ArticlePDF Available

Memory for own- and other-race faces: A dual-process approach

Authors:

Abstract and Figures

The current studies assessed the phenomenological basis of the cross-race effect by examining predictions of various social-cognitive mechanisms within a dual-process framework for both the perception (Experiment 1) and recognition (Experiment 2) of own- and other-race faces. Taken together, the current studies demonstrated that differential performance on own-race faces was largely due to qualitative differences in the encoding of facial information represented by a recollection process. Furthermore, false recollections with high ratings of confidence occurred more often when participants encoded and responded to unfamiliar other-race faces. The theoretical implications of these findings for the phenomenology of skilled perceptual-memory are discussed, and the applied consequences of the cross-race effect as an encoding-based phenomenon are considered. Copyright © 2005 John Wiley & Sons, Ltd.
Content may be subject to copyright.
APPLIED COGNITIVE PSYCHOLOGY
Appl. Cognit. Psychol. 19: 545–567 (2005)
Published online 14 March 2005 in Wiley InterScience
(www.interscience.wiley.com) DOI: 10.1002/acp.1097
Memory for Own- and Other-race Faces:
A Dual-process Approach
CHRISTIAN A. MEISSNER
1
*,
JOHN C. BRIGHAM
2
and DAVID A. BUTZ
2
1
Florida International University, USA
2
Florida State University, USA
SUMMARY
The current studies assessed the phenomenological basis of the cross-race effect by examining
predictions of various social-cognitive mechanisms within a dual-process framework for both the
perception (Experiment 1) and recognition (Experiment 2) of own- and other-race faces. Taken
together, the current studies demonstrated that differential performance on own-race faces was
largely due to qualitative differences in the encoding of facial information represented by a
recollection process. Furthermore, false recollections with high ratings of confidence occurred
more often when participants encoded and responded to unfamiliar other-race faces. The theoretical
implications of these findings for the phenomenology of skilled perceptual-memory are discussed,
and the applied consequences of the cross-race effect as an encoding-based phenomenon are
considered. Copyright #2005 John Wiley & Sons, Ltd.
The cross-race effect (CRE, or own-race bias, or other-race effect, as it has been variously
labelled) refers to the finding that faces of one’s own race are better remembered when
compared with performance on faces of another, less familiar race. Throughout the past
30 years, a host of studies have investigated this memory phenomenon focusing on its
practical and theoretical importance. Although some general agreement regarding the
mechanisms responsible for the CRE might have been expected from such a mass of
literature, debate continues among researchers regarding potential social-cognitive pro-
cesses governing the effect (see Meissner & Brigham, 2001; Sporer, 2001). The current
investigation seeks to inform this debate by framing the CRE within the constraints of a
dual-process theory of memory (Gardiner & Richardson-Klavehn, 2000; Kelley & Jacoby,
2000; Mandler, 1980; for a review see Yonelinas, 2002). Generally speaking, such theories
have distinguished between memory processes involving conscious-level conceptual
information that is elaboratively encoded and those involving fluent, perceptually-based
information that is believed to be encoded in an automatic, non-conscious manner. One
example of such a distinction involves the phenomenon of a person looking familiar, yet
we are unable to recollect details of who the person is or where we might have seen him or
her previously. In such an instance, memory is based largely upon fluency of familiarity,
Copyright #2005 John Wiley & Sons, Ltd.
*Correspondence to: Dr. C. A. Meissner, Department of Psychology, Florida International University, University
Park, Miami, FL 33199, USA. E-mail: meissner@fiu.edu
but is absent any episodic recollection of a prior encoding event. By exploring such
phenomenological differences in the processing of own- and other-race faces, the
current study attempts to distinguish between previous social-cognitive accounts of the
phenomenon.
POTENTIAL SOCIAL-COGNITIVE MECHANISMS
While a perceptual learning interpretation of the CRE has been central to many theoretical
explanations of the effect (Chance & Goldstein, 1996; Meissner & Brigham, 2001),
various social and cognitive mechanisms have been proposed to further explain partici-
pants’ superior performance on own-race faces. Sporer (2001) has recently reviewed many
of these proposals, and has summarized their potential influence within a simplified ‘in-
group/out-group’ model (IOM) of the CRE phenomenon. Below we briefly review several
of the proposed mechanisms.
Interracial contact
A number of researchers have posited that the quality and/or quantity of interracial contact
may play a significant role in the CRE. For example, researchers have proposed that
increased contact with other-race individuals may increase memory performance by: (a)
reducing the likelihood of stereotypic responses and increasing the likelihood that
individuals may look for more individuating information (Malpass, 1981); (b) influencing
individuals’ motivation to accurately recognize other-race persons through associated
social rewards/punishments (Malpass, 1990); or (c) reducing the perceived complexity of
unfamiliar other-race faces (Goldstein & Chance, 1971). Two major approaches to
investigating contact have been to examine groups of individuals differing in their degree
of other-race contact, or to assess individuals’ self-reported contact with other-race
persons. In general, researchers using the former approach have demonstrated that
individuals residing in integrated populations show less of a CRE when compared with
same-race individuals residing in more homogeneous populations (cf. Chiroro &
Valentine, 1995; Wright, Boyd, & Tredoux, 2003). With regard to self-reported interracial
experience, Meissner and Brigham’s (2001) meta-analysis indicated a small, yet sig-
nificant, correlation accounting for approximately 2% of the variance. In a related fashion,
several studies have demonstrated short-term improvement in recognizing other-race faces
utilizing various discrimination training procedures (Elliott, Wills, & Goldstein, 1973;
Goldstein & Chance, 1985; Lavrakas, Buri, & Mayzner, 1976; Malpass, Lavigueur, &
Weldon, 1973).
Cognitive disregard and categorization effects
A series of experiments by Rodin (1987) demonstrated that individuals will often
cognitively disregard, or render ‘invisible,’ others based upon various categorical dimen-
sions such as age, sex, or attractiveness. Rodin held that as ‘cognitive misers’ we often
attempt to rely upon such broad categorical judgments for determining whether further
interaction should occur. Other studies have also indicated that racial information may be
used in a similar fashion, and that prejudice or racial attitudes may moderate the extent to
which cognitive disregard occurs (cf. Sensening, Jones, & Varney, 1973). Similarly,
546 C. A. Meissner et al.
Copyright #2005 John Wiley & Sons, Ltd. Appl. Cognit. Psychol. 19: 545–567 (2005)
studies on the ‘out-group homogeneity effect’ have demonstrated that out-group members
are often perceived to be more homogeneous (in such attributes as attitudes, behaviour, or
appearance) than in-group members, and are represented in terms of stereotypic (as
opposed to individuating) characteristics (for a review, see Mullen & Hu, 1989). Recent
studies have indicated that this effect of racial stereotyping tends to be automatic
(Dasgupta, McGhee, Greenwald, & Banaji, 2000; Payne, 2001; Payne, Lambert, &
Jacoby, 2002), and under limited conscious control (however, see Monteith, Ashburn-
Nardo, Voils, & Czopp, 2002).
Consistent with the effects of cognitive disregard and out-group homogeneity, studies in
the face perception literature have demonstrated that other-race faces are racially classified
more quickly than own-race faces (Levin, 1996; Valentine & Endo, 1992). With regard to
the CRE, research by Levin (1996, 2000) and by MacLin and Malpass (2001, 2003) has
suggested that this automatic inclination to categorize other-race faces may distract from
the encoding of individuating information, and thereby lead to poorer performance when
recognizing other-race faces. In this ‘race-feature perspective,’ it is proposed that attention
directed towards the encoding of race as a facial feature may divert cognitive resources
that would otherwise be used to seek-out individuating information.
Levels of processing
In their ‘levels of processing’ approach to memory, Craik and Lockhart (1972) held that
memory strength for any given stimulus is a function of the depth to which the stimulus is
processed at encoding. ‘Depth’ is generally considered a continuum defined by the extent
to which various conceptual or semantic attributes of the stimulus are processed. Studies
attempting to vary depth of processing often include a manipulation of orienting (or
encoding) instructions. In both the word list and face memory literatures, participants’
recognition performance has been shown superior when ‘deep’ encoding instructions are
provided (Bower & Karlin, 1974; Sporer, 1991). As some authors have noted, it is
important to consider that levels of processing manipulations may actually vary the
amount of information encoded, rather than the depth of processing, per se (Sporer, 1991;
Winograd, 1981).
