What about the women?
tech.coursera.orgIt bothers me how some of those charts lend themselves to possibly wrong conclusions. For example women tend to score higher in courses with more women. But it seems unlikely that men suppress women in courses, thereby lowering their scores. For one thing, students don't tend to interact that much in those courses. And a couple of charts earlier they showed a strong difference in percentage of women in, say, humanities vs engineering. That seems like a much more likely cause.
At the end of the day, these are free courses, so it seems unlikely there are less women because they are excluded in some way. Lack of interest seems the more likely cause.
You're right that it would be wrong to conclude that men suppress women in courses -- certainly we found no evidence of that (I helped write the post). As we note in the post, while the correlation is unequivocal, causality is complex and worthy of further study.
This was flagged off the front page, which is rather unfortunate.
From the "Gender and Performance" chart, the distribution of the values appears to have a correlation of 0, so I'm not sure how a positive correlation is inferred.
Thanks for the question! The correlation here is r=0.38 with a p-value < 2e-15. We've added this note to the post.
(Note: I'm one of the authors of the post.)