Hacker News Reader

Sorry, but what part of "they only surveyed software developers in the United Kingdom" gives you the impression that their sampling might actually be good?

Worth noting that the study itself doesn't seem to mention their sampling methodology.

View on HN

Studies are never meant to be an ultimate answer to everything. And just because they're not the ultimate answer, doesn't mean that they are therefore worthless, bad or "bullshit".

It actually does mean that in most cases. Having false confidence in a hypothesis based on non-rigorous studies is much worse (and in some cases, a lot more dangerous) than withholding judgment until the data (all the data, not some cherry-picked slice of it) really is conclusive.

No. You perform studies to validate observations or theories and, if you are able to validate them, offer potential factors as contributing causes. A study where the data does not support the conclusion is worthless, and can in fact be harmful.

Reply on news.ycombinator.com