Sie sind auf Seite 1von 12

Law and Human Behavior, Vol. 7, No.

1, 1983

The Ability of Prospective Jurors to Estimate the


Accuracy of Eyewitness Identifications

J o h n C. B r i g h a m * a n d R o b e r t K. B o t h w e l l *

This study investigated whether people's "common knowledge" is sufficient to enable them to accurately assess
the validity of evidence based upon an eyewitness report. Prospective jurors were provided with descriptions of
actual research studies on eyewitness identification and asked to predict the research results. A sample of
registered voters, selected to represent the court's lists of prospective jury members, was presented with scenarios
derived from the method sections of previously conducted empirical studies of eyewitness identification involving
target-present lineups. Respondents' estimates of the hit rates in the three target-present situations were signifi-
cantly greater than the actual hit rates in each of the three cases. The mean percentage of respondents who
overestimated the hit rates, averaged across the three cases, was 83.7%. The results indicate that awareness of the
unreliability of eyewitness evidence does not appear to be part of the "common knowledge" of prospective jury
members. These data can be seen as refuting the claim made by some courts that expert testimony about
eyewitness evidence would not tell the jury members anything they did not already know.

INTRODUCTION

Erroneous eyewitness identification of the accused has long been recognized by legal
scholars as a crucial problem in the administration of justice (e.g., Borchard, 1932;
Frank & Frank, 1957; Frankfurter, 1927; Gardner, t933; Houts, 1956; Muensterberg,
1912). Despite numerous reported instances of misidentification (e.g., see Wall, 1965;
Sobel, 1972; Loftus, 1979) legal writers have asserted that most juries and many
judges are unaware of the sources of error in eyewitness testimony and place undue
faith in its veracity. Indeed, jurors appear to regard eyewitness evidence as one of the
most persuasive kinds of evidence that can be presented. Wall (1965, p. 19) has
asserted that "in general, juries are unduly receptive to identification evidence and are

*Department of Psychology, Florida State University, Tallahassee, Florida 32306. This research was supported
in part by National Science Foundation grant BNS 77-27476 to the first author. We are indebted to Melissa
WolfsKeil and David Ready for their comments on earlier versions of this paper. These data were presented at
the Twentieth International Congress of Applied Psychology, Edinburgh, Scotland, 1982.

19
0147-7307/83/0300-00t9503.00/0 9 1983 Plenum Publishing Corporation
20 BRIGHAM AND BOTHWELL

not sufficiently aware of its dangers; evidence of identification, however un-


trustworthy, is taken by the average juryman as absolute proof." Loftus' (1975) study
showed that simulated juries were significantly more likely to vote for conviction if
there was eyewitness corroboration of the crime (72%) than when there was no
eyewitness testimony (18%), even when the witness was discredited on the basis of
being legally blind (68%).
One possible reason laypersons place considerable faith in eyewitness identifica-
tion evidence is that they may subscribe to a version of the ancient theory of "eidola,"
which postulated that faint copies of objects can enter the mind directly (Neisser, 1967,
p. 3). Buckhout (1974) has suggested that most laypersons believe that the brain
records veridical perceptions of particular events in a manner similar to a videotape
recorder. Many people seem to believe that perceptions of particular events are stored
on something akin to memory "tapes." Hence, when an individual is asked to recall a
past event, it is assumed that the individual can simply play back the appropriate tape,
producing an exact duplicate of the original perception. Numerous studies have shown
the videotape recorder analogy to be extremely inaccurate (e.g., Bartlett, 1932; Miller,
1956; Neisser, 1967). Contrary to the analogy, research has demonstrated that prior
knowledge affects the ways by which information is encoded and retrieved, often
resulting in distortions in recall arising from constructive processes at time of acquisi-
tion and/or from reconstructive processes at the time of recall. The concept of recon-
structive memory is especially important because people may believe they are remem-
bering accurately when, in fact, they are not. Thus, people's confidence in their
memory is a questionable criterion for deciding whether those memories are accurate
(Bransford, 1979). In addition, one's expectations can strongly affect reconstructive
activities. Loftus and Palmer (1974), for example, have demonstrated that leading
questions that change such expectations can facilitate errors in reconstruction.
Pyschological research on eyewitness identification is typically reported in tech-
nical books or journals which are not generally available to, or read by, laypersons
(Katz & Reid, 1977). Hence the typical jury member must rely solely on his or her
"common sense" in evaluating eyewitness evidence. However, "common sense"
notions may not accurately reflect the true state of affairs. For example, many people
believe that memory for stressful events is better than memory for nonstressful events
(e.g., see Marshall, 1966; Brigham, 1981). Psychological research, however, has
suggested that under stress conditions the likelihood of erroneous perception increases
(e.g., see Bruner, 1958; Buckhout, 1975).
Another idea in which "common sense" assumptions may differ from empirical
findings concerns the relationship between witness confidence and accuracy. Although
it may seem logical that witnesses who are more confident of their identifications
would be more likely to be accurate, recent research suggests that the confidence-
accuracy relationship is not necessarily a strong positive one (see Deffenbacher, 1980;
and Leippe, 1980, for discussions of that issue). The concept of reconstructive mem-
ory implies that people's confidence in their memories may not be a good criterion for
deciding whether those memories are accurate.
Despite such suggestive examples of the discrepancy between common belief and
results of general psychological research, many courts have objected to the use of
expert testimony by psychologists on eyewitness identifications on the grounds that its
EYEWITNESS IDENTIFICATION 2i

