Future two-variable surveys like this should definitely consider taking Dave's Hacker Survey data (http://davelyon.net/mldata.html http://davelyon.net/mldata.ht

Future two-variable surveys like this should definitely consider taking Dave's Hacker Survey data (http://davelyon.net/mldata.html http://davelyon.net/mldata.html) into perspective. The 21-35 range seems to cover something like 1.5 sd on either side — around 85% of us if we assume normality. Investing more of these poll choices to age resolution within that range would have been interesting.