funny thing about big data and reporting
is that almost every report finds exactly what the entity requesting the report wanted.
I suspect any reasonable set of queries, assuming all data people are demanding is actually recorded, would show that for any group, they will show "underpaid".
Run it for women and it will "prove" they're underpaid.
Run it for any minority, and I suspect the same rules will provide the same output.
All those poor underpaid Indian DBAs.
Hey, I bet we'd find most exploited H1B's in tech are *male*.
so we'll filter and drill-down until we get the results that "satisfy" the recipient. Filter anyone with advanced degrees, filter anyone on a work visa, filter anyone who's been with the company less than a year. Filter out overtime pay, disciplinary or attendance issues, and anything the Union says is derogatory. Then take an average of pay from the owners of the company down to the kids in the mailroom, split by gender.
Viola! Politically acceptable results Unlocked!
(and that's pretty much how we do the major reporting by which Policy is generated here as well. If the results don't match the Politics, the report is rejected until it does. Any results that prove something politically dangerous, are "extraneous" and removed as to "not clutter" the charts and graphs. Perot would be proud.)