Data Corrections Policy

Data errors are inevitable in any data operation. What matters is how they are discovered, disclosed, and corrected. EB5Status follows a transparent correction process modeled on academic data repositories and statistical agencies, where every correction is documented, publicly accessible, and permanently preserved.

We believe that trust is built not by claiming perfection, but by demonstrating accountability. When we make an error, we fix it, explain what happened, and take steps to prevent recurrence. This policy describes that process in full.

Our correction process applies equally to all data published on EB5Status, regardless of trust tier, topic, or page. No data point is exempt from scrutiny, and no correction is too small to document.

We classify every confirmed error along two dimensions: type and severity. This classification determines the correction timeline, the scope of review, and the level of notification required.

Error Types

1. 1.Typographical errors. The underlying data is correct, but the displayed value contains a formatting mistake, transcription error, or rendering issue. For example, a processing time displayed as “18.5 months” when the correct value is “15.8 months” due to a digit transposition.
2. 2.Data errors. The underlying value itself is incorrect. This may result from a source capture mistake, a data entry error, or a misread of the original publication. The source data says one thing; our database says another.
3. 3.Methodology errors. The data inputs are correct, but the calculation, interpretation, or aggregation logic produces an incorrect result. This category includes formula errors, incorrect date range boundaries, and misapplied statistical methods.

Severity Levels

1. •Minor. The error does not change any conclusions or materially affect the meaning of the data. Examples include rounding differences, minor label inconsistencies, or formatting issues that do not alter the value.
2. •Significant. The error changes a specific data point in a way that could affect a reader’s understanding. Examples include an incorrect processing time for a specific form type, or a wrong cutoff date in a visa bulletin summary.
3. •Critical. The error affects multiple data points, invalidates a conclusion, or could mislead users making important decisions. Examples include a methodology error that propagates across an entire dataset, or a systematic source capture failure.

Errors reach our attention through three primary channels:

1. 1.Internal quality checks. Our verification process includes automated consistency checks that compare new data against historical baselines and flag anomalies. We also perform manual spot checks during each data collection cycle.
2. 2.User reports. Readers, attorneys, researchers, and other users who identify a potential error can report it directly. We treat every user report as a credible lead and investigate accordingly.
3. 3.Source data revisions. Government agencies occasionally revise previously published data. When USCIS, the Department of State, or another source agency updates a figure we have already captured, we treat the revision as a correction event and update our records to match the revised source.

Once an error is confirmed, we follow a structured correction workflow:

1. 1.Investigation. We verify the reported error against the original source document. If the error involves a derived metric, we re-run the calculation and compare the result against the published value.
2. 2.Source verification. We retrieve and re-examine the original government publication, FOIA response, or other source material. We do not rely on secondary sources or memory to confirm a correction.
3. 3.Correction implementation. The corrected value replaces the erroneous value on all affected pages. For methodology errors, we update the calculation logic and re-process all affected outputs.
4. 4.Correction notice publication. A formal correction notice is published on the Corrections Log with full details of the error, the fix, and the prevention measures taken.
5. 5.Changelog entry. The correction is recorded in the site’s data changelog with a timestamped entry linking to the correction notice.
6. 6.Subscriber notification. Users who downloaded or interacted with the affected data receive an email notification describing the correction and providing the updated values.

For critical errors, steps 3 and 4 are executed within 24 hours of confirmation. For significant errors, the target is 5 business days. Minor errors are batched and corrected on a weekly cycle.

Every correction published by EB5Status includes a structured correction notice. These notices serve as the permanent record of what changed and why. Each notice contains:

• •Date discovered. When the error was first identified or reported.
• •Date corrected. When the fix was deployed to the live site.
• •Description of error. A plain language explanation of what was wrong, including the original (incorrect) value and the corrected value.
• •Affected data points. A list of every page, chart, or export that contained the erroneous value.
• •Cause of error. The root cause, whether a data capture mistake, a calculation bug, a source misread, or a source revision by the originating agency.
• •Steps taken to prevent recurrence. Specific process or system changes implemented to reduce the likelihood of the same type of error occurring again.

Correction notices are never edited or deleted after publication. If a correction notice itself contains an error, a new correction notice is published referencing the original.

All corrections are recorded in our public Corrections Log, which serves as the authoritative record of every data change made after initial publication. The log is sorted in reverse chronological order, with the most recent corrections appearing first.

Corrections are permanent, publicly accessible, and linked from the affected data pages. When a data page has been corrected, the page displays a notice linking to the relevant correction entry. This ensures that anyone viewing the data can see not only the current (corrected) value, but also the history of changes.

We also maintain a Data Changelog that records all data updates, including both corrections and routine data refreshes from source agencies. The changelog provides a complete audit trail of every change to our published dataset.

Corrections Log | The complete, chronological record of all data corrections.

Methodology Overview | Our complete data methodology, including confidence labels, calculation descriptions, and non-negotiable rules.

Data Collection Process | How we gather and verify data from each primary source.

Data Changelog | A timestamped record of all data updates, corrections, and source refreshes.