subject matter is within the "common knowledge" of the average layperson. For
example, in the case of Commonwealth v. Jones, Chief Justice Tauro of the Massachu-
setts Supreme Court excluded a research psychologist's expert testimony arguing, in
effect, that neither a psychologist nor a psychiatrist could tell the jury anything the jury
did not already know:
9 . . the proferred testimony, concerning the psychological characteristics of recall and the dangers
attached thereto, would seem to be well within the experience of men, but, in any event, it is
certainly within the ken of a trial judge who must constantly deal with problems of identification.
The judge below hardly needed the assistance of a psychologist and a psychiatrist to make findings
concerning matters which the Supreme Court of the United States and the highest court of this
Commonwealth have treated in numerous opinions (quoted in Wilson, 1975, p. 73).

Researchers concerned with the factors that affect eyewitness testimony have disputed
such claims, but there is little direct evidence on the relationship between jurors'
"common sense" assumptions and the empirical findings in the area (see Wells, in
press, for a review). In an exploratory study, Porietas and Loftus (cited in Loftus,
1979, pp. 172-177) surveyed registered voters attending the University of Washington
to investigate what they knew about the factors that affect eyewitness testimony.
Students answered questions on the effect on eyewitness accuracy of (1) cross-racial
identification, (2) stress, (3) violence of the event, (4) weapon focus, (5) question
wording, and (6) new information. The responses suggested that common beliefs held
by people regarding eyewitness ability sometimes conform to the psychological re-
search but in other cases do not; the percentage of subjects who answered correctly (in
the researchers' judgment) ranged from 18% to 90% on the multiple-choice
questions.
Yarmey and Jones (in press) extended the findings of Porietas and Loftus to the
areas of time perception, police vs. civilians as eyewitnesses, decay of memory for
faces over time, the influence of mugshot identifications on later identifications from a
lineup, the relationship between eyewitness confidence and eyewitness accuracy, and
the effects of age of a witness on eyewitness testimony. In addition, Yarmey and Jones
surveyed various populations such as legal professionals, law students, other students,
and citizen jurors. The findings for citizen jurors add generalizability to the findings of
Porietas and Loftus. Approximately two-thirds of the citizen jurors were unaware that
eyewitnesses tend to overestimate the length of time involved in a witnessed crime.
Approximately two-thirds of the citizen jurors believed that police would be superior to
civilians as eyewitnesses. Fifteen percent thought that eyewitnesses' memory for faces
would be 90%-95% accurate several months after first seeing the face. Over half were
unaware that an eyewitness' identification of someone from a set of photographs is
likely to later produce an identification of the same person from a lineup (regardless of
whether the identified person is guilty or not). Approximately two-thirds of the citizen
jurors indicated that they believed in the existence of a positive relationship between
eyewitness accuracy and eyewitness confidence. Hence, there appear to be many areas
where the layperson's beliefs are discrepant with experimental findings (see also
Deffenbacher & Loftus, 1982, for further support of this general conclusion).
Wells (1978) has argued that although many psychological researchers have
assumed that judges and jurors weigh eyewitness testimony more heavily than they
should, "there is not empirical evidence to support the assumption that jurors and
22 BRIGHAM AND BOTHWELL

