3 ms·
The difference between the salaries of tab-users vs. space-users in David's report are too close to call, so I added bootstrapped 95% confidence intervals for t
by minimaxir 9y ago
The difference between the salaries of tab-users vs. space-users in David's report are too close to call, so I added bootstrapped 95% confidence intervals for the aggregate median salaries (which was easy to do since the code was open source): https://twitter.com/minimaxir/status/875386185350168577 https://twitter.com/minimaxir/status/875386185350168577
If the 95% confidence intervals for tabs and the 95% confidence intervals for spaces intersect at a point, there is a possibility for failing to reject the hypothesis that the difference between the two is nonzero at the alpha = 0.05 level. Since there is little overlap in most cases, the original hypothesis holds.