Based on a large data scrape from Edison research, a data analyst alleges that software glitches like the ones that occurred in Oakland county and Antrim county Michigan, could have happened all over the country. Other forensic data scientists analyzed the raw data from specific counties in battleground states, replicated the output, and corroborated that something ‘fishy’ was definitely happening with recorded vote tallies in many counties across the United States.
The analyst created a program to comb through the data to gather all instances where votes switched from Trump to Biden. in his breakdown, ‘lost votes’ means that the total amount of votes counted decreased by that amount throughout the counting. The analyst stated that he has performed this test in states that use Dominion Voting Systems so far.
The first image is the ratio change already detected for Antrim county.
He then analyzed the scrape from Antrim county and matched it to what he alleged was happening all over the country.
In the now confirmed software glitch that occurred in Antrim county, issues with the counting software caused results reported Wednesday morning to not match up with tabulators. Antrim County Clerk Sheryl Guy said results on electronic tapes and a computer were somehow scrambled after the cards were transported in sealed bags from township precincts to county offices and uploaded onto a computer.
A journalist for the New York Times appeared to have been the first to point out the nonsensical results in this specific county. Following the error being revealed, officials corrected the mistake and the county flipped from around 60% for Joe Biden, to 60% for President Trump.
Antrim County, Michigan makes no sense. pic.twitter.com/Gg4ktmLV0h
— Ben Rothenberg (@BenRothenberg) November 4, 2020
In Oakland County’s 15th county commission district, a fixed computer glitch turned a losing Republican into a winner. A computer error led election officials in Oakland County to hand an upset victory Wednesday to a Democrat, only to switch the win back to an incumbent Republican a day later. The incumbent, Adam Kochenderfer appeared to lose by a few hundred votes, an outcome that seemed odd to many in his campaign. After the apparent computer error was found and fixed, Kochenderfer ended up winning by over 1,000 votes.
Those darned algorithms again?
Is this still not news?
“A computer error led election officials in Oakland County to hand an upset victory Wednesday to a Democrat, only to switch the win back to an incumbent Republican a day later”https://t.co/wDrmy4PYxK
— Maajid أبو عمّار (@MaajidNawaz) November 10, 2020
Now this analyst argues that glitches like these may have been happening all over the country, not just Michigan. He added a list of all the states, and separated them by voting systems. No state uses exclusively one system, so there is some overlap in the breakdown. He then reordered the states, first by switched votes, then by lost votes.
Here is what he found.
The data he used is from Edison Research, it is used for election coverage by ABC News, CBS News, CNN, NBC, and many other outlets. It is also used for the website of the New York Times.
In another analysis done by data analyst Phil Evans and software engineer Bennie Smith, they noticed a possibly connected and somewhat disturbing trend in certain Michigan counties.
After scraping election data from several counties in Michigan, they noticed that the more the precinct is Republican, the higher the percentage of Trump votes were transferred to Joe Biden. In other words, in the most Republican precincts President Trump received a lower percentage of the Republican vote than in the least Republican precincts.
Other analysts downloaded the raw election data for Oakland County that the analysts used, and plotted the data for both candidates. There is in fact a clear and peculiar trend. As Republicanism increases, Biden’s percentage of individual ballots cast begins to outperform his percentage of straight-party ballots in a linear function. The inverse holds true for Trump’s votes.
This chart compares both candidates against republicanism which is the correct way to plot these two datasets. There is a clear reversal of the expected trend once Republicanism reaches around 25%.
Another analyst pointed out that the same phenomenon existed in the 2016 election in Oakland county in particular. In his opinion, this diminishes the likelihood of fraud. The analyst does acknowledge however that the phenomenon is strange to say the least.
Why would Republicans not vote for the Republican candidate in the most republican precincts?