I had held off from commenting on the recent polling showing either John McCain pushing Barack Obama off his pedestal or Obama pulling into his widest lead in the last month. I find this pair significant, not for their informative value as to who is ahead or behind, but for the demonstration of exactly how increasingly useless these polls are. This rare case shows us the same company conducting polling during the exact same three day period and coming up with a set of results that are 12 points apart. The polls claim to have a margin of error of +/- 2 and 4 points respectively. They aren’t even within the maximum stated margin of each other. (To be fair, Gallup puts a disclaimer on the poll showing McCain in the lead – which nobody seems to want to talk about – as follows:
Gallup editor Frank Newport tells Jill that “registered voters are much more important at the moment,” because Election Day is still 100 days away, but that the likely-voter result suggests that it may be possible for McCain to energize Republicans and turn them out this fall.
This is because the best results for McCain are based on “likely voters” as opposed to “registered voters.” During an interview this morning, one of the Gallup directors mentioned that “likely voters” is still a very ambiguous term this far out from the election, and could readily shift in either direction by late September.
As I have previously proposed, dramatic changes in the structure of modern society have not been taking into account by polling companies. Combine this with a desire to do polling every day of the year and you have a recipe for a flawed product which is of no use or interest to anyone but we politically obsessed members of the chattering class who have far too much free time to fill. Let’s take a brief look at some trends which should have us taking these numbers with large pillars of salt.
One of the big factors is that it’s getting harder to get people to even participate. This may be to lack of interest or enthusiasm, being too busy or whatever, but it seems to be the case. Chuck Todd of MSNBC was on the air this morning describing a conversation he had with some pollsters who talked about how difficult it was to find anyone to complete a one half hour survey covering most of the major issues. Even in some cases where respondents were willing to take the survey, they lost interest part way through, began passing on some questions or just hung up, forcing the pollster to use computer generated models to predict the rest of the responses. The majority of people who are willing to do these surveys come from older, white voters, which skews the results further.
While some advances are being made, polling companies still haven’t managed to overcome the gap in people who either use only a cell phone, or are home so infrequently and rely on their cell phone so much that they might as well not have a land line. Again, this factor skews towards on set of demographics over others, rendering results more questionable.
One of the biggest factors, which I harp on constantly, is the known fact that a very significant portion of the American electorate simply isn’t engaged in the election yet. For those of us who live, eat and breathe politics, we began fighting the 2008 presidential election approximately ten minutes after the polls closed on George W. Bush’s last victory. But out on Main Street America you will still find a possibly shocking number of people who couldn’t name both candidates right now, say nothing of speak intelligently to the issues. I take part in a weekly traveling social event which takes me to bars, pubs and taverns across New York and Pennsylvania. I make a point of being the annoying fellow who brings up politics at the bar and asking some general questions. The level of knowledge and interest I find is dismaying.
If the polls this year ever get back to some sort of meaningful level of accuracy, it won’t be until after the conventions are over and the debates on network television kick in. Perhaps at that point we can start debating what they mean. For now, you’d be better off with a maypole than a political poll.