judges are overbelieving of witnesses" (Wells, 1978, p. 1551, italics in original). The
present study is designed to yield data relevant to the validity of this assumption. By
providing prospective jurors with descriptions (sans results) of actual research studies
in this area and asking them to predict the research results, we can assess whether their
"common knowledge" is sufficient to enable them to accurately access what is likely
to happen in actual eyewitness situations. ~ The studies selected met several criteria:
they were current, they investigated questions of wide interest, and they employed
situations of high mundane realism. To increase the external validity of the results,
subjects were randomly selected from the same population from which the jury pool is
created.

METHOD

Subjects
Subjects were selected so as to yield a sample equivalent to the samples of
prospective jury members compiled by the courts in Leon County, Florida. The courts
randomly select from a list of all registered voters. Since this voter list was unavailable
to us, 238 persons were randomly selected from the telephone directory and contacted
by phone. One hundred seventy three (73%) of these indicated that they were regis-
tered to vote. Of the registered voters, 132 (76%) agreed to participate in the survey,
and questionnaires were sent to them. The cover letter described the survey as designed
to find out how well people can estimate the findings of research on eyewitness
identifications. Ninety completed questionnaires were returned, accounting for 69% of
those who agreed to participate and 52% of the total sample of registered voters
originally contacted.
The sample contained 47 (52%) males and 43 (48%) females; the mean age was
41 years old. Most were married (70%), white (94%), and relatively well educated
(80% had attended some college and 57% had received college degrees). For purposes
of comparison, 1981 Leon County statistics indicate that 46% of registered voters are
males, 80% are white, and the mean age is 38.9. Almost one-third (31%) of our
respondents said they had served on a jury before; most of these indicated that they had
served either once (54%) or twice (18%). Only 11% of those who had served on a jury
said that eyewitness identification had been an important part of the case(s) that they
observed.

Instrument
The questionnaire was designed to assess respondents' ability to estimate accura-
cy rates in eyewitness identification situations and, in addition, their own beliefs about
the effects of stress on witness accuracy, the relationship between witness confidence
and accuracy, and the weight which should be given to eyewitness evidence in com-
parison to other types of evidence.

IThis method has previously been employed only with college student subjects (see Wells, in press). The results
of these studies (i.e., that students' estimates are considerably discrepant from the experimental data) are
corroborated in the current research, which employed a more heterogeneous sample.
EYEWITNESS IDENTIFICATION 23

Two recently conducted research studies were described in the questionnaire, one
employing a staged crime and the other utilizing somewhat atypical visits to conven-
ience stores. In each of the studies the researchers had employed target-present lineups
in the identification phase. The setting and procedure of each study was described in
some detail via a one-page synopsis (380 to 530 words long) detailing the methods but
not the results of the study. To simplify matters, each description indicated that the
study had employed exactly one hundred subjects. After reading each scenario, re-
spondents were asked to estimate the number of subjects who were able to correctly
identify the target in the lineup.

