I enjoyed http://visualjournalism.com/another-beautiful-infographic-with-glaring-errors-goes-viral/2010/11/17/, so I thought I'd share it here.
Cute. It's sorta like framing the guilty subject: the research supports the conclusion they want, but not clearly enough for their liking, so they lie about what the research supports.
Presumably you all saw http://meta.wikimedia.org/wiki/Fundraising_2010/Banner_testing.
Nothing got more click-thrus than Jimmy, begging pathetically.