Interestingly, evidence has been mixed with regard to a levels of processing explanation
of the CRE. For example, Chance and Goldstein (1981) asked participants to describe
aloud their reactions to a series of own- and other-race faces. The authors then coded the
statements according to the presumed ‘depth’ that had occurred, and observed that own-
race faces were processed at a significantly deeper level than were other-race faces.
However, several studies have manipulated level of processing within the cross-race
paradigm and have failed to observe the predicted interaction (i.e. that own-race faces
might benefit more from a deep-processing instruction; see Burgess & Weaver, 2003;
Devine & Malpass, 1985).
Distinctiveness effects
With regard to memory for human faces, one quality of the face that has been shown to
influence recognition performance is that of distinctiveness. Numerous studies have
demonstrated that novel faces rated as distinctive or unusual are recognized significantly
better (i.e. they receive a higher proportion of hits and a lower proportion of false alarms)
than faces rated as typical in appearance (e.g. Brigham, 1990; Shepherd, Gibling, & Ellis,
Cross-race effect 547
Copyright #2005 John Wiley & Sons, Ltd. Appl. Cognit. Psychol. 19: 545–567 (2005)
1991; Valentine, 1991). Given the relevance of this stimulus property to memory perfor-
mance, researchers in the face memory literature have attempted to use distinctiveness as an
example of one characteristic by which individuals represent faces in memory (Byatt &
Rhodes, 1998; O’Toole, Deffenbacher, Valentin, & Abdi, 1994; Valentine & Endo, 1992).
Within the cross-race paradigm, researchers have asserted that attention to distinctive-
ness may differentiate performance on own- and other-race faces. For example, Chiroro
and Valentine (1995) examined the effects of race, typicality, and level of perceptual
experience within the cross-race paradigm. Based upon the assumptions of Valentine’s
(1991) Multi-Dimensional Space (MDS) Framework, Chiroro and Valentine predicted that
only individuals who had considerable previous experience with other-race faces (high
contact) would demonstrate distinctiveness effects for both own-race and other-race faces.
This was due largely to the notion that such individuals should be able to distinguish
between typical and distinctive other-race faces based upon features they had extracted
through prior experience. In contrast, low-contact individuals were predicted to demon-
strate no differences in performance according to the distinctiveness dimension of other-
race faces. Overall, Chiroro and Valentine’s results demonstrated the predicted interaction
such that distinctiveness effects for low-contact individuals were confined to own-race
faces. On the other hand, high-contact individuals demonstrated significant effects of
distinctiveness regardless of the race of the face.
Configural-featural hypothesis
One seemingly notable finding in the face memory literature has involved research on the
face inversion effect— the phenomenon that inverted (upside-down) photos of faces are
identified more poorly than inverted photos of other objects (e.g. houses, cars, etc.). In
early work on this effect, Yin (1969) concluded that face recognition was the product of a
unique system, different from systems responsible for recognizing other kinds of visual
stimuli. In contrast to this ‘neural specialization’ hypothesis, Diamond and Carey (1986)
later proposed that perceptual learning might be operating in face recognition. In several
experiments they showed that the inversion effect was not unique to faces, but rather
occurred when participants had a great deal of experience with the stimulus materials. In
particular, inversion appeared to disrupt the effectiveness with which individuals were able
to encode stimuli that were highly familiar to them. Diamond and Carey believed that this
effect stemmed from experienced participants’ reliance upon configural (or relational)
properties of the stimulus. Novice participants, on the other hand, relied upon only the
featural (or isolated) aspects of the face that were less influenced by inversion. A number
of subsequent studies have supported this general configural-featural hypothesis (see
Farah, Wilson, Drain, & Tanaka, 1998).
The notion that expertise leads to configural processing has also been applied to the
CRE. In particular, Rhodes, Brake, Taylor, and Tan (1989) proposed that greater
experience with own-race faces would lead to a larger inversion effect, due to an increased
reliance upon configural information. The encoding of other-race faces, on the other hand,
should not be as influenced by inversion due to the featural aspects that are relied upon. As
hypothesized, Rhodes and colleagues observed that own-race faces were significantly
more susceptible to inversion than were other-race faces for measures of both reaction
time and accuracy (see also, Fallshore & Schooler, 1995). However, several other studies
have failed to observe this interaction of inversion with the CRE (Buckhout & Regan,
1988; M. C. R. Burgess, unpublished doctoral dissertation, 1997).
548 C. A. Meissner et al.
Copyright #2005 John Wiley & Sons, Ltd. Appl. Cognit. Psychol. 19: 545–567 (2005)
In-group/out-group model
In his IOM account of the CRE, Sporer (2001) attempted to bring together these varied
social-cognitive explanations. Sporer’s IOM assumes that a default or automatic process
occurs when an individual encounters an own-race (or in-group) face, such that encoding
involves deeper level processing with the individual focused upon relevant, configural
properties, or dimensions of the face that are useful for distinguishing it from other similar
faces in memory (consistent with Valentine’s, 1991, MDS model). In contrast, when an
individual encounters an other-race (or out-group) face, racial characteristics first signal an
automatic categorization response. This categorization response may then be linked to
other cues to cognitively disregard such faces, and may result in attentional processes being
allocated elsewhere. This categorization may also signal that less effort should be extended
in the encoding process, thereby leading to shallow (or featural-based) encoding of the face,
and may signal stereotyping processes that lead to improper inferences regarding salient
characteristics of the face in memory. Given the greater homogeneity in representation of
other-race faces resulting from the lack of distinctiveness effects, Sporer’s IOM also
predicts an effect on response bias such that individuals will be more liberal in responding
to other race faces (consistent with the general empirical literature, see Meissner &
Brigham, 2001). Taken together, the model also proposes to account for other in-group/out-
group phenomena in the facial memory literature, such as the effects of age (Wright &
Stroud, 2002) and gender (Slone, Brigham, & Meissner, 2000).
A DUAL-PROCESS RECOGNITION PERSPECTIVE
Over the past 30 years, dual-process memory models have influenced our understanding of
recognition memory by demonstrating that two qualitatively distinct processes appear to
be operating. More specifically, these models distinguish between a recall-like process in
which episodic information is retrieved at the time of recognition, and a fluency-based
process in which general familiarity (in the absence of specific episodic information) is
used as a basis for recognition. A variety of such models have been proposed over the
years, including Jacoby and colleagues’ distinction between ‘recollection’ and ‘famil-
iarity’ processes (Jacoby, 1991; Kelley & Jacoby, 2000), Tulving, Gardiner, and collea-
gues’ distinction between ‘remember’ and ‘know’ judgments of phenomenological
experience (Gardiner, 1988; Gardiner & Richardson-Klavehn, 2000; Tulving, 1983),
and Mandler’s (1980) concepts of ‘integration’ and ‘elaboration’. Across these varied
dual-process approaches, there are some similarities in the proposed mechanisms. For
example, recollection (or ‘remember’ judgments, or elaboration) is generally influenced
by generative or semantic encoding, division of attention, speed of responding, novel
learning, and the effects of aging and amnesia. In contrast, familiarity (or ‘know’
judgments, or integration) is generally sensitive to fluency manipulations, forgetting
over short retention intervals, and changes in the response criterion (for a review, see
Yonelinas, 2002). Over the past decade, the role of recollection and familiarity has also
been demonstrated in the facial recognition literature (Bartlett, Strater, & Fulton, 1991;
Bastin & van der Linden, 2003; Mantyla, 1997; Mantyla & Cornoldi, 2002; Paller, Bozic,
Ranganath, Grabowecky, & Yamada, 1999; Parkin, Gardiner, & Rosser, 1995; Reinitz,
Morrissey, & Demb, 1994; Yonelinas, Kroll, Dobbins, & Soltani, 1999). Findings in such
studies have been generally consistent with those in the word recognition literature,
Cross-race effect 549
Copyright #2005 John Wiley & Sons, Ltd. Appl. Cognit. Psychol. 19: 545–567 (2005)
suggesting that recollection and familiarity processes operate in facial memory and that
such phenomena may be explained by more general cognitive mechanisms.
