19 Sept 2008

Surveys and Statistics

It has been said that 85% of all statistics are made up on the spot including this one.

For example consider this one

But this one from The Hindu Online edition is the real winner.

Majority of Indians prefer smoke free zones.

92%, 96%, 95% on various counts are just unrealistic. Not that I am for smoking in public places, but falsifying data to push for reforms?

The first rule for a survey is that it take a representative sample. If you go to a anti smoking convention and ask the attendees for their opinion that isn't really a survey. Unfortunately the only fine print available is

"Around 1030 persons aged between 15-64 were interviewed between August 9 and August 24, 2008, the survey added."

By the way, what does AROUND 1030 people mean? Do you not even know the sample population size?