Watch list hobbled by data errors

Four years after the federal government launched the interagency Terrorist Screening Center (TSC) and assigned it the daunting task of harmonizing more than a dozen separate watch lists, balky technology and quirky business practices still combine to introduce gaps and errors in the critical database.

For example, several known or suspected terrorists were not properly identified in the Terrorist Screening Database (TSDB), including 20 watch list records that were not made available to frontline officials, according to a report from the Justice Department's inspector general. 'We also found that the number of duplicate records in the database has significantly increased since our last review,' the auditors said.

The screening center has not completed the process of merging its source material, and the participating agencies have retained business practices that apparently prevent a standard approach, officials familiar with the systems said.
TSC Director Leonard Boyle said in an interview that each of the databases that originally contributed information to the TSDB operated according to distinct business rules, which has hampered consolidation efforts.

The report highlighted other TSDB technical failings. It cited watch list errors introduced by the TSDB's division into two components, one of which is used to accept information from outside sources and the other to relay data to external users.

In addition, terrorist watch lists still are influenced by the variations between agencies' definitions of what constitutes a terrorist case.

'Each organization has its own rules as to who can be considered a subject of a terrorist case,' Boyle said. 'The agencies are controlled by different definitions of what it means to be a terrorist.'

As for the technical problems, Boyle said, 'we are now at a transitional stage where we cannot export data from the component that does the import function, with the exception of data sent to [the FBI's National Criminal Information Center]. We expect to reach that stage [of consolidating the two components] in approximately six months.'

He noted that the system's two divisions could have been characterized as separate databases a year ago, but now they are more closely related components.

Unresolved discrepancies

The two components are supposed to contain identical information, but until recently, discrepancies between them went unresolved.

As of April, the TSDB included more than 700,000 records and was growing at a rate of 20,000 records each month, according to the Justice report.

The divisions in the critical terrorist watch list contribute to inaccuracies in the database, which frontline law enforcement officers and border officials use to screen people.

The IG found discrepancies in the number of files in each of the components. In their initial checks, the auditors found discrepancies ranging from 18 records to 38 records between the two databases that together form the TSDB watch list.

Later in the audit, center staff members told IG investigators that they had identified 2,682 records that were not being exported to any screening database used by frontline federal employees, such as immigration workers and consular officials.

As the TSC workers combed through the trove of mystery records, they found that 2,118 of them should not have been put on any watch list and should be removed from the system. During a subsequent manual review of the remaining 564 records, TSC employees identified eight individuals who had not been appropriately placed on the watch list and needed to be renominated for inclusion in the TSDB.

'Even a single omission of a suspected or known terrorist from the watch list is a serious matter,' the auditors said.

Boyle said center employees have started to reconcile the data contents of the two components every day to eliminate variations in their content.

The screening center's leadership has known since the organization's inception in 2003 that much of the data it uses is unreliable, according to various reports and public statements by government officials.

For example, one database that the TSC received in its early days lacked a field to identify the subject's gender because it was built with the assumption that all terrorists are men, FBI officials said.

The center's staff has combed through large parts of the data it inherited from other agencies to eliminate poison-pen data, among other inaccuracies. The auditors found in their recent report that TSC had not established a process for regularly reviewing the information in the database.

Checking no-fly list

Boyle said center officials recently have been combing over the data in the TSDB's No Fly List and have developed a plan for scrubbing other categories of information in the database.

But some proposed names have been entered in the database on the strength of nomination only by the FBI without participation by other overseers, the auditors said.

Law enforcement agencies at various levels of government can check TSC's records about a person via telephone inquiries in some cases. Some federal agencies, such as the State Department's consular offices and the Customs and Border Protection arm of the Homeland Security Department, have access to TSC records via linked databases.

Despite its repeated acknowledgements that the information in TSC systems is faulty, the FBI has at times claimed that press reports citing watch list inaccuracies are inaccurate.

The auditors said that TSC's operations had improved significantly during the past two years but that major technology and business process flaws hindered its effectiveness.

In a written response to the report, the FBI generally accepted the auditors' recommendations and described its evolving plans to improve TSC operations.
The Justice Department's inspector general's staff issued its first report on technology at the Terrorist Screening Center two years ago. At that time, the IG auditors made 40 recommendations for changes. Its recent report reflects some improvements in the center's operations, but the auditors identified 18 areas in which major improvements were needed.

Additional problems with the TSC's technology that the auditors identified included:

• Cumbersome FBI data-handling practices that create unnecessary errors, anomalies and inconsistencies in the records.
  
• The FBI practice of entering data about suspected terrorists into a downstream database, which prevents other agencies from reviewing the bureau's data.
  
• Needless delays in entering data into the terrorist watch list, which cause a significant vulnerability to the integrity of the consolidated database.
  
• Duplicate information problems with 6,262 records in the Terrorist Screening Database.
  
• Weak quality assurance methods.
  
• Tardy redress reviews.
  
• Delays in resolving data quality assurance problems that ranged as long as 329 days, with an average processing time of 80 days. As of February, TSC officials were working with 3,000 open quality assurance problems.
  
• Delays and inefficiencies in handling requests by specific people for correction or deletion of their watch list records.