Will you have Senate needles? Two of these states have interesting Senate races, too.
Unfortunately, no. We’re focusing our needle attention on the presidential race. We don’t get many opportunities to test this stuff, and we’re reducing the number of moving pieces to maximize our chance of success.
Can the needle make race calls?
No. The needle does not make race calls or offer definitive statements. As always, race calls or projections will be made by organizations like The Associated Press, CNN and Fox News.
Does 70 percent mean 100 percent? Does 90 percent mean 100 percent?
No. While most of us have a tendency to round probabilities up or down, 70 percent really means 70 percent. When the National Weather Service says the chance of rain is 30 percent, you still might bring an umbrella with you when you leave the house.
How does it work?
We start the night with a weak expectation of the final result for every precinct, by vote method. An average precinct in one of these states may have around 2,000 voters. If you’re in one of these three states, we have an estimate for how absentee, early and Election Day votes will break down in your neighborhood.
How do we make these estimates? We use our Times/Siena polls to estimate how each registered voter in these states will vote, based on demographic characteristics and whether that person voted early, according to state records. We adjust these estimates to match the pre-election polling, and then we aggregate our estimates up to the precinct, by vote method.
Next, the results. Some of our colleagues will be gathering and validating precinct results from hundreds of web pages across these states, in real time. We’ll pass those results off to the needle, which will compare the results with our pre-election expectations. If one candidate is doing better than we expect, our expectations for the remaining votes will shift accordingly. The same concept applies to turnout. If the Election Day vote is higher than expected, the needle will expect higher Election Day turnout elsewhere.
As the night goes on, the needle will start to make even more specific inferences. It might be able to glean, for instance, that Election Day turnout is higher in Black precincts than we expected, but not in white ones. Or it might figure out that Mr. Trump is doing better in mostly white working-class rural areas, but not elsewhere.