Tuesday, November 17, 2009
Poll: Miami Herald doesn't care about the integrity of polls
I'm putting off the Spence-Jones controversy until I've gathered some more resources. And it'll just take more time to write than I've had to give today.
Instead I just wanted to mention I poll I recently conducted here in North Miami Beach. I only polled myself and my mother, which estimating the population of NMB at 50k (it was less than 40 in 2000) means I polled about .004% of the population.*
Does that poll seem unsubstantial to you?
Enter the Miami Herald's coverage of a poll done in Cuba. This is breaking news.
Some group was able to conduct an opinion poll in Cuba with a sample size of 432. That's .0039% of the Cuban population. What did this fantastically obscure poll find? Apparently Cubans don't like the direction their country is going in.
Aside from pointing out the obvious, you shouldn't take this poll, or almost any poll you see in the mainstream media seriously because the opinions of less than one percent of the population do not represent the views of the whole.
One need not take a course in statistics to recognize the usefulness of this advice. While these 432 Cubans have a right to their opinions--I'm sure they're very good opinions--an individual trying to learn of the daily struggles and mindsets of Cubans should not take this information into account for said purpose.
The Herald includes the opinion of a Miami pollster, suggesting "general caution in interpreting results from the country." While this is perhaps a relevant point to make, I am more concerned with the lack of information about the procedures used by the polling group to gather the information and most importantly, about the insignificant pool of persons polled.
Moral of the story? Polls usually don't really tell us anything. They force a small group of people to be grouped together by vague opinions which they may not even hold.
I'm a big stickler for the misuse of polls, so look forward to more rants.
*This is a lie. I did not poll anyone.