What if the poles are right and the votes are wrong.
I will write one diary about the possibility of election fraud. Hopefully, this will help us all learn how to write about conspiracy theories using facts and data. I realize that this is not a pleasant topic and leads to despair and hopelessness, but why we cannot ever question our voting systems is beyond me.
To prove something is real first we must prove that it exists. A rhetorical question; Does anybody think that no ballot box has ever been stuffed enough to change an election?
So now we are just trying to determine how much it happens.
2004 Ohio was the watershed for me. A lot of Kossaks are still around from then so I will just cover a highlight. In 11 very conservative counties, more people voted for gay marriage and an African American judge than John Kerry. This was a strange outlier, as in all other Ohio counties Kerry received more votes than gay marriage or the judge. The reason that this was interesting to the investigators, was that a way of cheating in Ohio at the time was to misidentify the precinct when counting ballots at the county level. Since the presidential candidates rotated their positions on the punch card ballots, but the ballot propositions didn’t, a ballot counted in the wrong precinct would switch presidential votes but not votes for gay marriage or judges. These eleven counties had some of the largest increases in Bush Percentages in Ohio, and this evidence was good enough for a judge to hold the ballots in Ohio for years. Richard Hayes Phillips examined ballots in these 11 counties through the legal discovery process In “King Lincoln Bronzeville Neighborhood Assn vs. Blackwell,” and found physical evidence of fraud in all of them. From Seneca county, where not a single precinct’s vote count matched their pole book’s to 212 ballots in a row for Bush were discovered in Butler county. ( What Happened in Ohio Fitrakis, Rosenfeld,and Wasserman 2006)
https://neocon-panic-attacks.blogspot.com/2008/05/witness-to-crime-richard-hayes-phillips.html
My point is that what happened in Ohio in 2004 is always said to be impossible. Many people must have been involved. It was hardly a secret, but nobody was caught or punished (Some people were punished in Cayahuga county for a different fraud, Ohio, was full of them). Also the safeguards which are claimed to be in place were ignored, as they probably routinely are. Usually 3% of the ballots are recounted, but that 3% needs to be random, typically the precinct chosen was predetermined. When the counts didn’t match were found they were ignored or recounted. In Monroe County after the 3% hand counts did not match machines after 2 tries, a repairman from TRIAD was summoned to fix the voting machine. And finally, It’s hard to believe all this came out of nowhere then disappeared into nowhere.
Polls
Pollsters are smart, they do math. (I realize there are many people on Koss who are very good at manipulating data, and we are constantly looking at poll numbers, and having it explained to us whether they are accurate or not. ) For decades pollsters have chased a “Red Shift.” In other words polls always seem to be wrong in the same direction, which is they underestimate the final Republican Vote.This is evidence for election fraud, but a lot of energy is expended explaining how these very smart people can never get it right. The purpose of polls is to predict the final vote after all, not catch cheating, but they don’t seem to be able to do that.
This “Red Shift” became news in 2004, as its ubiquity of it became clear.
“In the five presidential elections from 1988-2004, there were 238state presidential exit polls,194 of which red-shifted from the Democrat to the Republican. Of the108 exit polls that exceeded the 2% margin of error, 99 red-shifted to the Republicans. Of the 65 that exceeded a very conservative 3% Margin of Error (MoE),all but one red-shifted to the GOP.
At the 95% confidence level, one would expect 12 of the 238 polls to exceed the MoE and be evenly split between the Democrat and the Republican.”
Edison,which did the exit polls for the TV networks studied the results, and they tried to determine the cause of this discrepancy.They settled on the “reluctant Bush voter hypothesis,” which was that Bush voters were reluctant to say that they voted for Bush in the anonymous exit polls.
The reluctant Bush voter hypothesis was a hypothesis by elimination. Edison/Mitofsky does not have any studies which show that Bush voters were refusing to take the poll in larger numbers than Kerry voters. What they did was eliminate the other hypothesis of sampling and populations errors.The shy voter explanation has been updated to Trump voters who are definitely not shy.
