Content uploaded by Christian A. Meissner
Author content
All content in this area was uploaded by Christian A. Meissner on Nov 06, 2018
Content may be subject to copyright.
This article was downloaded by: [Iowa State University]
On: 28 May 2014, At: 10:07
Publisher: Routledge
Informa Ltd Registered in England and Wales Registered Number: 1072954
Registered office: Mortimer House, 37-41 Mortimer Street, London W1T 3JH,
UK
Visual Cognition
Publication details, including instructions for authors
and subscription information:
http://www.tandfonline.com/loi/pvis20
Can I see your passport please?
Perceptual discrimination of
own- and other-race faces
Christian A. Meissnera, Kyle J. Susaa & Amy B. Rossa
a Department of Psychology, University of Texas at El
Paso, El Paso, TX, USA
Published online: 20 Sep 2013.
To cite this article: Christian A. Meissner, Kyle J. Susa & Amy B. Ross (2013) Can I see
your passport please? Perceptual discrimination of own- and other-race faces, Visual
Cognition, 21:9-10, 1287-1305, DOI: 10.1080/13506285.2013.832451
To link to this article: http://dx.doi.org/10.1080/13506285.2013.832451
PLEASE SCROLL DOWN FOR ARTICLE
Taylor & Francis makes every effort to ensure the accuracy of all the
information (the “Content”) contained in the publications on our platform.
However, Taylor & Francis, our agents, and our licensors make no
representations or warranties whatsoever as to the accuracy, completeness, or
suitability for any purpose of the Content. Any opinions and views expressed
in this publication are the opinions and views of the authors, and are not the
views of or endorsed by Taylor & Francis. The accuracy of the Content should
not be relied upon and should be independently verified with primary sources
of information. Taylor and Francis shall not be liable for any losses, actions,
claims, proceedings, demands, costs, expenses, damages, and other liabilities
whatsoever or howsoever caused arising directly or indirectly in connection
with, in relation to or arising out of the use of the Content.
This article may be used for research, teaching, and private study purposes.
Any substantial or systematic reproduction, redistribution, reselling, loan, sub-
licensing, systematic supply, or distribution in any form to anyone is expressly
Can I see your passport please? Perceptual
discrimination of own- and other-race faces
Christian A. Meissner, Kyle J. Susa, and Amy B. Ross
Department of Psychology, University of Texas at El Paso, El Paso,
TX, USA
(Received 5 September 2013; accepted 22 July 2013)
Psychological research has consistently demonstrated that individuals are better at
discriminating faces of their own race when compared with faces of another, less
familiar race. Given the racial/ethnic diversity of individuals screened by security
personnel at transportation and border checkpoints, it is important to understand
whether the cross-race effect may play a role in simultaneous perceptual discrimina-
tion tasks that mimic such screening operations. Three experiments assessed the
deleterious effects of cross-racial identification in this context. Results demonstrated
greater discrimination accuracy for own- versus other-race faces, and a propensity for
screeners to be overconfident in their decisions, particularly for other-race persons.
Further, perceived age differences between the target and his identification photo
and the use of a disguise were found to moderate cross-race effects during this task.
Keywords: Perceptual discrimination; Cross-race effect; Feature recruitment.
A daily challenge for travel document screeners is to assess whether
individuals entering a country are providing correct documentation of their
identity. In fact, thousands of times each day screeners at airports and
border checkpoints are tasked with determining and confirming identity in a
high-stakes environment. On a typical day in 2012, US Customs and Border
Please address all correspondence to Christian A. Meissner, Department of Psychology,
University of Texas at El Paso, 500 W. University Avenue, El Paso, TX 79968, USA. E-mail:
cmeissner@utep.edu
The first two authors contributed equally to the current manuscript. This material is based
upon research supported by the US Department of Homeland Security under Award Number
2008-ST-061-BS0001. The views and conclusions contained in this document are those of the
authors and should not be interpreted as representing the official policies, either expressed or
implied, of the U.S. Department of Homeland Security.
Visual Cognition, 2013
Vol. 21, Nos. 9 10, 12871305, http://dx.doi.org/10.1080/13506285.2013.832451
#2013 Taylor & Francis
Downloaded by [Iowa State University] at 10:07 28 May 2014
Protection admitted 963,121 passengers and pedestrians. At the same time,
they reported identifying an average of 66 fraudulent documents (http://
www.cbp.gov); however, an unknown number of falsified documents go
undetected. Some of these fraudulent documents consist of individuals who
are using a valid passport or identification card of another, similar looking
person*these are commonly referred to as imposters. Although inspection
for imposters may be challenging under the best of circumstances,
psychological research has demonstrated that identification of and discri-
mination between individuals of another, less familiar race or ethnicity is
significantly more difficult when compared with performance on faces of
one’s own race or ethnicity (Meissner & Brigham, 2001). Known as the
cross-race effect (CRE), this phenomenon has been the subject of more than
40 years of research and has been demonstrated in a variety of experimental
and forensic contexts (see Brigham, Bennett, Meissner, & Mitchell, 2007;
Tanaka, Heptonstall, & Hagen, this issue 2013). Given the relevance of this
phenomenon to national security issues, the current studies sought to better
understand the influence of the CRE in the seemingly simplistic task
presented to travel document screeners thousands of times each day. In
particular, our studies investigated the influence of the CRE in a perceptual
discrimination paradigm that mimics the travel document screening task by
assessing the phenomenology associated with the effect, and by examining
how variation in characteristics of the target person or identification image
tendered might moderate the CRE.
