2018 Midterm Predictions

Michael Darmousseh
Sep 4, 2018 · 3 min read

2 years ago I predicted Trump would win due to what I saw in polling errors when conducting robot vs live polling. This year is a little different, turnouts in midterms are much lower and tend to draw in fewer independents and more partisans. Also, races are highly localized except for a few. This makes them also impossible to predict, even with good data.

Instead of trying to predict the exact numbers, with 2 months still until the election, I want to give a range, and then 2 weeks before the election I will shorten it to a much smaller range. Before I predict, let’s start with the data. I’ve been following 538’s numbers very closely as well as the polls that show up on realclearpolitics. There are a few key races I want to focus on that I think give me a strong suspicion of what is going to happen.

First, California’s 48th district. Let’s look at the key demographics: Wealthy & Suburban. Went for Clinton by 2% in 2016, but the congressman is an incumbant and won fairly sizeably last time. Also, in a deep blue state, California. This is probably the most toss up of all races that I’ve seen. Demographic and regional trends point to a democratic landslide in this district, but the power of incumbancy and local policy first also matters. I believe that the democrat will still win this race by about 3–4% because of the other factors I will mention later.

Second, Kansas’ 2nd district. Demographics: A mix of urban, suburban and rural. 90% white. You might not believe that this is listed as a toss-up as Trump won the district 56% to Clinton’s 37%, but if there is a blue wave, then this is one of the districts that can turn. However, with no incumbant and republicans not really rallying around a single candidate, this is a district that can easily shift on turnout. Despite this, I believe that the republican candidate will win this by at least 8% if not much more than that.

In order to explain my belief on these two races, I want to look at a few data points that can be summarized as follows:

Democrats are going to have a wave in deep blue and deep red territory, but will not show up for the middle of the country.

Trump has brought more attention to the media and politics more than any recent president. Both his detractors and followers are highly motivated to participate in the upcoming election, especially the detractors. Unfortunately for democrats, detractors tend to be concentrated in two areas: Already highly blue states, like California and oregon will benefit and this might be enough to tip the house in the democrats favor, but the detractors aren’t significant enough in places like Kansas 2nd to overcome regional trends. The 7% lead that 538 gives to democrats is not enough to win a majority on it’s own due to self selection and urban rural divide.

For a specific analysis, I’d like to use a recent poll with a large sample size

If you look on the last page you will find a 7 point gap for democrats. However if you look at 2014 around the same time frame, you will also see a 5 point gap for democrats. This poll also has about a 2.8% margin of error, which means that the actual gap is 95% likely to be anywhere from 4 points to 10 points and a 5% chance that it’s outside that. Let’s ask the question, how did 2014 actually turn out? Republicans ended up with +6%. The polls were off by 11%. This was incredibly inaccurate. Why? Again, because these are very hard to predict. Do I believe this will happen again? No, 11% is an enourmous error and current events, such as the economy, foreign policy, and congressional favorability were very low in 2014 which drove up turnout by republicans. The economy is much more favorable today and is less likely to be a factor, instead Trump himself is likely to be a factor that will drive up turnout among democrats in certain locations, but I am not neccesarily convinced it will be enough. I believe the house is currently in a toss up with +/- 5 seats either way. This is close enough to basically have a divided house since coalitions with that small of a majority are almost impossible to form in the house.

TLDR; I predict that the house is a toss up with it going +/- 5 seats either way.

Michael Darmousseh

Written by

Engineer at DocuSign

Welcome to a place where words matter. On Medium, smart voices and original ideas take center stage - with no ads in sight. Watch
Follow all the topics you care about, and we’ll deliver the best stories for you to your homepage and inbox. Explore
Get unlimited access to the best stories on Medium — and support writers while you’re at it. Just $5/month. Upgrade