2022 Election prediction retrospective.

The 2022 midterm elections have come and gone, leaving behind a lot of data and assumptions going forward to the 2024 presidential elections. In this article, we’ll go over the results of the election and compare them to the predictions made by me (published here). In a following article we’ll talk in general about the trends and realities of the American electorate and its effect going forward and closer to the presidential election of 2024. I want to give a special shoutout to Logan Phillips (@LoganR2WH on Twitter), founder of racetothewh.com for having by far the most extensive and well researched forecast this year. Logan Phillips has been a major inspiration in improving my own forecasting skills and bettering myself ahead of the 2024 presidential race.

Senate

In my election forecasting, the Senate would be won by Democrats 50 to 49, with Georgia going to a run-off. My prediction was 100% correct in predicting the outcomes of every race. Georgia, as predicted, will be heading to a run-off after neither Warnock nor Walker managed to get passed 50% of the vote. Democrat Warnock ended up 0.9% ahead of Republican Walker, that being 0.2% higher than the margin I had forecasted. In Arizona, Democrat Kelley beat Republican Masters by 4.9%. While the I correctly predicted the outcome, Kelly performed substantially better (3.4%) than my forecasted margin of victory of 1.5%. In Nevada, Democrat Cortez Masto kept her senate seat by beating Republican Laxalt by 0.9%, this is 0.5% higher than the margin of victory I forecasted when correctly predicting the outcome. While last but not least, Democrat Fetterman won against Republican Oz by 4.8%, doing much better than my 0.5% expected margin. When we look at the margin of error in the 4 Senate battleground states, we see that I had a 2.1% average margin of error. My margin of error of 2.1% in these 4 races is lower/better than Logan Phillips Race to the White House 2.53% and Nate Silver’s 538 margin of error of 2.93%.

House of Representatives

Democrats had an outstanding year in the House of Representatives elections, managing to get 213 seats to Republicans 222. This is a 5 seat Democratic overperormance of my prediction which stated that Democrats would be holding 208 and Republicans winning 227. Me being off by 5 seats puts me ahead of other major forecasters like 538 (off by 8), Decision Desk (off by 10), Split Ticket (off by 12), Real Clear Politics (off by 22), Washington Post (24). There were only 2 other major forecasters that did better than my prediction in this area, those being Logan Phillips Race to the White house, only off by 1 seat, and the Economist, which was only off by 3 seats. Another way of measuring how good or bad a political prediction is, it’s by looking at the percentage of races called accurately. In the House of Representatives, I had a 96.6% accuracy. Essentially, I predicted 96.6% of races outcomes correctly. That comes out to 420 correct, and only 15 incorrect. My 96.6% accuracy is tied for first place with 538, followed by Race to the White House and the Economist with 96.3%, Split Ticket 95.4%, and Decision Desk with 94.7%. Ten of the fifteen campaigns wrongly predicted were races won by Democrats but predicted Republican, while the other 5 were the opposite, won by Republicans but predicted Democrat. Four of the five seats that were predicted as Democrat but were actually won by Republicans were in New York, showing the poor Democratic performance in that state. Those four races in New York are congressional districts 3, 4, 17, and 19. All four being major wins for Republicans who had little hopes of winning, with congressional district 17 being a major win for Republicans, ousting the third highest ranked Democrat at the time. If Democrats had kept these 4 seats, they would’ve been 1 seat away from the 218 needed to hold the House of Representatives. The map above shows the districts that were predicted wrongly (in red), and the ones predicted correctly (in green). Keep an eye out for a following piece on the general trends, realities of the American electorate, and their effect with regards to the presidential election of 2024.

Bibliography

8, N., 6, N. and 30, S. (no date) House 2022 election - national forecast, Race to the WH. Available at: https://www.racetothewh.com/2022/house/final (Accessed: November 23, 2022).

Gil, M.A.L. (2022) 2022 general election prediction, El Espirituano. El Espirituano. Available at: https://www.elespirituano.com/post/2022-general-election-prediction (Accessed: November 23, 2022).

NateSilver538 (2022) 2022 Senate election forecast, FiveThirtyEight. Available at: https://projects.fivethirtyeight.com/2022-election-forecast/senate/pennsylvania/ (Accessed: November 23, 2022).