This is a very poor article, masquerading as a scientific paper, but published in a political publication. The authors pick a few jargon-heavy effects, claim there's no evidence for those effects, and dismiss the entire argument, and fall back on the lazy "boys like things, girls like people" explanation, and is laced throughout with accusations of politically based dishonesty in opposing views. In other words, it's no scientific paper, but a political polemic.
Whether or not these few specific effects they claim to have debunked are real or not, to explain the gaps, you have to go far beyond a few narrow sociological effects. There's a self-reinforcing culture of same-hiring-same that glorifies algorithmic intuition as the end-all be-all of software development. There are cultural and fashion trends in the social media landscape that reinforce what's appropriate to be interested in based on your self-identity. The authors claim that gendered interest in CS and the like has remained "stubbornly low" because "women prefer working with people", ignoring huge evidence that even as CS has become far more human-centered and collaborative since the 80s, the numbers of women declaring it as a major have plummeted.
Whether or not these few specific effects they claim to have debunked are real or not, to explain the gaps, you have to go far beyond a few narrow sociological effects. There's a self-reinforcing culture of same-hiring-same that glorifies algorithmic intuition as the end-all be-all of software development. There are cultural and fashion trends in the social media landscape that reinforce what's appropriate to be interested in based on your self-identity. The authors claim that gendered interest in CS and the like has remained "stubbornly low" because "women prefer working with people", ignoring huge evidence that even as CS has become far more human-centered and collaborative since the 80s, the numbers of women declaring it as a major have plummeted.