3 ms·
It's so disappointing to see studies like these relied on when reported by news outlets. To start, we don't have the data or statistical code used to generate t
by RA_Fisher 10y ago
It's so disappointing to see studies like these relied on when reported by news outlets. To start, we don't have the data or statistical code used to generate the results. Consider this a basic unit test. Of course we'd want to further scrutinize the code and methodology, but we're not even there. :-(
I'd suggest a marker for quantitative research posted to HN along the lines of: runnable?
- porejide 10y agoCode: totally agreed. Data: tricky, because of privacy and HIPPA.
- basseq 10y agoYou could make this complaint about pretty much every academic study ever. Rarely is the underlying data, statistical models, etc. shared publicly. Should they be (for repeatability and peer review)? Sure, I buy that argument. But it's not a valid reason to criticize this particular study. And at least it's the source journal article, not a media post!
- RA_Fisher 10y agoI do make this argument for pretty much every quantitative study.