The science of polling has always mystified me, mainly because I have never had the time to buckle down and learn all the rules, eccentricities, ins, outs, ups and downs that come with it. So I present the following excerpt from an op-ed piece in the City Weekly (Salt Lake City, Utah) for discussion.

America is set to dump the party that’s hosted the past seven years, and the GOP brain trust is aware that polling data suggesting a close presidential race is wildly inaccurate. Pollsters can’t contact the cell phones of all those Obama-loving younger voters who don’t own landlines (the only ones pollsters are allowed to call). Those members of the youthful demographic who are being surveyed prefer Obama by 20 plus points. The very poor (primed to actually vote this time) and the highly educated (also strongly Obama) are likewise less reachable by telephone. --Jim Catano, City Weekly 9/4/08

Can this be true? Is it possible that modern day polling firms have not taken issues like this into account when doing their calculations?

In college, I had a political science class from a professor who was (and is) not only the prominent pollster in the area, but was also Karl Rove's political science college advisor. Now, I wanted to take what this professor said seriously, but given the fact that he namedropped faster than attendees of the Sundance Film Festival, I was inclined to dismiss him as an egocentric spawn of the devil. However one of the things he did hammer home was the point that because of little issues like those mentioned in the quote above, polling will never be as accurate as we would like.

Your thoughts? Is Mr. Catano's point valid?