This highlights a very specific cross-section of the programming community that isn’t necessarily representative. Users of minority programming languages and tools such as editors are going to be significantly underrepresented due to network effects.
The results very well follow the “developer evangelist”/cargo-culting and “underpaid enterprise code monkey” groups, however.
According to “Interview Pass Rates,” it looks like there needs to be an Emacs proficiency test as part of preliminary interviewing, because Emacs users are the most likely (by far) to pass interviews!
(The article is probably right that most Emacs users are more experienced, better programmers than users of other editors, because there’s practically no bandwagon to hop onto)