Studies investigating the CRE from both basic ‘‘recognition memory’’
and applied ‘‘eyewitness identification’’ paradigms have demonstrated a
consistent advantage in the recognition/identification of own-race faces when
compared with memory for faces of another race (Meissner & Brigham,
2001), and archival analyses have demonstrated the importance of the CRE
effect in real cases of eyewitness identification (Behrman & Davey, 2001).
Furthermore, researchers have documented this phenomenon both within
the US using Caucasian, African-American, Hispanic, and Asian-American
participants, and cross-nationally with participants from Great Britain,
South Africa, Germany, Turkey, and various nations in the Middle East (for
reviews, see Meissner & Brigham, 2001; Sporer, 2001).
A number of studies have examined the variety of social, cognitive, and
perceptual factors that influence the CRE, yet the precise mechanisms at
play remain a matter of debate. In a recent theoretical review, Hugenberg,
Young, Bernstein, and Sacco (2010) proposed a categorization-individuation
model as a framework for understanding the CRE. This model accounts for
a long history of empirical studies demonstrating the importance of
perceptual learning and social-cognitive processes in explaining the super-
iority of own- versus other-race recognition. Namely, their integrative model
suggests that the CRE results from the interplay of social categorization,
1288 MEISSNER, SUSA, ROSS
Downloaded by [Iowa State University] at 10:07 28 May 2014
perceiver motivation, and perceiver experience with other-race faces. Social
categorization processes are the core of the model, suggesting that the CRE
occurs as individuals attend to ‘‘identity’’ characteristics of own-race/
ethnicity faces, whereas ‘‘category’’ characteristics capture attention for
other-race faces (Hugenberg et al., 2010; see also Hugenberg, Wilson, See, &
Young, this issue 2013). In a path analysis comparing the relative
contribution of processes underlying the CRE, Susa, Meissner, and de
Heer (2010) observed that both social categorization and higher-level
cognitive processes each play a role in mediating the CRE in recognition
memory. These processes also appear to be moderated by the interracial
experience of perceivers and their motivation to individuate other-race faces.
The effects of individuating interracial experience in reducing the CRE have
been examined in numerous studies. For example, Bukach, Cottle, Ubiwa,
and Miller (2012) demonstrated that greater self-reported individuating
experience was associated with a reduced CRE in holistic face processing.
Hancock and Rhodes (2008) also demonstrated that greater self-reported
interracial contact was associated with a reduction in the CRE in face
recognition.
Several other studies have demonstrated the importance of perceptual
learning and cognitive processing underlying the CRE (Marcon, Susa, &
Meissner, 2009; Meissner, Brigham, & Butz, 2005). These studies suggest
that one’s perceptual experience with a given race leads to the development
of feature recruitment strategies and representational structures (i.e.,
encoding strategies) that facilitate perceptual discrimination and memory.
For example, studies by Furl, Phillips, and O’Toole (2002) demonstrated that
experience with other-race faces led to a feature set that optimized the
encoding of information (Caldara & Abdi, 2006; see also Anzures, Quinn,
Pascalis, Siater, & Lee, this issue 2013; O’Toole & Natu, this issue 2013;
Tanaka et al., this issue 2013).
Social categorization and higher-level representational structures are also
believed to underlie the perceptual discrimination of faces, the task that most
closely resembles the operational duties of travel document screeners who
examine passports or identification cards to assess whether they match the
tendering individual. To date, only a handful of studies have documented the
CRE using some variant of a ‘‘perceptual discrimination’’ paradigm. This
research includes studies that have assessed the sequential presentation of
stimuli in a perceptual identification or visual search task (Lindsay, Jack, &
Christian, 1991; Marcon, Meissner, Freuh, Susa, & MacLin, 2010; Walker &
Hewstone, 2006; Walker & Tanaka, 2003), as well as a simultaneous face-
matching task (Megreya, White, & Burton, 2011; Sporer, Trinkl, &
Guberova, 2007). For example, a study by Lindsay et al. (1991) observed
the CRE in a sequential presentation perceptual discrimination task. Over
the course of 50 trials, a photograph of a face was tachistoscopically
CAN I SEE YOUR PASSPORT PLEASE? 1289
Downloaded by [Iowa State University] at 10:07 28 May 2014
presented for 120 ms, followed by the previously presented face and a
matched foil. Results of this study demonstrated that the CRE was present
such that White participants performed significantly worse on trials
involving African-American faces relative to trials involving White faces.
A more recent series of studies conducted by Marcon et al. (2010)
demonstrated a robust CRE using a perceptual discrimination paradigm
that involved a sequential ‘‘visual search’’ task requiring participants to
identify a person among a group of photographs. Finally, Sporer et al. (2007)
also demonstrated a strong CRE using the simultaneous presentation of
stimuli in a visual search task. In their study, Turkish children responded
more quickly than Austrian children when matching Turkish faces.
Given the practical importance of identifying known or suspected
imposters and concurrently minimizing the inconvenience of ‘‘misidentifica-
tion’’ for innocent persons, it is important to document whether the CRE
may play a role in perceptual discrimination tasks that mimic travel
document screening procedures. Most prior studies have demonstrated the
CRE in perceptual discrimination using paradigms that sequentially present
stimuli; two studies have used a simultaneous presentation procedure in the
context of a visual search task (Megreya et al., 2011; Sporer et al., 2007). The
current study is the first to explore the simultaneous presentation of a single
target and matched foil in a manner that models the inspection process of
travel document screeners.
