplistic take on these results reads quite differently: ‘women fail to win’; ‘women more often step off the career ladder to raise children’; ‘they thereby put themselves at a disadvantage’; ‘female academics are not pushy enough’; ‘unwillingness by women to self-cite’: all these phrases perpetuate the myth that somehow, it’s us.  Women are just doing it wrong.

The thing that really gets me about the Economist article, though, is not the incomplete discussion of the research or the lack of thought beyond proximate causes: it’s the dog whistle tactic of the mention of Larry Summers, and the not so subtle admiration of his daring in making the comments that he did.  The juxtaposition of this research on gender bias with the recollection of that particular episode is not accidental.

I know I don’t need to say it – but don’t read the comments.  On the other hand, the article by Professor Walter et al is very readable, and contains a lot more thoughtful discussion: for example, is the citation gap decreasing over time?*  Recommended reading for anyone interested in this stuff.

*Answer: it could be, but it could also be that there is a time-lag for the citation gap to show up.  More research needed!

Fallibility is not a failing

I’ve been thinking about why I started this blog, and my initial reservations: firstly, that I would run out of material once I had finished translating my original talk into blog posts.  The more significant concern was that I would end up overwhelmed and discouraged by the whole business – as rereading my first post has just reminded me.

So I thought I should write an update on progress, since it’s been about six months: writing this blog has made me feel much, much better about the stupid business of sexism in science.  Yes it is there, and it is real, and it does bad things.  So I do get grumpy about it sometimes.  But I also think that there is something hugely positive in spending time thinking about why science is sexist, and trying to understand what is really going on – it may be no coincidence that my favourite thing in all of science, is a well constructed thought experiment.  What would science look like, if the sexism were entirely eradicated?  It’s actually quite a nice thing to ponder.

More pragmatically, I have found that it gets easier to call people out on casual sexism, and to do this without labelling them as sexist.  It turns out this is a really helpful thing to be able to do! It also appears that people are getting used to me bringing these issues up, and even anticipating my comments before I make them.  This is no bad thing.

So what is my conclusion?  Own your biases, people!  It makes everything much easier, from deciding what coffee to order when out of town, to understanding that men who have worked in a male-dominated environment all their lives may not be choosing their language with the intent to exclude you. And it is kind of nice, explaining, and seeing them get it.

If there is such a thing as a tenet of faith in science, it has to be that fallibility is not a failing. Just something to learn from.

Grumpy about Gillard

I’ve been thinking about recent events over the Tasman a little bit in the last week.  I know very little about the difference between Julia Gillard’s and Kevin Rudd’s politics, and from what I can tell they are about equally well-behaved as politicians.  But the recent sexist attacks on Gillard do add an additional level to the story.

I have read well reasoned comment from people who are delighted to see the back of Gillard, and the nature of politics is such that there’s no need for any sexism to account for that.  But my problem now is the extent to which I can see that sexist outcomes occur despite a lack of directed sexism by individuals, simply due to the culture of the society that we live in.

Those who know the politics better than me will be able to form an opinion on whether the spill was sexist, or not.  But I suspect there are few women who would rule it out completely: it’s one of the fun games you get to play as a woman, this second guessing of the way in which you are judged by others.

I don’t mean to suggest that self-doubt is limited to women, not at all.  In fact, there is a lot being written now about imposter syndrome in science, and the need for openness about the emotional challenges we face, that is really reassuring.  But there is a woman-specific aspect to this, even if the only cause of it is our own subconscious biases.

A recent example from politics in New Zealand was the reaction to Nanaia Mahuta’s complaint that long hours at Parliament were difficult with a baby needing breastfeeding.  She got rather seriously slammed in the media for saying it, with the accusation being made that she was playing politics. How unwomanly. Using her family for political gain!  I was reminded of this by a recent comparison of Australian magazine covers highlighting that Rudd had brought his daughter into the campaign to combat the effect of his opponent’s three daughters…

Ugh. What’s good for the gander is clearly not good for the goose.

