this is why understanding sample size, selection bias, and confidence intervls matters - the difference between actual statistics and the marketing version that gets passed off as "research".