One interesting aspect of perceptual face discrimination worth noting is
the surprising inaccuracy and overconfidence demonstrated by participants
in such tasks. Using own-race faces, Kemp, Towell, and Pike (1997)
conducted a perceptual identification experiment in a real-world setting to
assess how well supermarket cashiers could correctly link photographs on
credit cards with customers. Kemp et al. found that cashiers’ decisions to
accept or reject the cards were correct only 67.4% of the time, suggesting that
correctly identifying own-race faces proved to be a difficult task. This
experiment documented the existence of the difficulty of face matching
within a perceptual framework; the current studies sought to extend the
findings to include own- and other-race faces and to assess the relationship
between screener confidence and accuracy.
If a CRE exists within the passport screening context, might certain
factors exacerbate the likelihood of observing such an effect? Prior research
by Marcon et al. (2010) suggests that as the difficulty of feature recruitment
increases (either by taxing memory processes at encoding or retention, or
increasing perceptual confusion at test), the CRE will be more likely in a
perceptual discrimination task. In the current studies we assess whether
certain stimulus factors that challenge feature recruitment might moderate
the CRE in the document screening context. Three experiments were
conducted. Experiment 1 examined the nature of the CRE during this
1290 MEISSNER, SUSA, ROSS
Downloaded by [Iowa State University] at 10:07 28 May 2014
task and also assessed the relationship between screener confidence and
accuracy. Experiments 2 and 3 assessed factors that may moderate the CRE
during this task, including the perceived age difference between the target
person and the identification image tendered (Experiment 2) and the impact
of extraneous paraphernalia or attempts to disguise identity (i.e., hat and
sunglasses; Experiment 3).
GENERAL METHOD
Overview
The current experiments sought to simulate the travel document inspection
task experienced by security screeners in a controlled laboratory setting. All
experiments used a similar methodology involving the simultaneous
presentation of a photograph of a target person and a scanned passport
image either from the target (i.e., matched trials) or from another individual
(i.e., mismatched foils; see Kemp et al., 1997). Across 80 trials, participants
were asked to respond as quickly, but as accurately, as possible in
determining whether the two images were of the same person (or whether
they represented different individuals). Each trial remained in-view until the
participant responded, and participants were unrestricted in their response
times.
Given prior research on the CRE using perceptual discrimination
paradigms (e.g., Lindsay et al., 1991; Marcon et al., 2010; Megreya et al.,
2011; Sporer et al., 2007) and studies such as Kemp et al. (1997)
demonstrating surprising inaccuracy in a face identification paradigm
similar to that of the document screening context, we predicted that the
CRE would be evidenced in participants’ performance in a simultaneous
perceptual discrimination task, and that participants would be significantly
less calibrated in their estimates of confidence when identifying other-race
(relative to own-race) face pairings. Further, we predicted that conditions
that challenge feature recruitment processes would be more likely to evidence
the CRE, including the use of aged photographs and disguises by a traveller.
Stimuli
All trials were presented on computers using MediaLab software. Instruc-
tions and images were displayed on 20-inch LCD monitors at 1280 1024
resolution. Face images were five inches high on the screen with avisual angle
subtending 11 degrees. Approximately 100 African-American faces and 100
Mexican-American faces (and their corresponding passport photos or
identification cards) of college-aged males were used as stimuli across the
three experiments. Our stimuli consisted of two photographic images that
CAN I SEE YOUR PASSPORT PLEASE? 1291
Downloaded by [Iowa State University] at 10:07 28 May 2014
were gathered from university students. One image was a photograph taken of
a student under standardized conditions, where the student was smiling and
wearing everyday clothing. The second image involved a scan of the student’s
passport or identification photograph. For mismatched trials, we paired the
photograph of the student with a scanned passport photograph of a high
similarity other (a ‘‘matched foil’’). To select matched foils, a small sample of
Mexican-American (n3) and African-American (n3) participants were
recruited to provide ratings of confusability (1 ‘‘not confusable’’, 7
‘‘very confusable’’) to pairs of faces within our database. Matched foils
showed no significant differences in ratings of confusability by Mexican-
American and African-American participants, t0.81, ns.Figure 1 provides
an example of a Mexican-American matched foil trial.
Participants
Participants included Mexican-American individuals recruited through the
University of Texas at El Paso who received course credit for their
participation. The university student population (which is reflective of the
Paso del Norte region) is composed primarily (83%) of Mexican-Americans
and Mexican nationals. Less than 3% of the population is African-
American, suggesting that our participants’ perceptual experience for this
sample of faces will be quite minimal. Unfortunately, such a small African-
American population also limited our ability to recruit a comparison sample
in the present studies, although the CRE literature suggests rather robust
effects that are conditioned more on prior perceptual experience than on
Figure 1. Example of a Mexican-American matched foil trial. To view this figure in colour, please see
the online issue of the Journal.
1292 MEISSNER, SUSA, ROSS
Downloaded by [Iowa State University] at 10:07 28 May 2014
racial or ethnic featural variation or physiognomy (see Meissner & Brigham,
2001). The potential limitation associated with our sampling is further
addressed in the General Discussion.
Procedure
Participants were presented with a series of simultaneous perceptual
discrimination trials in which they attempted to determine whether the
identification image tendered by a particular individual matched that
individual. A total of 80 trials were completed by participants in each
experiment. Across trials, we systematically varied the race of the target
individual, as well as whether the target and identification image were of the
same individual or that of a matched foil (two different people). Participants
responded on each trial by pressing a key to signal that the target and
identification image matched or did not match. Response keys were
counterbalanced across participants. Participants were instructed to respond
on each trial as quickly, but as accurately, as possible. In Experiment 1, each
trial was followed by a confidence rating, in which participants estimated
their confidence on a scale from 50% to 100% (with increments of 5%).
Following the confidence rating (for Experiment 1) or after each trial (for
Experiments 2 and 3), participants viewed a 2000 ms pattern mask prior to
the subsequent trial. Following 10 consecutive trials, participants received a
1 min rest period before continuing with the task.
