There have been two interesting news stories on elections in the past week as well as an interesting inside baseball geek out concerning how to model and predict Senate elections that could be either interesting outliers, or harbingers of change. The two interesting stories are the Democratic Parties of Kansas and Alaska happily seeing their …
Fundamental divergence; oddity or realignmentPost + Comments (32)
The poll aggregators are showing significantly better results for Democrats than the sites whose models including significant ‘fundamental’ weights. Sam Wang at Princeton Election Consortium is a aggregator with no special sauce predictor and has a good post on this:
I categorized models as “Fundamentals-based (Type 1)” and “Polls-based (Type 2)”. The major media organizations (NYT, WaPo, 538) have all gone with a hybrid Type 1/Type 2 approach, i.e. they all use prior conditions like incumbency, candidate experience, funding, and the generic Congressional ballot to influence their win probabilities — and opinion polls. What does that look like…
Senate Democrats are doing surprisingly well. Across the board, Democratic candidates in the nine states above are doing better in the polls-only estimate than the mainstream media models would predict. This is particularly true for Alaska, Arkansas, and North Carolina….
It is nearly Labor Day. By now, we have tons of polling data. Even the stalest poll is a more direct measurement of opinion than an indirect fundamentals-based measure. I demonstrated this point in 2012, when I used polls only to forecast the Presidency and all close Senate races. That year I made no errors in Senate seats, including Montana (Jon Tester) and North Dakota (Heidi Heitkamp), which FiveThirtyEight got wrong.
The fundamental predictors put a significant thumb on the scale by looking at factors such as incumbency, cash raised, candidate quality and the generic tilt of a state. These models would say that a generic Republican in Arkansas (a state that went big time for Romney and it generally conservative) should be significantly advantaged over a generic Republican candidate in Maine. The pure poll aggregators basically say that all of the fundamentals will be baked into the polls by the summer before the election, so polls tell a simpler story with far fewer confounding variables than a fundamental based approach.
If my memory serves me right, the Democrats have signficantly overperformed the fundamentals at the Senate level in 2010 (Nevada, Colorado, Delaware), 2012 (Indiana, Montana, North Dakota, Missouri) and seemingly so in 2014 at this time. John Kerry also overperformed his fundamentals in 2004 although it was an insufficient overperformance for a win. If my memory is right, is there a small but consistent error factor in the fundamental measures that is failing to pick up on the basic fact that there is a slight but persistant Democratic candidate quality edge that is being driven by the basic fact that the Republican primary electorate is batshit insane? And that primary electorate has a persistant ability to advance candidates who are weaker than they “should” be because they turn off some marginal members of the typical Republican winning coalition into either non-voters or hold your nose Democratic voters?