How might a dual-process perspective account for the CRE in facial recognition? At
present, no studies have attempted to blend such an explanation with that of previously
suggested social-cognitive explanations. However, given that many of the social-cognitive
mechanisms proposed in the CRE literature have been shown to demonstrate certain
patterns of influence on recollection and familiarity, they may be separable via predictions
regarding their influence on each process. Thus, examining the CRE within the context of
a dual-process perspective may provide insights into the specific social-cognitive mechan-
isms that are likely to be responsible for the phenomenon. Drawing from a review of the
dual-process literature by Yonelinas (2002), we discuss below each of the proposed social-
cognitive explanations with regard to their dual-process predictions (see Table 1).
First, some researchers have suggested that individuals may generate a categorization
response when presented with other-race faces (Levin, 1996, 2000; MacLin & Malpass,
2001, 2003; Sporer, 2001). Studies within the stereotyping literature have suggested that
such responses tend to be rather automatic (Dasgupta et al., 2000), and are thus supported
by an ‘accessibility bias’ that is under limited conscious control (Payne, 2001; Payne et al.,
2002). Consistent with this notion of biased responding, such a stereotyped encoding
process for other-race faces might suggest a greater influence of familiarity (or fluency-
based) processes on the recognition of other-race faces when compared with performance
on own-race faces.
In contrast, categorization may result in cognitive disregard and thereby divert
attentional resources during the encoding process. Studies both in the word list and facial
recognition (Parkin et al., 1995; Reinitz et al., 1994) literatures have demonstrated that
limiting attentional resources at study generally reduces the influence of recollection
processes, but leaves intact familiarity-based responding. In a similar manner, studies that
have manipulated effortful encoding (e.g. via a manipulation that asks participants to
‘generate’ a certain encoding context for a word versus simply ‘reading’ the word in a
given context; cf. Jacoby, 1991), and those that have attempted to increase semantic or
meaning-based encoding (e.g. via a manipulation that encourages a ‘deep’ versus
‘shallow’ level of processing; cf. Rajaram, 1993) have demonstrated large effects on
recollection processes. With regard to the CRE, such mechanisms might suggest that own-
race faces would yield a rather substantial increase in recollection when compared with
memory for other-race faces.
As discussed previously, researchers have attempted to model facial recognition
memory based upon various perceptual characteristics, most notably including perceived
facial distinctiveness (Byatt & Rhodes, 1998; O’Toole et al., 1994; Valentine & Endo,
Table 1. Dual-process predictions generated from social-cognitive mechanisms of the own-race
bias in memory for faces
Social-cognitive mechanisms Predicted effect Predicted effect on
on ‘recollection’ ‘familiarity’
Racial stereotyping Own-race ¼Other-race Own-race <Other-race
Encoding-based processing Own-race >Other-race Own-race ¼Other-race
(attentional resources, effortful/
semantic encoding, distinctiveness)
Configural-featural processing Own-race ¼Other-race Own-race >Other-race
550 C. A. Meissner et al.
Copyright #2005 John Wiley & Sons, Ltd. Appl. Cognit. Psychol. 19: 545–567 (2005)
1992). Furthermore, research has shown that individuals are more attuned to the
perception of distinctiveness when processing own-race faces (Chiroro & Valentine,
1995). Within the dual-process memory literature, research by Mantyla (1997) has
suggested that encoding faces based upon distinctive characteristics (as opposed to
more ‘relational’ encoding discussed previously) has a similar effect to that of other
encoding-based manipulations namely, distinctiveness processing largely increases
recollection-based responding (see also, Mantyla, 1993; Rajaram, 1993, 1996). Given
that individuals are more attuned to distinctiveness in own-race faces, one might expect
recollection to play a larger role in memory for own-race faces, similar to that of other
attentional and encoding-based manipulations.
Finally, studies from the inversion paradigm have suggested that faces may be
processed based upon two types of informationconfigural (or relational) and
featural (or isolated) aspects of the face (Farah et al., 1998). This research has also
suggested that the degree of configural versus featural processing may be a function of
perceptual expertise, and that the deficit in memory for other-race faces may be due to
individuals’ inability to process such faces in a configural manner (Fallshore & Schooler,
1995; Rhodes et al., 1989). A recent study by Yonelinas and colleagues (1999)
examined the effects of face inversion (and the configural-featural distinction) within
the context of a dual-process model of recognition memory. Their results demonstrated
that familiarity, but not recollection, was significantly influenced by the inversion
manipulation, thereby implicating its role in the configural representation of human faces.
As such, the configural-featural hypothesis would indicate that own-race faces should
yield a greater degree of familiarity-based responding due to a reliance on configural
aspects of the face.
The current studies were designed to examine the influence of recollection and
familiarity in memory for own- versus other-race faces, and thus to distinguish between
competing social-cognitive explanations of the CRE phenomenon. First, as will be
discussed below, Vokey and Read (1992) have demonstrated that rated distinctiveness
may be decomposed into two orthogonal components, namely face-specific episodic
information (i.e. memorability) and context-free familiarity. While both components have
been shown to contribute to recognition, Mantyla (1997) has demonstrated that the effects
of rated distinctiveness appear to be related to judgments of recollection. In a similar
manner, Deffenbacher, Johanson, Vetter, and O’Toole (2000) have shown that the
memorability component of rated distinctiveness is primarily influential during the
encoding of faces, and thus might have implications similar to other encoding-based
manipulations in the dual-process framework. Furthermore, O’Toole and colleagues
(1994) have shown that judgments of memorability tend to be associated with localized
distinctive features, while the familiarity component may be related to configural
properties. Our initial study attempted to replicate and extend this research to the CRE
using a perceptual rating paradigm.
In a second study, we then assessed the contributions of recollection and familiarity to
memory for own- and other-race faces by examining phenomenological judgments of
Remember-Know (cf. Gardiner & Richardson-Klavehn, 2000) and assessing the relation-
ships between rated perceptual components of memorability and familiarity estimates and
recognition performance on own- and other-race faces (cf. Mantyla, 1997; Vokey & Read,
1992). We hypothesized that if the CRE were largely the result of encoding-based
processes (e.g. attentional resources, effortful or semantic processing, distinctiveness
effects, etc.), then phenomenological judgments should indicate greater recollection-based
Cross-race effect 551
Copyright #2005 John Wiley & Sons, Ltd. Appl. Cognit. Psychol. 19: 545–567 (2005)
responding for own-race faces. In contrast, if the CRE were a function of perceptual
categorization of other-race faces, then greater familiarity-based responding might be
expected for other-race faces. Finally, if the CRE were the result of reliance upon
configural versus featural facial information, then phenomenological judgments should
indicate greater familiarity-based responding for own-race faces. As can be seen, these
varied predictions generally stand in opposition to one another. Thus, it was believed that a
dual-process paradigm might be useful for better understanding the mechanisms respon-
sible for the CRE.