Data analysis
Accuracy of response was assessed across trials. In Experiment 1, the
relationship between confidence and accuracy was assessed using estimates
of calibration, over/under confidence, and resolution (see Jonsson &
Allwood, 2003; Juslin, Olsson, & Winman, 1996; Krug, 2007; Olsson,
2000). Calibration reflects the agreement between confidence level and
accuracy level, where perfect calibration occurs when subjective confidence
levels match objective accuracy levels. The formula for computing calibra-
tion was identified by Jonsson and Allwood (2003) as follows:
calibration ¼1
nXt
t¼1ntrtm ct
ðÞ
2;
where nis the total number of questions answered, tis the number of
confidence classes used, c
t
is the proportion of correct answers for all items
in the confidence class r
t
,n
t
is the number of times the confidence class r
t
was
used, and r
tm
is the mean of the confidence ratings in confidence class r
t
.
CAN I SEE YOUR PASSPORT PLEASE? 1293
Downloaded by [Iowa State University] at 10:07 28 May 2014
Over/underconfidence reflects the confidence expressed in participants’
decisions relative to their identification accuracy. Perfect calibration and no
over/underconfidence are found when the respective equations equal zero.
Resolution reflects the ability of participants to distinguish between
occasions when the target is the same versus a matched foil within each
confidence class. In other words, within each confidence class high resolution
occurs when greater subjective confidence is assigned to correct answers,
whereas a lower confidence rating is assigned to incorrect answers.
Resolution is computed as follows:
Resolution ¼1
nXt
t¼1ntctcðÞ
2;
where cis the proportion of all items for which the correct alternative was
selected.
To achieve maximal resolution, a participant must assign lower con-
fidence estimates to all questions answered incorrectly compared with the
questions answered correctly within each confidence class.
Identification accuracy for all experiments was measured by separating
the proportion of correct identifications on target trials (‘‘hits’’) and the
proportion of false identifications on matched foil trials (‘‘false alarms’’)
across the experimental manipulations. These estimates were then used to
compute discrimination accuracy (A
z
):
Az¼/d0
ffiffiffi
2
p
;
where
d0¼ZHZFA;
and response criterion (c):
c¼0:5zFA þzH
ðÞ:
Inferential statistical tests were used to assess perceptual discrimination
performance across the signal detection measures of accuracy and con-
fidence measures as a product of the experimental manipulations. Analyses
will focus on the composite measures of discrimination accuracy and
response criterion.
EXPERIMENT 1
The purpose of Experiment 1 was to document the nature of the CRE in a
simultaneous perceptual discrimination task and to determine the relation-
ship between confidence and accuracy for both own- and other-race face
1294 MEISSNER, SUSA, ROSS
Downloaded by [Iowa State University] at 10:07 28 May 2014
pairings (cf. Juslin et al., 1996; Weber & Brewer, 2003). We predicted that the
CRE would be evidenced in this perceptual discrimination task such that
participants would demonstrate better identification performance for own-
vs. other-race faces. We also predicted that participants would be better
calibrated and less overconfident when responding to own- versus other-race
faces.
Method
Participants
.Fifty Mexican-American undergraduate students (27
females; mean age 19.88 years) from the University of Texas at El Paso
participated in this experiment.
Design
.A one-way repeated measures design was used to assess the
influence of race of face (Mexican-American vs. African-American) on
signal detection estimates of discrimination accuracy and response criterion.
As described earlier, measures of confidence were also assessed for each face
pairing, and these ratings were used to estimate calibration, over/under
confidence, and resolution (see Jonsson & Allwood, 2003).
Stimulus materials and procedure
.The stimulus materials and
procedure used for Experiment 1 are consistent with that described in the
General Method section.
Results and discussion
Table 1 summarizes participants’ performance across the signal detection
and confidence/accuracy measures. Analysis of discrimination accuracy (Az)
indicated a significant main effect for race of face such that performance was
greater for own- relative to other-race face pairings, t(49) 5.35, pB.001,
d0.76. There was no significant difference between own- and other-race
TABLE 1
Descriptive Statistics for Identification Performance and Confidence in Experiment 1
Own-Race Other-Race
M SD 95% CI M SD 95% CI
Disc. Accuracy (Az) .83 .11 [.80, .86] .72 .13 [.68, .76]
Response Bias (C).23 .40 [.34, .12] .14 .34 [.23, .05]
Calibration .05 .04 [.04, .06] .07 .05 [.06, .09]
Over/Under .10 .12 [.07, .14] .18 .14 [.14, .21]
Resolution .03 .02 [.02, .03] .03 .02 [.02, .03]
Note.CIconfidence interval of the mean.
CAN I SEE YOUR PASSPORT PLEASE? 1295
Downloaded by [Iowa State University] at 10:07 28 May 2014
face pairings with regard to response criterion, t(49) 1.12, p.27, d
0.16.
Calibration curves for own- and other-race face pairings can be found in
Figure 2. Consistent with predictions, analyses indicated that participants
were significantly better calibrated for own- versus other-race face pairings,
t(49) 4.08, pB.001, d0.57, and significantly more overconfident for
other- versus own-race face pairings, t(49) 5.97, pB.001, d0.85. In
contrast, there was no significant difference between own- and other-race
face pairings with regard to resolution, t(49) .79, p.44, d0.15.
Overall, the results suggest that the CRE is prevalent in a simultaneous
perceptual discrimination task that mimics the travel document inspection
process performed by security screeners on a daily basis. Although this task
may appear quite simple, caution should be taken in using confidence as an
indication of accuracy, particularly for other-race faces.
