witnesses can overcome the weapon focus effect,
which implies that weapons probably do not capture
attention automatically.Kerri L. PickelSee alsoEstimator and System Variables in Eyewitness
Identification; Eyewitness Descriptions, Accuracy of;
Juries and Eyewitnesses; Juries and Joined Trials; Stress
and Eyewitness MemoryFurther Readings
Loftus, E. F., Loftus, G. R., & Messo, J. (1987). Some facts
about “weapon focus.”Law and Human Behavior,
11 (1), 55–62.
Pickel, K. L. (1999). The influence of context on the
“weapon focus” effect. Law and Human Behavior,
23 (3), 299–311.
Steblay, N. M. (1992). A meta-analytic review of the weapon
focus effect. Law and Human Behavior, 16(4), 413–424.WITNESS MODEL
Formal (mathematical and computer simulation) mod-
els have been developed and applied in a wide range of
areas in psychology. The application of formal models
can be helpful for clarifying theoretical assumptions,
generating precise predictions, and testing the ade-
quacy of theoretical explanations by comparing the-
ory-generated predictions with human-generated data.
This entry describes a computational model called the
WITNESS model that has been developed for eyewit-
ness identification. The WITNESS model assumes that
lineup members are compared with an error-prone
memory representation of the perpetrator and that
identification decisions are based on a weighted com-
bination of absolute and relative match information.
Eyewitness identification research has a clear and
immediate real-world application. Given the converg-
ing evidence that mistaken identification is the primary
factor contributing to wrongful conviction, it is impor-
tant to understand the causes of mistaken identification
and develop techniques to minimize its occurrence.
Toward that end, eyewitness identification research has
provided several important insights into the factors that
cause or are associated with eyewitness errors and has
been instrumental in efforts to reform the procedures
by which identification evidence is obtained.It is also important to understand the psychological
processes that underlie eyewitness identification deci-
sions and errors. Here, the questions are not about
what happens, but rather why what happens happens.
Toward that end, it can be useful to develop and test
comprehensive, mathematical models of the memory
and decision processes that underlie eyewitness iden-
tification. One model that has recently been developed
for eyewitness identification decisions is called the
WITNESS model. This entry will present a brief
overview of the WITNESS model and how it is
applied to eyewitness identification.The WITNESS Model
The WITNESS model makes a few simple assumptions
about the memory and decision processes that underlie
identification decisions and produces response proba-
bilities that can be compared with data. The compari-
son between responses generated by the model and
responses generated by human witnesses provides a
means of evaluating the assumptions of the model.
These assumptions are straightforward. First, the
model assumes that any event or stimulus may be rep-
resented as a vector of features,f 1 ,f 2 ,f 3 ,...,fM. For
eyewitness identification, the critical stimulus is the
perpetrator of the crime, represented in the model as a
vector P. Memory, denoted as a vector M, is an
incomplete and error-prone representation of the per-
petrator. Specifically, in the model, a given feature of
the perpetrator Pjis stored correctly with probability s
and stored incorrectly with probability 1−s.
Lineup members are also represented as vectors of
features. An important aspect of eyewitness identifi-
cation generally and the WITNESS model specifically
concerns the similarity relationships between lineup
members (denoted by vectors L 1 ,L 2 ,L 3 ,...,LN), the
perpetrator (P), and the witness’s memory of the
perpetrator (M).
According to the model, when the lineup is pre-
sented, the witness compares each lineup member
with his or her memory of the perpetrator. The result
of this comparison process is a number of match val-
ues, each one indicating the similarity between lineup
member (Li) and the witness’s memory (M) of the
perpetrator, denoted m(Li,M). Thus, for a six-person
lineup, there will be six match values.
Of course, these match values do not specify the
decision that the witness will make. The witness must
apply a decision rule that considers the match values864 ———WITNESS ModelW-Cutler (Encyc)-45463.qxd 11/18/2007 12:44 PM Page 864