Verified by 
Key points
- People who consider the same information, but in a different order, often reach different opinions.
- In forensic labs, poor information sequencing can prejudice analysts and result in costly errors.
- New research-based tools can help forensic analysts sequence information effectively.
- These procedures stand to increase the reliability and transparency of forensic analysts' decisions.
Note: This post was co-authored with Adele Quigley-McBride, Ph.D. (Duke University), Brandon L. Garrett, J.D. (Duke University), and Tiffany Roy, M.S., J.D. (ForensicAid).
Contrary to popular belief, analyses of forensic evidence (e.g., fingerprints, firearms, bloodstain patterns) are typically performed not by computers, but by human analysts—and as such, they are subject to human limitations and biases. To date, forensic errors have contributed to hundreds of known wrongful convictions resulting in thousands of years of wrongful incarceration, leading researchers to wonder: Why did these mistakes happen, and how can we prevent them in the future?
Over the last decade, researchers have made great strides in measuring the frequency of forensic errors, identifying the psychological causes of those errors, and proposing some solutions. What has been missing, however, is concrete guidance on how forensic laboratories—many of which are badly understaffed and/or underfunded—can take practical steps to minimize error.
In our latest publication, we offer a simple solution—one that fits on a single sheet of paper. To be exact, we drew upon both seminal and contemporary research to develop a worksheet that forensic labs and analysts can use to document and optimize their decision-making processes. The worksheet is free to download and can be tailored to virtually any forensic discipline. Below, we describe its development and use, beginning with the classic studies that laid its foundation.
A Primer in Primacy Effects
Dating back to Ebbinghaus’ groundbreaking studies of memory in the 1880s, psychologists have long known that information appearing earlier in a sequence has a stronger impact on people’s memories, judgments, and decisions—a phenomenon called the primacy effect. In one classic study, for example, participants judged someone described as “intelligent, industrious, impulsive, critical, stubborn, and envious” more favorably than someone described as “envious, stubborn, critical, impulsive, industrious, and intelligent,” even though it was the same list of traits in reverse order. Primacy effects can also affect how we interpret subsequent information; in that same study, people tended to interpret “critical” as either a positive (i.e., meticulous) or negative (i.e., judgmental) trait, depending on which traits came before it.
The Problem: Order Effects in Forensics
Informed by studies such as these, a seminal 2013 article explained that some forensic errors may result from the amount of information that analysts receive and the order in which they receive it. In theory, a forensic analyst’s job is to give an independent opinion of materials for which they have expertise (e.g., pattern evidence, bodily fluids, firearms). In practice, however, forensic analysts often also receive extraneous and prejudicial information from police, lawyers, and other investigators—such as information about a suspect’s race and criminal history—that can taint their subsequent analysis of the forensic evidence in undesirable ways.
There is now ample evidence that extraneous information can bias forensic analyses of all kinds. In one study, fingerprint analysts reached different opinions of the same fingerprints if told about the suspect’s confession or alibi ahead of time. In another study, crime scene investigators interpreted the same death differently if told in advance that the deceased had a history of depression or domestic violence. Further evidence comes from real-world errors, such as the FBI’s high-profile misidentification of Muslim-American attorney Brandon Mayfield as the perpetrator of the 2004 Madrid train bombings—a mistake that an independent review later attributed to “confirmation bias.”
The Solution: Information Management
The solution to this problem may seem obvious: Ensure that forensic analysts don’t receive any biasing information prior to their analysis. This is more difficult to achieve in practice, however, as potentially biasing information is unavoidable or even necessary in some forensic disciplines.
Last year, psychologists introduced a procedure called Linear Sequential Unmasking–Expanded (LSU-E), which aims to balance idealism and pragmatism in forensic decision-making. As it stands, forensic analysts often receive and consider information in an incidental and indiscriminate fashion. In contrast, LSU-E provides guidance on how to thoughtfully sequence information so that analysts do not consider potentially biasing information until after critical steps of the analysis are complete.
For example, crime scene investigators may receive biasing information from police before analyzing a scene, but that does not happen systematically—sometimes police have left the scene before the analyst arrives, and different officers may impart different information. LSU-E provides a framework which analysts can use to make and document their decisions about what information to consider and in what order, such that their initial opinion should be based solely on the most relevant and objective information. Then, the analyst may consider other information step-by-step, adjusting their opinion as appropriate, and documenting any such adjustments.
The Challenge: Getting Laboratories on Board
To encourage and facilitate the use of LSU-E, we recently developed and published a worksheet that analysts can use to record the sources of information that guided their analysis and when they considered each. The worksheet walks analysts through their consideration of the information available to them, asks them to rate each item along three criteria (objectivity, relevance, and biasing power), and encourages them to describe any steps taken to mitigate its potential biasing effect.
In addition, we developed training materials and examples of how to rate and consider different kinds of information that are often available to analysts. In theory, this worksheet should also help non-experts (e.g., judges and juries) make informed decisions about the trustworthiness of the resulting forensic opinion by increasing transparency about the basis for the analyst’s opinion, whether and how their opinion evolved over time, and whether they took appropriate steps to avoid bias.
Historically, some forensic labs have embraced procedural changes to combat bias and error, while others have resisted—perhaps because they lacked concrete guidance on how to do so. We hope that our worksheet and supporting materials will encourage more labs to follow suit. Notably, our worksheet represents a novel application of research-based solutions to a real-world problem and an interdisciplinary effort between three psychologists, a legal scholar, and a forensic analyst—individuals from very different backgrounds who share the common goal of maximizing the validity and transparency of forensic opinions.
