OH NBC/Marist: Tie

[NOVA Green]

Hmmmm....

I wonder if we will continue to see these types of trends in post Labor Day polls now that voters are starting to tune into the election.

I would still give DeWine an edge here, being Ohio and all, not to mention being a Household name whose been around for decades in state politics (Still remember sipping iced tea and eating cookies his wife baked on his front porch back in the early '90s in SW-OH, when he was just a lowly US-House member).

Still, been a long time "Politician" can cut both ways, especially if the voters in Ohio haven't really seen the "hope and change" they were expecting after Trump got elected President....

[NOVA Green]

I really wish we knew more about the polling methodologies that even "higher quality" polling firms are using to model the 2018 electorate in various elections.

By all indications based upon actual votes cast in Special Elections, history would tend to suggest that it is a bit different than in recent Midterm elections (2010 and 2014).

How are polling methodologies and models being adjusted? Are LV screens consistent between polling organizations? What types of internal feedback loops are in place to detect outliers and if so, how are these measured and quantified?

I guess that's one of the things I like about the NYT/Sienna House polls, is that at least they are being more transparent, and additionally have some type of "internal review" that flags races where the variance is higher than modeling suggests in order to explore *WHY* the variance exists, and is the variance a function of polling models versus actual polling responses.