managers hired women because they expected programming to be a low-skill clerical function, akin to filing, typing, or telephone switching. Assuming that the real “brain work” in electronic computing would be limited to the hardware side, managers reserved these tasks for male engineers.
As the intellectual challenge of writing efficient code became apparent, employers began to train men as computer programmers. Rather than equating programming with clerical work, employers now compared it to male-stereotyped activities such as chess-playing or mathematics.
Then, in the late 1960s, a series of activities was set in motion that tended to exclude women: all-male professional associations etc. Today, the stereotypes created back then still have an effect:
Today, we continue to assume that the programmers are largely anti-social and that anti-socialness is a male trait. As long as these assumptions persist [...] the programming workforce will continue to be male-dominated.
Back when I was taught at university, I noticed the effect of those stereotypes: There are many more women in degrees such as “Bio-Informatics” and “Media Informatics”, presumably because they are considered more interesting and as existing in a less anti-social environment. However, students often end up taking an “Informatics-only” job, and are just as qualified as “pure” Informaticians: The latter have usually studied a bit more Math and core computer topics, the former have other useful skills, but they are all capable programmers. Moreover, because you have some freedom regarding what you want to focus on during your studies, the boundaries between the variants of Informatics are blurred, anyway.