I wondered about writing about this on this blog, which I am trying to reserve for thoughts directly relevant to the situation of women in science. While sexism in science has some similarities with the sexism experienced by women in politics, or business, or who work as professional cleaners for that matter, it is also quite distinct.   But here’s the big thing that women in politics and science have in common: the regular and continual assessment of performance that scientists undergo in research funding applications, and politicians face in the polls, but more directly in the court of public opinion (whatever that is exactly).   Both of these seem to me to be related forms of peer review: politicians  have relatively massive peer bases, but the anonymity of the reviewers in both cases has the same effect.

If every single time you undergo one of these public evaluations you are playing the guessing game as to whether those evaluating you are sufficiently aware of the biases of the society that you live in, then…  the playing field is not level, even in the absence of overt discrimination.  So can we please get past saying that skewed gender ratios are okay because the women are choosing to leave?

To finish on a more positive note: a lovely analysis of Kevin Rudd’s suits, ties, and a couple of (tame) body parts.

Necessary evils

The difficulty in challenging boys-only reading clubs is that when it comes down to it, this is as simple a form of affirmative action as you can imagine.  In the face of a lack of reading achievement, this is a laudable initiative.  At least, if I am going to argue, as I do, that the reason that science is sexist is simply because it is sexist, in the sense of gender inequality tending to reinforce itself through the propagation of stereotype, then I have to be receptive to initiatives that challenge the status quo, right?

I think this is logical.  It is also the bit that makes me uncomfortable, and sometimes, rather queasy.  There is no way, as a woman in science, that I want to be in the position of arguing for special measures to promote women in science.  This makes me profoundly uncomfortable.  On the other hand, what are we supposed to do, in order to effect change?  Wait for men to make the case for us?

The only way to even approach the argument is to make it clear that it isn’t about you.  This is hard.  I had dinner with a departmental visitor, a few weeks ago, who casually tossed out “of course, it’s a lot easier for women academics” in the middle of a conversation about career options, while I had my mouth full.  I stared at my plate until the conversation took itself elsewhere.  Should I have said something?  I’m sure it wasn’t directed at me: the problem is, I have trouble forgetting off the cuff comments of this sort.

One of my favourites dates back a couple of years: I was on a committee discussing possible names to put forward for various research awards.  A woman’s name was mentioned; it turned out she had already been given the award several years previously.  “Oh yes,” opined an elderly member of the committee, “women always get these awards about five years earlier than they would if they were a man”.

There are things that you can’t unhear.

There was no evidence for anything of the sort.  And for once, there were half a dozen women in the room — but none of us said a word.  The mere possibility of affirmative action is sometimes used — I think unconsciously — as a way of silencing anyone who has the potential to benefit from it.  It may be unintentional, and it may be innocent — it is still a huge psychological hammer when you get hit with it.

So the problem with affirmative action, if I am to summarise?  It devalues the achievements of individuals.  It reduces personal satisfaction and motivation, and can lead to division and ill-will.  It hurts most the people that it intends to help.  And it may, for all that, be a necessary evil.

Sitting on the gender fence

A while ago, I tweeted the following:

(the tweeted link no longer works, but it is here)

…and had a friend respond with the (I imagine sincere) query: “what starts?”

I’ve thought about how to best answer this, and decided it merits a post.

The story was very straightforward really: school decides boys need encouragement to read more, and start a club to encourage them.  So far so good.  The club is boys only.  Here I wrinkle my nose and wonder if this is really necessary.  Then a child is quoted in the paper saying that the club is cool, specifically because it excludes girls.

Children say these things.  This is the world they are brought up in, and this is not so surprising, nor anything to get upset about.  The fact that this quote was what the story was built around?  That does bother me.

The reference to feeling ill was meant rather indirectly: not so much that the story itself was upsetting, but more that I knew in advance that commenting that I didn’t like it would be seen as overly sensitive. Let me put this straight: I approve of anything that encourages children to read more, though I have to admit that personally I never saw that it had much to offer as a social activity.  But if a club encourages children to read, then that is all to the good.

Why the need for gender segregation?

