The Truth Behind Standard Criminal Checks

What employers should know.

By Genie Tyburski

Genie Tyburski is a Web research applications specialist at Ballard Spahr Andrews & Ingersoll.
She manages The Virtual Chase at www.virtualchase.com.

Chris O’Keefe, a former technology manager for the Teachers Insurance and Annuities Association of America-College Retirement Equities Fund, knows the cost of making a poor hiring decision. According to an April 25 Newsweek article, O’Keefe allowed a new contract employee from one of the investment management firm’s preferred vendors to have access to sensitive client information. While employees in the department routinely handle confidential information, no one was aware the new worker was a felon convicted of conspiring to commit insurance fraud.

Should they have known? As the Newsweek article points out, this is one instance where a simple Google search on the individual’s name — Sonia Radencovici — easily uncovered Securities and Exchange Commission filings that summarize the criminal charges. An adept researcher could have followed the lead to find documents about her conviction and sentencing filed in a federal district court in Connecticut — all for about 16 cents plus the cost of the researcher’s time. What went wrong in this particular incident?

Tek Systems, the contractor that placed Radencovici with TIAA-CREF, hired Kroll Inc., a reputable investigative research firm, to conduct the criminal check. Kroll defended its failure to find the relevant criminal conviction by asserting that Tek requested a “standard” check. This industry standard includes conducting an investigation in counties where the research subject has lived or worked in the past seven years. The time period is not arbitrary. It comes from section 605 of the Fair Credit Reporting Act (15 U.S.C. §1681c), which precludes consumer reporting agencies, such as Kroll, from sharing information about arrests (convictions excepted) that are older than seven years except when the annual salary of the prospective hire equals or exceeds $75,000. However, it’s important to note that state law also might have an impact on the type of information reported, as well as the time period covered.

Limiting the investigation to certain counties is a matter of practicality. Other than the FBI’s National Crime Information Center, which only law enforcement and criminal justice agencies can access, no national criminal research system exists.

In recent years, some public records vendors have touted their criminal records databases as a nationwide system on par with the NCIC. But the claim is questionable. First, many of these systems don’t cover all states. Second, coverage isn’t consistent from state to state. In some states, the databases contain county court filings and state corrections records, but in others they comprise only corrections records. Since corrections records are created when someone goes to prison, these databases sometimes fail to provide information about misdemeanors or felonies of a nonviolent nature. Third, when county court records are available for a state, the database can exclude coverage of certain counties. It might provide limited date coverage or contain records that have not been updated for several months. Finally, the criminal record-keeping process is imperfect and sometimes results in missing or inaccurate disposition data.

President and CEO of the Employment Screening Resources group Carl R. Ernst and pre-employment screening attorney Les Rosen, contend that use of these commercial research systems alone fail to meet the due diligence requirements of the FCRA for employment purposes. The FCRA requires that Consumer  Reporting Agencies “follow reasonable procedures to assure maximum possible accuracy of the information” (§1681e). Sometimes the databases provide more than what a credit reporting agencies might report (e.g., information about an old arrest). Other times, the commercial databases lack identifying information, such as a date of birth or Social Security number, or information about the disposition of a matter.

In some cases, vendors themselves forbid the use of their databases for FCRA regulated purposes. For example, Thomson West forbids use of Crim, its criminal records on Westlaw databases, for FCRA regulated activities such as checking the background of an employee or candidate for employment. West isn’t the only vendor to disallow such uses. Other databases you can’t use for FCRA activities include LexisNexis criminal records, ChoicePoint’s AutoTrackXP and Accurint.

However, if the contract permits such use and the researcher understands the limitations of the databases, Ernst and Rosen said they favor using commercial databases to supplement offline research. In fact, in their 1996 cowritten article, “Issues and Opportunities in Pre-employment Screening,” Ernst and Rosen argue that “failure to utilize a database with broader geographic coverage than searches of local courts and statewide databases is ‘negligent’ because the information is readily available and casts a much wider net than traditional criminal history searches.”

This begs the question, should Kroll researchers have searched a database, such as the U.S. Party/Case Index? While it doesn’t cover every federal court, nor does it provide identifying information in accordance with policy set by the Judicial Conference, it serves as an inexpensive, ready-to-use tool for querying multiple federal courts across the nation. If necessary, the researcher could verify potentially relevant information at the courthouse.

In light of the Radencovici incident, Newsweek suggests “standard” checks should include searching Google. While Newsweek might not understand the FCRA requirements, it’s hard to argue with the results in this case. Google uncovered the relevant information — Kroll didn’t.

But the solution isn’t that simple. Since the FCRA mandates that credit reporting agencies provide current, accurate and relevant information, use of Google for pre-employment screening could significantly add to the cost of the research for little value in return. You would have to verify potentially relevant matches, which might be easy with a name such as Radencovici, but would be increasingly difficult with more common names.

Rather, the incident underscores the need for education and improved communication between employers and researchers. Employers, when faced with choices that affect the coverage and cost of the research, should not read the menu from right to left. While education and improved communication in the present example might not have changed the outcome, it could have put the employer in a better position to make an informed decision about the initial trust it should place in a new employee.

Entire contents copyright © 2005 James Publishing, Inc.

All Rights Reserved.