Today I heard about something that really tickles my fancy. 7-Eleven customers can choose between Obama or McCain coffee cups and the results of the informal poll are tallied daily. Apparently this poll correctly predicted the results of the last two elections.
In the 2000 7-election their George W Bush coffee cup outsold the Al Gore cup by just 1%. The 2004 7-election results tracked identically with published national results. 51% for Bush and 40% for Kerry
As a former marketer, I think this is a brilliant marketing campaign. Kudo's to the marketing team at 7-Eleven. You can follow the results daily.
I hope that this year, before all the journalists start commenting on polls, that they do a short statistics course so that they can correctly report on polls. I become rather irate when I read reports that claim that candidate x has overtaken candidate y in the polls by 2 percentage points, only to discover that the margin of error is 6% and the difference between the candidates is only 2%. How difficult is it to understand the concept of a margin of error? Then there are all the gleeful reports where the journalists haven't checked the study design...







