So the New York Times had an article about a poll they conducted recently about texting and driving. They found that 97% of the people surveyed believed that it should be against the law.
However, this line drew some attention on Gizmodo:
The nationwide telephone poll was conducted Oct. 5-8 with 829 adults and has a margin of sampling error of plus or minus 3 percentage points.
Apparently some people saw the 829 figure and were confuzzed. "Number too small. All numberz bad. Survey bad."
I attempted to address the issue with a brief description of sampling, but it didn't do much good, and of course because this is teh innernets, stamping out ignorance is like putting out a grease fire with water: if you don't bring an incredible amount of water, you'll be overwhelmed by the fire, but either way, there won't be much left when you're done.
So I had my say, and now I'm moving on.
Not quite sure why the Times said 3 percentage points, though. I got 1.5 from this online calculator, but then I wonder if there is somehow a one-tailed vs. two-tailed issue here. Just curious ...