As the elections neared elections, the New York Times made an interesting experiment: conducting a survey of dozens of neighborhoods and showing users the results in real time. Partnering with Siena College's survey team, the idea was that users would get a better understanding of how the election results will depend on the choice models and the number of respondents when the poll is presented in a kind of election night, in which the results are presented be observed format.
At the same time, of course, the team also produced a plethora of survey results in districts critical to the control of the house. ON ton of results. There were 90 separate house surveys, including nearly 30 districts that were interviewed twice.
The results? Well, somehow everywhere, as one might expect when questioning different neighborhoods.
Forty-four polls had a Republican in the lead. Forty-one had a Democratic leadership. Five were tied.
However, there is a more interesting way to look at these results: compare them to the expectations that can be expected in each district given how the district voted in 2016. When we do that, the diagram looks like this.
Here the pattern is slightly different: 59 districts were more democratic than the vote in 2016, while 30 districts were more Republican than 2016. (If we compare ourselves with 2012, the difference is about the same.)
As we go deeper, it gets even more interesting. If we superimpose these results on a simple trend line, we see that the shift from 2016 to the Republicans in early October has become a little friendlier, and then returned to the Democrats at the end of the election. This first period coincides with the battle for the appointment of Brett M. Kavanaugh to the Supreme Court.
Patrick Ruffini of the data company Echelon Insights came in a different way to a similar finding. He compared the daily answers in the Times poll with the first 10 days of surveys. Find In early October, the Republicans were doing much better than at the end of September. By the end of the month, the pendulum had repulsed the Democrats.
We can get a similar sense of how the survey has changed over time as we look at the races the Times has polled several times.
Here are the races where the Republican improved between the first and second Times Siena polls. In seven of the ten races, the Republican ended with a lead; In four of these races, they were behind for the first time.
Here are the races that have become friendlier for the Democrats.
In the course of the Times polls, nineteen races to the Democrats became friendlier. 14 of them ended up with the leadership of the Democrats; Six of the 14 who were led by a Republican to a Democratic.
If you are curious about how much this says about the Democrats who have recaptured the house, you should not focus on the fact that the results themselves involved a split. Instead, focus on the fact that these were almost all districts that were last held by Republicans.