Assessing the generic Congressional poll

In case you haven’t noticed, I love collecting and analyzing polls—they tell me the score in the game of politics. Like I did in 2008, this year I’ll collect polls and do analyses for the presidential races, the Senate races, and the gubernatorial races.

What I won’t do (with a few exceptions) is analyze races from the House of Representatives.

Why ignore House polls? Because there are too damn many House races and too few polls released for them. That’s why.

Seriously, this is a hobby, and I can hardly keep track of the relatively small number of Senate and gubernatorial races without having to keep track of an additional 435 House races. Instead, I follow a few races of interest, mostly from Washington state. And I follow the “generic congressional polls.”

Generic congressional polls ask something like, “If the election for the U.S. House was today, would you vote for the Republican candidate or the Democratic candidate.”

Empirically, subtle swings in the results of these polls seem to be amplified into larger electoral swings. They somehow capture the mood of the electorate.

For the past year there have been over 80 generic congressional polls taken on a national sample. Here is what they look like in aggregate over the past year:


The trend is encouraging for Democrats, who have gone from a deep deficit in late 2010 to what looks like a tie in early 2012.

Notice anything funny about the graph?

Yeah…it’s pretty obvious, isn’t it. The “green” pollster seems to have a (roughly) 5% bias in favor of Republicans.

Wanna guess who that pollster is? If you guessed Rasmussen, you’re right. Here are the Rasmussen polls shown alone:


If you are a Republican, perhaps this is the only pollster you watch. If so, then things looked okay until today, when Rasmussen reported for the first time in a year a net advantage for the generic Democratic candidate. But don’t panic, my Republican friend…I mean, it could just be a fluke.

Or not. Here are all the other pollsters except Rasmussen:GenericCongress30Dec11-30Jan12Congress-Rasmussen

If we are to take the collective findings of eight other pollsters over Rasmussen, the generic Democratic House candidate has led the generic Republican House candidate since sometime in August 2011.

In fact, you might say that, without the Rasmussen results, Republicans might have been in a state of panic for the past six months—instead of just the past 12 hours.


  1. 1

    Roger Rabbit spews:

    Last week, an NBC/WSJ poll showed nearly identical percentages of Democrats, Republicans, and independents — about 55% — want to replace every single member of Congress.

  2. 2

    Roger Rabbit spews:

    Just when you thought Republicans couldn’t get stupider …

    “The White House is criticizing comments by the Republican National Committee chairman comparing President Barack Obama to the Italian cruise ship captain who allegedly abandoned his sinking ship.

    “RNC Chairman Reince Priebus said Sunday on CBS’s ‘Face the Nation’ that Obama was ‘our own little Captain Schettino.’ … Former RNC Chairman Michael Steele told MSNBC the analogy was ‘unfortunate.’

    “In response, White House press secretary Jay Carney said: ‘If you are so desperate for attention that you make an analogy that Michael Steele deems inappropriate, you know you’ve probably gone too far.'”


    Meanwhile, after proving they can’t run an election even in their own caucuses, Republicans apparently have fired their Iowa state chairman.

  3. 4

    rhp6033 spews:

    # 2: Gee, and I thought the only job I would have been willing to give Republicans was to manage their own caucuses. Apparantly that’s not even open to them. What other jobs are they qualified for? It defies the imagination.

    Of course, with the Republicans you should always consider that corruption is a more likely cause than mere incompetence. Occassionally, with Republicans, you get corruption AND incompetence.

  4. 5

    rhp6033 spews:

    Those of us without the patience or advanced math skills to do this kind of poll analysis really thank you for your efforts, Darryl.

    Just out of curiosity – what changes have been made to polling over the past decade to account for the migration of households from land lines to cell phones? My assumption is that this tends to disproportionately exclude younger adults from surveys, unless adjustments are made. It seems that the “angry (older) white males” are the ones who still insist on keeping a land lines, whereas younger people don’t see the point in paying for a land line when they already have a cell phone.

  5. 6


    Zotz & rhp6033…my pleasure!

    Just out of curiosity – what changes have been made to polling over the past decade to account for the migration of households from land lines to cell phones?

    Some pollsters now do a cell phone sub-sample–e.g. SurveyUSA. Other pollsters target democratic characteristics, and adjust their results for deviations from their ideal.