I remember being told, at age 11 or 12, the standard trope about single sex education: it results in improved outcomes for girls, and even though boys do better in co-ed schools, single sex schools result in better outcomes overall.  I wasn’t a hugely critical 12 year-old, and I received this piece of information with a straightforward feeling of pleasure to be on the right side of the gender fence, for once.

I have already referred to studies that show that performance can vary in the presence of the other sex.  So yes, I think it entirely likely that single sex education resulted in significantly improved outcomes for girls, in coping with and moderating a sexist society. The differentiated choices of girls at single-sex vs co-ed high schools in studying physics is a clearly demonstrated modern example.  But here’s the rub: just as it is impossible to have a level playing field in science in the absence of gender equality, it is impossible to know whether single sex education would have any positive impact in a world in which the sexes were effectively equal.

To put this all more bluntly: single sex schooling only puts off the time when women are forced to confront the fact that stereotypes are real, and that they do real damage.  Worse than that; is it possible that segregating the sexes actually does more harm through the reinforcement of stereotype that the good it achieves through insulation from reality?

I wonder: how do those 12 year old boys feel on being told that they are going to single sex schools for the sake of their sisters’ education?  Or do they get told a different story?

Sticking with initials

I had a minor moral dilemna last week.  Or perhaps it’s just a moment I’m not very proud of.

My most recent paper has been a real pain to get published.  It’s exactly the sort of science that gets me excited, but where to publish has not been obvious, and the existence of previously published work which is just plain wrong doesn’t help.  But this is nothing so strange, and my PhD student has been relentlessly positive and willing to learn from each revision how to strengthen the points that we are making.  So so far, no problem at all.

Last week we had reached the 3rd submission of this paper, and neither of us had any inclination to change anything further.  Yet I hesitated before emailing through the instruction to submit, and added:

“Oh, and change names to initials only – it is what I have previously gone with.”

I have always used my first initial, by preference — partly because I am lucky enough not to have an overly common last name, and partly because of precedent — I don’t recall ever thinking all that much about it.  But a week or so before, something had crossed my radar, and this caused a certain self-consciousness.  I come across discussion of women in science very frequently now, but so far, I haven’t had a lot to add to my previous posts.  Then I saw a recent discussion of the Matilda Effect (in different places, but probably first via Athene Donald on twitter (see her take here).

I’m not sure how much attention I would have given it if I hadn’t been in the middle of dealing with referees comments on the previous submission of my paper, which may have impacted on the way I read the following paper.

In The Matilda Effect in Science Communication: An Experiment on Gender Bias in Publication Quality Perceptions and Collaboration Interest, Knobloch-Westerwick and colleagues examine the effect of role congruity on our evaluations of others.  In a stereotypically male field, are men’s achievements valued more highly?  And are women equally advantaged in fields we consider feminine?

I won’t spell out the conclusions of the paper in any great depth.  But in a task where 243 communication scholars were asked to evaluate the quality of a conference abstract, those abstracts apparently authored by men were rated more highly than those authored by women.  The effect was stronger if the topic of the abstract was stereotypically ‘male’, rather than ‘female’ or gender-neutral.  This all seems unsurprising, given everything else — as for the Moss-Racusin study,  the men and women participating were demonstrably equally biased.

Probably the most surprising line in the paper was the follow-on from a summary of all the evidence demonstrating bias against women in science (or stereotypically male fields).  Framed as an exception, the authors stated that in fact women would “likely benefit from a blind peer review process”.  Leaving aside the fact that blind peer review is very uncommon (I can’t think of any examples in fact), this supposed advantage for women is actually just a statement about what might level the playing field.

So should I have instructed my student to change our names to initials only?  I don’t know.  Given that I have always used initials, I don’t see why not.  But I do know that now, I feel weirdly guilty about it.  Am I obliged to always be visibly a woman?  I can see how this is something that I could learn to resent…

The super short version

Infographic on women in science.  It’s cool.

Why might you want to disbelieve the Moss-Racusin study?

