Research statistical analysis

[danny-andrews]

Find the best method for normalizing stats like this. Is taking the median time better than the mean? Or am I losing important information. Maybe I could just use a benchmarking library and get more statistical results, but I feel like my homegrown solution is fairly good. link link2

[danny-andrews]

I believe the best measure of dispersion is the median absolute deviation.

[danny-andrews]

Takeaways:
Mann–Whitney U test (Wilcoxon rank-sum test)
Bootstrapping
Using minimum rather than mean or median