EXPERIMENT 1: PERCEPTUAL RATINGS
OF OWN- AND OTHER-RACE FACES
As discussed previously, a variety of studies have demonstrated that unfamiliar faces rated
as distinctive or unusual are recognized significantly better than faces rated as more typical
in appearance (e.g. Brigham, 1990; Shepherd et al., 1991; Valentine, 1991). Conceptua-
lizations of facial distinctiveness, though, have become increasingly complex in recent
years. For example, Vokey and Read (1992) conducted factor-analytic work demonstrating
that distinctiveness (or typicality) seemed to be composed of two orthogonal components,
namely memorability and context-free familiarity (with distinctiveness crossloading on
both factors; see also, Morris & Wickham, 2001; O’Toole et al., 1994). Vokey and Read
hypothesized that the familiarity arising from some specific prior exposure (e.g. a face
seen before) is not intrinsically separable from the structurally induced familiarity that
arises from a lifetime of experience with similar exemplars. Furthermore, the authors
proposed that this ‘general familiarity’ could be shown to include ratings of familiarity,
attractiveness, and likability. Memorability, on the other hand, was defined by participants’
beliefs about how ‘easy to remember’ a face would be, and was believed to account for
identity-specific episodic information in the face. With regard to recognition performance,
Vokey and Read asserted that a distinctive or unusual face is better recognized because it is
both low on familiarity (as it resembles few other faces in the recognizer’s experience) and
high in perceived memorability (because distinctive attributes of the face provide episodic
information). Conversely, a typical face is likely to be poorly remembered due to its high
similarity to other known faces and its low memorability (or its difficulty in obtaining
uniqueness in memory).
Distinctiveness has sometimes been discussed in the literature as if it were an inherent
quality of a face, but it is important to note that what is really involved is perceived
distinctiveness— that is, familiarity and memorability as they relate to the perceiver’s
experiences and perceptions. While a distinctive or unusual feature might make some
faces seem memorable to virtually all perceivers, other properties of faces may be
perceived very differently depending upon one’s life experiences (e.g. differential
expertise in dealing with own- and other-race people) and one’s representational structure
for encoding different kinds of faces, such as Black faces and White faces (cf. Chiroro &
Valentine, 1995; Valentine & Endo, 1992). Furthermore, researchers have suggested that
memorability and familiarity components may be related to the phenomenological basis of
recollection and familiarity, respectively (Mantyla, 1997; Vokey & Read, 1992). Thus, we
believe that component ratings may provide a means for exploring pre-experimental
perceptual variance involving recollection and familiarity-based responding to own- and
other-race faces.
552 C. A. Meissner et al.
Copyright #2005 John Wiley & Sons, Ltd. Appl. Cognit. Psychol. 19: 545–567 (2005)
The vast majority of work on facial distinctiveness has involved White participants
rating White faces. It remains unknown whether Blacks’ ratings of Black faces would
show the same component structure, or whether ratings of other-race faces would show the
same structure as ratings of same-race faces when a full cross-over design is investigated.
Preliminary data suggest that this issue merits further study. For example, O’Toole et al.
(1994) found that Whites’ ratings of White faces showed the same factor structure
identified by Vokey and Read (1992); however, for Whites’ ratings of Japanese faces,
distinctiveness was related to memorability, but not to familiarity.
The present study assessed the degree to which pre-experimental differences in stimulus
representation exist within the CRE paradigm. Both White and Black college students rated
photographs of White and Black male faces on measures of distinctiveness, familiarity,
likability, attractiveness, and memorability. Ratings were examined for differences in
overall magnitude and component structure across own- and other-race faces. While we
expected to replicate the Vokey and Read (1992) factor structure for own-race faces, it was
unclear whether this structure should be expected for other-race faces given the results of
O’Toole and colleagues (1994). Our default assumption was that familiarity and memor-
ability components would be evidenced in the perception of other-race faces as well. If
these components relate to those conceptualized within a dual-process perspective, then
differences in the magnitude of ratings provided for own- versus other-race faces may be
predicted by the social-cognitive mechanisms believed operating in the CRE. Furthermore,
as will be examined in Experiment 2, the relation between perceptual ratings of memor-
ability and familiarity and subsequent recognition performance on own- and other-race
faces should provide insights as to the mechanisms operating in the CRE.
Method
Participants
The sample consisted of 50 White and 50 Black students from an Introductory Psychology
course. Participants were tested in mixed-race groups ranging in size from two to five. The
mean age of the participants was 19 years and 75% were female.
Materials
Photographs of 80 White and 80 Black college-aged males were randomly selected from
our pool of White and Black faces. Photographs were standardized across all targets such
that each was wearing an identical burgundy-coloured sweatshirt, standing in front of a
grey background, in a full-frontal, non-smiling pose. Photographs were transformed to
colour slides, which were projected onto a screen subtending approximately 10 14
visual degrees.
In an attempt to replicate the conditions of Vokey and Read (1992) and O’Toole et al.
(1994), we utilized the following five rating indices across faces: (a) distinctiveness; (b)
likability; (c) attractiveness; (d) memorability; and (e) familiarity. Each characteristic was
assessed on a labelled 7-point scale. There has been some variability in how these
dimensions have been defined by earlier researchers. With regard to the measure of
distinctiveness, we utilized the rating scheme employed previously by O’Toole et al. and
Valentine and Bruce (1986), namely: ‘How difficult would it be to pick this person out of a
crowd?’ on a scale with 1 ¼‘Not Difficult’ and 7 ¼‘Very Difficult’. The likability (‘Does
this person look as if he would be likable?’; 1 ¼‘Not Likable’ and 7 ¼‘Very Likable’),
attractiveness (‘Is this person attractive looking?’; 1 ¼‘Not Attractive’ and 7 ¼‘Very
Cross-race effect 553
Copyright #2005 John Wiley & Sons, Ltd. Appl. Cognit. Psychol. 19: 545–567 (2005)
Attractive’), and memorability (‘Would this face be easy to remember?’; 1 ¼‘Easy to
Remember’ and 7 ¼‘Difficult to Remember’) measures were those used by Vokey and
Read (1992). Finally, the familiarity measure was that used previously by Vokey and Read
(1988) and O’Toole et al., namely: ‘Is this face confusable with someone you know?’ on a
scale with 1 ¼‘Not Confusable’ and 7 ¼‘Very Confusable’.
Design
A2(2) mixed multivariate design was used with race of participant as the between-
subjects factor, race of face as the within-subjects factor, and the five face-rating indices as
dependent measures. Presentation of faces was blocked by race, the order of which was
counterbalanced across participant groups. Within each block of race, the presentation of
faces was randomized and counterbalanced to control for the effects of order and
interference. Finally, participants rated each face on all five rating scales, consistent
with the methodology utilized by Vokey and Read (1992). Although asking participants to
rate faces on all five scales may inflate the correlations between ratings, our intent was to
follow previous methodologies and replicate such findings in the cross-race domain. In
light of this concern and to control for any response biases, the order of presenting the five
rating scales was randomized for each target face presented and across participants.
Procedure
Upon entering the lab, participants were familiarized with the general face-rating
procedure. Participants were then presented with each face for 15 s. They were instructed
to view each face for several seconds prior to rating it on the various indices. A short break
was given to participants following the first block (race) of faces.
Results
Ratings of own- and other-race faces
Participants’ ratings for each face were subjected to a 2 ( 2) mixed multivariate analysis
of variance (MANOVA), with race of participant as the between-subjects factor, race of
face as the within-subjects factor, and face-rating index as the dependent measures.
Results indicated a significant multivariate race of participant by race of face interaction,
F(5, 94) ¼19.43, p<0.001, across the five rating scales. A significant race of face main
effect was also observed, F(5, 94) ¼3.08, p<0.05; however, no differences were observed
across participant race, F(5, 94) ¼0.96, ns. All subsequent univariate analyses were
corrected for alpha inflation by utilizing a modified criterion of p<0.01. Univariate
analyses indicated a significant CRE for both White (Cohen’s d: distinctiveness ¼0.26,
memorability ¼0.35, familiarity ¼0.61, likability ¼0.36, and attractiveness ¼0.58) and
Black (Cohen’s d: distinctiveness ¼0.77, memorability ¼0.70, familiarity ¼0.30,
likability ¼0.44, and attractiveness ¼0.38) participants on each of the five rating
scales, Fs(1, 96) >12.30, 0.13 <MSEs<0.30, ps<0.001. Table 2 displays the pattern
of results indicating that own-race faces were perceived as yielding greater memorability
and greater familiarity. Although effect size comparisons indicate some variation in
the magnitude of the CRE for memorability and familiarity ratings, there were no
statistical differences in the effects, Zs<0.45, ns. (It should be noted that blocking faces
by race did not significantly influence participants’ perceptual ratings either as a main
effect, F(15, 270) ¼0.97, ns, or as an interaction with the observed CRE,
F(15, 270) ¼0.28, ns.)