Scenarios
The first scenario described the high-seriousness, knowledge-after condition of
the Leippe, Wells, and Ostrom (1978) study in which midwestern introductory psy-
chology students were exposed to a staged calculator theft and were subsequently
asked to make eyewitness identifications from a photo lineup. The theft occurred while
five students, one of whom was a confederate, were waiting in a reception room while
the experimenter prepared materials elsewhere. After everyone was seated, the experi-
menter entered and said: " I have a student on the phone from an earlier session who
says she left a bag here. Has anyone seen it?" The confederate pointed to a brown
paper bag placed on a nearby table and asked if that was it. The experimenter replied
that it was, and left the room to complete the phone call. After the experimenter left,
the confederate grabbed the paper bag, dropped it (to make sure all the students were
aware of the theft), picked it up, and quickly left the room. The thief was seen by the
other students for a total of two to three minutes. Approximately 1 minute later the
experimenter returned and, after being informed of the apparent theft, exclaimed,
"You're kidding! That was her calculator!" After implying that he would attend later
to the matter of the theft, the experimenter took the students to a second room where
they filled out a questionnaire. They then were taken individually to a third room where
they were exposed to a six-photo target-present lineup. After reading the scenario of
this study, our survey respondents were asked: ~ of the one hundred students who
were exposed to the procedure just described, approximately how many would you say
were able to correctly identify the thief? (His picture was included among the six
photos). Answer with a single number between zero and one hundred in the space
provided below."
The second scenario described a field study carried out by Brigham, Maass,
Snyder, and Spaulding (1982) in which white convenience-store clerks were asked to
identify from two photo lineups a white and a black male confederate-customer who
had been in the store two hours earlier. The confederate-customers' visits to each store
were somewhat unusual and took 3 to 4 minutes, more time than average visits by real
customers. The visits took place 5 minutes apart. One customer asked for a pack of
cigarettes and paid with only pennies; the other customer carried a product up to the
counter, suddenly discovered that he did not seem to have enough money, fumbled in
his pocket for more change, started to the door, found enough change, came back to the
counter, and finally bought the product. After either interaction, the customer asked
directions to the airport, the bus station, the hospital, or a shopping mall, whichever
was fairly distant from the location of the store. The order in which the customers
24 BRIGHAM AND BOTHWELL

appeared, type of interaction they initiated, and locations to which directions were
asked, were randomly determined for each store visit.
Two hours after the customers' visits, two young white males dressed in suits
entered the store and described themselves to the clerk as "law interns" with a local
law firm. The "law interns" told the clerk that either of two people they were
interested in may have been in the store within the last 24 hours and then displayed two
sets of six photos each, one set of whites and one of blacks. Each set of photos included
a photo of the customer who had been in the store two hours earlier, as well as photos of
five other individuals who were quite similar to the customer in age, hairstyle, color-
ing, and general appearance. In the present survey, two questions followed this
scenario, asking respondents how many of the (white) clerks would have accurately
identified the white customer and how many would have accurately identified the black
customer.
The final section of the questionnaire was composed of three 7-point Likert items
which assessed the respondents' beliefs about several issues concerning eyewitness
identification in general: the impact of witness arousal on accuracy, the relationship
between an eyewitness's confidence and accuracy, and the value that jury members are
likely to put on eyewitness evidence as compared to other types of evidence. Subjects
were also asked to estimate the percentage of time that a positive identification from a
live or photo lineup by an eyewitness to an actual crime is likely to be correct. Finally,
demographic data were collected from the respondents; they were also asked how
many times they had served on a jury and whether eyewitness identification had been
an important part of the case(s) they observed.

RESULTS

The survey respondents' mean estimates of the eyewitness accuracy were com-
pared with the actual research findings by means of t-tests. The percentages of esti-
mates which were greater than or less than the actual research findings were also
calculated for each scenario. In response to the scenario describing the calculator theft,
the respondents significantly overestimated the percentage of students who were able
to correctly identify the confederate-thief It (89) = 20.72, p < .001]. Our respon-
dents estimated that 70.6% (S.D. = 26.6) of the students would have been able to
correctly identify the confederate-thief, but Wells and his associates found only 12.5 %
able to do so. Ninety-one percent of our respondents gave estimates that were greater
than 12.5%.
In response to the convenience-store scenario, voters again significantly overesti-
mated the percentage of white clerks who were able to correctly identify the target
person. This held both when the target person (customer) was black and when he was
white. The voters estimated that 51.1% (S.D. = 29.6) of the clerks exposed to the
black target person would have been able to correctly identify him out of the photo
lineup presented two hours later; actually only 32.3% of the clerks were able to make a
correct identification [t (87) = 5.94, p < .001]. Seventy percent of the respondents
gave estimates that were greater than the actual finding.
Turning to predicted recognition of the white customer, survey respondents
EYEWITNESS IDENTIFICATION 25

