Naim Kabir
Nov 11, 2020

--

Re the Y-axis, I think it's just the difference between % votes for Trump MINUS % votes for R Senate/House candidates.

So a high value meant a precinct had more votes for Trump relative to other Republicans, and a negative value meant a precinct had lower vote turnout for Trump relative to other Republicans.

With tha context:

1. The Y axis data is just a measure of Trump votes relative to Republican votes. If everyone who voted for Republicans down-ballot also voted for Trump 100% of the time, we'd expect this to be 0.

2. You're correct, positive values reflect people who voted for Trump but not necessarily Republican down-ballot candidates.

3. Remember the Y axis is Trump vote % - Downballot-R % . It's certainly possible to have a negative value if there are people who voted for R down-ballot candidates but voted Independent or Democrat for the Presidential ticket.

I'm not sure if that addresses your concerns but hopefully adds some helpful detail to clarify the picture here?

--

--

Naim Kabir
Naim Kabir

Written by Naim Kabir

Engineer. Focused on experimentation, causal inference, and good software design. naimkabir.com

Responses (1)