554 C. A. Meissner et al.
Copyright #2005 John Wiley & Sons, Ltd. Appl. Cognit. Psychol. 19: 545–567 (2005)
Interracial agreement in ratings across race of face
We also assessed whether participants of both races would agree as to rankings of the full
set of faces on the five scales. Correlations were calculated across the 80 faces between the
mean ratings given by Whites and by Blacks. With the exception of the familiarity scale,
White and Black participants demonstrated high general agreement in their ratings of both
races. For White faces, between-race rs(80) were: memorability ¼0.75, p<0.001;
distinctiveness ¼0.69, p<0.001; attractiveness ¼0.87, p<0.001; likability ¼0.83,
p<0.001; and familiarity ¼0.52, p<0.001. For Black faces, between-race rs(80) were:
memorability ¼0.88, p<0.001; distinctiveness ¼0.87, p<0.001; attractiveness ¼0.82,
p<0.001; likability ¼0.66, p<0.001; and familiarity ¼0.24, p<0.05.
Principal components analysis of own- and other-race face ratings
In an attempt to replicate the results of earlier studies with regard to the component
structure of rated distinctiveness, a principal components analysis, with Varimax rotation,
was conducted on the ratings for own- and other-race faces, separately. As displayed in
Table 3, two components with Eigenvalues >1.00 emerged from the analyses, consistent
with previous research (Vokey & Read, 1992). Given our attempts to replicate previous
research, a rather stringent criterion loading of 0.50 was established for evaluating the
contribution of each rating scale. The two components for each cell appeared to converge
in generating two independent components, which have been labelled by previous
researchers as memorability (memorability and distinctiveness) and familiarity (attrac-
tiveness, likability, and familiarity). However, counter to the findings of Vokey and Read
(1992), rated distinctiveness failed to crossload on the two components. Similar principal
component analyses were also conducted for White and Black faces separately as rated by
Table 2. Ratings of distinctiveness, memorability, familiarity, likability, and attractiveness with
reference to own- and other-race faces (Experiment 1)
Own-race faces Other-race faces
MSD M SD
Distinctiveness 4.83 0.96 4.34 0.92
Memorability 5.06 0.84 4.59 0.88
Familiarity 2.12 0.82 1.76 0.80
Likability 3.73 0.58 3.49 0.60
Attractiveness 2.39 0.70 2.06 0.70
Table 3. Exploratory principal component loadings for ratings of distinctiveness, memorability,
familiarity, likability, and attractiveness and across own- and other-race faces (Experiment 1)
Own-race faces Other-race faces
Memorability Familiarity Memorability Familiarity
Distinctiveness 0.98 0.07 0.97 0.09
Memorability 0.98 0.04 0.98 0.04
Familiarity 0.15 0.84 0.22 0.87
Likability 0.09 0.83 0.16 0.84
Attractiveness 0.02 0.91 0.12 0.78
Proportion of variance 39% 45% 35% 47%
Bold values refer to primary perceptual dimensions for each component.
Cross-race effect 555
Copyright #2005 John Wiley & Sons, Ltd. Appl. Cognit. Psychol. 19: 545–567 (2005)
White and Black participants. Although not displayed, results replicated those of the
combined analyses in each of the four cells of the design.
Discussion
Taken together, the results of Experiment 1 indicated that own-race faces were perceived
as exuding greater perceptual memorability and familiarity than other-race faces (see Table
2). Such a pattern of effects is consistent with encoding-based manipulations in the dual-
process literature demonstrating that superior encoding leads to increases in both
recollection and familiarity-based responding. The greater familiarity that was perceived
for own-race faces may also be indicative of configural processes that have been shown to
operate in own-race face recognition. While the CRE has been demonstrated with
perceptual tasks (see Lindsay, Jack, & Christian, 1991), it remains to be seen whether
the current pattern of results might also be responsible for the recognition memory deficits
observed when responding to other-race faces.
Across the ratings of Black and White faces there was also a surprising degree of
interracial agreement, with the exception of the familiarity scale. However, given the low
degree of interracial contact that generally exists, such an absence of agreement on the
familiarity scale (which asked participants whether the face was ‘confusable with someone
you know’) may reasonably be understood. With regard to the representation of rated
distinctiveness, both own- and other-race faces generally yielded the same component
structure involving memorability and familiarity (see Table 3). Our results also replicated
the pattern observed by O’Toole et al. (1994) in which rated distinctiveness (or typicality)
failed to crossload on the familiarity component.
EXPERIMENT 2: DUAL-PROCESS ESTIMATION FOR
RECOGNITION OF OWN- AND OTHER-RACE FACES
Based upon the social-cognitive explanations proposed by previous researchers (see
Meissner & Brigham, 2001; Sporer, 2001), it is possible that one of three general patterns
may emerge when a dual-process perspective is applied to the CRE (see Table 1). In the
following study, we tested dual-process predictions of the CRE by assessing the contribu-
tions of recollection and familiarity to performance on own- and other-race faces in a
recognition memory paradigm. First, we took an experiential approach by asking partici-
pants to provide phenomenological judgments of recognition using Remember, Know, and
Guess alternatives (cf. Gardiner & Richardson-Klavehn, 2000). Second, using a stimulus-
based approach (cf. Mantyla, 1997), perceptual ratings of memorability and familiarity
from Experiment 1 were used to predict recognition performance on own- and other-race
faces in Experiment 2. Taken together, it was believed that these two approaches would
converge to isolate the contributions of recollection and familiarity in the CRE paradigm,
and thereby distinguish between competing social-cognitive accounts of the phenomenon.
Method
Participants
The sample consisted of 32 White and 32 Black students from an Introductory Psychology
course. Participants were tested in mixed-race groups ranging in size from two to five. The
mean age of the participants was 19 years and 77% were female.
556 C. A. Meissner et al.
Copyright #2005 John Wiley & Sons, Ltd. Appl. Cognit. Psychol. 19: 545–567 (2005)
Materials
The 80 White and 80 Black male photographs used in Experiment 1 were also used here.
Photographs presented at study involved the targets wearing their everyday clothes,
standing in front of a grey background, in a full frontal, smiling pose. Photographs
presented at test involved the targets wearing a burgundy-coloured sweatshirt, standing in
front of a grey background, in a full frontal, non-smiling pose. These changes in photos
were implemented such that the only remaining variability across study-test involved the
invariant features of the target’s face. Photographs were projected onto a screen to
approximately life size, subtending approximately 10 14 visual degrees.
Design
A2(22) mixed repeated measures design was used, with race of participant as the
between-subject factor, race of face as the within-subject factor, and a repeated test of
participants’ recognition for own- and other-race faces. In addition, participants responded
to characteristics of their memory for a given face at test via the Remember-Know-Guess
procedure. Each testing involved a mixture of own- and other-race faces, and the order of
presentation for faces was randomized and counterbalanced across participant groups at
both study and test, with the restriction that no more than three faces of the same race were
presented in sequential order.
Procedure
Upon entering the lab, participants were familiarized with the general recognition task.
For each study phase, participants were presented with 20 Black and 20 White faces for 3 s
each. They were instructed to study each face for a later memory test. Following the study
phase, participants were given a 5 min distractor task involving a digit search puzzle.
Testing resumed thereafter and involved the presentation of 40 Black and 40 White faces
(40 New, 40 Old) in a randomized sequential order for 7 s each. Participants were asked to
indicate whether the face was ‘New’ or ‘Old’, and to provide an estimate of their
confidence on a 6-point Likert scale with end-point labels of ‘sure new’ and ‘sure old’.
If they reported that an item was Old, they were also asked to describe their memory for
the face using the Remember, Know, and Guess responses provided. Following comple-
tion of the first recognition task, participants were given a brief break, after which the
study phase for the second recognition task was begun. Procedures of the second
recognition test replicated that of the first with a different set of Black and White faces.
