... you know, because of the polls, here's a little reminder.A little word about polls. Polls are models, too. It's not like everybody who is going to vote is going out every day and voting in these polls.
Pollers poll a sample population, and then guess how well that sample represents probable voters, and what percentages of probable voters. To describe that, they ascribe weight factors to the numbers they collected in the sample and try to project what the actual voting population will do.
They are affected both by their inputs and their assumptions and their accuracy depends on how well the poller matched them with each other and how valid the assumptions he came up with actually are.
They can provide useful information, but they should not be confused with reality. They are mathematical descriptions about what the modeler believes describe reality. You know, like climate models. :-)
No comments:
Post a Comment