Hacker News Reader

HNHacker News

Top New Best Ask Show Jobs

Comment by darzu | Hacker News Reader

Parent Full thread

darzu·

The sampling biases are _really_ important here, stated another way:

Of developers who use WakaTime, most use Windows.

Of developers who answer StackOverflow surveys, most use Linux.

These aren't contradictions.

View on HN

Sure, but "Of devs who answer StackOverflow surveys, most use Linux" is just wrong. To prove it, add up some top IDEs used that don't even run on Linux [0]:

+ Visual Studio (32.15%)

+ Notepad++ (27.71%)

+ Xcode (10.53%)

= 70.39%

That leaves only 29% left for Linux, but they claim Linux usage is 40% [1].

Unfortunately, Linux is still 3rd in primary OS usage and this survey question is flawed.

[0] https://survey.stackoverflow.co/2022/#section-most-popular-t...

[1] https://survey.stackoverflow.co/2022/#section-most-popular-t...

You can't add %'s like that - the question allowed users to select multiple values.

Yes, you can add percents like that because they share the same base. Thanks, that makes sense... accepting multiple values for a "what's your primary OS" question is the problem.

Edit: Sorry, I thought you meant only the OS question allowed multiple answers. You're right, makes sense they would allow multiple answers for IDE too.

Still, they shouldn't call it "primary OS" if multiple answers are treated with the same weight.

People can select multiple editors, so you can't sum them like that. E.g. people on windows could use Visual Studio AND Notepad++. All this says is that at least 10% use Mac because xcode only runs on the Mac, and at least 32% use Windows.

Oh, all the questions accepted multiple answers? Ok makes sense now... then they can't add up to 100. Still the OS question is flawed because multiple answers means it's not their primary OS.

Reply on news.ycombinator.com