Why compare a survey (things that developers say) with hard pull request data (things that developers have actually done)?
Why compare a survey (things that developers say) with hard pull request data (things that developers have actually done)?
GitHub is (or was, at some point) written in Rails and attracted the Ruby community early. Rails is developed on GitHub, as are many other major Ruby projects, including the Ruby language itself.
Python is hosted on its own Mercurial repository (though it is shifting to GitHub later this year) and that community never had the tribal draw to move its development from things like mailing lists into GitHub. Patches are still handled on the Python bug tracker, etc.
As for individual project source management, some projects will have all contributors with their private fork and have them make individual pull requests as part of the code review process, some will give a core of contributors direct access to make raw pushes, and so on.
Because of not every project uses GitHub and because GitHub's unmanaged pull request object is not a reliable unit of measurement, it's better to trust a survey that directly reflects real-world observances than to make extrapolations based on an unreliable metric.