Thursday, November 1, 2018
Is there a Blue Wave? (11/1/18)
In the past I generated models based on polling data to predict election outcomes. However, models can only be as accurate as the polling data I am trying something different this year. I am only going to look at early voting numbers. A few states provide some demographic information on their early voting totals such as party affiliation, ethnicity, and gender. By comparing early voting numbers with prior years we may be able to obtain some good information. For instance, in 2016 the early voting numbers in Nevada and North Carolina told me the average polling numbers in these states was wrong. And that ended up being correct. The average of polls in Nevada gave Trump a slight lead of between 1 and 2 points and Clinton a 2 to 3 point advantage in North Carolina. However, voting trends from 2012 to 2016 showed higher Democratic turn out in Nevada and higher Republican turn out in North Carolina. The data was correct and the polls were wrong. Why? Because the sample size is enormous in states that allow early voting. In 2016 here are some averages for national early voting: Democrats = +7.5 (% of Democrats - % of Republicans) Female = +14.1 (% of Females - % of Males) White = +39.1 (% of Whites - % of Blacks - % of Hispanics) In 2018 the averages look like: Democrats = +4.9 (% of Democrats - % of Republicans) Female = +9.2 (% of Females - % of Males) White = +38.7 (% of Whites - % of Blacks - % of Hispanics) Here are a few State trends that have key races from 2016 to 2018: Arizona +2.7 R Florida +3.1 R Colorado +0.7 D Georgia +2.3 B, +3.3 M North Carolina +2.8 D, +3.5 W, +3.3 M Nevada +3.5 R West Virginia +5.4 R For example, North Carolina means there is 2.9% more Democrat turnout, 5.2% more white turnout, and 4.7% more male turnout in 2018 than 2016. Higher Democrat turnout is obviously good for Democrats, but a higher white and male turnout is more favorable to Republicans. Overall, since blacks vote in higher percentage for the Democratic candidate and males lean towards Republicans I give a slight advantage to Republicans even though there is a significant higher Democrat turnout, but North Carolina has a high Democrat populous. Conclusion: Things are trending very fast towards the Democrats. Ideology turnout is down to 2.6%, 0.4% in favor of a higher minority turnout, and the gender turnout is down to 4.9% in Male turnout. Every data point moved the democrats way except Nevada. I suspect at this time things favor Republicans in Arizona, Florida, and Nevada whereas Democrats have the edge in Colorado, Georgia, and North Carolina.