According to the graph in the middle of the article (unless I'm reading it wrong) it would seem that there are more male geniuses.
The graph then, would seem to indicate that the sort of tests that determine a child's eligibility for entrance into the gifted programs in New York either a) do not track with IQ tests or b) filter out children with IQs lower than say 95 (therefore making more girls eligible than boys).
I had the same initial reaction to the graph, but she seems to be using the graph to show that reality may be at odds with the enrollment numbers. Data from 1932 does seem quite out of place in an article addressing "recent trends".
She goes on to suggest that the gifted programs are more targeted to girls - things like verbal-oriented tests and needing to "sit still" for periods of time.
The 1932 data described in the article is important because it covered nearly the entire population of a country for a particular age group -- uneven data collection could not have skewed the results. Uneven data collection is the great enemy of psychological public policy, since politicians can so conveniently project their dispute onto the people that were not sampled. The 1932 Scotland data slams that door shut.
Other testing programs since then have showed a slow, steady rise in the average IQ, called the Flynn effect. However the relative shapes of the male and female histograms have been quite constant. (Ditto for the relative shapes of the racial curves.)
How would filtering out children with IQs below 95 have any effect whatsoever on the demographics of the children who pass eligibility tests for gifted programs?
The graph then, would seem to indicate that the sort of tests that determine a child's eligibility for entrance into the gifted programs in New York either a) do not track with IQ tests or b) filter out children with IQs lower than say 95 (therefore making more girls eligible than boys).