{"id":623,"date":"2018-11-04T22:07:25","date_gmt":"2018-11-05T06:07:25","guid":{"rendered":"http:\/\/blogs.reed.edu\/politika\/?p=623"},"modified":"2018-11-09T09:40:04","modified_gmt":"2018-11-09T17:40:04","slug":"reed-students-forecast-the-2018-midterm-forecasts-gop-will-lose-30-seats","status":"publish","type":"post","link":"https:\/\/blogs.reed.edu\/politika\/2018\/11\/04\/reed-students-forecast-the-2018-midterm-forecasts-gop-will-lose-30-seats\/","title":{"rendered":"Reed Students Forecast the 2018 Midterm Election: GOP Will Lose 30 seats"},"content":{"rendered":"

Seventy-eight Reed College students have made a forecast of the outcome of the 2018 midterm congressional elections, and their forecast is for a Democratic takeover of the House, with an average predicted seat change of -30 for the GOP.\u00a0<\/strong><\/p>\n

Reed College students taking Political Science 260:\u00a0American Politics and Policy <\/a>with Professors Paul Gronke<\/a> and Chris Koski<\/a>\u00a0have been working for the past week on a “fundamentals” forecasting model of the 2018 midterms, drawing on insights from articles in\u00a0PS: Political Science and Politics<\/em><\/a>\u00a0and a wonderful Twitter thread about “silly” variables.<\/p>\n

Kudos to\u00a0G. Elliott Morris<\/a>, Brad Spahn<\/a>, Derek Willis<\/a>, and Deborah Sharon<\/a> for participating in the thread and helping to get our students excited about the project!<\/p>\n

Now to the forecasts! The students estimated a model based on two variables, presidential approval (collected from “a credible source”, ha ha evil professor, what constitutes “credible” you ask), a measure of the economy, and one “silly variable”.<\/p>\n

Student estimate the GOP seat loss to be between 30 and 40 seats, but there are a few, shall we say, outliers? I expect that the 170 seat loss is an overly optimistic Democrat, along with those students how predicted a 30-40 seat gain.\u00a0\"\"<\/p>\n

Or perhaps a math error? We shall investigate tomorrow in class.<\/p>\n

When we remove outliers (any forecast < -50 seats and > -10 seats), the average student forecast is a 31 seat loss for the GOP<\/strong>, which would of course result in a Democratic takeover of the House.<\/a><\/p>\n

There seems to be no patterns of unusual forecasts by section. (We allowed students to collaborate in groups of up to 3. The blank values are student teams who combined work across sections.)\"\"<\/p>\n

I know you all have been waiting for the silly variables, so I won’t hold you in suspense any longer. A number of these additional predictors sound distinctly non-silly. Apparently, irony is dead in 2018.<\/p>\n

The list of silly variables (not Silly Party):\u00a0<\/p>\n