Skip to content

Addressing Bias in Latent Print Processing- CSAFE Examines Crime Laboratory Evidence Submission Forms

Evidence submission forms are critical to accurate forensic evidence analysis, but learn how they may unknowingly bias examiner conclusions

Proper handling of forensic evidence is critical in a criminal investigation.  Evidence preservation, the integrity of analytical results and the admissibility of evidence in court are all crucial to justice.

Individuals requesting analysis of evidence must keep accurate records. Many crime laboratories use very specific evidence submission forms to help ensure timely and accurate testing.

However, the type of information each crime lab requests about evidence before analyzing can vary widely. Moreover, it’s possible that evidence submission forms themselves can influence the results of an investigation with potentially biasing information.

CSAFE Collects Evidence Submission forms from Crime Labs across the U.S.

Approximately 75 percent of accredited crime laboratories from across the United States have contributed to CSAFE research to investigate latent print processing evidence submission forms.  Researchers have collected 105 case processing forms from U.S. crime labs, and are analyzing what type of information requested such as a suspect’s criminal history, race or age.

How Much Information is Necessary to Compare Latent Prints?

Latent print examiners clearly need a certain amount of information to conduct their analysis.  But CSAFE researcher from University of Virginia Dr. Daniel Murrie says they may not need all of the information crime labs are currently collecting.

“Any information about the case that isn’t necessary to do the forensic task at hand is known as ‘context irrelevant information,’” Murrie explained.

For example, is it necessary for a latent print examiner to have information about the victim when analyzing prints, or could that unknowingly influence the decision?

“When task-irrelevant information suggests that a particular conclusion is true—such as two prints matching or not matching—it is possible that these phenomena will influence how forensic scientists perceive and evaluate the evidence they are analyzing,” Murrie said.

Murrie explains that an examiner may approach a latent print comparison very differently with knowledge such as a suspect has already confessed or instead has a reliable alibi.

“Distinguishing between task-relevant and irrelevant information is important because the forensic science community is always striving to increase the standardization and reliability of its analyses,” CSAFE researcher Dr. Sharon Kelley said. “It is difficult to achieve reliability if different examiners rely on different amounts and types of potentially biasing information to inform their analyses.”

CSAFE researchers say that it’s possible that even information that is very relevant to the analysis can be biasing from time to time, such that examiners might benefit from looking at that information only after making other, preliminary determinations—a process known as linear sequential unmasking.

Consistent with recommendations from the National Commission on Forensic Science, it is imperative that forensic disciplines identify what information is critical to the completion of a particular forensic science task (e.g., latent print comparison, ballistics, arson investigation) and what information is superfluous, and can therefore be omitted.

What We are Learning About Evidence Submission Forms

Preliminary results of the CSAFE study indicated that nearly all forms requested information about the offense, including offense type, location and description, police or incident report and reporting or arresting officer.  Similarly, nearly all forms requested information about the suspect, including name, age or date of birth, sex, race and criminal history information. Victim information was also routinely collected, including name, age or date of birth, sex and race.  Although not a majority, several forms included questions that are potentially task-relevant to certain determinations but have the potential to bias forensic analysts, such as “Were the suspect and victim acquainted with each other?”  Approximately 18% of forms requested information likely task-irrelevant to forensic science work such as “Is the suspect a flight risk?”

“What we’re learning is there are a lot of blank spaces where what gets written in about the offense or about the case could really differ based on the person completing the form,” Kelley said.  “We would like to see how much of that is completed and how many details about an offense get written in as another way to get a sense of what information should get stripped out before an analyst actually sees it.”

Strategies to Reduce Bias in Latent Print Analysis

CSAFE researchers say that a context information management procedure can help reduce bias in examiners by decreasing unnecessary exposure of task-irrelevant information.

“This may be accomplished by restricting access to information according to an examiner’s duties,” Murrie said.  “For example, a case manager may have broad access to case information in order to complete administrative duties whereas a latent print examiner may only have partial access to information—specifically, they would have access only to information that is absolutely necessary to the specific analysis at hand (i.e., is task-relevant).”

So, instead of a latent print examiner seeing all of the information about a case, including the evidence submission form, information about the suspect or victim, the type of offense, etc., he or she would only be given the actual prints.

Another potential way to manage context is to adapt existing Laboratory Information Managements Systems (LIMS).  For instance, LIMS could limit access to information based on someone’s role in a lab or someone’s role in a particular case.

CSAFE researchers plan to expand their study to include collection of completed evidence submission forms.  Join UVA researcher and postdoctoral fellow Brett Gardner for a presentation at AAFS in February 2018 to learn more.  Interested in diving deeper into the role human factors and cognitive psychology play in forensic science? Visit the CSAFE blog!

FROM THE BLOG