EXPERIMENT 2
In the United States (and in many other countries), passports are valid for 10
years for adults over the age of 16. For children 15 years or younger,
passports are valid for 5 years. Several states have even longer expiration
dates for other forms of identification; for example, the state of Arizona
affords a 12-year expiration to driver’s licence photographs. Given that an
individual’s appearance can change dramatically over the course of decade,
Experiment 2 investigated whether the CRE observed in the travel document
Figure 2. Calibration curves representing the confidence/accuracy relationship for own-race and
other-race face pairings.
1296 MEISSNER, SUSA, ROSS
Downloaded by [Iowa State University] at 10:07 28 May 2014
inspection task might be moderated by the perceived age difference between
the target person and his identification image. Namely, Experiment 2
assessed participants’ performance on a simultaneous perceptual identifica-
tion task involving own-race and other-race persons, and manipulated the
perceived age of the identification photo. Consistent with Experiment 1, we
predicted a main effect for race, such that identification performance for
own-race face pairings would be greater than other-race face pairings. We
also predicted a main effect for perceived age, such that identification
performance would be better for ‘‘recent’’ age pairings relative to ‘‘distant’’
age pairings. Finally, we predicted that age differences between the target
person and the identification image would moderate identification perfor-
mance, such that when the identification image age was ‘‘distant’’ there
would be larger cross-race differences in identification performance.
Method
Participants
.One hundred Mexican-American undergraduate students
(69 female; mean age 20.17 years) at the University of Texas at El Paso
participated in this study.
Design
.A22 repeated measures factorial design was used to assess
the influence of race (own- vs. other-race) and the perceived age difference
between matched images (recent age vs. distant age) on identification
performance. Target (same person) and matched foils (different people)
trials were included to obtain estimates of discrimination accuracy and
response criterion across each manipulation.
Stimulus materials and procedure
.This experiment followed the
basic procedure outlined in the General Method section. However, unique to
this experiment was the perceived age difference manipulation of the target
pairings that were perceived to be either recent in age or distant in age (see
Figure 3). Unfortunately, we were unable to determine the precise age of the
passport or identification photographs that were collected; thus, we
evaluated perceived age differences by asking 10 participants (five Mex-
ican-American and five African-American) to rate the age differences
between 50 pairings of matched stimuli. Prior studies suggest that perceptual
estimates of age are a good approximation of chronological age (Burt &
Perrett, 1995). Participants focused only on same-race pairings and were
unaware of the purpose of the study. Age difference ratings were then
averaged across the raters of each race. The 10 pairings of each race that
were rated to be most ‘‘recent’’ and the 10 pairings that were rated to be
most ‘‘distant’’ for each race of face were used in the experiment as our
manipulation. The average perceived age of all of the ‘‘recent’’ pairings was
CAN I SEE YOUR PASSPORT PLEASE? 1297
Downloaded by [Iowa State University] at 10:07 28 May 2014
8 months, and the average perceived age of all ‘‘distant’’ parings was 28
months.
Results and discussion
A22 repeated measures ANOVAwas conducted to assess the influence of
the race of face manipulation (own- vs. other-race) and the perceived age
difference between target images (recent age vs. distant age) on identification
performance. Table 2 summarizes participants’ performance across each
measure. Since only ‘‘matched’’ (same person) face pairings received the
perceived age difference manipulation, the findings for this effect in
discrimination and response criterion are largely driven by ‘‘hit’’ rates.
Figure 3. Example of a Mexican-American distant age trial. To view this figure in colour, please see
the online issue of the Journal.
TABLE 2
Descriptive statistics for identification performance in Experiment 2
Own-race Other-race
M SD 95% CI M SD 95% CI
Recent age
Disc. accuracy (Az) 0.82 0.13 [.79, .85] 0.82 0.13 [.79, .85]
Response bias (C) 0.23 0.50 [.33, .13] 0.41 0.54 [.52, .30]
Distant age
Disc. accuracy (Az) 0.66 0.17 [.63, .69] 0.59 0.16 [.56, .62]
Response bias (C) 0.22 0.46 [.13, .31] 0.19 0.47 [.10, .28]
CI confidence interval of the mean.
1298 MEISSNER, SUSA, ROSS
Downloaded by [Iowa State University] at 10:07 28 May 2014
With regard to discrimination accuracy, results indicated a significant
main effect for race of face in the predicted direction. Namely, identification
performance was significantly greater for own-race (relative to other-race)
face pairings, F(1, 99) 4.56, p.035, h
p
2
.04. As predicted, results also
indicated a main effect for perceived age, as identification performance was
significantly greater for ‘‘recent’’ age pairings compared to ‘‘distant’’ age
pairings, F(1, 99) 304.731, pB.001, h
p
2
.76. Finally, these main effects
were qualified by a significant Race Perceived age interaction, F(1, 99)
8.57, p.004, h
p
2
.08. Simple effects analyses indicated no significant
CRE for identification images that were perceived to be recent in age, t(99)
0.41, p.68, d0.04; however, when faces were perceived to be distant
in age, participants demonstrated significantly greater identification perfor-
mance for own- versus other-race faces, t(99) 2.83, p .006, d 0.30.
With regard to response criterion, results indicated a significant main
effect for race, such that participants were more conservative in their
judgements towards own- versus other-race face pairings, F(1, 99) 5.61, p
.02, h
p
2
.05. There was also a significant main effect for perceived age,
indicating that participants were more conservative in their judgements for
face pairings that were distant in age, relative to those recent in age, F(1, 99)
303.06, pB.001, h
p
2
.75. These main effects were qualified by a
significant interaction, F(1, 99) 9.51, p.003, h
p
2
.09. Simple effects
analyses indicated that when the identification image was perceived to be
recent in age, participants were significantly more conservative in their
judgements (more likely to say there was not a match) for own- versus other-
race face pairings, t(99) 3.26, pB.002, d0.33. However, when an
identification image was perceived to be distant in age, there was no
significant difference between own- and other-race face pairings, t(99)
0.59, p.56, d0.06.
