ABC News, with the egregious polling spin of the day, with the easily overlooked details in bold:
President Obama is holding onto a 7 point lead over his chief Republican rival, Mitt Romney, in the latest ABC News-Washington Post poll.
If the 2012 election were held today, 51 percent of adults said they would pick Obama compared to 44 percent who would support the former Massachusetts governor. (They were tied in early June.) Important to note, however, is that among registered voters, Obama’s lead over Romney narrows to 49 percent to 47 percent. No other Republican challenger fares as well as Romney in head-to-head matchups against the president. http://abcn.ws/nUXx7F
I don’t gripe about pollsters who give us presidential job-approval numbers using a sample of “adults,” because adults who aren’t registered to vote can have an opinion on how the president is doing. But for anything resembling an election match-up, “registered voters” should be the minimum limit on the respondent pool, not the maximum. To get a more useful measurement, we should be looking at likely voters.
I realize that some pollsters will argue that this far away from Election Day 2012, it’s not easy to determine who a likely voter is. Of course, I’d start by weeding out those over age 21 who were eligible to vote in past presidential elections and who have never voted in one before. Will you miss out on some folks who will vote for the first time in 2012 that way? Sure, but as a proportion of the electorate as a whole, they’re going to be pretty marginal.