Secret Data But No Verified Ballot - What Is Data Transparency's Blind Spot?

Data transparency's blind spot is the lack of raw, machine-readable source data that would let anyone verify election results, leaving only certified summaries. In Ghana, a country of over 35 million people, similar transparency gaps have sparked public outcry over mineral licences.

The Opaque Process Falsely Called 'What Is Data Transparency'

Last spring, I was sitting in a café in Leith watching a live-stream of a state election night. The presenter flashed a colourful PDF that listed the final tallies for each precinct, then moved on to the next slide. No one asked where the numbers came from; no one demanded the underlying files that the machines had produced. That moment reminded me of a pattern I have observed across many public sectors - the promise of "transparency" is often fulfilled by a single, polished document that cannot be audited.

When state officials release aggregated summaries instead of the granular, time-stamped data from voting machines, they are practising public disclosure theatre. The certification document lists the final numbers, but the chain of custody from ballot to count remains hidden. Researchers who wish to probe anomalies are forced to accept the state’s internal aggregation as gospel. In my experience, that practice erodes trust faster than any single counting error could.

One comes to realise that without a government data transparency standard for election data, the certified results become an article of faith rather than a demonstrable fact. The "certified" label suggests rigorous verification, yet the underlying data is never offered for independent scrutiny. The situation is comparable to the Extractive Industries Transparency Initiative, where the absence of raw licence data has long hampered civil-society monitoring. The parallel is stark: both arenas promise openness while delivering a black box.

During a conversation with a colleague who works for a local watchdog, she told me that the agency had spent months filing Freedom of Information requests simply to obtain the timestamps of when each precinct’s machines uploaded their results. The response? "Those files are proprietary and cannot be released for security reasons." That answer is the very definition of a transparency façade - it pretends to protect the public while keeping the data out of reach.

Key Takeaways

• Certification documents lack raw, machine-readable data.
• Aggregated PDFs hide the audit trail.
• Without standards, researchers cannot verify anomalies.
• Public disclosure theatre erodes electoral trust.
• Open data is needed for genuine government transparency.

How 'Certified' Results Violate Real Government Data Transparency

When I visited a county clerk’s office in the Highlands a few weeks ago, I asked to see the cast-vote records - the digital equivalents of each paper ballot. The clerk smiled politely and said, "Those are internal files, not for public release." The statement illustrated a fundamental breach of the data and transparency act principles that should govern public administration: the chain of custody from ballot to final report must be visible.

Election certification that lacks a transparent chain of custody creates an accountability gap where discrepancies can be explained away as "procedural" rather than investigated. Without mandated, structured open government data feeds from voting systems, third-party analysts face a "garbage in, gospel out" problem. They must trust the state’s internal aggregation, which defeats the purpose of oversight and mirrors the challenges highlighted in the Exploring California’s Generative AI Training Data Transparency Act, which stresses that access to underlying data is the only way to validate algorithmic outcomes.

In practice, the certification document becomes the final word. It sidesteps the requirements that a data and transparency act would impose - namely, that the underlying dataset be as accessible and trustworthy as the conclusion. I was reminded recently of a case in Canada where changes to citizenship law were published only as a summary, forcing lawyers to chase the original legislation to understand the impact. The parallel is clear: publishing only the headline numbers without the arithmetic invites suspicion.

One colleague once told me that in his audit of a mid-term election, the only data he could obtain were PDFs of totals. He spent weeks reconstructing the timeline of when each precinct uploaded its results, relying on press releases rather than system logs. The effort highlighted the inefficiency and opacity of current practices - a waste of public-service resources and a disservice to voters.

The Legal Fiction of Public Disclosure Without Source Data

Many state public disclosure laws for elections were drafted in the era of paper ballots. They require officials to "make available" the results, but they do not define what format or granularity that availability must take. This legal vacuum is now weaponised to withhold digital source data under the guise of security.

When a state releases a PDF of the final tallies, it can legally block access to the cast-vote records or system logs that would prove those totals are correct. The result is a system where the public sees the headline number but is barred from seeing the arithmetic. It is a modern twist on the old adage "the numbers are there, you just can't see them".

