2020 Presidential Primary and General Election: Post-Election Automated Ballot Tabulation Audit
An independent, automated audit relies solely on the use of independent software to tabulate ballot images. The results from the independent tabulation are then compared to the tabulation results from the voting system. Any variances between the two tabulations are easily identified and resolved. For the 2020 Presidential Primary and General Elections, this office contracted with Clear Ballot, a Boston-based elections technology company which offers the ClearAudit software product. ClearAudit is the only currently available, market-ready software product that can perform an independent automated post-election tabulation audit using ballot images from another voting system.
The post-election tabulation audit is conducted using ballot images. The use of ballot images allows elections officials to maximize the technological functions of the new voting system while minimizing human error and eliminating chain of custody issues by using securely stored ballot images, rather than voted paper ballots. The use of ballot images removes the need for election officials to physically handle or count voted ballots unless a petition for recount or other judicial challenge is granted.
A discrepancy variance level of 0.5% was established for the post-election tabulation audit. If there is a variance between the voting system results and the audit results greater than 0.5% in any given contest and the variance cannot be explained, additional auditing will be performed before the election results are
To prepare for the post-election tabulation audit, SBE provided Clear Ballot with:
- PDF files of all ballot styles for all 24 counties;
- Pre-election reports for all precincts1 showing no votes cast;
- Information to assign ballot styles to precinct; and
- Files needed to process ballots used with the ExpressVote ballot marking device.
1Because the primary election was primarily a vote by mail election without precinct-based polling places, ballots were sorted into three precincts. All vote by mail and absentee ballots were placed into the ABS1 precinct; all in-person election day ballots (except for provisional ballots), were placed into the EVC-1 precinct; and all provisional ballots were placed in the PROV precinct. At the current time, we do not have the ability to place absentee, vote by mail, early voting centers, or election day voting center ballots into princints based on where the voter lives, but we are investigating that option for future elections.
Upon receipt of the data, Clear Ballot:
- Created a Ballot Definition Files for each county;
- Validated each Ballot Definition File; and
- Created a ClearAudit database for each county.
The local boards of elections send to Clear Ballot the images of ballots cast during early voting and on election day. Once Clear Ballot receives the ballot images, Clear Ballot:
- Transfers the ballot images from each county into its corresponding audit database;
- Tabulates the ballot images from election day and early voting;
- Resolves unreadable ballots;
- Performs an audit database review; and
- Sends a Preliminary Statement of Votes Cast to SBE.
For the General Election, once SBE receives the Preliminary Statement of Votes Cast, SBE provides Clear Ballot with election results from early voting, election day and all mail-in ballots counted prior to election day. This is Phase 1 of the audit. The delay in sending the Phase 1 results to Clear Ballot is intentional. It creates a “blind” audit, meaning that Clear Ballot provides its results without knowing the results from the voting system. Clear Ballot uses the results from Phase 1 to create various reports comparing the sets of results.
After the remaining mail-in ballots and provisional ballots have been counted, the local boards of elections send to Clear Ballot images of all ballots cast in the election. This is Phase 2. Clear Ballot adds the remaining mail-in ballots and all provisional ballots into the appropriate database, tabulates the ballot images, resolves unreadable ballots, and generates a new Comparison of Votes Cast that includes all ballots cast in the election.
For the Primary, once SBE received the Preliminary Statement of Votes Cast, SBE provided Clear Ballot with election results from all ballots counted at the vote by mail canvass through and including election day, as well as all ballots cast at vote centers on election day. This was Phase 1 of the audit. Clear Ballot used the results from Phase 1 to create various reports comparing the sets of results.
After all vote by mail and provisional ballots were been counted, the local boards of elections sent to Clear Ballot images of all counted vote by mail and provisional ballots. This was Phase 2. Clear Ballot added these ballot images into the appropriate database, tabulated these ballot images, resolved unreadable ballots, and generated a Comparison of Votes Cast for the vote by mail and provisional ballot canvasses.
Reports Produced by Clear Ballot
With this audit solution, election officials and other interested individuals can sort contest, ballot and precinct reports, see images of contests and ballots, and provide detailed information about how each ballot image was adjudicated. Clear Ballot produces for each county four audit reports.
- Comparison of Cards Cast for each canvass: This report compares the number of ballots counted on election day, during both mail-in canvasses, and during the provisional canvass against the number of ballots tabulated by Clear Ballot. This ensures that the same number of ballots were tabulated by both systems.
- Comparison of Ballots Cast by Precinct: This report compares the number of ballots cast in each precinct against the number of ballots tabulated during the audit. This is another way to ensure that the same number of ballots are tabulated by both systems.
- Comparison of Votes Cast: This report compares the results from the voting system against the audit results and identifies possible discrepancies by candidate or choice.
- Contest Vote Discrepancy Threshold Report: This report shows - by contest - the number of vote differences between the two systems and the vote difference as a percentage. Before the audit performed, SBE determined that a percentage of 0.5% or higher would trigger an additional review, which could include a manual review of voted paper ballots.1
2020 Presidential General Election Audit Results
2020 Presidential General Election Phase 1 reports for each county are available below. Phase 2 reports will be uploaded when they are available. The State Board of Elections also provides a web portal where the ballots and the audit results can be reviewed.2
2020 General Election: Post-Election Tabulation Audit Reports - Phase 2 - All Ballots
2020 General Election: Post-Election Tabulation Audit Reports - Phase 1 - Early Voting, Election Day and partial Mail-In Ballot Results
2020 General Election: Post-Election Tabulation Audit Reports - Preliminary Statement of Votes Cast
1There is a known difference in how the voting system and Clear Ballot tabulate write in votes in a contest where the voter votes for more than one candidate. Due to this difference, the vote discrepancy in these contests is sometimes higher than the threshold. In order to combat false discrepancies as a result of this difference in tabulation, there are two versions of the Contest Vote Discrepancy Threshold Report- one with write-ins, and one without. For the purposes of determining if a vote discrepancy is over the threshold, the version without write-ins is used.
2The web portal for the 2020 General Election will be available sometime in December 2020. The portal currently shows ballots and audit reports from the 2020 Special Primary and Special General Election in Congressional District 7, and from the 2020 Presidential Primary Election.
2020 Presidential Primary Election Audit Results
2020 Presidential Primary reports for each county are available below. The State Board of Elections also provides a web portal where the ballots and the audit results can be reviewed. To access the portal, click here.
Important Note: Due to a printing error, some voters in Baltimore City's First City Council District received ballots that contained the correct contests but did not align with the ballot scanners. The affected contests (City Council District 1 and Judge Of The Circuit Court Judicial Circuit 8- Both on the Democratic ballot only) on these ballots had to be manually duplicated so that they could be read by the voting system. In order to be able to conduct the automated tabulation audit, Phase 1 was broken up into Phase 1.0 and Phase 1.5.
Phase 1.0 contains the original ballots as scanned as they were. The voting system was unable to read the affected contests. The audit tabulation software was also unable to read some of those ballots. Normally when the audit system can not automatically adjudicate a ballot, it is manually adjudicated, by a person. For the affected ballots in Phase 1.0, the affected contests on these ballots were not manually adjudicated. This will show that the voting system and the audit system both were unable to read the incorrect ballots.
In Phase 1.5, the affected contests on ALL ballots in the City Council District 1- Democratic contest were manually adjudicated. The Phase 1.5 tabulation was compared to the voting system results that included the duplicated ballots. This will show that 1) the duplicated ballots were duplicated correctly, and 2) that the voting system correctly tabulated the results.