Daily Mail's Misleading Headline Shows Shows How Not to Use "Polling"
From unrepresentative reader polls to sensationalised headlines, you don't have to look far to find poor uses of polling in the British media. However, few instances are as egregious as today's Daily Mail front page. The Mail led with the remarkable revelation that a "snap poll" revealed "one-in-six people know who scandal-hit BBC star is". It then extrapolated from this that "millions" can name the unnamed BBC presenter. How can the Mail be confident of this fairly strong conclusion? Did their poll deploy random sampling, effective weighting or rely on a broad sample size? Well... No.
Only after flicking over to page 4 does the Daily Mail reveal the true nature of their "indicative rather than scientific" straw poll. They recruited journalists to approach punters in the street and ask about the story - their conclusion that "millions" know the BBC presenter's identity relies on 49 responses.
What's more, there seem to be no effort to ensure random sampling or weight the sample to be representative - they have no basis to apply the results to the national population. Still it could be worse. It's not like this is being written on the front-page of the country's most circulated newspaper...