Results
Own- and other-race recognition performance
Table 4 provides estimates of performance on measures of hits, false alarms, discrimina-
tion accuracy (A0), and response criterion (B00
D
). A 2 ( 22) mixed MANOVA assessed
the influence of race of participant, race of face, and repeated testing on the proportion of
hit and false alarm responses. A significant multivariate CRE, or race of participant race
of face interaction, was obtained, F(2, 61) ¼18.66, p<0.001. A significant multivariate
main effect of repeated testing was also present, F(2, 61) ¼20.22, p<0.001. All other
effects were non-significant, Fs(2, 61) <1.36, ps>0.25. Univariate tests on the CRE
indicated a significant race of participant race of face interaction on both hit,
F(1, 62) ¼7.05, MSE ¼0.01, p<0.01, and false alarm responses, F(1, 62) ¼30.42,
MSE ¼0.01, p<0.001, resulting in the ‘mirror effect’ pattern demonstrated across
Cross-race effect 557
Copyright #2005 John Wiley & Sons, Ltd. Appl. Cognit. Psychol. 19: 545–567 (2005)
previous studies (Meissner & Brigham, 2001). Effect sizes were somewhat larger for
White than for Black participants on both hits (Cohen’s d¼0.32 vs. 0.28, respectively)
and false alarms (Cohen’s d¼0.92 vs. 0.61, respectively); however, this variation in effect
size for hits and false alarms was not statistically significant (Zs¼0.08 and 0.49, ps¼0.94
and 0.62, respectively). Univariate tests on the effect of Time were significant for both hit,
F(1, 62) ¼20.60, MSE ¼0.01, p<0.001, and false alarm responses, F(1, 62) ¼30.33,
MSE ¼0.01, p<0.001. The pattern of means increased from Time-1 to Time-2 for both
hit (Ms¼0.64 and 0.69; SDs¼0.12 and 0.13, respectively) and false alarm (Ms¼0.22
and 0.30; SDs¼0.08 and 0.12, respectively) estimates, suggesting a shift to a looser
criterion of responding at Time-2.
A2(22) mixed MANOVA analysis of signal detection estimates of discrimination
accuracy (A0) and response bias (B00
D
) was also conducted. A significant multivariate CRE,
or race of participant race of face interaction, was obtained, F(2, 61) ¼18.36, p<0.001.
In addition, a significant multivariate main effect of repeated testing was also observed,
F(2, 61) ¼21.94, p<0.001. All other effects were non-significant, Fs(2, 61) <1.46,
ps>0.20. Univariate tests on the CRE indicated significant race of participant race of
face interactions on both A0,F(1, 62) ¼30.55, MSE ¼0.01, p<0.001, and B00
D
estimates,
F(1, 62) ¼5.81, MSE ¼0.13, p<0.05. The CRE again was larger for White than for Black
participants on estimates of A0(Cohen’s d¼0.91 vs. 0.60, respectively) and B00
D
(Cohen’s
d¼0.39 vs. 0.21, respectively); however, this variation in effect size for A0and B00
D
was
not statistically significant (Zs¼0.46 and 0.34, ps¼0.64 and 0.73, respectively).
Univariate tests on the repeated measures effect were significant only for B00
D
estimates,
F(1, 62) ¼42.94, MSE ¼0.02, p<0.001. As expected, participants’ response criterion
became significantly less conservative from Time-1 to Time-2 (Ms¼0.31 and 0.01;
SDs¼0.31 and 0.42, respectively).
Reliability of recognition performance for own- and other-race faces
Given the repeated design employed in the present study, we were interested in assessing
the reliability of face recognition for own- and other-race faces, as well as the reliability of
the CRE over time. With regard to participants’ discrimination accuracy (A0) scores,
performance on both own- and other-race faces was significantly reliable, rs(64) ¼0.36
and 0.38, ps<0.01, respectively, but of small-to-moderate size. The consistency of
participants’ response bias (B00
D
) on own- and other-race faces was also significant across
testing sessions, rs(64) ¼0.56 and 0.48, ps<0.001, respectively. Finally, with regard to
the reliability of the CRE itself, differences in own- and other-race performance on A0and
B00
D
were not significantly related over time, rs(64) ¼0.20 and 0.21, ps¼0.11 and 0.09,
respectively. However, the magnitude of this reliability estimate for A0is comparable to
that observed by Slone et al. (2000) in a similar analysis: r(127) ¼0.21, p<0.05.
Table 4. Recognition performance for own- and other-race faces (Experiment 2)
Own-race faces Other-race faces
MSDM SD
Hits 0.69 0.17 0.64 0.17
False alarms 0.22 0.13 0.30 0.14
A00.81 0.10 0.75 0.12
B00
D
0.21 0.42 0.10 0.44
558 C. A. Meissner et al.
Copyright #2005 John Wiley & Sons, Ltd. Appl. Cognit. Psychol. 19: 545–567 (2005)
Remember-know judgments across own- and other-race faces
When participants responded ‘Old’ to a face at test, they were asked to describe their
memory for the target using the Remember-Know-Guess procedure (see Gardiner &
Richardson-Klavehn, 2000). Given the general assumptions of independence governing
the dual-process model (cf. Kelley & Jacoby, 2000), estimates of Know were corrected for
independence for each subject using the independence remember/know procedure (i.e.
K
C
¼K/[1 R]; see Yonelinas & Jacoby, 1995), and this corrected version was used in all
subsequent analyses. Table 5 provides estimates of Remember, Know, and Know
C
judgments on hit and false alarm responses for own- and other-race faces.
A2(22) mixed MANOVA examined the proportion of Remember and Know
C
responses on correct identifications (i.e. ‘hits’) across race of participant, race of face, and
repeated testing. A significant multivariate CRE, or race of participant race of face
interaction, was obtained, F(2, 61) ¼26.93, p<0.001. All other effects were non-
significant, Fs(2, 61) <2.60, ps>0.08. Univariate analyses of the CRE demonstrated a
significant race of participant race of face interaction only on Remember judgments,
F(1, 62) ¼33.72, MSE ¼0.02, p<0.001. Consistent with the recognition data, this CRE
was somewhat larger for White than for Black participants (Cohen’s d¼0.86 vs. 0.60,
respectively); however, the variation in effect sizes was not statistically significant
(Z¼0.38, p¼0.70). Although Know
C
judgments were numerically greater for own-
race faces, this effect failed to reach the conventional level of statistical significance,
F(1, 62) ¼3.14, MSE ¼0.04, p¼0.08. The effect sizes were small, but equivalent, for
White and Black participants (Cohen’s ds¼0.11), and power to detect the effect was quite
limited with the present sample (1 ¼0.16).
Remember-Know
C
judgments for incorrect identifications (i.e. ‘false alarms’) demon-
strated a similar pattern. A 2 ( 22) mixed MANOVA indicated a significant multi-
variate CRE, or race of participant race of face interaction, F(2, 61) ¼3.39, p<0.05. A
significant effect of repeated testing was also present, F(2, 61) ¼21.94, p<0.05. All other
effects were non-significant, Fs(2, 61) <1.89, ps>0.15. Univariate analyses of the CRE
demonstrated a significant race of participant race of face interaction that was isolated to
Remember judgments, F(1, 62) ¼6.88, MSE ¼0.04, p<0.01. This CRE was somewhat
larger for Black than for White participants (Cohen’s d¼0.42 and 0.27, respectively),
although this variation in effect sizes was not statistically significant (Z¼0.30, p¼0.76).
