The special thing about the election campaign 2018: how little has changed? This is the lesson from George Washington University's Third Political Survey published today. I think most people know that we live in a world of strong partiality, which is becoming increasingly clear in the vote in Congress. But how much partiality has helped to stabilize people's preferences in this election season has surprised me anyway.
The GW Policy Survey is unusual in that it polled three times the same group of registered voters in 2018: May, July and now October 17-25. A total of 2,321 respondents participated in all three surveys, or 73 percent of the 3,150 respondents originally interviewed in May.
Take the "generic election" question, where respondents must choose between the Democratic and Republican candidates in their district. In May, the gap between these respondents was 46 percent for the Democrats and 37 percent for the Republicans. About 13 percent were undecided.
Five months later, the split was virtually identical: 48 to 40 percent. The percentage of undecided was lower (7 percent), but a shift from a draw to a "decision" did not really benefit either party.
Another view: 96 percent of those who voted for the Republican candidate in May also opted for the Republican candidate in October. And 97 percent of those who voted for the Democrat in May did so in November.
Overall, the sample consisted 80 percent of partisans whose loyalty has not changed. This was not much less than the 90 percent that consistently endorsed or disapproved of President Trump during this period. Consistent partiality is simply the norm.
Democrats, in addition to their lead in the generic vote, have also pushed other dimensions forward and done so again in this October poll:
Since July, these figures have changed only by a few percentage points. (See the July numbers here.) In the October survey, for example, the advantage of the Democratic candidate for work permits (54 percent versus 46 percent for the Republican) was only slightly better than in July (52 percent versus 48 percent).
The same stability was found in a range of policy issues, including economics, health care, taxation and immigration. Despite months of efforts by both sides to shift opinions on these issues, they have remained relatively constant.
For the GOP in particular, consistent reports on economic growth have not improved the public's view of the economy. In October, 39 percent said the economy would improve (as opposed to getting worse or staying the same). In May, it was 37 percent.
One last thing we have been following: the participation and participation of voters. We asked not only if they wanted to vote, but also if they had taken other measures in the past three months – for example sign a petition or talk to someone about how they should vote. We also asked if they were contacted by a political party.
Overall, there was little evidence of increased participation, with the exception of a small increase in the percentage that talked to someone about voting behavior. As in previous GW policy polls, Democrats and Republicans alike were likely to vote, but Democrats had an advantage in some other areas:
Of course, there are reservations. Perhaps those who have participated in our survey over the past five months have a strong opinion that is less likely to change. In addition, this national sample can not record changes in individual districts. It also does not capture opinions about the Senate races, where the forecast has really changed.
However, the results of this survey are in line with what other surveys and forecasts on the fight for the House show: Overall, the situation has not changed so much:
The only remaining question is how much Democrats can turn favorable national conditions, polls, and fundraising into a house majority and, if so, how big a majority is.