Taken together, Experiment 2 demonstrated that perceived age moderates
the CRE in identification performance. The results suggest that travel
document screeners may be especially susceptible to making an error in
judgement with other-race face pairings when there is a substantial perceived
distance between the age of photograph and the individual tendering the
passport. Nonetheless, caution should be used as we were only able to
implement a manipulation of perceived age differences in our stimuli, not
actual age difference.
EXPERIMENT 3
Several studies have demonstrated that the use of disguises can hinder the
accuracy of face identification (Righi, Peissig, & Tarr, 2012; Terry, 1993).
Such disguises may be as simple as hats or glasses that obscure upper regions
CAN I SEE YOUR PASSPORT PLEASE? 1299
Downloaded by [Iowa State University] at 10:07 28 May 2014
of the face, regions that appear most important for discriminating faces.
Experiment 3 assessed identification performance on own- and other-race
faces, but also assessed whether the presence of a disguise would moderate
the CRE. We predicted a main effect for race of face, consistent with prior
studies. We also predicted a main effect for the use of a disguise, such that
identification performance would be better for those who were not disguised
compared to those who were disguised. Finally, we predicted a significant
interaction such that the use of disguises would particularly disrupt
performance on other-race faces.
Method
Participants
.One hundred Mexican-American undergraduate students
(71 female; mean age 20.49 years) at the University of Texas at El Paso
participated in this study.
Design
.A22 repeated measures factorial design was used to assess
the influence of race (own- vs. other-race) and the use of disguises (no
disguise vs. disguise) on identification performance.
Stimulus materials and procedure
.Stimuli and procedures largely
followed the General Method section. Unique to this experiment was the
manipulation of disguise. Half of the stimuli presented included a target
absent any disguise; the other half included the target wearing a disguise (i.e.,
a baseball cap and sunglasses; see Figure 4).
Figure 4. Example of a Mexican-American disguised trial. To view this figure in colour, please see
the online issue of the Journal.
1300 MEISSNER, SUSA, ROSS
Downloaded by [Iowa State University] at 10:07 28 May 2014
Results and discussion
Table 3 summarizes participants’ performance across all measures. With
regard to discrimination accuracy, a 2 2 repeated measures ANOVA
observed significant main effects for both race of face and the use of a
disguise. As predicted, identification performance was significantly greater
for own- versus other-race face pairings, F(1, 99) 45.06, pB.001, h
p
2
.31, and for targets who were not wearing a disguise relative to those who
were wearing a disguise, F(1, 99) 125.27, pB.001, h
p
2
.56. Further,
results indicated that these main effects were qualified by a significant
interaction, F(1, 99) 31.49, pB.001, h
p
2
.24. Simple effect analyses
indicated that when the target was not wearing a disguise there was no
difference in identification performance between own- and other-race face
pairings t(99) 1.64, p.105, d0.16. However, when the target was
wearing a disguise, identification performance was significantly greater for
own-race, compared with other-race, face pairings, t(99) 7.66, pB.001, d
0.81.
With regard to response criterion, there was a significant main effect for
both race of face, F(1, 99) 33.058, pB.001, h
p
2
.25, and the use of
disguises, F(1, 99) 35.777, pB.001, h
p
2
.27. These effects suggest that
participants are more conservative in their judgements (less likely to say that
the target and identification photo are a match) when the target comes from
another race than their own race, and when the target is wearing a disguise as
opposed to not wearing a disguise. There was no significant interaction, F(1,
99) 0.30, p.59, h
p
2
B.01.
In summary, Experiment 3 demonstrated that wearing a disguise can also
exacerbate the CRE in a simultaneous perceptual discrimination task,
leading to particularly poor performance in other-race contexts. The robust
main effects of both the CRE and the use of disguise also demonstrate that
certain conditions can undermine the accuracy of screeners at security
TABLE 3
Descriptive statistics for identification performance in Experiment 3
Own-race Other-race
M SD 95% CI M SD 95% CI
No disguise
Disc. accuracy (Az) 0.82 0.13 [.79, .85] 0.79 0.13 [.76, .82]
Response bias (C) 0.32 0.54 [.43, .21] 0.12 0.54 [.23, .01]
Disguise
Disc. accuracy (Az) 0.73 0.13 [.70, .76] 0.57 0.16 [.54, .60]
Response bias (C) 0.10 0.48 [.19, .01] 0.14 0.42 [.06, .22]
CI confidence interval of the mean.
CAN I SEE YOUR PASSPORT PLEASE? 1301
Downloaded by [Iowa State University] at 10:07 28 May 2014
checkpoints. Recent research suggests that disguises may disrupt partici-
pants’ ability to encode holistic information about a face and this may
account for the interaction between the CRE and disguise given potential
configural-featural differences in face processing across own- and other-race
faces, respectively (see Righi et al., 2012).
GENERAL DISCUSSION
Across three experiments, we demonstrated that the cross-race effect (CRE)
is prevalent in simultaneous perceptual discrimination tasks that mimic
travel document screening contexts. Namely, when individuals are tasked
with determining whether an identification image matches the tendering
person, performance is significantly better for own- versus other-race
persons. In Experiment 1, we demonstrated that individuals demonstrate
overconfidence in this seemingly simple task, particularly when the target is
of another, less familiar race. Experiments 2 and 3 also found that certain
conditions can moderate the CRE in perceptual identification, including the
perceived age difference between the target and the identification photo and
the use of a disguise.
