Wednesday, January 12, 2011

Three-Dimensional Mental Rotation

This is the fourth post in my series concerning the science of sex differences. The earlier ones can be found here, in the order of the links, including an introduction to the whole series.

This post, and the next one are about the "big guns" of those who advocate innate differences for why women do worse in mathematics and sciences, "worse" being interpreted in various ways, such as lower average test scores, smaller extreme tails on those tests, fewer women entering engineering or science and so on.

Three-dimensional mental rotation is a common test, and one which consistently shows gender differences. Men and boys do better on it, on average, than women and girls. It may be one candidate (together with tests of some verbal abilities in the opposite direction) for a biological sex difference on cognition. On the other hand, we don't test newborn babies on this ability, and neither have we controlled for the influences of the environment before such tests are interpreted.

Let's take one step backwards to notice that tests like this one are tests, not actual evidence for the innateness of whatever is tested, only evidence for differences in average test scores. Let's also note that another spatial test which shows gender differences, the object-location-memory test, gets very little attention, perhaps because it shows a consistent female edge.

Back to three-dimensional mental rotation. The test usually consists of pictures, one showing a structure comprised of building block-type cubes at various angles, and a group of other pictures showing how that initial structure might look after it was rotated some amount of degrees. Only some of that group show the correct outcome, and the task is to match those to the original picture. Here is an example of the test:

Building blocks! Which are now stocked in boys' aisles at toy stores. Lise Eliot (Pink brain. Blue Brain. p. 218)on this topic:
Whether it emerges in infancy or late preschool, the sex difference in mental rotation is known to grow larger through childhood and adolescence. Spatial skills, like all mental tasks, improve with practice, so it's likely that boys' mental rotation ability is honed by the many hours they clock in spatial pursuits. Boys begin with trucks and blocks in pre-school and continue with their games of catch, hoops, and all manners of virtual driving and shooting games, so it's not too surprising that the sex that spends most of its leisure time watching objects flip, rotate, and fly through space ends up a lot better in mental rotation.
Though I thought that girls are every bit as likely to play catch and hoops, Eliot does have a point. In fact, the birth of video games (such as Tetris to which I was addicted some time ago) may work to further strengthen this difference because most games are for boys.

On the other hand, something fascinating has cropped up in that context: A very short intervention seems to have the power to change the mental rotation scores of women. Rebecca Jordan-Young (Brainstorm, p. 287-8):
Feng, Spence and Pratt (2007)identified a basic information-processing capacity that underlies spatial cognition and showed that differences in this capacity (the distribution of spatial attention) are related to differences in the higher-level process of mental rotation ability. They then showed that a remarkably brief intervention -- just ten hours of practice with an action video game -- caused "substantial gains in both spacial attention and mental rotation with women benefiting more than men." The ten-hour training did not completely eliminate the sex-difference, but it came extraordinarily close -- the mean scores after training were no longer statistically distinguishable between males and females.
This suggests that the mental rotation test is amenable to learning. What it measures can't therefore be defined as something which cannot be changed. Yet that seems to be the implication of those who explain the scarcity of women in science and engineering by the results of the mental rotation test. The underlying concept must be that the test measures something innately different and stable.

What has always struck me about this test and the focus on traditionally male fields when discussing it is that there is one traditionally female field which requires excellent mental rotation skills: Dressmaking.

The pattern pieces are two-dimensional, the final product distinctly three-dimensional, and anyone trying to alter patterns or making new ones must have good three-dimensional mental rotation skills.

Eliot mentions other typically female professions in which such skills are necessary: interior decoration and fashion design (Pink Brain. Blue Brain. p.230)
....a recent study of university students in London found that female fashion-design majors performed even better than male engineering and computing majors on a test of mental rotation.
Whether their performance was a result of the studies they were taking or whether this group already had strong mental rotation skills before their studies is unclear from this quote in Eliot's book. Yet knowing the answer to that question would be of great value. If the test results of these female fashion-design majors were high due to the content of their studies, then practice clearly plays a very strong role in the mental rotation scores.

Given the great focus on this one particular test, it may come as a shock to you that this just might be the only test for which schools offer no formal training. Eliot (Pink Brain. Blue Brain. p.230):
Every other ability -- verbal, math, music, interpersonal, and even kinesthetic (athletic) -- has its own slot in the curriculum or its own box for teachers to check off on report cards. Not so for spatial skills, for which the only real training comes from the video games, building toys, and targeting sports that are almost exclusively the preoccupation of boys.
Mental rotation tests are also susceptible to the stereotype threat. This refers to the priming effect of evoking certain stereotypes on a test-taker. An example from Cordelia Fine's Delusions of Gender (p. 27-8):
People's mental rotation ability is malleable; it can be greatly enhanced by training. But there is a far quicker, easier way to modulate mental rotation ability. By now you already know what these methods involve: manipulating the social context in such a way that it changes the mind that is performing the task. For example, you can feminize the task. When, in one study, participants were told that performance on mental rotation is probably linked with success on such tasks as "in-flight and carrier-based aviation engineering...nuclear propulsion engineering, undersea approach and evasion, [and] navigation," the men came out well ahead. Yet, when the same test was described as predicting facility for "clothing and dress design, interior decoration and interior design...decorative creative needlepoint, creative sewing and knitting, crocheting [and}flower arrangement, this emasculating list of activities had a draining effect on male performance.

Alternatively, instead of changing the gender of the task, you can keep the task the same but push gender into the mental background. Matthew McGlone and Joshua Aronson, for example, measured mental rotation ability in students at a selective liberal arts college in the northeastern United States. One group was primed with gender, while another group was primed with their exclusive private-college-identity. Women who had been induced to think of themselves as students at a selective liberal arts college enjoyed a performance boost, scoring significantly higher than gender-primed women. Likewise, Markus Hausmann and colleagues found that although gender-stereotype-primed men outperformed gender-stereotype-primed women, men and women primed with an irrelevant (geographical region-based) stereotype performed similarly on the mental rotation task.

Why such a long post on one specific cognitive test? Because this is the test which shows the largest and most consistent differences between men and women, and because this is the test which is usually mentioned in explaining a possible ability-related reason for women's scarcity in science, mathematics and engineering.

That even this test lends itself to learning and is affected by the stereotype threat emphasizes the need to keep societal influences in mind when the essentialists employ this particular argument.