Essentially, they argue, because StumbleUpon users are so disproportionately active, any sample that misses some of them will vastly underrepresent their impact in driving traffic.

FORBES: Does StumbleUpon Really Drive More Traffic Than Facebook?