An interesting thing happened, when I started talking about gender equality in science for a sustained period of time, with a broad range of people in the scientific community.  Consistently, the same issues were brought up as reasons to disbelieve, or to query the main conclusions, of the Moss-Racusin study.

A contender for first place would be this:

“The numbers show that the women are worse!  They are more sexist than the men!”

So no, according to the authors, the difference between the behaviour of the male and female faculty members is statistically insignificant, making them equally sexist.  But this is something that people seem to want to believe, often coupling it with a story about a particular woman who has made it to a senior position in science, but isn’t very nice.  She has done this by stomping on others, actively discouraging the promotion of less senior women about her, and so on.

I have very little time for this point of view, though I sort of see the appeal of the argument.  I can see that it feels that way, simply because we tend to expect more of women who are in these positions.  (It’s the small numbers problem again.)  There is a lot of literature about the difficulties that women have in making a go of leadership positions, not because of any inability to perform the job, but because of the way that they are judged for doing the job:

“If they act consistently with female stereotypes, their competence is questioned. But if their behavior is consistent with that of the stereotypical male, they’re viewed as being too tough.”

From here.

Another thing that was brought up more than once, was the view that the problem has gone.  It’s in the past, because

“In my research group, there are 10 girls and I’m the only man”

From a postgrad.  It’s not a perfect quote but the use of ‘girls’ and ‘man’ I remember.

Or else people — usually not biologists — feel the need to point out that in biology, in contrast to the physical sciences, women outnumber men.  True at undergraduate level often enough, and even at postgrad: but that’s not what we are talking about here.  At senior levels in biology, the women disappear very, very quickly.  I get that there is a need to point out that things are improving, and yes, if you compare to 50 odd years ago, I think this is true.  But have we seen progress in the last 20 years or so?  I think maybe not, with the exception of where specific programmes and policies have been put in place to effect change.

On the web, people are more forthright (or perhaps they know less about science that the people who’ve been to my seminars; dunno)

So, there’s this

or this comment,

or this comment on why the gender pay gap is unfair to men:

“You could look at it that way. You could also ask the question, “Why are so few men getting degrees in teaching?” And I’d say that a large part of that is because most K-12 teaching jobs pay pretty poorly, and men in our society are under pressure to get jobs that can provide for a family. So men who might be interested in teaching end up going into some other field where they can get better pay. Gender biases don’t just hurt women.”

Or this one, which echoes the point I made earlier:

“Look, I’m about as much of an open-minded, equal rights supporting liberal as you’re going to find most of the time, but there comes a time when making such excuses stops being convincing. My biology classes in college generally had more women than men, but I would never consider complaining about that, or claiming there was a gender bias.”

It’s a complicated issue, and by that I don’t mean to wave a magic wand over the different opinions that people have on (a) the existence of gender bias, and (b) the importance of it.  But I don’t think that casting doubt on the validity of this particular study gets you anywhere; yes, lots of people make different career choices, yes, people who are sufficiently self-aware are probably somewhat less sexist than the average.  But the results of this study make an awful lot of sense when compared to the statistically meaningful and persistent gender bias in science.  Maybe it can give us an idea or two of what to do about it.

On Bias

So what is bias, and why do we have it?  Well, the analogy that I like to use looks like this:

I drink black coffee.  I live in Wellington at the bottom of the North Island of New Zealand.  We have good coffee.  Everywhere.  So I order long blacks, and I expect them to arrive shorter than long and covered in crema.

I used to live in Auckland.  I know Auckland, I know coffee in Auckland.  I know what it is like in general: specific cafes have changed a lot since I lived there.  When I visit Auckland, I order flat whites.  This is practical; a milky coffee conceals a number of sins (though by no means all!) in the roasting and grinding and extraction of espresso.  Thus, by ordering a flat white, when I don’t have any particular knowledge about the competence of the barista, I minimise my chances of disappointment.

