In the next few days, Decision Desk HQ is slated to roll out an official forecasting tool for the 2018 midterm elections to the United States House of Representatives. That tool — a statistical model, which we ominously will call “the model” from here on out — is built by me, G. Elliott Morris. We wanted to take this time to introduce some of my past work and hopefully provide a bit of background information before we (almost officially) kick off the 2018 House elections season.
I first dipped my toes into election forecasting in the Fall of 2015 when Vermont Senator Bernie Sanders became a formidable non-Hillary Clinton option for the 2016 Democratic nominee for president. The adventure was as much a learning one as it was a consumer one; the state-by-state forecast of Clinton’s chances of winning helped me develop a deeper understanding of statistics and computer programming. I never meant to start a blog that would receive much attention, but the following general election dropped that opportunity at my doorstep (largely due to the kind folks at this website!).
Naturally, I developed a model to forecast the 2016 US Presidential election (it gave Trump a 15% chance) to build out more of my forecasting skills. At the point I launched the live model in July of 2016, it had become an obsession of mine. I was on the cusp of predicting an entire presidential election, I thought. That didn’t happen — but I learned quite a bit and gained a few loyal fans. I also co-authored a book chapter on the election, which you can check out here.
In more recent months I have built models that (successfully) predicted the Presidential and Parliamentary elections in the Netherlands, France, and the United Kingdom. I’ve written some lower-level pieces on the 2018 midterms and read up on some past forecasting methods for them.
The result of those labors is a forecasting model that uses insight from my successes and failures to bring you a reliable estimate of the 2018 House midterm election. It will appear on DDHQ in the coming days. Of course, you can view the live model I’ve already published on my blog here.
If you want to get in touch about the forecast model, feel free to contact me via twitter or email. Please note: when I’m not forecasting elections I am in my final year earning degrees in government, history, and computer science so responses may take a few hours. (In that vein of thought, if you want to get to know me even better the “About” page on my blog is packed with information).
Over the next 14 months, I hope you will join us as we monitor all 435 House midterm elections from every angle. We will be publishing daily rundowns of House election news — who’s running, who’s not, who’s earning the big bucks, etc. I’m excited to be watching carefully to provide quantitative analysis for the Decision Desk HQ!