George,

The latest Quinnipiac/NYTimes/CBS poll in Virginia shows no change since Sept. 19 in the Obama-Romney race. Indeed, Obama picked up a point. The matchup is now Obama 51-Romney 46. This is among likely voters.

The Quinnipiac University polling organization conducts highly reliable polls and was chosen a few months ago by the New York Times and CBS News to be their pollster. (I am also very familiar with this poll since I trained its director when he was a PhD student of mine at the University of Connecticut.)

The latest Quinnipiac poll also found virtually no change since Sept. 19 in the Obama-Romney races in Wisconsin and Colorado. This lack of change since the Big Debate in these battleground states does not surprise me. Both candidates had appeared in person many times in these states and were well known by many before the debates. Thus, many had already made their evaluations and had decided who they were going to vote for.

Given the lackluster performance of Obama in the Big Debate, one would expect less support for him after the debate in national samples including all states, not just the battleground states. Gallup does a daily tracking poll and averages results over 7 days. The latest Gallup tracking poll which encompasses 7 days since the debate shows Obama ahead of Romney among registered voters 50-45. This spread has remained the same since mid-September.

But then there is the Pew poll that was released on Monday showing a huge change after the Big Debate. The mid-September Pew poll had Obama ahead among likely voters 51-43. In a poll conducted from October 4 through October 7 (just after the debate), Romney was ahead 49-45 among likely voters. A huge change. Pew is a highly reliable pollster, so this result is truly puzzling. It may be that Pew happened to have drawn a random sample that was at one of the tails of the hypothetical normal distribution of sample means. (I'm sure I just lost you unless you had a good statistics course and understood it.) More simply, the sample that Pew drew may have been an outlier-a rare possible sample that is far from the true mean.

As for Tim Kaine, he continues to hold a comfortable lead in the latest Quinnipiac poll, Kaine 51-Allen 44 among likely voters. This is the same as the last Quinnipiac poll in mid-September.

Of course, it takes the efforts of campaign workers to make sure that "likely" voters become "actual" voters.

David RePass (Electronic mail, October 11, 2012)