November 13, 2001
Study says police lineups often miss the mark
A new California State University, Sacramento
study is critical of on-the-scene, eyewitness identification
of crime suspects, and also casts doubt on the reliability
of other types of police lineups.
The study, by CSUS psychology professor Bruce Behrman and
student researchers, is the largest of its kind ever carried
out. It appears in the most recent issue of the Journal of
Law and Human Behavior.
Data came from Sacramento city crime records. It was gathered
with significant assistance and support from the Sacramento
City Police Department, which is intent in improving its lineup
"Our findings are consistent with the hypothesis, which
was that 'field show-ups' may bias the eyewitness," Behrman
says. "Most people will identify the suspect as the criminal,
and some may do so even in cases where the police decide not
to make an arrest."
Behrman and the students examined five years of crime data
from more than 700 police lineups. They looked at "field
show-ups," in which eyewitnesses identify a suspect near
the crime scene, as well as photographic lineups and live
For each case, they noted the eyewitness response along with
a ranking of how much physical evidence police had gathered:
substantial, minimal or none. The suspect's eventual trial
or release was not considered.
In field show-ups, eyewitnesses identified the suspect as
the criminal about 78 percent of the time, regardless of the
level of evidence police ever acquired. In many cases, police
decided not even to book the people identified by eyewitnesses.
Behrman says that if the field show-ups were truly effective,
the percentage of eyewitness identification level should have
increased at better evidence levels. The findings suggest
eyewitnesses are strongly biased by seeing a suspect in police
custody near the scene of a crime.
The study shows photographic and live lineups were better,
though far from bullet-proof in court.
In these lineups, 65 percent of eyewitnesses make an identification
of a suspect for whom police had "substantial" evidence.
Far fewer identified suspects when the evidence was "minimal"
or "none." This "real life" finding is
consistent with previous laboratory findings made by Behrman
and other researchers throughout the country.
However, when just one week has passed, Behrman's team found
that eyewitness identification of suspects with "substantial"
evidence drops to 33 percent. In other words, memories fade
The team also found that eyewitnesses tend to be more likely
to finger suspects of another ethnic group; victims are no
more accurate than other witnesses; and identification accuracy
doesn't seem to be influenced by whether a weapon was used.
Behrman, a professor at CSUS since 1967, has been studying
eyewitness memory since the late 1980s. He often consults
with area law enforcement agencies, and is a frequent expert
witness in court. He and his students have just begun another
study focused on whether the confidence level of an eyewitness
indicates the eyewitness's accuracy.
further information send E-Mail to firstname.lastname@example.org or
Public Affairs (916)
Index of Stories
Return to CSUS Home Page