I got called tonight by some presidential survey, and since I have 5 kids, and was pretty tired, I almost just said, "no thanks" but decided to take it just because I thought most people like me probably would not take it. This got me thinking about inherent biases in these polls based on non-random samples. Is it even possible to get a random sample, even if you try to match up the self-reported (or even registered) Democrat, Republican and Independents? Surely certain demographics of people must be less likely to answer these polls. In fact, I'd think it's a lot of people like me who don't answer it - borderline middle-aged, conservative, family-type, good job, content with my life. Obviously a message board is not a random sample, but just out of curiousity, say you got the call that said, "I'm Penelope and I'm with the independent XYZ survey organization and I'm taking a survey of your opinions on the upcoming elections." What is your political leaning and would you answer this telephone survey. I'm new and didn't find anything specifically like this on search, so sorry if this is a repeat and I just missed it.