The Generic Ballot poll is probably the most accurate thing we can use right now to predict the outcome of the 2018 midterms. In fact, the generic ballot use used heavily in my 2018 Senate Predictions, which weighted the polls at 35%. However, as we all know, polling is no clear prediction. The 2016 polls were off just enough that it seemed terrible, and polls taken just before the Alabama Special Election showed Roy Moore leading. So, just how accurate is the generic ballot poll, especially this early out?
A generic ballot poll normally asks just one, non-specific question: Which party would you support in a Congressional election? The question mentions no candidates or specific races. For the most part, the results of this poll are pretty accurate. The margins of the generic ballot have consistently equated to the margins in the following midterms. Generic ballot polls taken on final days before midterms are extremely accurate, only ever deviating by about 2 points maximum. However, we’re not days away from the 2018 midterms. We’re months away. So, how will current generic ballot results play out? Well, according to an article by FiveThirtyEight, the White House’s party tends to do worse than the results of polls taken months out before the midterms. So, we can expect that Republicans, being the party in the White House, will do worse than what is shown right now. At the moment, Democrats hold about a 7 point lead over Republicans on the generic ballot, down from a nearly 14 point lead in late 2017.
If historical results prove to be predictive (which isn’t always true), then we should see Democrats getting above a 7 point margin come November.