estimated that 68.9% (S.D. = 26.1) of the store clerks exposed to the white target
person would have been able to correctly identify him from the photo lineup. Brigham
et al. (1982) had found that only 31.3% of the clerks were able to make a correct
identification in this case [t (88) = 13.44, p < .001]. Ninety percent of our respon-
dents gave estimates that were greater than the actual finding. Contrary to several
earlier studies (e.g., see Brigham & Barkowitz, 1978), strong evidence of an "own-
race bias" had not been found in the convenience store study; there had been no
significant difference between the white clerks' ability to correctly identify black target
persons and white target persons. However, our survey respondents thought that the
white clerks should have been significantly more able to correctly identify the white
target person than the black target person [t (87) = 6.87, p < .0011.
In addition to being asked to estimate eyewitness identification hit rates in specific
situations, subjects were asked to estimate the percentage of eyewitness identifications
that are correct in general. Sixty-three percent indicated that they thought more than
50% of eyewitness identifications that are made are correct. Twenty-eight percent of
the respondents responded that about 50% are correct. Only 9% of our respondents felt
that less than 50% are correct. Note that the mean estimated hit rates for the three

I ~ Actual
Research
Results

Prospective
Jurors'
80% Mean Estimate

70.6% 68.9%
70% ~-

60% ~-
Percentage 51.t %
of 50% -
Eyewitnesses
Making 40% -
Correct 32.3% 31.3%
w / / i /

Identifications 30% - f /ll~


r V////
r ,q##//
r lJ If/Ill
r
20% - lr,~r'jjt rJ.r/j
I"/'11.~
r/,#/t v.//Z,4
10% _ 1 [ ~
v,cf/j

Staged
N_ Black
F."/A
White
__
Theft Customer's Customer's
(Leippe et.al, store v i s i t store visit
1978) (Brigham et.al, (Brigham et.al,
1982 ) 1982)
Fig. 1. Prospective jurors' accuracy in estimating research results.
26 BRIGHAM AND BOTHWELL

target-present lineups were all over 50% (70.6% for the calculator theft, 51.1% for
white clerks identifying black target persons, and 68.9% for white clerks identifying
white target persons) while actual accuracy rates found in the studies were all less than
35%.
The findings above indicate that most people overestimate actual research hit
rates. A related question concerns the effect of individual difference characteristics
(sex, education, juror experience, age) on ability to estimate the research findings.
None of these individual difference variables significantly affected mean estimated hit
rates for the first scenario. Sex, however, affected variability in estimating hit rates,
with females (S.D. = 30.5) significantly more variable in their estimates than males
[S.D. = 20.0; Fm~x (2,42) = 2.36, p < .05]. Looking at proportion of subjects who
over- or underestimated the hit rates, it was found that sex and juror experience
affected the proportion of subjects who overestimated the hit rate. None of the males
gave estimates equal to or less than the actual finding while 17% of the females did so
(z for proportions = 2.90, p < .005). Contrary to what might have been expected,
none of those with juror experience gave estimates equal to or less than the actual
finding whereas 12% of those without juror experience did so (z for proportions =
2.82, p < .005).
A sex difference occurred again on respondents' estimates of the percentage of
white clerks who were able to correctly identify the black target person. Males gave a
mean estimated hit rate of 56.4% whereas females gave an estimate of 42.4% [t (85) =
2.23, p < .05]. Only 20% of the males gave estimates equal to or less than the actual
finding, whereas 41% of the females did so (z for proportions = 2.27, p < .05). None
of the other individual difference variables (level of education, juror experience, age)
significantly affected mean estimated hit rates or percentage overestimating. None of
the individual difference variables significantly affected estimated accuracy rates or
percentage overestimating for the clerks' identification of the white customer.

Supplementary Data
Respondents also expressed beliefs about the effects of arousal and confidence on
eyewitness identifications and whether jury members placed too much or too little
value on eyewitness identification evidence. The majority of respondents (55%) felt
that an emotionally aroused witness is less likely to make an accurate identification,
but almost one-third of the respondents (31%) endorsed the opposite viewpoint. This
trend was prevalent regardless of sex, education, juror experience, or age. A majority
of the respondents (56%) suggested that the accuracy-confidence relationship is a
positive one, but a substantial minority (42%) felt that there is no consistent relation-
ship between witness confidence and accuracy. There were no significant differences
according to sex, juror experience, or age on this question, but more educated respon-
dents were somewhat more skeptical of a positive accuracy-confidence relationship
[7~2 (2) = 5.29, p < . 10]. Eighty percent of the respondents without any college
education, 60% of those with some college education, and 47% of those with college
degrees believed in a positive accuracy-confidence relationship.
The majority of respondents (58%) felt that jury members place too much empha-
sis on eyewitness identification evidence, but over one-third (37%) thought that jury
EYEWITNESS IDENTIFICATION 27