Am I shallow?  Perhaps I am shallow.  But the analogy is this: if I rely on the stereotype that it is hard to get a decent long black in Auckland, I manage to order coffee pretty well successfully, most of the time, with a minimum of effort.  If I care to put the thought into checking out the brand of beans, the machine, and what is being delivered to the person next to me, I might choose to order a long black if that is what I really want.  But that takes some doing.  So the take away message is this: stereotypes are useful.  They are time savers!  Our brains rely on them for all sorts of things.  But they can also deceive.

And when it comes to people?  Whether it is judging the heights of men and women, or the relative athleticism of different ethnicities (sorry for paywall), we all have a lot of experience with people.  This puts us at the mercy of our stereotypes.  If you don’t admit to them?  Well, you can get yourself into a merry knot, trying to explain away your actions.

How about this?

How much damage could that do, do you reckon?  But it’s actually worse than that.  The factor that may matter most, is not other people’s biases against you, for whatever reason: it’s that you may be biased against yourself.

Women perform worse in mathematics tests if they are reminded before taking them that they are women.  This is stereotype threat, and I can tell you, as a woman: this is a hell of a lot scarier than any worst case scenario of being discriminated against by others.

Yes, that is right: your own head has it in for you.  At least, it does some of the time.  If you’re a woman, it may give you the warm fuzzies when you start to sew…

Why does this matter?  Well, this is why we talk about the importance of role models.  The PNAS study, that showed the existence of bias in scientists: that doesn’t demonstrate conscious discrimination.  Unconscious, yes.  But the important thing, to me, is not the discrimination:  it’s the connection between discrimination, and the choices people make,  if I can hark back to Larry Summers’ three things.  Could it be that these factors are not so independent?

So when you see questions like these:

“Are sexist attitudes still to blame – or is it a fear of being thought uncool?“

It might be worth questioning whether or not this dichotomy is valid.

What do you mean, we’re all sexist?

Yes, I mean I’m sexist.  I try not to be, but I know that deep down I am.

But this isn’t about me (whew!).  This is what it’s about: a study of real people, showing unequivocally that changing the gender of a job applicant from female to male improves their chances of being hired.  Not just because they’re male, of course; purely because they are more competent and easier to mentor – as judged from a c.v. where only the name (and consequently, gender) is changed.

Their increased competence, as illustrated by the all important gender of their given name, confers certain financial advantages: a salary ~12% higher than would be offered to the woman.

The people making these judgements?  Faculty members at American universities, in Physics, Chemistry or Biology.  Both women and men, and here’s the punchline: both men and women are equally sexist!

Hooray for gender equality!

Read the paper (it’s open access!) if you are interested in the statistics.  If it really bothers you that the y-axis on the bar graph of salary doesn’t start at zero, send a friendly email to the authors.  One thing that talking about this piece of work has taught me?  Physical scientists can be really nasty about social science:

“This just seems to cast doubt on the scientific rigor of gender studies. If women social scientists are behind these studies and this is how women do science, maybe this is why women are not in the higher levels of physical sciences and engineering.”

(in comments on this assessment of the paper).

So yes, it is only one study.  It has limitations.  It does not contain the wavefunction of the universe, nor any discussion of what you would do with it if you had it…

What it is, is consistent.  With studies on the effect of race: Are Emily and Greg More Employable than Lakisha and Jamal?; and on the effect of gender on employment outside science: Constructed Criteria: Redefining Merit to Justify Discrimination.

There’s a lot more literature out there, if you start looking.  Most of the studies are done on postgraduate students who participate in these studies for credit in their programmes, which is why it is significant that the PNAS paper demonstrates bias in actual people who make actual hiring decisions.  But what struck me about the second of these papers, was a  small detail: the title refers to the fact that when hiring a police chief, the male candidate is preferred, but the criteria which are assessed as being important for the job change depending on whether the candidate is male or female, in order to justify the decision.  If the participants are asked to state clearly their criteria for the job beforehand, a significant part of the bias goes away.

Then there is this:

Participants who rate themselves as highly objective exhibit a significant amount of bias.

Participants who rate themselves as not so objective appear to make relatively unbiased judgements.

The value of self-awareness, anyone?