FIVE STATES. FIVE ELECTION INVESTIGATIONS. THE DOCUMENT TRAIL HAS BEGUN.
Mind Bend Theory has contacted election authorities across Texas, Georgia, Indiana, Iowa, and Idaho. We are requesting the original records behind reported 2026 election discrepancies—and asking officials to show exactly how every ballot was accounted for.
Mind Bend Theory | Independent Election Forensics | November 12, 2026
Why We Launched This Investigation
An election result is more than a number displayed on a website. Behind that number should be a verifiable chain of records: voter check-ins, ballots received, ballots counted, machine reports, reconciliation worksheets, corrections, and final certification.
When those records do not initially agree, the public deserves an explanation supported by evidence.
Following the November 3, 2026 midterm elections, Mind Bend Theory (MBT) identified multiple publicly reported election-administration discrepancies across five states.
These incidents involve different systems and different circumstances. Some concern ballots initially left uncounted. Others involve votes counted by machines but omitted from public reporting. Another involves a scanner malfunction and differences between voter check-ins and ballots recorded.
We are not claiming these incidents prove election fraud, coordinated manipulation, or an incorrect final outcome. We are investigating whether the documentary record fully explains what happened.
That distinction is central to our work.
1. TEXAS — Statewide and County-Level Reconciliation
Who we contacted: Texas Secretary of State, Texas Department of Public Safety, Tarrant County Elections, and Bee County Elections.
Where: Texas, with particular attention to Tarrant and Bee counties.
What we requested: Original and corrected election reports, county canvass records, voter-registration processing information, reconciliation documentation, and publicly releasable audit trails relevant to our review.
Why we contacted them: MBT is examining whether differences between published election reports, registration statistics, voter participation records, and county reconciliation totals can be explained through official documentation.
Our investigation also includes questions raised in a whistleblower-related submission. We are treating those claims as allegations requiring independent verification, not established findings.
Tarrant County publishes time-stamped election-night results and identifies additional audit and reconciliation reports, providing a basis for comparing the original reporting sequence against later records.
Current status: Requests submitted. Bee County has acknowledged our request and advised that final election processing and canvassing must be completed before it can prepare a cost estimate. The county also identified risk-limiting audit activity scheduled for November 19–20, with its hand-count audit to follow.
What we still need to establish: Whether each apparent discrepancy reflects a reporting definition, timing difference, correction, administrative error, or a genuinely unresolved reconciliation exception.
2. GEORGIA — SEMINOLE COUNTY
Who we contacted: Seminole County Elections Clerk.
Where: Seminole County, Georgia — State House District 154.
What happened: Georgia's Secretary of State reported approximately 388 more Election Day voter check-ins than ballots initially reported. State officials said an election-results memory card had not been uploaded into the county's election-management system.
When the additional results were incorporated, the apparent lead changed from Democrat Harry James to Republican Gerald Greene. The corrected county results have since been certified, while the state's investigation into the reporting failure remains open.
The county's elections supervisor has publicly disputed aspects of the state's characterization of the memory-card issue.
What we requested: Original and corrected precinct results, memory-card inventory and custody documentation, election-management-system import records, tabulator reports, validation logs, and incident documentation.
Why this matters: This case demonstrates that a reporting omission can change which candidate appears to be leading an election.
We want to determine whether the failure was an incomplete upload, an import-validation problem, or another procedural issue—and what independent records confirm the corrected totals.
Current status: Formal request sent. Awaiting records.
3. INDIANA — CLARK COUNTY
Who we contacted: Clark County's election administration through its official county email channel.
Where: Jeffersonville High School voting location, Clark County, Indiana — State Senate District 45.
What happened: Election Day problems included a ballot scanner malfunction and electronic pollbook concerns.
Published preliminary results showed Republican Chris Garten leading Democrat Nick Marshall by 297 votes.
A reported comparison identified 1,640 ballots cast against 1,612 electronic pollbook check-ins, a difference of 28. That discrepancy alone is smaller than the candidate margin, but it does not independently resolve questions about the scanner malfunction.
On November 10, the Clark County Election Board voted against a proposed physical-ballot count at four affected precincts. Democratic officials have discussed pursuing a formal recount or election contest, while Garten has expressed confidence in the reported result.
What we requested: Scanner incident records, machine and memory-device handling documentation, electronic pollbook reconciliation, precinct ballot totals, chain-of-custody records, and election-board review materials.
Why this matters: A voter check-in discrepancy and a scanner malfunction are not necessarily the same problem. They must be evaluated separately.
We want to establish whether the number of physical ballots, recorded voters, and tabulated votes can all be independently reconciled.
Current status: Formal request sent. Awaiting records.