Pollster attempts to fix
There are basically two ways to alter poll results to match “recorded votes.” Since polls are stratified type statistics a pollster could simply weight Republicans higher or assume there are more Republicans. Also for opinion polls you could simply add another fudge factor called likely voters. For exit polls one could adjust the weights of the type of voters or the way they voted.
The“likely voter” fudge factors seem to be the choice of pollsters.
https://fivethirtyeight.com/features/registered-voter-polls-will-usually-overrate-democrats/
https://fivethirtyeight.com/features/what-pollsters-have-changed-since-2016-and-what-still-worries-them-about-2020/
It is a truism that Republican voters are more likely to vote, so traditional polls, which weight registered voters not “likely voters” give the democrats an edge of about 2.6 % according to Nate Silver. Also, pollsters change from registered voters to “likely voter” polls near the election so it looks like some movement has occurred, when it may have not. For the 2020 election even more fudges were added for education and geography. Still the polls were “wrong.” Pollsters can’t catch up with the change in the votes. I would point out two things. One, on the question of whether the polls are wrong or the votes, matching the polls to the votes after the election to answer that question is a tautology. There is no independent evidence offered to pick vote counts over polling (there’s plenty the other way, if only that there are so many independent polls). The fact that the pollsters don’t seem to be able to keep up is a problem especially in this election, where everybody voted, and the argument that democrats didn’t come out is specious.
By 2004 the weighted pole sample by race used in the exit poll was (Codered p 176) already more republican than the electorate compared to weights measured by ANES, an independent measure of the electorate. The table below shows the same comparison for racial break down of the electorate.
category
|
Exit Poll wt.
|
Independent ANES wt
|
White
|
80
|
70
|
African American
|
10
|
16
|
Latino
|
8
|
8
|
Asian
|
1
|
3
|
Other
|
2
|
0
|
Native American
|
0
|
4
|
Exit polls come in two flavors. The raw exit polls which are supposed to be secret, but always seemed to leak, and the final exit polls, which are adjusted to match the final vote count. These adjusted exit polls are used by pundits to explain things like “56%” of white women voted for Trump.
Richard Charnin in his book “Proving Election Fraud” (Authorhouse 2010) on page 52 has the figure above. This figure points out a physical impossibility in the adjusted exit poll data. The calculation leading to this conclusion is simple. The exit poll asked the question “who did you vote for in the last election?” Total votes for the Presidential candidate can be separated into those who did not vote in the last election, those who voted for the presidents opponent, and those who voted for the president. He estimates a 1.2 percent per year death rate, then adds up all the sources of votes from the adjusted exit polls. The results in the figure are the number of votes needed from returning Party voters divided by the number of Party voters still alive 4 years later. Any number above 100% is not possible, in five of ten elections the number of returning Republicans is impossible, in the adjusted exit poll survey. This should not be a problem, if the underlying vote counts make sense. It appears that the exit polls stopped asking who the voter voted for in the last election after 2008. In the pollsters defense, it would have been very difficult to make the Bush election plausible, Bush lost a majority of new voters and people who said they didn’t vote in 2000 by large margins according to the same polls. So those numbers could not be increased, without making Kerry a clear winner. Somehow Bush got 12 million new recorded votes in 2004 over 2000, which was basically a tie, to win the popular vote count.
Jonathan Simon and Code Red
After 2004 there was a lot of interest in electronic voting machines. Jonathan Simon, B. Odell, Dale Tavris and Josh Mitteldorf, did a study of the 2006 election (p203 Code Red by Jonathan Simon www.codered2016.com. They purchased exit poll data from Survey USA, for congressional, governor and senate offices. They tested the theory that if there was election fraud, it would concentrate on close elections. So elections were chosen at random where a single ballot would contain both competitive and noncompetitive races. In this way response rates would be removed from the data because the same voter response would cover both races. The dependent variable was the difference between the exit poll estimate and the recorded vote. The average vote shift in the competitive races was 3.6% to the Republican side and in the noncompetitive races the average difference between the recorded official vote count and the recorded vote count was 1.7% to the Democrats. The average of the paired differences competitive vs non competitive for the same ballots was slightly larger at 5.47 percent. These differences were statistically significant. There were some races that did not red shift at all. The worst was a shift of 16% in Henry County Missouri. The 2006 election was considered a big democratic win, yet it was easy to detect fraud.
