I’m still looking for a serious numbers-cruncher who thinks Romney has a good chance of winning
That headline above is not intended to be snarky. I really do want to find a reputable polling analyst — not just some pundit plying anecdotal observations and wishful thinking — who actually likes Mitt Romney’s chances of garnering at least 270 electoral votes.
If only to provide a little balance on this avowedly liberal blog, I’m looking for some counterpoint to all of the numbers-based projections that President Obama is likely to be re-elected.
That’s why I jumped all over it last night when Nate Silver described a reasonably plausible scenario in which Romney emerges victorious (see HERE). But Silver also noted that Obama still seems likely to win.
Let’s be honest. Romney can, in fact, win. But more than a few state polls will have to be wrong for that to happen. History suggests that such will not be the case. But new historical precedents come along every now and again, even in presidential politics. No African-American had ever been elected president — until one was. We’re told that a Republican can’t win the White House without Ohio, but that, too, might happen some day.
Anyway, I continue my search for an analysis of current polling data that makes the Mittster the favorite. And the argument has to involve credible arithmetic.
Drew Linzer over at Votamatic, a guy who’s even more optimistic than Nate Silver about Obama’s prospects, is out today with a POST that at least raises the possibility that some polls are wrong:
The accuracy of my election forecasts depend on the accuracy of the presidential polls. As such, a major concern heading into Election Day is the possibility that polling firms, out of fear of being wrong, are looking at the results of other published surveys and weighting or adjusting their own results to match. If pollsters are engaging in this sort of herding behavior – and, as a consequence, converging on the wrong estimates of public opinion – then there is danger of the polls becoming collectively biased.