Skip to main content
ABC News
Why Generic Ballots May Underestimate Democrats

In the last couple of posts on our House forecasts, I’ve noted that the generic Congressional ballot tends to show worse results for Democrats than other indicators like polls of individual House races. The generic ballot, as of last week, pointed toward a Republican lead of 7 or 8 percentage points among likely voters, according to our estimates. Local polls are a little bit less straightforward to read because you have to make a lot of adjustments for polls that are (i) partisan, (ii) bad, or (iii) both. But in general — while there is a lot of variance from district to district — the local polls seem to point toward a House vote that would be about evenly divided.

My interest in this question was piqued yesterday when I came across a poll of likely voters in New York’s 24th Congressional District, in which a vulnerable Democratic incumbent, Michael Arcuri, has a 48-to-40-point lead over his Republican challenger, Richard Hanna. The 24th is one of those districts that Republicans might need to win if they are going to make monumental gains in the House: in presidential elections, it has been slightly more Republican-leaning than the country as a whole, and Mr. Arcuri has been in office for less than four years. So while I could show you a half-dozen similar districts in which the Republicans have strong numbers, this particular result ought to be modestly comforting to Democrats.

What’s interesting about this poll, however, is that it also asked respondents whether they would rather see Democrats or Republicans control the House. And there, voters gave Republicans a 3-point advantage, 46 to 43. So, even though a plurality of voters in this district want to see Republicans in charge of Congress, a plurality also wants to re-elect the Democrat, Mr. Arcuri.

Technically, a question about which party the voter would rather see control Congress is not a generic ballot question. Instead, a generic ballot question is something like this:

“If the election for Congress were held today, would you vote for the Democratic candidate in your district or the Republican candidate in your district?”

That is, a generic ballot question asks voters which party’s candidate they’re planning to vote for in their district. But the candidates usually aren’t named — instead “the Democrat” is pitted against “the Republican.”

How poll respondents actually interpret this question is open to debate, however. They may read the question as though it were asking about which party they’d rather see control Congress, or how they might vote if an average Democrat were matched up against an average Republican — or simply which party they like better. Also, they may not be aware of which candidates are actually running for the House in their district — or, they may know the candidates, but not know which parties they represent. (Candidates from both parties sometimes try to conceal their partisan identity — Democrats more so than Republicans in this sort of political environment.) So, while the generic ballot can be a good gauge of the overall political mood — and can sometimes be superior to local polls early in the political cycle when one or both candidates in a particular district are unknown to voters — it cannot always be interpreted literally.

There is another set of polls, however, that come closer to providing a side-by-side comparison of the generic ballot and the same House races with the names of the candidates revealed. These are the set of polls from the American Action Forum, a conservative-leaning 501(c)(3) organization that has released polls in 31 House districts so far.

The American Action Forum pollsters asked two relevant questions of their respondents. First, they asked:

In general, would you prefer that the next congressman from your Congressional district be a Republican or a Democrat?

Then, however, they asked another question in which the identities (and parties) of the candidates were revealed, e.g.:

If the election for Congress were being held today and the candidates were Spike Maynard, the Republican, and Nick Rahall, the Democrat, for which candidate would you vote?

Republicans did better on the first set of questions, which asked voters whether in general they would prefer to see a Democrat or a Republican elected in the district. On average, over the 31 districts, Republicans led on this question by 6 points: 39 to 33.

When the candidates were named, however, the Democrats’ gap was lessened. They trailed by an average margin of 2 points, 43 to 45. (Results from the individual districts that American Action Forum tested are shown below.) That might imply that the generic ballot overestimates Republicans’ standing by about 4 points, at least in swing districts.

Results of American Action Forum Polls in 31 Congressional Districts: “Generic Ballot” versus Ballot with Named Candidates

A few caveats (as usual) are in order here. First, the set of questions that American Action Forum asked first — what I refer to as the “generic ballot” in the table — does not quite make for a perfect, side-to-side comparison with the question in which the candidates were named. A voter might in general prefer to want to vote for a Republican, but might vote for the Democrat anyway because he happens to like that candidate. Second, in American Action Forum’s “generic ballot” question, a lot of the poll respondents were undecided or explicitly told the pollster that their ballot preference depended on the identities of the candidates, which creates some additional ambiguity.

Still, the notion that specific Democratic candidates do slightly better than generic ones would square with what we’re seeing elsewhere in the data. It also arguably squares with the respective strategies of the two parties, as Republicans are generally trying to nationalize the race, while Democrats — lacking much in the way of a coherent national message — are trying to localize it. Mr. Arcuri, for instance, the Democratic incumbent we discussed above, did not vote for the Democrats’ health care bill the second time it came up for a vote in the House (although he did vote for it the first time), which might immunize him from some “generic” Republican attacks on “Obamacare.”

None of this is likely to save Democrats from having a rather poor November. But, it could make the difference between their losing around 55 seats in the House, which is about what you get if you look at the generic ballot and ignore all other indicators, or more like 40, which is about what you get when you look solely at local indicators and ignore the generic ballot. Our forecasting model, which looks at some combination of the two, now pegs’ Democratic losses at around 45 seats but with a large amount of uncertainty on either side of that estimate.

Nate Silver is the founder and editor in chief of FiveThirtyEight.