How AsylumTracker turns public records into research pages. Updated 2026-07-01.
Every data-led page identifies its source family and snapshot date. EOIR court outcomes, court-nationality subsets, ZIP assignments, and ICE detention records can have different coverage windows and publication rules; they are not merged into a single implied time period.
The asylum grant rate is granted / (granted + denied). Procedural outcomes such as termination or withdrawal are shown separately and do not enter that denominator. Nationality pages require at least 200 published decided outcomes. Nationality-and-court pages are stricter: the nationality must have at least 5,000 published decided outcomes, each pair at least 300, and only the three highest-volume qualifying courts are published.
Analysis must describe what a dataset contains without implying causation or an individual prediction. Comparisons state their denominator, threshold, and scope. Pages do not rank lawyers, recommend a legal strategy, or present historical statistics as legal advice.
The interactive tool can update independently from the static research pages. Static pages show their source snapshot date, and each monthly report preserves that date so later updates do not silently change a previously published report. The data manifest also records the current dataset version.
Send correction requests to [email protected] with the URL, disputed value, comparison source, and observation date. Reproducible data or calculation errors are corrected in the source pipeline and regenerated pages; interpretive clarifications are dated on the affected page or next snapshot report. Requests to alter accurate public statistics for reputation management are not treated as factual corrections.