An alternative manner to assess the influence of recollection and familiarity is to
compute estimates of A0for judgments of both Remember and Know, and subsequently to
assess differences in performance across own- versus other-race faces. A 2 ( 22)
Table 5. Estimates of remember and know for recognition of own- and other-race faces
(Experiment 2)
Hits False alarms
Own-race Other-race Own-race Other-race
faces faces faces faces
MSDMSDMSDM SD
Remember 0.57 0.23 0.44 0.26 0.15 0.23 0.23 0.22
Know 0.28 0.13 0.34 0.15 0.44 0.28 0.37 0.24
Know
C
0.65 0.31 0.60 0.32 0.51 0.44 0.52 0.41
Cross-race effect 559
Copyright #2005 John Wiley & Sons, Ltd. Appl. Cognit. Psychol. 19: 545–567 (2005)
mixed MA NOVA was again used to estimate differential performance. Results indicated a
significant multivariate CRE, or race of participant race of face interaction,
F(2, 61) ¼14.32, p<0.001. All other effects were non-significant, Fs(2, 61) <1.43,
ps>0.20. Univariate analyses of the CRE demonstrated a significant race of partici-
pant race of face interaction that was isolated to A0for Remember judgments,
F(1, 62) ¼28.86, MSE ¼0.02, p<0.001. The effect was comparable for White and Black
participants (Cohen’s d¼0.75 vs. 0.77, respectively), and demonstrated that A0for
Remember judgments was greater for own-race faces (M¼0.78, SD ¼0.12) than for
other-race faces (M¼0.68, SD ¼0.14).
Memorability-familiarity components and recognition of own- and other-race faces
Perceptual ratings of memorability and familiarity in Experiment 1 indicated that own-
race faces were perceived as exuding greater memorability and greater familiarity when
compared with other-race faces. While the CRE has been demonstrated in perceptual tasks
(cf. Lindsay et al., 1991), we were interested in whether such stimulus attributes might
relate to participants’ ability to recognize faces, and further whether such attributes might
capture the same essence of episodic recollection and fluent familiarity as that described in
the dual-process literature. Mantyla (1997) has previously demonstrated that the effects of
rated distinctiveness appear to be related to judgments of recollection from a stimulus-
level analysis. As a manner in which to confirm the role of recollection processes observed
in Experiment 2, we conducted a stimulus-level correlational analysis linking the
components of memorability and familiarity collected in Experiment 1 to the estimates
of recognition for each face collected in Experiment 2. Similar to the regression analyses
conducted by Vokey and Read (1992), we computed partial correlations between
estimates of memorability and familiarity, and recognition performance as measured by
proportion of hits and false alarms, as well as the signal detection parameters A0and B00
D
.
This pattern of partial correlations for own- and other-race faces replicated the findings of
Vokey and Read, particularly with regard to hit and false alarm responses. As displayed in
Table 6, while memorability was significantly related to hit responses, both memorability
and familiarity combined to predict false alarm responses. Although Vokey and Read
found that both components combined to predict discrimination accuracy (A0), the
familiarity component in our analyses failed to reach the conventional level of significance
(ps<0.10). Finally, while neither memorability nor familiarity significantly predicted
own-race response criterion, familiarity did contribute to other-race response criterion.
When partial correlations were tested for any differential CREs, results replicated
findings from the Remember-Know
C
responses presented previously. In particular,
Table 6. Partial correlations observed between rated estimates of memorability/familiarity in
Experiment 1 and estimates of recognition accuracy in Experiment 2
Own-race faces Other-race faces
Memorability Familiarity Memorability Familiarity
Hits 0.42* 0.07 0.22* 0.11
False alarms 0.38* 0.19* 0.20* 0.23*
A00.42* 0.14 0.28* 0.15
B00
D
0.01 0.11 0.08 0.21*
Note:*p<0.01.
560 C. A. Meissner et al.
Copyright #2005 John Wiley & Sons, Ltd. Appl. Cognit. Psychol. 19: 545–567 (2005)
memorability was more highly correlated with responding for own-race than for other-race
faces in both hits, Z¼2.31, p<0.05, and false alarms, Z¼2.07, p<0.05. A similar
pattern was observed for memorability and estimates of A0; however, the difference in the
size of the relationship between own- and other-race faces failed to reach the conventional
level of significance, Z¼1.90, p¼0.06. Overall, familiarity was not differentially
diagnostic across own- and other-race faces, Zs<1.50, ps>0.14.
Discussion
Experiment 2 examined the CRE from a dual-process perspective by way of several
approaches. First, an experiential approach was taken by asking participants to provide
phenomenological judgments of recognition using Remember, Know, and Guess alter-
natives (cf. Gardiner & Richardson-Klavehn, 2000). Our results demonstrated that
memory for own-race faces involved greater recollection of information from the prior
study episode when compared with performance on other-race faces. While own-race
faces also showed a numerical increase in familiarity-based responding when compared
with other-race faces, this difference was not statistically significant. Second, using a
stimulus-based approach (Mantyla, 1997; O’Toole et al., 1994; Vokey & Read, 1992),
perceptual ratings of memorability and familiarity from Experiment 1 were used to predict
recognition performance on own- and other-race faces in Experiment 2. While the overall
patterns of correlation between memorability and familiarity estimates and recognition
performance replicated those observed previously by Vokey and Read, our results further
demonstrated a differential effect of memorability on the recognition of own- versus
other-race faces. Consistent with prior analyses, this stimulus-level analysis indicated that
memorability significantly influenced both hit and false alarm responses to own-race
faces, whereas familiarity failed to differ as a function of race of face. Taken together, the
results of Experiment 2 appear to be consistent with the effects of attentional and
encoding-based mechanisms, but are largely inconsistent with the possible role of
stereotyping or configural-featural information in differentiating performance on own-
and other-race faces.
GENERAL DISCUSSION
Although 30 years of research has examined this phenomenon, no one theoretical account
has generally been accepted (see Meissner & Brigham, 2001; Sporer, 2001). The current
studies attempted to inform this debate by framing the CRE within the constraints of a
dual-process framework. In general, such theories (Gardiner & Richardson-Klavehn,
2000; Kelley & Jacoby, 2000; Mandler, 1980; Yonelinas, 2002) have typically distin-
guished between memory processes involving conscious-level conceptual information that
is elaboratively encoded and those involving fluent, perceptually-based information that is
encoded in an automatic, non-conscious manner. It was believed that identifying the
contributions of recollection and familiarity to the CRE would provide insights regarding
possible social-cognitive mechanisms.
Overall, the results of our studies have demonstrated that the CRE appears to be due to a
greater reliance upon recollection for own-race faces, in which individuals qualitatively
encode more information about own-race faces, creating a more diagnostic representation
for subsequent identification. While familiarity was also greater for own-race faces in
Cross-race effect 561
Copyright #2005 John Wiley & Sons, Ltd. Appl. Cognit. Psychol. 19: 545–567 (2005)
several analyses, this effect was rather weak in comparison. Such a pattern of effects is
quite consistent with attentional and encoding-based manipulations in the dual-process
paradigm (see Yonelinas, 2002), including divided-attention, generation, and semantic
encoding tasks. In addition, the role of recollection is consistent with the differential
effects of distinctiveness observed in the CRE paradigm (Chiroro & Valentine, 1995;
Mantyla, 1997). If the effect of categorization for other-race faces was considered as an
attentional diversion to encoding (Levin, 1996, 2000; MacLin & Malpass, 2001, 2003),
then the current pattern of effects would also support such a mechanism. In contrast, the
primary role of recollection in the CRE does not appear to provide support for differential
reliance upon configural versus featural information (Yonelinas et al., 1999) or for the role
of stereotyping other-race faces.
The phenomenology of skilled perceptual-memory
More generally, the current findings may extend a dual-process interpretation to other
perceptual learning phenomena. In particular, the role of recollection in differentiating
performance on own- and other-race faces further validates the influence of conscious-
level, conceptual processes in skilled perceptual-memory, a proposition set forth early in
the skilled memory research of Chase and Simon (1973). This finding, however, does not
negate the potential role of a more pure perceptual advantage in the performance of skilled
participants. For example, Reingold and colleagues (Reingold, Charness, Pomplun, &
Stampe, 2001) have observed a perceptual-encoding advantage for chess masters invol-
ving a larger visual span (compared with novice players) when responding in a check-
detection paradigm. Interestingly, this perceptual advantage was observed only when
chess masters were provided a ‘meaningful’ board arrangement. In contrast, when
participants were provided a ‘scrambled’ arrangement of pieces, the differences in visual
span across the skill levels were negated. Thus, it appears that conceptual and perceptual
processes may act in concert with one another to provide performance advantages in
perceptual-memory tasks (see also, Charness, Reingold, Pomplun, & Stampe, 2001).