During my research, I interviewed a former election official who explained that the state's IT department classifies raw voting machine logs as "critical infrastructure" and therefore exempt from FOIA requests. That classification, while arguably protecting against cyber-threats, also creates a de-facto transparency ban. The official lamented that the law was never updated to reflect the digital nature of modern elections.

In the same vein, the Canadian citizenship law changes 2026 - what's new and who benefits illustrate a similar pattern: legislative updates are announced in glossy briefs, while the underlying policy documents remain sealed.

Because the law treats the certification document as "public disclosure", it satisfies the letter of the statute but not its spirit. The public is left with a mockery of transparency in state government for our most critical democratic function - the election.

A Blueprint for Real Open Government Data in Elections

True open government data for elections requires more than just publishing totals. It means releasing anonymous, time-stamped cast-vote records and corresponding ballot images in a non-proprietary format alongside the certification document. When I visited an open-data pilot in a small Midwestern county, the officials showed me a public repository where every vote was represented as a JSON line - each entry included a precinct identifier, a timestamp, and a hash of the ballot image. The dataset was downloadable, searchable, and could be fed into independent verification scripts.

Adopting a standard such as the Election Results Data Format (ERDF) would turn certification from a narrative into a verifiable dataset. The format specifies field names, data types, and encoding, allowing for automated consistency checks. Below is a simple comparison of the current "PDF-only" approach versus an ERDF-enabled open-data model.

This shift would move the burden of proof from sceptics to the certifying authority, requiring them to demonstrate - with data - that the announced results are an accurate tabulation of voter intent. As a journalist who has spent years chasing source documents, I can attest that the mere existence of a reliable dataset transforms the narrative from "we trust the officials" to "we can verify the officials".

Implementing such a system also addresses security concerns. By publishing anonymised, hashed ballot images, the data remains useful for verification while protecting voter privacy. The open-data approach does not weaken the system; it strengthens it by exposing any irregularities early, before they become points of contention.

What 'What Is Data Transparency' Should Never Be In Elections

Data transparency must never be reduced to a single PDF of final numbers. It is a continuous, data-rich process that proves integrity at every step, from ballot scanning to certification. When I was researching the recent state-wide audit, I found that the only publicly available artifact was a 12-page PDF. No raw logs, no machine output - just a summary.

Transparency should never be an excuse to withhold source data under claims of proprietary systems or security. Those walls are the very things that breed distrust and obscure errors. A colleague once told me that the vendor of the voting machines insisted that their software output be kept confidential, citing intellectual property. The vendor’s stance turned a public service into a private monopoly, a scenario that any open-government advocate would find unacceptable.

For election administrators, transparency should never end at certification. It must provide the forensic tools for anyone to replicate the count, transforming public disclosure from a performative act into a foundational proof. In practice that means publishing the full audit trail - timestamps, hash values, system logs - in an accessible format.

One comes to realise that without these tools, the public is left with a façade: a certified result that looks official but cannot be independently confirmed. The solution is not to tighten secrecy but to open the data, allowing the democratic process to be examined, challenged, and ultimately trusted.

Frequently Asked Questions

Q: Why is raw voting data essential for transparency?

A: Raw data lets anyone verify that the tally matches the votes cast. Without it, certification documents are untestable, leaving room for error or manipulation. Open data therefore underpins confidence in election outcomes.

Q: What legal obstacles prevent the release of source data?

A: Many state disclosure laws were written for paper ballots and do not define digital data formats. Officials often cite security or proprietary software concerns to withhold logs, creating a legal fiction of compliance.

Q: How does the Election Results Data Format improve auditability?

A: ERDF specifies a non-proprietary, machine-readable schema for each cast-vote record, including timestamps and hashes. This enables independent analysts to run consistency checks and reproduce the count without needing internal software.

Q: Can open election data coexist with voter privacy?

A: Yes. By anonymising identifiers and publishing only hashed ballot images, the data remains verifiable while protecting individual votes. The balance mirrors other open-data initiatives that safeguard personal information.

Q: What steps can states take today to enhance transparency?

A: States can start by publishing time-stamped cast-vote records in open formats, adopt a standard like ERDF, and amend FOIA statutes to include digital election logs as public records. These moves provide immediate auditability.