4. IOWA — JOHNSON COUNTY
Who we contacted: Johnson County government, requesting routing to the Auditor's Office and Elections Division.
Where: Johnson County, Iowa.
What happened: Approximately 2,017 absentee ballots were initially overlooked during election-night counting.
According to the county, unopened ballot boxes remained in secure storage and were discovered during post-election work on November 5.
The ballots were subsequently counted, with the county reporting 36,138 absentee ballots in total. Officials said the additional ballots did not change any election outcomes and that ballot custody remained intact.
What we requested: Original and corrected absentee results, ballot inventory and storage records, discovery reports, chain-of-custody documentation, audit instructions, reconciliation worksheets, and corrective-action records.
Why this matters: Unlike Georgia and Idaho, this incident involved ballots that had not initially been tabulated.
The central question is how a ballot inventory process could overlook approximately 2,017 ballots and which reconciliation procedures should have identified the omission before election-night reporting.
Current status: Request submitted through Johnson County's official contact form. Awaiting acknowledgment from the appropriate records custodian.
5. IDAHO — LATAH COUNTY
Who we contacted: Latah County Elections Department through the county's official departmental contact form.
Where: Latah County, Idaho.
What happened: Results from 2,343 early-voting ballots were omitted from the initial public election-night totals, even though county tabulators had already counted the ballots.
The omitted ballots represented approximately 12.9% of the county's final reported 18,104 ballots.
A full paper-ballot recount on November 10, overseen by the Idaho Secretary of State's Office, confirmed that the corrected results did not change the outcomes of any races or ballot measures.
What we requested: Original and corrected results, tabulator reports, election-reporting transfer and import logs, ballot reconciliation records, recount worksheets, state correspondence, and final canvass documentation.
Why this matters: Counting votes correctly and publishing those votes correctly are two separate responsibilities.
This incident provides an opportunity to examine how results can exist in a tabulation system without appearing in the public reporting system—and how post-election controls detect that failure.
Current status: Request submitted through the official Elections Department contact form. Awaiting acknowledgment and records.
Why We Are Comparing These Five States
These are not five identical incidents.
They represent distinct points where an election process can encounter problems:
StatePrimary issue under reviewTexasElection-reporting, registration, and reconciliation questionsGeorgiaElection-results upload and reporting discrepancyIndianaScanner malfunction and electronic pollbook reconciliationIowaPhysical ballot inventory and initial counting omissionIdahoTransfer of tabulated results into public election reporting
The comparison allows MBT to ask a broader question:
Are election safeguards consistently capable of detecting, explaining, and correcting discrepancies before final certification?
In Iowa and Idaho, officials reported that corrective procedures preserved the original outcomes. In Georgia, correcting the reporting omission changed the apparent leader. In Indiana, the close race and disputed procedures remain subjects of scrutiny.
These distinctions matter. A reporting correction is not automatically evidence of an improperly counted election. Equally, an unchanged outcome does not eliminate the need to understand a documented failure.
What We Are Doing With the Records
MBT will examine the available evidence using a structured forensic process.
First — Preserve the original reporting record. We will retain original publications, corrected figures, official statements, dates, and source references.
Second — Reconstruct the numerical differences. We will compare ballots received, ballots accepted, voter check-ins, ballots tabulated, published totals, and final canvass results wherever comparable records exist.
Third — Identify the documented failure point. We will determine whether the available evidence supports a ballot-handling, equipment, software-transfer, reporting, or procedural explanation.
Fourth — Test the official explanation. We will compare incident reports and public statements against source documents, identifying what is independently verified and what remains unresolved.
Fifth — Publish findings with clear evidence classifications. Confirmed facts, official explanations, conflicting accounts, and unanswered questions will be distinguished from one another.
Where original electronic records are available, MBT intends to preserve their integrity using file hashes and an evidence ledger.
What Happens Next?
As of November 12, MBT has submitted requests across all five states.
We are now awaiting acknowledgments, fee estimates, records releases, and any additional audit or recount findings.
We will not present pending requests as completed investigations. Nor will we declare misconduct without supporting evidence.
If the documents confirm an administrative mistake that was properly corrected, we will report that.
If they reveal unresolved discrepancies, inconsistent explanations, or weaknesses in reconciliation controls, we will document those findings as well.
Our Position
Election integrity is not a partisan standard.
A discrepancy involving a Republican candidate deserves the same scrutiny as one involving a Democrat. A corrected result deserves verification regardless of which party benefits.
The objective is not to produce a predetermined political conclusion.
It is to establish a record that voters, journalists, election administrators, and independent researchers can examine.
Five states. Multiple documented discrepancies. Formal records requests submitted. Independent verification underway.
MIND BEND THEORY
Follow the evidence. Verify the record. Let the findings speak.