Wisconsin 2016
In the 2016 election counties in Wisconsin partially counted by touchscreen voting machines showed a statistically significant unexpected increase in votes for Donald Trump over Hillary Clinton. This was noted in many places at the time, and Hillary did join the law suit to recount Wisconsin. Since it is impossible to recount touch screen voting machines, these were obviously not recounted. (A problem for Trump this year in Georgia.)
These machines were suspicious for another reason, in that they were programmed by a company called Command Central. The programming provided by that company was contained on separate flash drives one to test the equipment and one to count the votes, and their office was in the same building as Michele Bachman’s (ooh spooky).
https://wcmcoop.com/2012/05/22/meet-command-central-the-people-in-charge-of-wisconsin-voting-machines/
The effect was discounted by many as probably due to hidden demographic variables, which were not named. My first KOS diary was an examination that showed that the effect was robust against income, population density, and race. If your are interested, Hillary Clinton appeared to win in Wisconsin in the raw exit poll by 5%.
https://www.dailykos.com/stories/2018/7/15/1780105/-Another-conspiracy-theory
Things that make you go hmm
For the Virginia 2017 election the method of counting was changed from touch screen to paper ballots. That November democrats flipped 15 seats in the house in a surprise upset. The winning streak continued this year and Democrats now have a trifecta in that they control all of the state government. This election was considered a bellwether for the presidential election, but it didn’t work out that way in 2020. Maybe the big effect was the switch to paper ballots.
https://www.washingtonpost.com/local/virginia-politics/paper-ballots-make-a-comeback-in-virginia-this-fall/2017/10/07/724106ec-a92b-11e7-850e-2bdd1236be5d_story.html
https://www.liberty.edu/champion/2017/10/opinion-virginias-switch-to-paper-ballots-during-the-election-can-help-curve-hacking/
Recount rules that seem to be more about hiding fraud than anything else are abundant
Florida’s recount law show below only allows for recounts of ballots the machine rejected, you’re not allowed to check the machine counts.
The close vote margin for Florida elections is less than or equal to 0.5% of the total votes cast for a given office or measure. These recounts are initiated automatically by election officials. The laws specifies that candidates, including candidates for retention to a judicial office, and ballot measures are all subject to recounts.Both primary and general elections are eligible for recounts. The candidate or candidates who are defeated by the close-vote margin may waive the recount if a request is made in writing to the canvassing board. Section102.141(7).
As noted in the counting methods above, the close-vote margin first initiates a retabulation of the paper ballots. If there tabulation results in a vote margin that is less than or equal to 0.25%, a hand recount is conducted. However, the hand recount is only for those ballots containing under votes or over votes, and is required only if the candidate or candidates decline to waive the manual count and the collective total of under votes and over votes is enough to alter the result of the election. Section102.166(1).
Republican legislatures made it harder to recount votes in Wisconsin and Michigan. Of the 5 states that appeared to be flipped on Hillary Clinton in the exit polls Pennsylvania, North Carolina and Florida were basically unrecountable due to restrictive laws. Now recounting is harder in the other 2. It is easy to attribute bad motives to these laws.
One other thing needs to be noted. Gerrymanders lead to close elections in some districts. This actually requires more stealing, because those are subject to wave type elections.
Plebeian Summary
Avoiding the most obvious cause of the polls being wrong is not a good idea, because sometimes if it walks like a rotten fish it and smells like a rotten fish; it is a rotten fish. Sometimes it isn’t, but asking keeps people honest. It is possible that Biden really did win big enough that it couldn’t be stolen.
I realize there are a lot of problems with the above analysis, the main one being that it is hard to believe that over 30 years there has been massive election theft, with no conspirators. Although interestingly enough Blackwell, the guy in Ohio, showed up in the Trump campaign. For a long time I thought this was like a poll of how many people cheat on exams. It really wasn’t a big conspiracy more like a statistical sampling of little ones. With the advent of electronic voting machines, maybe there is one big one.