Mr. Tierney, let's indeed discuss women in science.
First, let me start by saying that I applaud the discussion -- all potential explanations for a complex issue and all evidence need to be considered, even the ones that are not popular in the media or not "politically correct." I also believe that Larry Summer's now infamous comments about the possibility that biological differences account for the dearth of women scientists and technologists was, similarly, in the spirit of intellectual debate.
The problem with the biology argument that "boys are just more likely to be born good at math and science" isn't that it's not "politically correct" -- it's that it assumes that we can take away the power of societal influences, which have much more solid evidence than the biology hypothesis. Tierney makes the point himself in his article -- in order provide evidence for biological differences, he cites a longitudinal Duke study which shows that the highest achievers in SAT math tests (above 700), which counted 13 boys for every girl in the early 80s, became a ratio of 4 boys to 1 girls in 1991, "presumably because of sociocultural factors." Hmm, isn't this actual evidence that biology is not what is at play here? If it is possible to reduce the gender achievement gap in math by 3 thanks to "sociocultural factors", I rest my case, sociocultural factors are indeed extremely powerful.
The Duke study also notes that the 4/1 achievement gap at the highest score hasn't changed in the last 20 years despite ongoing programs to encourage girls in math and science, whereas the highest achievers in writing ability (SAT above 700) shows a ratio of 1.2 girls for every boy, slightly favoring girls. However, if the premise is that boys are inherently "better" at math, and girls are inherently better at writing, why would the achievement gap be so large in math and negligible in writing? The stagnant 4 to 1 ratio is not evidence that there is an innate biological difference in math aptitude, but rather confirmation that persistent sociocultural barriers remain -- that is, science and math are still thought of as male domains.
Research shows that math and science are indeed thought of as stereotypically male domains. Project Implicit at Harvard University studied half a million participants in 34 countries and found that that 70 percent of respondents worldwide have implicit stereotypes associating science with male more than with female. Years of research by Claude Steele and Joshua Aronson and their colleagues shows that implicit stereotypes affect girls' performance in math -a phenomenon called "stereotype threat". When girls receive cues that "boys are better at math," their scores in math suffer; one study in a classroom setting showed that the difference in performance between boys and girls in math SAT scores was eliminated by simply having a mentor telling them that math is learned over time rather than "innate".
The problem is, girls are routinely getting the message that they don't belong in math and science, further undermining their performance (and Mr. Tierney's article isn't exactly helping in changing the stereotype for the general public). The result of this implicit (unconscious) stereotype is that parents, teachers, and school counselors are less likely to encourage girls to pursue math and science than they are boys. These girls are then less likely to seek advanced math classes and would be unlikely, without those opportunities, to make it to the above 700 SAT math score