Here’s a remarkable thing about the 2018 election campaign: how little has changed. That is the lesson from the third George Washington University Politics Poll, which was just released today. I think most people know that we live in a world of strong partisanship, which is increasingly evident in congressional voting. But how much partisanship helped stabilize people’s preferences in this election season still surprised me.
The GW Politics poll is unusual because it has interviewed the exact same group of registered voters three different times in 2018: May, July and now Oct. 17-25. All told, 2,321 respondents took all three polls, which is 73 percent of the original 3,150 respondents first interviewed in May.
Take the “generic ballot” question, which asks respondents to choose between the Democratic and Republican candidates in their district. In May, the split among these respondents was 46 percent for the Democrat and 37 percent for the Republican. About 13 percent were undecided.
Five months later, the split was virtually identical: 48 percent to 40 percent. The percent who were undecided was lower (7 percent) but any shifts from undecided to “decided” didn’t really advantage either party.
Another way of looking at this: 96 percent of those who chose the Republican candidate in May also chose the Republican candidate in October. And 97 percent of those who chose the Democrat in May did so in November.
Altogether, 80 percent of this sample was composed of partisans whose loyalty did not change. This wasn’t much lower than the 90 percent who consistently approved or disapproved of President Trump across this time period. Consistent partisanship is simply the norm.
In addition to their lead in the generic ballot, Democrats have also led on several other dimensions and did so again in this October poll:
But, again, since July these numbers have only changed by a few percentage points at most. (See the July numbers here.) For example, in the October poll the Democratic candidate’s advantage in job approval (54 percent vs. 46 percent approval for the Republican) was only slightly better than in July (52 percent vs. 48 percent).
The same stability was evident across a range of political issues, including the economy, health care, taxes and immigration. Despite months of efforts by both parties to shift opinions on these issues, they have remained relatively constant.
Perhaps most notably for the GOP, consistent reports of economic growth have not improved the public’s perceptions of the economy. In October, 39 percent said that the economy was getting better (versus getting worse or remaining the same). In May, it was 37 percent.
A final thing that we have been tracking: voters’ involvement and participation. In addition to asking whether they intended to vote, we have asked whether they had taken other kinds of action in the past three months — like signing a petition or talking to someone about how they should vote. We also asked whether they had been contacted by a political party.
Overall, there was little sign of increased participation, with the exception of a small increase in the percentage who had talked to someone about how to vote. As in previous GW Politics Polls, Democrats and Republicans were equally likely to say they intended to vote, but Democrats had an advantage in several other areas:
Of course, there are caveats. Perhaps the people who remained in our survey over these past five months tend to have strong opinions that change less frequently. Moreover, this national sample cannot capture any changes in individual districts. Nor does it capture opinions about the Senate races, where the forecast really has shifted.
But the results of this poll are in line with what other polls and forecasts are showing about the battle for the House: Overall, things just haven’t changed that much:
The only remaining question is how much Democrats can turn favorable national conditions, polls and strong fundraising into a House majority and, if so, how big a majority.