members place the right amount of emphasis on eyewitness identification. No signifi-


cant differences occurred by sex, juror experience, or age, but there was some tenden-
cy [Z2 (2) = 5.82, p < . 10] for more educated respondents to be more skeptical; 69%
of those with college degrees thought that jury members place too much emphasis on
eyewitness identification while only 43% of those without college degrees felt
likewise.

DISCUSSION

Wells (1978) has pointed out that many research psychologists have tended to
assume in the absence of data that jurors believe that eyewitnesses are less fallible than
they really are. The present study provides considerable empirical support for this
assumption: subjects significantly overestimated eyewitness identification hit rates for
all three situations which used target-present lineups.
Survey respondents' tendency to overestimate eyewitness hit rates in target-
present situations appears to be a generalized phenomenon. In response to the general
question about overall accuracy of eyewitness identifications, two-thirds of the respon-
dents thought that in general eyewitness identifications are correct more than 50% of
the time. This trend was prevalent regardless of sex, education, age, or juror expe-
rience. Some evidence of a sex effect was found; there was a tendency for males to
overestimate to a greater extent than females. This effect reached significance twice; in
the staged theft case and in the convenience-store situation where clerks were identify-
ing black customers. Although the failure of juror experience to attenuate the overesti-
mation effect may at first seem surprising, it becomes less so given the fact that only
11% of the respondents who had served on a jury said that eyewitness identification
had been an important part of the case(s) that they observed. Hence, fewer than 4% of
our total sample had had personal experience as jurors in cases involving disputed
eyewitness evidence. The majority of respondents, while showing a tendency to
overestimate the accuracy of eyewitness identifications, felt that jury members place
too much emphasis on eyewitness identifications.
A comparison of the survey respondents' beliefs about the accuracy of eyewitness
identifications with the views of law enforcement personnel, prosecuting attorneys,
and defense attorneys assessed in statewide surveys (Brigham, 1981; Brigham &
WolfsKeil, in press) suggests that the public tends to fall between prosecuting attor-
neys and defense attorneys in their views on such matters. Law enforcement personnel
tend to be more similar to prosecuting attorneys than to defense attorneys in their
beliefs. For example, 40% of the defense attorneys, 56% of the present respondents,
73% of the law enforcement personnel, and 75% of the prosecuting attorneys thought
there was a positive relationship between certainty and accuracy in eyewitness identifi-
cations. In a similar vein, 21% of the law enforcement personnel, 24% of the prosecut-
ing attorneys, 58% of the present respondents, and 89% of the defense attorneys
thought that jury members place too much emphasis on eyewitness identification.
However, 31% of the prosecuting attorneys, 55% of the present respondents, 67% of
the law enforcement personnel, and 81% of the defense attorneys thought that emo-
tionally aroused eyewitnesses are less likely to be accurate than nonaroused eyewit-
28 BRIGHAM AND BOTHWELL

