Friday, November 06, 2020

How Fraud/Cheating Affects My Presidential Election Models

One comment I've had from several people when I point out that e.g. Michigan, Wisconsin (and likely Pennsylvania) went as I predicted is that that's because of cheating/fraud.

They're not necessarily wrong -- but the fact that they're not necessarily wrong tends to support my models, not contradict them.

First, a disclaimer: My "models" for predicting elections are not notebooks full of mathematical calculations. The only real math is "here are the results from last time." Everything else is looking at the trends in play and making educated guesses as to how those trends will affect future results.

When it comes to cheating/fraud in statewide votes for presidential elections in battleground states, I assume it's a wash -- that is, that Republicans will be about as successful at suppressing Democratic votes as Democrats will be at manufacturing Democratic votes -- unless I see clear evidence that there's some kind of change in motion.

I don't assume that in "safe" states. The reason those states are "safe" is that the party in power has the clear upper hand and either doesn't need to cheat or could easily out-cheat the opposition party. It's only when a state looks competitive that I see any need to really consider cheating/fraud as a factor.

My predictions this year missed at least one state, probably two. And they probably missed precisely because I mis-underestimated the cheating success of the party in power.

In Florida, I assumed that the GOP would have, at best, only partial success in suppressing the Democratic vote by defying the will of the voters on restoring former felons' voting rights, ensuring fewer polling places in likely Democratic areas, etc. I also assumed that the "anti-Castro Cuban" (aka "save our sugar subsidies and CIA money") lobby would only have limited success in either getting its vote out, or just plain manufacturing that vote, for Trump. I was clearly wrong.

In Georgia, I assumed that the GOP would be very successful in stealing the 2020 election just like it stole the 2018 midterm, through mass de-registration of likely Democratic voters and such. It looks like I was wrong about that, too.

In both Georgia and Florida, I assumed that the Democratic vote-manufacturing scams would function about as well as usual. I've not seen any reason to believe I was wrong on that count.

No comments: