It's interesting all the same to see AOG expound the idea of familiar role models being important to a child's development or motivation in a particular area.
It's a very difficult question to answer, as there are dozens of potential contributing factors. It varies between subjects and between schools, from city to city, based on household income, family background and so on. These other factors tend to dominate over gender differences anyway, so the media is slanting one particular feature while ignoring many of the rest.
Incidentally, the gender gap almost disappears when comparing only the very top A* grade, but can be observed in the A*-C pass range, so that suggests that there's a sharp peak in high-achieving (and, sadly, low-achieving) boys compared to the rest of the field.
Whatever the reason for the difference, we shouldn't generalise.