I was browsing through the back catalogue of Language Log today (because I enjoyed their latest releases, so to speak), and I found an article from October 2010 describing a study that purported to provide evidence to the effect that women are more apologetic than men. The paper, by Karina Schumann and Michael Ross, was described in detail by Mark Liberman in his post so I won’t bother to reiterate what he said in too much depth. In a nutshell, the researchers had thirty-three female and thirty-three male participants complete daily online questionnaires related both to instances that occurred each day in which they were transgressors of apology-worthy acts and to instances in which they received apologies for others’ transgressions. They found that members of the female cohort reported being subject to more apology-worthy transgressions, receiving more apologies, committing more apology-worthy transgressions, and giving more apologies than the males. They suggest that this is because males have a higher threshold in determining which actions and behaviours are worthy of apology–in other words, it will take a more egregious transgression to prompt a male to apologize than it will for a female.
I don’t have any particular qualms with the study itself, and I’m inclined to believe that the study was conducted as carefully as Dr. Liberman says it was (he knows more about statistics than I do, after all); it appears that among the participants they selected, the females really did exhibit a lower threshold for apology-worthy actions than the males did. And it’s very possible that the authors’ speculations about the sources of this disparity, namely “that women might perceive more offenses because they are more focused on the experiences of other people and on maintaining harmony in their relationships,” or “that men have a higher threshold for both physical and social pain,” could be the case. My problem in particular has to do with why we find this kind of study interesting, and what we hope to do with the knowledge that we derive from it.
If we compare this study to a hypothetical study that includes blondes versus brunettes, or black people versus white people, or people from San Francisco versus people from New York, we might find that the San Franciscans, the black people, or the brunettes apologize more than their counterparts, and we could speculate that San Franciscans are more apologetic because they have a lower threshold for physical and social pain than New Yorkers, or because they are more focused on the experiences of others and on maintaining harmony in their relationships. But why, when we search the scientific literature for research dealing with differences between groups of people, do we find a paucity of articles comparing New Yorkers to San Franciscans or blondes to brunettes? What is it about men and women that makes people so motivated to systematically codify all the differences between them? I don’t know the answer to this question, but I imagine that it’s either a response to or an assertion in favour of the idea that men and women should be differentiated from one another. And a political will apparently exists to promote this type of differentiation–a research study comparing New Yorkers to San Franciscans would likely be rejected by NSERC or its American equivalent as frivolous, but a study comparing men to women is worth a few thousand tax dollars.
Liberman acknowledges that studies dealing with differences between groups generally have a bad track record in terms of their role in perpetuating harmful stereotypes. He hints at this by noting at the end of his post that the study does not include information about within-group differences so that a comparison can be made with the across-group differences. This track record arises from a tendency for people to slide into generalizations about the group members. When looking at populations, statisticians can say that on average, this population is so-and-so, even while acknowledging that the individuals within the group are different from one another. But when looking at results of a well-done and careful scientific study that talks about men and women as discrete groups, it’s still hard not to apply the results, which refer to an imaginary entity (the average or prototypical member), to all of the members as individuals. Schumann and Ross do this, to some extent, when they speculate about the causes for their findings. When Liberman suggests that it would be good for the papers to pick up this story because it’s not the usual stereotyping fluff, he overlooks the role of this article and others like it in producing a conception of gender that posits men and women as categories of people with more across-group differences than within-group differences, and this conception of gender is what makes gender stereotyping possible (just like it makes it possible to stereotype blondes or New Yorkers). Liberman helpfully indicates as much in another post of his that he links to as part of a warning against misinterpretation of generic plurals:
Most members of the general public don’t understand statistical-distribution talk, and instead tend to interpret such statements as expressing general (and essential) properties of the groups involved. This is especially true when the statements express the conclusions of an apparently authoritative scientific study, rather than merely someone’s personal opinion, which is easy to discount.
His injunction against “the use of generic plurals to express statistical differences” was presumably intended for discussions of statistical differences in public arenas, because the article in question–which, remember, he would have liked to see picked up by the press–uses generic plurals with aplomb: “The diary data suggest,” the authors write, “that women offer more apologies than men do […].”
When I talk about “discrete groups,” I’m referring to the problems of clearly delineating who is a man and who is a woman. A hundred and fifty years ago scientific researchers had no problems differentiating between black people and white people, but nowadays we’re more aware of the fact that categories of race aren’t as clear-cut as they claimed they were (even at the genetic level). A lot of work has been done to show that the same is true with categories of gender and sex, but this work is hampered by actively malicious treatments of sex and gender differences (like what Liberman refers to as “gender-stereotyping fluff”) as well as by uncritical acceptance of gender as a set of two distinct categories (like the study in question). Both types of research make gender stereotyping possible, because they are both used for that purpose. Clearly, this study, as much as it is exemplary in its execution, is still treading on the wrong side of the problem / solution divide.