Wrote a small program to calculate polling averages. Here's the output for September polls. Positive numbers indicate a Biden lead, negative indicate a Trump lead. This is not intended to be a substitute for things like RCP or FiveThirtyEight, but just something compact and easy to read.
- Code: Select all
Maryland 32.0
California 26.7
Maine CD-1 26.1
Vermont 23.3
Washington 22.0
New Jersey 18.3
Maine 16.3
New Mexico 15.0
Oregon 12.0
Virginia 9.5
Colorado 8.8
Minnesota 8.5
United States 7.3 <-
Nevada 6.6
Wisconsin 6.6
Michigan 6.6
Pennsylvania 6.0
Nebraska CD-2 6.0
Arizona 4.3
Maine CD-2 3.7
New Hampshire 3.0
Florida 1.9
North Carolina 1.5
Ohio 1.1
Iowa -0.3
Georgia -0.5
Texas -2.0
Kansas -5.0
South Carolina -6.0
Louisiana -6.0
Missouri -6.5
Montana -7.0
Indiana -14.0
Utah -18.0
Kentucky -19.5
Oklahoma -24.4
Idaho -25.2
Each pollster's polls are averaged together, then these averages are weighted (according to 538's pollster ratings) and averaged together for each state. A-rated pollsters get a weight of 4, B a weight of 3, C a weight of 2, D a weight of 1, and unrated or F-rated pollsters get thrown out. I consider "A/B" pollsters as B-rated (same thing goes for B/C or C/D, they're C and D respectively). I could also post an unweighted version if you guys want that.
I'm not a statistician or anything, so don't crucify me too hard if I'm doing something dumb. I'll post this maybe every few weeks or something. Will note any changes I make here.