Categories
Random Work

How to lie with statistics

This Slate article shows how the medical profession lies with statistics by using qualifiers that virtually no one understands, like “relative risk.” How big of a difference is it? Read to find out.

I also like eMarketer Daily‘s usual habit of overinterpreting differences. I still remember nearly a year ago they published a report on how different ethnicities use cell phone technologies differently. The “findings” talked about how 28% one group (African Americans or Hispanics, I forget which) used text messaging or the Internet on their cell phones as if it were a big difference from the second group, wherein “only” 26% did. There was no mention of the margin of error for the study.

I especially liked eMarketer Daily’s report a few months later that ClickZ lies with statistics. As I e-mailed to a coworker, that’s really the pot calling the kettle black.

Don’t worry, I promise to post about Hayden later today. I know it’s what you really want to see. AND I promise to post about the quest for the I<3L&O mug tomorrow.