4 ms·
Why do you consider the survey design "sloppy"? E.g., doom-emacs is a well-known emacs distribution; there nothing weird if 18% of survey respondents use it.
by HerrMonnezza 6y ago
Why do you consider the survey design "sloppy"? E.g., doom-emacs is a well-known emacs distribution; there nothing weird if 18% of survey respondents use it.
Whether the survey respondents are a representative sample of the general emacs population is a different question, and one which is very hard to answer given that we have no other data about people's preferences in emacs. Hopefully, the survey will become more popular with time, so we'll see results approaching the "true" values.
- spenczar5 6y agoThat's exactly what I mean by "sloppy" - it's not clear what question this thing answers, because the question wasn't carefully written down ahead of time, so there is no way to really generalize anything here. Because different response methods are pooled together in results without any reweighting, it's not clear how much bias has been introduced vs a hypothetical global emacs population measure. What conclusions do you think you can draw from this? I don't know how to draw any that aren't hedged heavily. 18% for doom seems much higher than I expected - is that because I underestimate it, or because of how the survey was conducted? There's no way to tell, here. Good survey design _does_ exist; it's an extremely intensely studied practice. It's not surprising that a casual internet survey doesn't follow anything resembling modern surveying techniques, but it does make results unclear.
- brundolf 6y agoThis isn't a voter poll, bud. It's an informal survey of a text editor's userbase. It can be pretty imprecise and still usefully inform the project.