Last month’s Washington Post/ABC poll delivered some bad news for Democrats from a sample that should have favored them (+6 among adults, +5 among registered voters). Today’s WaPo/ABC poll shows them bouncing back. Guess what changed? Oh, let’s not always see the same hands …
Less than a month before the midterm elections, the political landscape remains strongly tilted toward Republicans, although Democrats have made modest improvements with voters since their late-summer low point, according to a new Washington Post-ABC News poll.
Democrats have cut in half the GOP‘s early-September advantage on the question of which party’s candidates voters say they will support on Nov. 2. They have also made small gains on the question of which party people trust to handle big issues, such as the economy and health care.
Voters give Democrats a significant edge as the party that would do a better job in helping the middle class, which has been a key campaign message from the White House in recent weeks.
President Obama‘s approval rating has rebounded to where it was in July after hitting an all-time low a month ago. Also, in some state races, Democratic candidates have taken the lead over their Republican opponents or narrowed GOP advantages.
And how did the Democrats manage this rather remarkable comeback? Well, the WaPo/ABC pollster managed to find their usual sample gap. They went from a 31/25/39 D/R/I split in September in the general sample and 31/26/37 among registered voters, to 33/23/29 in the general sample and 34/25/37 among registered voters. That nine-point advantage to Democrats among RVs is almost twice what it was in the previous sample.
To believe that this represents the electorate, we would have to believe that (a) Democrats have had a big month in attracting voters to their banner, (b) Republicans somehow lost a bunch of voters in the same period, and (c) Democrats now have an advantage outstripping their 2008 situation when they won the presidency by seven points in the popular vote. Not even their own poll supports any of those conclusions, and both Gallup and Rasmussen this year put the partisan ID split among the general population at between 1.5 and 3 points.
Another point: the article discusses the “likely voters” that give the GOP a six-point edge on the generic ballot. That’s down from 13 points in September, by the way, which comes from having three points more Democrats and one point fewer Republicans in the sample, as well as two points fewer independents, but that’s not the main issue with this calculation. Nowhere in the polling data does it show how the pollster determined the likelihood of respondents to vote — or what the sample breakdown of that group actually is.
This isn’t a predictive model for a midterm election in four weeks. It’s a narrative support device, conducted to allow media outlets to report that Democrats are making The Big Comeback and help them rally the troops.