Politics polls are notoriously inaccurate for some of same reasons you mention ie detection. As a former political reporter, I became aware of this a long time ago. Getting a true sample of voters who will go to the polls is difficult, and what is reported re polls often is inaccurate because people often don’t want to participate or are not reachable by cell phones, etc. Too much attention is paid to flawed polls vs issues pre-election. Detection of the popular vote ahead of an election is hard to accomplish.
Yes, pollsters don't seem to understand the critical importance of getting a random sample of the population they are trying to characterize, and of course, that is the point of my essay. So getting responses from people with landlines that are willing to answer their questions, not to mention are home when they call and don't have caller ID does not allow a meaningful characterization of, for example, American voters in general. Meaningful, truly random sampling is the core of meaningful conclusions. No amount of arm waving by pollsters will transform non-random, self-selected polls into valid characterizations. I guess if you want to know how people with landlines, willing to respond to polls made up of loaded questions, who are willing to interrupt their dinner, and don't have second or third jobs, and whose kids aren't screaming for attention feel about something, by all means, you should schedule an expensive poll.
Politics polls are notoriously inaccurate for some of same reasons you mention ie detection. As a former political reporter, I became aware of this a long time ago. Getting a true sample of voters who will go to the polls is difficult, and what is reported re polls often is inaccurate because people often don’t want to participate or are not reachable by cell phones, etc. Too much attention is paid to flawed polls vs issues pre-election. Detection of the popular vote ahead of an election is hard to accomplish.
Yes, pollsters don't seem to understand the critical importance of getting a random sample of the population they are trying to characterize, and of course, that is the point of my essay. So getting responses from people with landlines that are willing to answer their questions, not to mention are home when they call and don't have caller ID does not allow a meaningful characterization of, for example, American voters in general. Meaningful, truly random sampling is the core of meaningful conclusions. No amount of arm waving by pollsters will transform non-random, self-selected polls into valid characterizations. I guess if you want to know how people with landlines, willing to respond to polls made up of loaded questions, who are willing to interrupt their dinner, and don't have second or third jobs, and whose kids aren't screaming for attention feel about something, by all means, you should schedule an expensive poll.