A thread on how I think about US Senate elections, since I think it's useful. Many people tend to focus on the Senate as a whole and how it shifts in a given election. However, I think the results are much more understandable when looking at the three Senate classes separately.
Nov 3, 2020 · 4:15 AM UTC
1
3
This year's US Senate elections are for Class II (other than the special elections in Arizona and Georgia). This class was last elected in 2014 (electing 22 Republicans and 11 Democrats), and before that in 2008 (20D, 13R) and 2002 (21R, 12D). Those are pretty big swings.
1
This isn't unusual for Senate classes in recent years; see the maps I made at dbaron.org/senate-classes/ .
(These are what I look at when I want to think about Senate elections.)
1
This way of thinking helps explain some people's confusion in 2018 when, in a Democratic year, the Senate shifted a few seats towards the Republicans.
2018 was a huge Democratic win in the senate (23D/10R), but it was replacing the 2012 class which was 25D/8R.
1
In this class in particular, the 2014 elections involved Republican pickups in Alaska, Montana, Colorado, South Dakota, Iowa, Arkansas, Louisiana, West Virginia, and North Carolina.
1
Going back to 2008, the Democrats picked up seats in Alaska, Oregon, Colorado, New Mexico, Minnesota, New Hampshire, Virginia, and North Carolina.
(Note that the Alaska, Colorado, and North Carolina seats were lost in 2014.)
1
And going back further to 2002, the Republicans picked up three seats (Minnesota, Missouri, and Georgia), and the Democrats picked up one seat (Arkansas).
(The Republicans would lose that Minnesota seat in 2008, and the Democrats would lose the Arkansas seat in 2014.)
1
1
So big shifts are possible. The range of scenarios in @FiveThirtyEight's Senate forecast at projects.fivethirtyeight.com… is not unusual
1