For the Defense - Vol. 7, Issue 4 - 6

" There Is No 'Right' Answer " :
The Perils of Cognitive Bias in
Autopsy Decisions
I
am a bias expert, but I hate the word " bias. " To most people,
it connotes malfeasance and evokes defensiveness-but to
psychologists like me, it means something much different.
We have known for over a century that all humans are subject
to cognitive bias, such that people with differing mindsets
often interpret the same information in different ways.1
In
various studies, for example, people judged the same hill as
steeper if wearing a heavy backpack, judged the same water
bottle as closer if thirsty, and judged two people as looking
more alike if falsely told that they were related.
Like a reflex, cognitive bias is natural, subconscious, and
even adaptive, but it can also impede our search for the truth,
which is paramount in criminal investigations. A decade ago,
my colleagues and I published a seminal paper in which we
explained that cognitive bias can undermine judgments
of forensic science evidence, especially when analysts are
exposed to extraneous case information.2
In one early study,
for example, fingerprint experts unwittingly changed 17%
of their prior opinions of the same prints after being led to
believe that the suspect had confessed or provided an alibi.3
Since then, research in this area has flourished, with studies
showing that cognitive bias can impact judgments of DNA,
toxicology, arson, anthropology, handwriting, blood spatter,
digital, and firearms evidence, among others.4
Accordingly,
the National Commission on Forensic Science has implored
laboratories to " take appropriate steps to avoid exposing
analysts to task-irrelevant information. " 5
6 For The Defense l Vol. 7, Issue 4
The Study: Cognitive Bias in Manner of Death Decisions
We asked members of the National Association of Medical
Examiners ( " NAME " ) to read several clinical vignettes and
determine how the person in each vignette had died (i.e.,
suicide, natural, accident, homicide, or undetermined). One
vignette described a three-year-old girl who was found
unresponsive at home, taken to the hospital with a head
injury, and died shortly thereafter. To test for cognitive bias,
each respondent read one of two versions of this vignette,
between which we changed the girl's race and caretaker but
made no other changes.
We received responses from 133 NAME members-many of
whom judged the death as " undetermined, " which was the
most justifiable decision in light of the provided information.
Among the other respondents, we found clear evidence
of bias: If told that the girl was white and found by her
grandmother, they more often ruled the death an accident
(28%) than a homicide (13%)-but if told that she was Black
and found by her mother's boyfriend, they more often ruled
Jeff Kukucka
As evidence mounted across disciplines, I worked with my
colleagues in the OSAC for Forensic Science to develop and
promote safeguards against bias, but one group remained
obstinate-namely, death investigators, who insisted that
manner of death decisions require unlimited access to
information and are not subject to cognitive bias. So, we
decided to test that possibility.
For the Defense - Vol. 7, Issue 4

Table of Contents for the Digital Edition of For the Defense - Vol. 7, Issue 4
