I'd be surprised if anyone would go to the trouble of collecting the data for the purpose of discriminating.

The idea behind the benchmark was: we have no idea how well or badly we're doing, so let's look at some graphs. We knew some people wouldn't answer, but if enough people did, then at least you could say "x% of people in the company are y characteristic but they're only z% of senior management - what can we do to improve that?