Tuesday, January 08, 2002

How Much Faith Should you Place in that Poll?

It's not uncommon to come across a poll on the internet, or in a newspaper, and read that a certain percentage of the public believes X when you believe Y. You survey friends and acquaintances, and most people you know disagree with the results of the poll. Does that mean the poll is invalid?

You see a poll on the internet, and a lot of people have taken part in the poll. The results are totally ridiculous. How scientific is that poll anyway?

How much information does the person reporting about the poll tell you about it? How much faith should you place in the poll. Aren't there guidelines anywhere that should tell people about how a poll should work?

An excellent resource on polls exists on the pages of the National Council on Public Polls (NCPP), called 20 Questions A Journalist Should Ask About Poll Results. They discuss many of the things that a journalist should look for when reporting upon a poll. Even if you're not a journalist, this is an excellent set of questions and answers to look over, because it can give you a sense of the types of things a reporter should be telling you about a poll as they are reporting on it.

The problem with internet polls? Usually the people taking the polls are the ones electing to take polls. There's no selection process on the part of the pollsters to try to get opinions from a broad range of people with different opinions. If you have a small community, and the poll results show that most people voted (only once), then your poll results are going to be pretty good. But when you poll a thousand people to try to gauge the opinions of a few million, it is important to try to get opinions from a broad spectrum of participants.

The next time you see a poll that is used to back an opinion, such as a presidential approval rate on certain issues, take a look at the poll objectively, and see if it follows the guidelines from the NCPP.
- William Slawski