I don't consider it plausible that undocumented migrants vote in anywhere near sufficient numbers to have an appreciable chance of having an impact on US election results, and it's not even obvious to me that their votes - being heavily selected for propensity for fraud - would favor Dems, as opposed to, you know, the party that actually glorifies cheating and criminality.
It's definitely an out there theory -- estimating how much they tilt elections is difficult as it relies on precisely estimating several statistics that are unknown.
The main source for the high estimates for non-citizen voting is from some papers published by Richman with the CCES. The issue is that it's a online opt-in sample, which we know tends to have a lot of bogus respondents(it's also intended for only citizens). And there's some reliability issues with partially longitudinal survey. Sample sizes are also pretty small, but it's worthwhile doing a more updated analysis with multiple election cycles.
I’ve also developed a new prediction model for the 2024 U.S. Presidential election. This model focuses on using only pro-Trump partisan polls to see how they compare to high-quality polls that generally give Kamala Harris a 62.4% winning probability. My aim was to address claims from Democrats suggesting a "flood" of pro-Trump polls skewing major prediction sites.
Key Findings:
- High-quality polls estimate Harris’s winning probability at 62.4%, while pro-Trump partisan polls suggest Trump’s chances are at 70.5%.
- The primary reason for these differing outcomes? Substantial differences in Midwest polling predictions.
- I also included a "pro-Trump Midwest bias" model based on prior polling errors, which shows a closer race, with Trump at 51.6% winning odds.
- In battleground states, both models align when predicting a Trump lead, but they diverge in Harris-led states, making this pro-Trump partisan model a more intensified version of the Midwest-bias approach.
It's a reasonable first model. The difficulty comes from the correlated errors. If polls are wrong in state X, they are more likely to be wrong in the same direction in nearby states. It's possible to model this by simulating a bunch of correlated normal distributions based on the assumed correlations of the errors (which you can obtain from the maps also), transforming the data into the vote % etc. I guess not worth the effort considering all the other people are already doing this kind of thing but better.
My guess is just an average of election betting odds 61.5% trump and nate Silver's model(old 538 one), which is 55%. Both of them have some bias towards dems but historically nate silver is somewhat biased more so trump has a ~60% chance of winning the election.
That IA’s do or don’t vote—or that if they do vote, it is not in appreciable numbers, is a rabbit hole we’d best avoid. To me this is a meaningless and obscuring argument. The primary issue is, “Why can IA’s vote at all!” In short, these are potential holes in the current voting system that must be corrected. The current system is broken and a rational voter can only assume such flaws as purposeful in nature. There lies the problem—if elections have such flaws they can never be considered determinative so that after an election, one side or the other will never accept such as the true will of the people. In that lies the crisis of democracy. There is no reason for the current rules for exercising the franchise—except to allow deceit and fraud. For example, few other Western democracies allow mail-in ballots. Why do we? How is it no proof of citizenship is required for voting for Federal office? I can go on, but to what effect. I myself, no longer vote due to the repeated election anomalies experienced, and as such do not consider myself bound to the so called “will of the people”.
I don't consider it plausible that undocumented migrants vote in anywhere near sufficient numbers to have an appreciable chance of having an impact on US election results, and it's not even obvious to me that their votes - being heavily selected for propensity for fraud - would favor Dems, as opposed to, you know, the party that actually glorifies cheating and criminality.
Singular anecdote, etc., etc., but the one American that I know for sure was an "illegal" at one point is an avid Trump supporter. https://x.com/powerfultakes/status/1776477016386068534
It's definitely an out there theory -- estimating how much they tilt elections is difficult as it relies on precisely estimating several statistics that are unknown.
The main source for the high estimates for non-citizen voting is from some papers published by Richman with the CCES. The issue is that it's a online opt-in sample, which we know tends to have a lot of bogus respondents(it's also intended for only citizens). And there's some reliability issues with partially longitudinal survey. Sample sizes are also pretty small, but it's worthwhile doing a more updated analysis with multiple election cycles.
For the previous election, even Richman's own analysis and various court cases found that the percentage of non citizens that voted is too low to effect the election result in Arizona(https://www.washingtonpost.com/politics/2024/03/06/truth-about-noncitizen-voting-federal-elections/)
I’ve also developed a new prediction model for the 2024 U.S. Presidential election. This model focuses on using only pro-Trump partisan polls to see how they compare to high-quality polls that generally give Kamala Harris a 62.4% winning probability. My aim was to address claims from Democrats suggesting a "flood" of pro-Trump polls skewing major prediction sites.
Key Findings:
- High-quality polls estimate Harris’s winning probability at 62.4%, while pro-Trump partisan polls suggest Trump’s chances are at 70.5%.
- The primary reason for these differing outcomes? Substantial differences in Midwest polling predictions.
- I also included a "pro-Trump Midwest bias" model based on prior polling errors, which shows a closer race, with Trump at 51.6% winning odds.
- In battleground states, both models align when predicting a Trump lead, but they diverge in Harris-led states, making this pro-Trump partisan model a more intensified version of the Midwest-bias approach.
Check out the full breakdown and analysis here: https://kaijaeger.substack.com/p/prediction-update-and-a-new-model
It's a reasonable first model. The difficulty comes from the correlated errors. If polls are wrong in state X, they are more likely to be wrong in the same direction in nearby states. It's possible to model this by simulating a bunch of correlated normal distributions based on the assumed correlations of the errors (which you can obtain from the maps also), transforming the data into the vote % etc. I guess not worth the effort considering all the other people are already doing this kind of thing but better.
riggers
My guess is just an average of election betting odds 61.5% trump and nate Silver's model(old 538 one), which is 55%. Both of them have some bias towards dems but historically nate silver is somewhat biased more so trump has a ~60% chance of winning the election.
That IA’s do or don’t vote—or that if they do vote, it is not in appreciable numbers, is a rabbit hole we’d best avoid. To me this is a meaningless and obscuring argument. The primary issue is, “Why can IA’s vote at all!” In short, these are potential holes in the current voting system that must be corrected. The current system is broken and a rational voter can only assume such flaws as purposeful in nature. There lies the problem—if elections have such flaws they can never be considered determinative so that after an election, one side or the other will never accept such as the true will of the people. In that lies the crisis of democracy. There is no reason for the current rules for exercising the franchise—except to allow deceit and fraud. For example, few other Western democracies allow mail-in ballots. Why do we? How is it no proof of citizenship is required for voting for Federal office? I can go on, but to what effect. I myself, no longer vote due to the repeated election anomalies experienced, and as such do not consider myself bound to the so called “will of the people”.