What kind of bias will the polls have this year?

[Xing]

Do you think that the polls this year will overestimate Republican or Democratic support? For reference, polls had a fairly strong Republican bias in 2012, but a strong Democratic bias in both 2014 and 2016.

I think it’s likely that the polls have at least a slight Republican bias this year, since Democrats have more enthusiasm, though I could see polls being pretty accurate as well. I think some states like NV will have a Republican polling bias regardless of the bias of national polls (they overestimated Trump and Heck in NV despite underestimating Republicans elsewhere.)

[UncleSam]

Polls have generally been pretty accurate for the major races thus far, outside of VA-Gov. so I agree - could see Republican bias, could see pretty accurate. Barring an enthusiasm boost I doubt polls will overestimate Ds unless turnout models get really screwy from the polls’ side.

[Dr Oz Lost Party!]

I have a strong feeling they're underestimating the Democrats.

[Holy Unifying Centrist]

I don't think the polls are wrong w/ the house races. House dems aren't doing well in polls because 1) the dem party is very unpopular 2) No one knows who 80% of the challengers are. Name ID matters a lot in early polls. Incumbents are a lot more well known. And democrats will have a much easier job defining themselves and branding themselves effectively b/c they won't be compared to the unpopular, republican controlled Congress and presidency.

This isn't just a D thing - polls could be underestimating Senate republican challengers in WV/MT since no one really knows who Jenkins or Rosendale or whoever are.

Only polls I expect to be close to the actual result rn are North Dakota, Florida, and Tennessee, b/c everyone knows both of the candidates in those races pretty well. Also polls could be underestimating Beto b/c no one knows who Beto is

I do expect that once campaigns start, many dem challengers in the house (and rosen / possibly beto in the senate) will start to pull ahead.

[Yellowhammer]

In most cases, a democratic bias, like in 2016, but probably a slight republican bias in NV.

[Crumpets]

The polls for individual races have been pretty accurate so far, but haven't been in line with what you would expect the generic ballot to be at for those races to end up as they did. So maybe come election night the Senate races will be pretty spot on while Democrats really overperform in the House where predictions are based more on national polls than individual districts?