From a theoretical perspective, our results support previous studies
demonstrating the CRE within a perceptual identification framework.
Further, our stimulus manipulations were intended to challenge efforts at
feature recruitment in a manner that might differentially influence own-
versus other-race face performance (cf. Marcon et al., 2010). Consistent with
the hypotheses, individuals’ ability to discriminate faces was more robust to
changes in stimulus conditions when faces emanated from an expert stimulus
class (own-race faces).
From a practical perspective, our studies demonstrate the challenges faced
by security screeners who inspect travel documents in search of imposters.
Given the variety of ethnicities that a screener may be exposed to, the
fallibility of identifying imposters at border and transportation checkpoints
is of genuine concern. Although our studies involved college student samples
who are less experienced with the screening task, it is also important to note
that the paradigm we implemented was absent any other environmental or
contextual factors that could challenge or disrupt the identification
process*for example, screeners often assess multiple pieces of information
(e.g., authenticity of the passport or identification document, validity of a
visa or airline/train ticket, etc.) and may engage the target individual in a
series of questions designed to verify identity or assess trustworthiness. These
factors will be important to assess for their influence on the accuracy of
person (or imposter) identification. In addition, closed-circuit video is
becoming very common in Europe and North America as a method for
1302 MEISSNER, SUSA, ROSS
Downloaded by [Iowa State University] at 10:07 28 May 2014
assisting person identification and crime prevention; however, significant
challenges have been noted in correctly identifying the face of a target
individual from video, particularly if the video is partially obscured (Burton,
Wilson, Cowan, & Bruce, 1999). In the current national security context,
facial recognition systems (be they human or automated) will require
continued empirical investigation to improve identification accuracy. With
regard to human screeners, researchers should also focus on the development
of training programmes to assist individuals with improving discrimination
accuracy to include perceptual discrimination tasks (see Tanaka et al., this
issue 2013).
The current experiments represent one of the first assessments of the CRE
in a simultaneous perceptual discrimination task that models the document
screening context; further research will be necessary to replicate and extend
our findings. As noted previously, our sampling was limited to a population
of Mexican-American participants in the Paso del Norte region of the
United States. Although we were able to recruit small samples of African-
American participants to provide assessments of our stimuli and ensure
consistency in our stimulus sampling and manipulations, we were unable to
recruit samples of African-American participants for the current studies. The
CRE literature suggests a rather robust effect that is observed across a
variety of racial and ethnic groups and is more likely conditioned on
perceptual experience (see Anzures et al., this issue 2013; O’Toole & Natu,
this issue 2013; Tanaka et al., this issue 2013), cognitive processes (see
Hayward, Crookes, & Rhodes, this issue 2013), or motivational processes
(see Hugenberg et al., this issue 2013), rather than featural variation or
physiognomy across race. As such, we expect the current findings to be
moderated by similar factors if explored across a range of racial or ethnic
participant samples.
Overall, the current research suggests that individuals are overconfident
and subject to systematic errors in simultaneous perceptual discrimination
tasks that mimic security screening operations. The well recognized cross-
race effect was evident in participants’ performance, and this effect was
further exacerbated by stimulus conditions that challenged efforts at feature
recruitment. Given the hundreds of thousands of screening tasks conducted
daily by security personnel, further research understanding the theoretical
and practical implications of such findings appears warranted.
REFERENCES
Anzures, G., Quinn, P. C., Pascalis, O., Slater, A. M., & Lee, K. (2013). Development of own-
and other-race biases. Visual Cognition. doi:10.1080/13506285.2013.821428
Behrman, B., & Davey, S. (2001). Eyewitness identification in actual criminal cases: An archival
analysis. Law and Human Behavior,25, 475491. doi:10.1023/A:1012840831846
CAN I SEE YOUR PASSPORT PLEASE? 1303
Downloaded by [Iowa State University] at 10:07 28 May 2014
Brigham, J. C., Bennett, L., Meissner, C. A., & Mitchell, T. (2007). The influence of race on
eyewitness memory. The handbook of eyewitness psychology: Vol II. Memory for people (pp.
257281). Mahwah, NJ: Lawrence Erlbaum Associates.
Bukach, C. M., Cottle, J., Ubiwa, J., & Miller, J. (2012). Individuation experience predicts other-
race effects in holistic processing for both Caucasian and Black participants. Cognition,123,
319324. doi:10.1016/j.cognition.2012.02.007
Burt, D. M., & Perrett, D. I. (1995). Perception of age in adult Caucasian male faces: Computer
graphic manipulation of shape and colour information. Proceedings of the Royal Society of
London,259B, 137143. doi:10.1098/rspb.1995.0021
Burton, A. M., Wilson, S., Cowan, M., & Bruce, V. (1999). Face-recognition in poor quality
video: Evidence from security surveillance. Psychological Science,10, 243248. doi:10.1111/
1467-9280.00144
Caldara, R., & Abdi, H. (2006). Simulating the ‘‘other-race’’ effect with autoassociative neural
networks: Further evidence in favor of the face-space model. Perception,35(5), 659670.
doi:10.1068/p5360
Furl, N., Philips, P. J., & O’Toole, A. J. (2002). Face recognition algorithms and the other-race
effect: Computational mechanisms for a developmental contact hypothesis. Cognitive
Science,26(6), 797815. doi:10.1016/S0364-0213(02)00084-8
Hancock, K. J., & Rhodes, G. (2008). Contact, configural coding and the other-race effect in
face recognition. British Journal of Psychology,99,4556. doi:10.1348/000712607X199981
Hayward, W. G., Crookes, K., & Rhodes, G. (2013). The other-race effect: Holistic coding
Differences and beyond. Visual Cognition. doi:10.1080/13506285.2013.824530
Hugenberg, K., Wilson, J. P., See, P. E., & Young, S. G. (2013). Toward a synthetic model of own
group biases in face memory. Visual Cognition. doi:10.1080/13506285.2013.821429
Hugenberg, K., Young, S. G., Bernstein, M. J., & Sacco, D. F. (2010). The categorization-
individuation model: An integrative account of the other-race recognition deficit.