nesses. These data suggest that prejudices underlying common beliefs about the
accuracy of eyewitness identification effect the way that eyewitness testimony is
interpreted; this may seriously aggrevate the problem of erroneous identification in the
criminal justice system.
Woocher (1977, p. 1015) has pointed out that "the theory upon which expert
testimony is excepted from the opinion evidence rule in that such testimony serves to
inform the jury about affairs not within the full understanding of the average man." If
one makes the value judgment, as we do, that justice is best served when jurors'
weighing and evaluation of evidence is closely tied to the validity of that evidence,
then the present results are troubling. On the average, more than 80% of prospective
jurors overestimated the likely accuracy of eyewitness evidence in the three incidents
we described to them. The mean overestimation percentage (83.7%) found in the
present study might have been considerably different had other studies been described
to the subjects. However, some evidence as to the generality of these findings can be
derived from confidence intervals. The 95% confidence interval for the present data
ranges from 54.3% to 100%, yielding support for the oft-cited (and until now, empiri-
cally unsupported) generalization that "most" jury members will overestimate the
likelihood that eyewitness evidence will be accurate.
If we consider the 83.7% mean overestimation percentage found here as a rough
estimate of the general likelihood of overestimation, the implications for the legal
system are sobering. Some have argued that in cases involving disputed eyewitness
evidence defense attorneys could somehow screen out prospective jurors who are
especially overbelieving during the voir dire (jury selection) process. Demographic
characteristics of jurors would not seem to be of much value in such a screening since
the present data indicate that education, previous juror experience, and age were not
systematically related to the tendency to overestimate. Sex showed some effect; males
were somewhat more likely than females to overestimate in 2 of the 3 situations. Given
the base rate probability of .837 overestimation, the likelihood of randomly getting a
jury where 6 (or more) of the 12 members do not overbelieve eyewitness evidence
would be less than 1 in 100 (p = .007). In fact, there is a 69% likelihood that at least 10
of the 12 randomly selected jurors would be overestimators. Thus, even if a foolproof
criterion for identifying "overestimators" were found, and if an attorney chose to
utilize some or all of the peremptory challenges for this purpose, the attorney's task
would still be monumental. If an attorney used 5 challenges to eliminate 5 overestima-
tors from the jury, the likelihood of ending up with a jury wherein 6 or more of the 12
members would not be overestimators is still less than 5% (p = .046). If an attorney
used 10 challenges in this manner, the probability of ending up with a 12-person jury
where half or more members are not overestimators would be slightly less than 14%.
Even if an attorney was allotted 20 challenges and used them all in this manner,
chances are less than 50% (p = .425) that the resulting jury would contain 6 or more
members who did not overestimate. Hence, the voir dire procedure alone does not
appear sufficient to enable attorneys to create juries which will evaluate eyewitness
evidence appropriately.
It should be noted that the three situations described in the present study involved
target-present lineups. It remains an open question as to how accurate prospective
jurors would be in predicting eyewitnesses' ability to realize that the criminal is not
EYEWITNESS IDENTIFICATION 29

present (i.e., make a correct rejection response) in a target-absent lineup. We did not
analyze research examples using target-absent lineups in the present study for two
reasons. There have been relatively few studies which have utilized target-absent
lineups (cf., Malpass & Devine, 1981) in situations of high mundane realism. Further,
it was felt that the use of six scenarios would be likely to have a strong negative impact
on the questionnaire return rate. However, research which utilizes such target-absent
scenarios would be of considerable value. It should be noted that in actual cases
involving disputed eyewitness identifications the prosecution is likely to assume that a
target (criminal)-present lineup has been used while the defense asserts that a target
(criminal)-absent linup was employed.
In conclusion, the present data refute the claim that expert psychological testimo-
ny on eyewitness identifications would not tell the jury members anything they do not
already know. Not only do jury members overestimate the accuracy of eyewitness
identifications in target-present lineups, they also appear unaware, to some extent, of
the sources of error associated with this type of evidence. Therefore, the present data
indicate that the testimony of an expert on these matters would not invade the province
of the jury. Rather, such testimony would aid the jury in its evaluation of evidence and
would thereby further the cause of justice.

