An analysis on our model and the predicted 'Blue Wave'

    The past year’s political whirlwind has come to a peak before the midterms in an expected Blue Wave. ‘Blue Wave’ has been all over social media and the news. No matter your political affiliation, the last year was tumultuous for everyone. In this essay, we want to dive deeper into which districts ORACLE is predicting will flip from Republican over to Democrat and identify the reasons why our model predicts such shifts. As we identify trends between districts, we will try to discover the overall cause of this ‘Blue Wave’ and determine if it truly qualifies as a historic anomaly. We’ll also look at our model and see where there could be flaws that are misrepresenting the current race. These are our conclusions.

    To start, the national mood in our model (named ‘bigmood’) is leaning Democrat, due to the generic ballot. ‘Bigmood’ is essentially the difference between the generic ballot and the prediction of our fundamentals only model. This serves as a sort of correction factor for our fundamental prediction, since our ‘SEER’ prediction, if it is off, should be consistently off in all districts.  This generic ballot lean towards Democrats, which is the premise for the predicted ‘Blue Wave’. We just need to make sure that ‘bigmood’ isn’t influencing our model in the wrong direction and giving us false expectations about Democrats winning. To find districts that flipped from 2014 or 2016 in 2018, we created an Excel spreadsheet of all 435 congressional districts (excluding Pennsylvania due to excessive gerrymandering) and our model’s predictions for those districts. We then investigated those ‘flips’ to ensure they weren’t simply the product of the bigmood carrying too much weight. Those that flipped parties -- defined as a Democrat winning a simple majority of the vote one year and a Republican winning a simple majority of the vote or vice versa -- were assigned an arbitrary value of 1 for flips between 2014 and 2018 or 2016 and 2018, and all other districts were assigned a 0. Sorting the Excel sheet by our column of binary values yielded 38 districts that ORACLE predicts will flip this year.

    Our most consequential finding is that every district that shifted to the Republican party in 2016 is predicted to switch back in 2018. Out of the 38 districts we investigated, seventeen traditionally had Democratic representatives yet elected a Republican candidate in 2016. All 17 of these districts are predicted to move back to having a Democratic representative, which shows that only 21 of our districts flip due to abnormal electoral shifts; in other words, 17 of our 38 districts flip regularly, while 21 of them are flipping this year due to unusual factors, such as scandals, a particularly low presidential approval rating, or intense campaigning and fundraising by Democrats. We also looked into which (if any) districts have had candidate related scandals that influenced their voting turnout. What we discovered is that only a few flipped districts have had scandals. In CA-48, it was revealed by the FBI that the Republican incumbent, Dana Rohrabacher, is viewed by Russia as a viable US intelligence source. This caused Democrats and Republicans from CA-48 to to jump ship, and is reflected in the latest polls. Rohrabacher’s Democrat opponent, Harley Rouda, is up by three points. The Republican incumbent in IA-01 failed to report that he is the CEO of an internet company that helps companies bury any FDA citations below positive search results. Clearly, since only those two flipping districts have scandals, our model’s excluding of scandals is not impacting our predictions: our inclusion of polls into our final prediction accounts for shifts in public opinion based on scandals.

    ‘Bigmood’ and ‘SEER’ have different predictions and only one of those districts is a toss-up, Minnesota-07. SEER predicts it will go Republican, ‘bigmood’ leans Democrat, proving that demographics play a big role in predicting results. SEER will be off because it heavily weights the party of the president, and thus isn’t taking into account the President’s popularity. RealClearPolitics has Democrats up 7.6% in the generic ballot for 2018, however in 2016 the end generic ballot, their prediction had Democrats up by .6% but Republicans ended up winning by 1.1%. The generic ballot asks constituents which party they are most likely to vote for, and can be used to see which way the nation is swinging. Generally, the nation tends to vote for the opposite party than the President as reflected in the generic ballot, and ORACLE predicts the current election will fit right into this pattern. The main drawback with considering the ‘SEER’ prediction alone is that it’s predicting the winning party based on the 2012, 2014, and 2016 election, when the President was a Democrat and the political climate was significantly different.

    The rest of the districts we investigated have no candidate-related scandals, and many of the districts that ORACLE flipped this year, flipped in 2014. These districts are not a part of a bigger movement, they are just always very close races. In fact, every district that flipped in 2016, ORACLE predicts will flip back this year.

    For example, Florida has a lot of competitive redistricting, meaning that congressmen are constantly changing the districts boundary lines. Due to this push and pull from both parties, there are various Democratic districts that are currently represented by Republicans. We did discover a trend in Florida. FL-02, -06, -07, and -18 flipped this year and also flipped from 2014 to 2016.

Conclusion

    Dubbed ‘The Blue Wave’, the rise of Democratic candidates has been a hot topic in the media, but with little substance. Though ORACLE predicts that Democrats have a 76 percent chance of taking the House, it seems this predicted ‘wave’ has more to do with the historical pattern of flipping and basic demographics than a movement. District flipping is common in the grand scheme of politics, and this year is comparable to many before. Numerous districts are so gerrymandered that past electoral history is not an accurate reflection of the current situation, plus, the current President has a record low approval rating. Though Democrats are pulling ahead, and ORACLE does predict them as winning, this election fits right into the pattern of past midterm elections.

Appendix/Districts Examined: