November 18 news:

All of the methodology and technical notes from this report apply here. In particular, the information below is the sole opinion of America Counts. It does not necessarily reflect the opinion of any organization, including government agencies.

2:14pm: North Carolina currently has 3,546 comparisons available, from certain precincts from these 10 counties: Wake, Pitt, Lee, Iredell, Henderson, Buncombe, Durham, Granville, Guilford, and Haywood. Most of them match exactly, and the rest still need to undergo the basic checks for typical transcribing, aggregating, and matching errors. Based on what we’ve seen so far, our hypothesis is that once this check is done, it will result in all exact matches, but we won’t know for sure until it’s complete. So as of yet, the analysis has turned up no evidence of error in the North Carolina vote reporting that we have checked. More updates on North Carolina to follow soon.

1:46pm: Florida results milestone—after a delicate process of staging, transcribing, aggregating, matching, and discarding contest-choice-level official vote totals that were redacted for being too small, we ended up with 399 valid comparisons, all of which show no discrepancy. All of these comparisons are from Broward County, because this is where all of our transcribable Florida submissions are from. With 20 more hours of strenuous work, we may be able to gain 10-15 additional valid comparisons, but this is not worth it.

We have covered election day voting, for all the contests (with few exceptions, including that Broward County seems to have decided to not report Hollywood Question 6), for these precincts: D001, D003, R040, R044, T005, V011, V012, V024, V028, W006, X008. 

This analysis constitutes independent evidence that the vote reporting was done accurately in these areas.

There were several more precincts for which we received transcribable submissions, but it appears that one of the poll tapes was missing, since the aggregated poll tape totals were short of the official totals by about one poll tape’s worth of data. The missing poll tapes may have not been posted for some reason or blown away or something like that, meaning that the Actual Vote user wasn’t aware at the time that they were not capturing the full story. We are going to check with the Broward County Supervisor of Elections, but they may not respond in a timely fashion.

6:27am: Georgia results milestone—after transcribing many blurry recordings, strategically choosing to disregard values that couldn’t be read, and a complicated aggregation/matching/winnowing process, we have ended up with 326 comparisons to the official results, all of which show no discrepancy. The number of raw transcribed values that go into this aggregated 326 number is much higher. The total number of transcribable submissions we received is much higher than the number of transcribable submissions that make up the 326, since we had to discard so many due to the Georgia vote reporting process making them impossible to compare to the official results (see below for more on this topic). And there are many values that were very hard to transcribe because they were too blurry. With 20 more hours of strenuous work, we may be able to gain 10-15 additional valid comparisons, but this is not worth it. So we’re stopping the Georgia analysis here for now.

We have covered election day voting, for all races (with few exceptions) from these counties and precincts:

• DeKalb County: AA, AH, CE-CZ, DG-TG, GD, HA, HF, KG, MQ, MS, RD, SF
• Fulton County: 08D, SS19B-SS26

This analysis constitutes independent evidence that the vote reporting was done accurately in these areas.

For Gwinnett County election day voting, we received and transcribed submissions from these 10 precincts: 007, 019, 023, 046, 047, 053, 056, 081, 103, and 113. But as far as we can tell at this point in the reporting process, Georgia has decided to redact literally every single race from the official results in these counties, by whatever rule they’re using to preserve voter anonymity. After all the work in Gwinnett County up to this point, it’s disappointing that this is the thing that’s blocking us from gaining any election transparency benefit in this county.

A similar thing is true for Fulton County precinct groups 08F1-08N1-08N2, RW02-RW03-RW03B, SS03-SS03A-SS07A-SS07B-SS07C-SS07D, SS20-SS29B-SS29C, SS09A-SS16, SS09B-SS13A-SS13B, and SS04-SS081-SS29A. In particular, each of these groups had at least one precinct with completely redacted totals in the official results, meaning that we have no valid comparison available for any precinct in that group.

In a situation like this, the best we could ever do is if, for each precinct, we had poll tapes for early voting, election day voting, provisional voting, and absentee by mail voting that each had precinct-level totals. We did not receive anything but election day poll tapes from Gwinnett County, though, since to get poll tapes from the other voting methods, one would have to use a technique other than visiting polling places on election night.

We received many early voting poll tapes (called “advanced voting” in Georgia), but we weren’t able to do any meaningful comparison with them (see below for more on this topic).