REFERENCES

Bartlett, F. C. Remembering: A Study in Experimental and Social Psychology. London: Cambridge Univeisity
Press, 1932.
Borchard, E. M. Convicting the Innocent. Hamden, Connecticut: Archon Books, t932.
Bransford, J. D. Human Cognition: Learning, Understanding, and Remembering. Belmont, California: Wads-
worth Publishing Co., 1979.
Brigham, J. C. The accuracy of eyewitness evidence: How do attorneys see it? Florida Bar Journal, I981,
55(10), 714--721.
Brigham, J. C., & Barkowitz, P. Do "They all look alike?" The effect of race, sex, experience, and attitudes on
the ability to recognize faces. Journal of Applied Social Psychology, 1978, 8, 306-318.
Brigham, J. C., Maass, A., Snyder, L. D., & Spaulding, K. The accuracy of eyewitness identifications in a field
setting. Journal of Personality and Social Psychology, 1982, 42, 673-681.
Brigham, J. C., & WolfsKeil, M. P. Opinions of attorneys and law enforcement personnel on the accuracy of
eyewitness identifications. Law and Human Behavior, in press.
Bruner, J. Social psychology and perception. In E. Maccoby, T. Newcomb, & E. Hartley, Readings in social
psychology (3rd edition). New York: Holt, 1958.
Buckhout, R. Eyewitness testimony. Scientific American, 1974, 231(6), 23-31.
Buckhout, R. Nearly 2000 witnesses can be wrong. Social Action and the Law Newsletter, 1975, 2(3), 7.
Deffenbacher, K. Eyewitness accuracy and confidence. Law and Human Behavior, 1980, 4, 243-260.
Deffenbacher, K. A., & Loftus, E. F. Do jurors share a common understanding concerning eyewitness behavior?
Law and Human Behavior, 1982, 6, 15-30.
Frank, J., & Frank, B. Not Guilty. New York: Doubleday, 1957.
Frankfurter, F. The Case of Sacco and Vanzetti: A Critical Analysis for Lawyers and Laymen. Boston: Little,
Brown, 1927.
Gardner, D. S. The perception and memory of witnesses. Cornell Law Review, 1933, 391-401.
Houts, M. From Evidence to Proof. Springfield, Illinois: Charles C. Thomas, 1956.
Katz, L. S., & Reid, J. F. Expert testimony on the fallibility of eyewitness identification. Criminal Justice
Journal, 1977, 1, 17%206.
Leippe, M. R. Effects of integrative memorial and cognitive processes on the correspondence of eyewitness
accuracy and confidence. Law and Human Behavior, 1980, 4, 261-274.
30 BRIGHAM AND BOTHWELL

Leippe, M. R., Wells, G. L., & Ostrom, T. M. Crime seriousness as a determinant of accuracy in eyewitness
identification. Journal of Applied Psychology, 1978, 63, 345-351.
Loftus, E. F., & Palmer, J. C. Recognition of automobile destruction: An example of the interaction between
language and memory. Journal of Verbal Learning and Verbal Behavior, 1974, 13, 585-589.
Loftus, E. F. Eyewitness testimony: Does the malleable human memory interfere with legal justice? SocialAction
and The Law Newsletter, 1975, 2, 5ft.
Loftus, E. F. Eyewitness Testimony. Cambridge, Massachusetts: Harvard University Press, 1979.
Malpass, R. S., & Devine, P. G. Eyewitness identification: Lineup instructions and the absence of the offender.
Journal of Applied Psychology, 1981, 66, 482-489.
Marshall, J. Law and Psychology in Conflict. New York: Anchor, 1966.
Miller, G. A. The magical number seven, plus or minus two: Some limits on our capacity for processing
information. Psychological Review, 1956, 63, 81-97.
Muensterberg, H. On the Witness Stand: Essays on Psychology and Crime. New York: Doubleday, 1912.
Neisser, U. Cognitive Psychology. New York: Appleton-Century Crofts, 1967.
Sobel, N. R. Eyewitness Identification: Legal and Practical Problems. New York: Clark Boardman, 1972.
Wall, P. M. Eyewitness Identification in Criminal Cases. Springfield, Illinois: Charles C. Thomas, 1965.
Wells, G. L. Applied eyewitness-testimony research: System variables and estimator variables. Journal of
Personality and Social Psychology, 1978, 36, 1546-1557.
Wells, G. L. How adequate is human intuition for judging eyewitness testimony? In Wells, G. L., & Loftus, E.
F., (Eds.), Eyewitness testimony: Psychologicalperspectives. New York: Cambridge University Press, in
press.
Wilson, C. R. Psychological opinions on the accuracy of eyewitness testimony. The Judges Journal, 1975, 14~
72-74.
Yarmey, A. D., & Jones, H. P. T. Is eyewitness identification a matter of common-sense? In B. R. Clifford & S.
Lloyd-Bostock (Eds,), Evaluating Witness Research: Recent Psychological Research and New Perspec-
tives. New York: Wiley, in press.

Das könnte Ihnen auch gefallen