After EOIR Fixes Most Egregious Data Errors, TRAC Releases New Asylum Data—But with a Warning
TRAC will begin once again releasing asylum data received from the Executive Office for Immigration Review (EOIR). After repeatedly uncovering issues in the quality of EOIR's data, the Transactional Records Access Clearinghouse (TRAC), a research data center at Syracuse University that studies the federal government, published its first report describing the problems it had uncovered. These included nearly a million filings by immigrants previously present in the court files TRAC received that had gone missing. The resulting public outcry caused the EOIR to restore most of these records but persistent problems remained: each month, new records continued to disappear. While fewer in number—several hundred to several thousand—the persistence of this problem was troubling as was EOIR's lack of apparent concern.
Then in June, there was a huge jump in records that suddenly disappeared from the case-by-case court records TRAC had received, this time involving asylum records. TRAC suspended updates to its "Immigration Court Asylum Decision" tool as the data was too flawed to use. After raising concerns to Director McHenry and the EOIR, the agency found and restored the most egregious bulk of the records, but the underlying problem of smaller numbers of records disappearing each month remained.
After monitoring EOIR's data for quality and completeness for several additional months, TRAC has determined that it will again begin releasing the data through TRAC's asylum tool later this week, but with a warning to researchers, practitioners, and the public about records that are still missing. We plan to post a running tally of asylum applications that have disappeared on the asylum tool itself to keep the public appraised of this problem. TRAC also remains troubled about the lack of apparent concern at EOIR headquarters over data management and the integrity of official court records.
TRAC's Asylum Data Tool Back Online Soon
For many months TRAC found growing mistakes with the EOIR's monthly data releases. TRAC reached out to the agency directly each month to resolve these issues, but became increasingly concerned that these repeated problems were a canary in the mine of larger data management issues. After the agency dug in its heels claiming that FOIA did not require it "to certify the accuracy" of the data it released, TRAC alerted the public to our concerns with a report on October 31, 2019 and wrote to EOIR Director McHenry. TRAC published two more reports in November and December documenting ongoing unresolved issues with disappearing records and again wrote Director McHenry. During 2020, TRAC continued to publish shorter updates in its subsequent reports. A timeline of reports related to the EOIR's data is provided below.
Although the missing of a million records was a serious concern, the nature of the records that had disappeared did not initially impact the data used in TRAC's asylum tool. This changed in May 2020. The data released by the EOIR through the end of April 2020 was missing so many records related to asylum relief that the total number of historical asylum cases appeared to actually decline. After careful consideration, TRAC decided to suspend updates to the online asylum decisions tool, effectively freezing the tool as of the end of March 2020. TRAC released a report on its decision on June 3 and sent yet another letter to EOIR Director McHenry.
Following TRAC's written complaints, EOIR restored most of the missing asylum records. Unfortunately, asylum records continue to disappear on a smaller scale each month. TRAC has continued to seek answers as well as to check the reliability of the EOIR's monthly releases of Immigration Court data while the asylum tool remained frozen.
Since TRAC found that the bulk of the missing records from April 2020 that caused us to suspend our asylum data tool have now been restored and have not disappeared in subsequent releases, the EOIR's asylum data once again meets the minimum threshold that allows us to release asylum data to the public. We are excited to announce that updated data will once again be shortly available through our Asylum Decisions tool.
However, we want to emphasize that this does not mean that the fundamental problems with the EOIR's asylum data or the agency's broader data management practices have all been resolved.
Thousands of EOIR records on applications for relief are still missing and the number grows each month. Although the number is not significant enough relative to the large number of overall asylum applications to seriously undermine the usefulness of TRAC's asylum tool, this does mean that enduring issues of data quality remain a real concern.
A timeline of TRAC's public reports and findings related to the EOIR's data management practices is available here:
TRAC Urges Caution with EOIR Data
In our first public report on the EOIR's missing records in October 2019, we opened with our central concern about the agency's data management practices quoted here in its entirety:
"Policymakers and the public routinely put their faith in federal agencies to provide complete and accurate information about their work. The value of government transparency is even higher in the area of immigration law and the Immigration Courts, which have become topics of considerable concern for Americans from all walks of life and for all three branches of government. In the present context, TRAC views concerns about EOIR's data inconsistencies—outlined below—as substantive, ongoing, and in need of prompt attention. Of greatest concern is the lack of commitment from EOIR to ensuring the public is provided with accurate and reliable data about the Court's operations."
This remains TRAC's central concern about the EOIR's data.
The EOIR's recent data issues may signal a culture shift at the agency, which, while not perfect, nonetheless had a reputation for transparency. In fact, in 2012, TRAC went on the record in a press release to defend the EOIR after an Office of Inspector General report issued during the Obama administration criticized the agency's handling of statistics. TRAC even went so far at the time as to urge other federal agencies to emulate the EOIR's commitment to public transparency. Today the EOIR's commitment to transparency is much less certain.
TRAC's concerns about the EOIR's data extends beyond asylum data. For instance, TRAC recently found that the EOIR misrepresented its data to support a controversial rule change that would eliminate administrative closure as a docket management tool for immigration judges. Specifically, the agency claims to show low numbers of case completions during the Obama Administration and high numbers of case completions during the Trump Administration. In reality, the data behind this argument artificially eliminates cases that were administratively closed. Its argument also fails to recognize that average annual case completions per Immigration Judge have actually declined rather than increased during the past four years.
TRAC takes no position on whether administrative closure should be eliminated or not. But the agency's misrepresentation of its own data to support a rule change is precisely the kind of concern we had in mind nearly a year ago when we expressed concerns about the trustworthiness of the EOIR's data.
The EOIR uses its own data to evaluate immigration judges' performance, support policy changes, and report on its work to Congress. The public should be concerned if the EOIR is mismanaging or misrepresenting its data. There is still evidence that it is doing both.