An unscientific sampling, Nov. 15, 2005
A cynical reader questioned me Monday about our Web poll.
&uot;Why don’t you print how many people are responding to your poll?&uot; he wrote. &uot;Do you have something to hide?&uot;
His point was that he did not trust the poll findings if, for example, only 25 people responded. I think he believes I’m trying to tilt it to benefit the worldwide liberal media conspiracy.
Not so. The number of responses frequently varies. When we have a question about a local issue, the response is generally high, between 150 and 300 people. When it’s a state or federal issue, the response drops off to almost nothing.
Once when we asked people whether teachers should get pay raises, overnight we had more than 700 responses, almost all for it. Conversely, our last poll, about the Plamegate investigation, generated only about 25 responses.
I posted the current poll last night. About 20 hours later, only 9 people had responded, and 67 percent of them (6) are in favor of allowing the CIA to torture people… I would not have believed there were six people in Suffolk in favor of torturing other human beings, but that’s a subject for another day.
We’d like to have a local question on our poll every day, but because of my limited intellectual capacity, that’s not always possible. I’d love to get suggestions, though. Send them to firstname.lastname@example.org, or post them below if you like.
It’s interesting that in a country of probably 200 million adults, scientific polls are often based on about 1,200 responses. That’.000006 percent (six one-millionths of the population). If I used that same size sample to survey a city like Suffolk of 75,000 people, that would be 45 one-hundredths of one person. So maybe 25 people is not an unrepresentative sample. I don’t know.
I jest somewhat, of course. Our poll is not scientific and makes no claims to be. It’s there for fun, to engage people in our Web site and what is going on in the community. I hope you’ll join the fun.
Until now, we’ve only been changing the poll weekly, but plans on our redesign call for changing it daily and posting the results (along with the total number of ballots cast, Mr. Pocklington), on our front page each day.