Psychological Review,117(4), 1168 1187. doi:10.1037/a0020463
Jonsson, A. C., & Allwood, C. M. (2003). Stability and variability in the realism of confidence
judgments over time, content domain, and gender. Personality and Individual Differences,34,
559574. doi:10.1016/S0191-8869(02)00028-4
Juslin, P., Olsson, N., & Winwan, A. (1996). Calibration and diagnosticity of confidence in
eyewitness identification: Comments on what can be inferred from the low-confidence
accuracy correlation. Journal of Experimental Psychology: Learning, Memory, and Cognition,
5, 13041316. doi:10.1037/0278-7393.22.5.1304
Kemp, R., Towell, N., & Pike, G. (1997). When seeing should not be believing: Photographs,
credit cards and fraud. Applied Cognitive Psychology,11, 211222. doi:10.1002/(SICI)1099-
0720(199706)11:3 < 211::AID-ACP430 > 3.0.CO;2-O
Krug, K. (2007). The relationship between confidence and accuracy: Current thoughts of the
literature and a new area of research. Applied Psychology in Criminal Justice,3,741.
Lindsay, D. S., Jack, P. C., Jr., & Christian, M. A. (1991). Other-race face perception. Journal of
Applied Psychology,76, 587 589. doi:10.1037/0021-9010.76.4.587
Marcon, J. L., Meissner, C. A., Frueh, M., Susa, K. J., & MacLin, O. H. (2010). Perceptual
identification and the cross-race effect. Visual Cognition,18, 767779. doi:10.1080/
13506280903178622
Marcon, J. L., Susa, K. J., & Meissner, C. A. (2009). Assessing the influence of recollection and
familiarity in memory for own- and other-race faces. Psychonomic Bulletin and Review,16,
99103. doi:10.3758/PBR.16.1.99
Megreya, A. M., White, D., & Burton, A. M. (2011). The other-race effect does not rely on
memory: Evidence from a matching task. Quarterly Journal of Experimental Psychology,64,
14731483. doi:10.1080/17470218.2011.575228
1304 MEISSNER, SUSA, ROSS
Downloaded by [Iowa State University] at 10:07 28 May 2014
Meissner, C. A., & Brigham, J. C. (2001). Thirty years of investigating the own-race bias in
memory for faces: A meta-analytic review. Psychology, Public Policy, and Law,7,335.
doi:10.1037/1076-8971.7.1.3
Meissner, C. A., Brigham, J. C., & Butz, D.A. (2005). Memory for own- and other-race face: A
dual process approach. Applied Cognitive Psychology,19, 545567. doi:10.1002/acp.1097
O’Toole, A. J., & Natu, V. (2013). Computational perspectives on the other-race effect. Visual
Cognition. doi:10.1080/13506285.2013.803505
Olsson, N. (2000). A comparison of correlation, calibration, and diagnosticity as measures of
the confidence-accuracy relationship in witness identification. Journal of Applied Psychology,
85, 504511. doi:10.1037/0021-9010.85.4.504
Righi, G., Peissig, J. J., & Tarr, M. J. (2012). Recognizing disguised faces. Visual Cognition,20,
143169. doi:10.1080/13506285.2012.654624
Sporer, S. L. (2001). The cross-race effect: Beyond recognition of faces in the laboratory.
Psychology, Public Policy, and Law,7, 170200. doi:10.1037/1076-8971.7.1.170
Sporer, S. L., Trinkl, B., & Guberova, E. (2007). Matching faces: Differences in processing speed
of out-group faces by different ethnic groups. Journal of Cross-Cultural Psychology,38, 398
412. doi:10.1177/0022022107302310
Susa, K. J., Meissner, C. A., & de Heer, H. (2010). Modeling the role of social-cognitive
processes in the recognition of own- and other-race faces. Social Cognition,28, 523537.
doi:10.1521/soco.2010.28.4.523
Tanaka, J. W., Heptonstall, B., & Hagen, S. (2013). Perceptual expertise and the plasticity of
other-race face recognition. Visual Cognition. doi:10.1080/13506285.2013.826315
Terry, R. L. (1993). How wearing eyeglasses affects facial recognition. Current Psychology:
Developmental, Learning, Personality, Social,12, 151162. doi:10.1007/BF02686820
Walker, P. M., & Hewstone, M. (2006). A perceptual discrimination investigation of the own
race effect and intergroup experience. Applied Cognitive Psychology,20, 461475.
doi:10.1002/acp.1191
Walker, P. M., & Tanaka, J. W. (2003). An encoding advantage for own-race versus other-race
faces. Perception,32, 11171125. doi:10.1068/p5098
Weber, N., & Brewer, N. (2003). The effect of judgment type and confidence scale on confidence-
accuracy calibration in face recognition. Journal of Applied Psychology,88(3), 490 499.
doi:10.1037/0021-9010.88.3.490
CAN I SEE YOUR PASSPORT PLEASE? 1305
Downloaded by [Iowa State University] at 10:07 28 May 2014