Similar to that of other studies in the skilled memory literature, the CRE demonstrated
in the present research provides only a ‘snapshot’ of performance differences in memory
for own- and other-race faces. Further research is required to better understand the changes
in phenomenological responding that develop throughout the skill acquisition process
leading to successful (and replicable) differentiation. Reder, Angstadt, Cary, Erickson, and
Ayers (2002) have recently demonstrated such a multi-trial training paradigm to examine
frequency-based effects in recognition. Gauthier, Tarr, and colleagues have also used a
training paradigm to explore the development of object recognition expertise and its
relation to the processing of human faces (Gauthier, Williams, Tarr, & Tanaka, 1998;
Rossion, Gauthier, Goffaux, Tarr, & Crommelinck, 2002). Linking such a paradigm with
the assessment of both performance-based and phenomenological responses should permit
researchers to gain a better understanding of the mechanisms underlying the perceptual
learning process.
Applied considerations
The present experiments, in concert with other recent research, lead us to the proposition
that the CRE occurs because people encode more qualitative information about own-race
faces, information that is diagnostic for subsequent identification situations. If the CRE
562 C. A. Meissner et al.
Copyright #2005 John Wiley & Sons, Ltd. Appl. Cognit. Psychol. 19: 545–567 (2005)
were seen as an encoding-based phenomenon, then factors present at encoding would be
particularly important in determining the magnitude of the CRE that would be likely to
occur. Consistent with this perspective, our recent meta-analysis (Meissner & Brigham,
2001) found that an encoding-based factor, study time, was a significant moderator of the
CREnamely, the CRE was stronger, largely due to higher false alarm rates, in
conditions that employed shorter study times. Given this perspective, other attentional
factors, such as the ‘weapon focus’ effect (Steblay, 1992), may also be found to exacerbate
the CRE. The role of encoding-based factors would seem to correspond closely to a
perceiver’s ‘opportunity to observe’ the target’s face, which was listed by the US Supreme
Court in Neil v. Biggers (1972) as one of five factors to be taken into account in evaluating
the likely accuracy of an eyewitness’s identification (Brigham, Wasserman, & Meissner,
1999). So for this factor (but not several others), the Court’s ‘educated guess’ appears to
have been supported by subsequent research.
Another important factor in identification situations is the confidence that an eyewitness
expresses in his/her identification. A witness’s ‘degree of certainty’ was also one of the five
factors listed in Neil v. Biggers, and subsequent research has shown that mock jurors are
indeed strongly influenced by the degree of confidence/certainty that a witness expresses
(e.g. Cutler, Dexter, & Penrod, 1990). However, research examining the utility of
confidence as a postdictor of accuracy has generally demonstrated a weak relationship
between degree of certainty and identification accuracy (see meta-analyses by Bothwell,
Deffenbacher, & Brigham, 1987; Sporer, Penrod, Read, & Cutler, 1995). Yet, there do
appear to be some conditions in which a stronger relationship may exist. For example,
when conditions vary widely across witnesses, confidence has been shown to be associated
with identification accuracy (Lindsay, Nilson, & Read, 2000). Furthermore, initial
judgments made with very high confidence have been shown to be quite diagnostic of
witness accuracy (Juslin, Olson, & Winman, 1996). In contrast, Wright, Boyd, and
Tredoux (2001) have observed that the confidence-accuracy relationship is significantly
weaker for other-race than for own-race identifications. The current study extends this
finding by demonstrating that individuals experience a greater proportion of false
recollections for other-race facesnamely, incorrect identifications that are made with
high confidence. This differential rate of false recollections and the lack of diagnosticity
for other-race confidence ratings would seem to be valuable information for those working
in the justice system who must decide how much credence to assign a disputed
identification.
As far as the reliability of the CRE, this has only been assessed in two published studies,
to our knowledge— Slone et al. (2000) and the present Experiment 2. On both occasions,
the effect showed only a small degree of test-retest reliability: r(127) ¼0.21 in Slone et al.:
r(64) ¼0.20 in the current study. This low level of reliability, coupled with the absence of
relationships to individual measures of memory or potentially relevant attitudes (except
for degree of interracial contact for Whites; see Slone et al., 2000), indicates that there is
no measure at present that would be forensically useful in predicting which individuals are
most likely to manifest a strong CRE in face identification. It remains to be seen whether
other potentially useful individual-difference variables can be identified.
Conclusions
The current analysis interpreted the CRE in terms of a dual-process framework, and attem-
pted to distinguish between competing social-cognitive accounts of the phenomenon.
Cross-race effect 563
Copyright #2005 John Wiley & Sons, Ltd. Appl. Cognit. Psychol. 19: 545–567 (2005)
Taken together, our findings suggested that the CRE occurs as a result of superior
encoding-based processing of own-race faces, particularly with regard to the greater
qualitative selection of facial information and the allocation of sufficient attentional and
cognitive resources to their storage in memory. In contrast, the role of racial stereotyping
and the potential for a configural-featural encoding distinction between own- and other-
race faces was not supported by these studies. From a practical standpoint, identifying the
CRE as an encoding-based phenomenon provides insight to factors present at encoding
(e.g. study time or distractions to attention such as ‘weapon focus’) that may be important
in determining the potential magnitude of the CRE in everyday practice.
ACKNOWLEDGEMENTS
The authors would like to thank Vicki Bruce, Steve Lindsay, Otto MacLin, Roy Malpass,
and Sigi Sporer for helpful comments on early drafts of this article. This research was
supported by a grant from the National Science Foundation to the second author.
REFERENCES
Bartlett, J. C., Strater, L., & Fulton, A. (1991). False recency and false fame of faces in young
adulthood and old age. Memory & Cognition, 19(2), 177–188.
Bastin, C. & van der Linden, M. (2003). The contribution of recollection and familiarity to
recognition memory: a study of the effects of test format and aging. Neuropsychology, 17(1),
14–24.
Bothwell, R. K., Deffenbacher, K. A., & Brigham, J. C. (1987). Correlation of eyewitness accuracy
and confidence: optimality hypothesis revisited. Journal of Applied Psychology, 72, 691–695.
Bower, G. H., & Karlin, M. B. (1974). Depth of processing pictures of faces and recognition
memory. Journal of Experimental Psychology, 103, 751–757.
Brigham, J. C. (1990). Target person distinctiveness and attractiveness as moderator variables in the
confidence-accuracy relationship in eyewitness identifications. Basic & Applied Social Psychol-
ogy, 11, 101–115.
Brigham, J. C., Wasserman, A. W., & Meissner, C. A. (1999). Disputed eyewitness identification
evidence: important legal and scientific issues. Court Review, 36, 12–25.
Buckhout, R., & Regan, S. (1988). Explorations in research on the other-race effect in face
recognition. In M. M. Gruneberg, P. E. Morris, & R. N. Sykes (Eds.), Practical aspects of
memory: Current research and issues: Memory in everyday life (Vol. 1, pp. 40–46). New York,
NY: John Wiley & Sons.
Burgess, M. C. R., & Weaver, G. E. (2003). Interest and attention in facial recognition. Perceptual &
Motor Skills, 96, 467–480.
Byatt, G., & Rhodes, G. (1998). Recognition of own-race and other-race caricatures: implications for
models of face recognition. Vision Research, 38, 2455–2468.
Chance, J. E., & Goldstein, A. G. (1981). Depth of processing in response to own- and other-race
faces. Personality & Social Psychology Bulletin, 7, 475–480.
Chance, J. E., & Goldstein, A. G. (1996). The other-race effect and eyewitness identification. In S. L.
Sporer, R. S. Malpass, & G. Koehnken (Eds.), Psychological issues in eyewitness identification
(pp. 153–176). Mahwah, NJ: Erlbaum.
Charness, N., Reingold, E. M., Pomplun, M., & Stampe, D. M. (2001). The perceptual aspect of
skilled performance in chess: evidence from eye movements. Memory & Cognition, 29, 1146–
1152.
Chase, W. G., & Simon, H. A. (1973). Perception in chess. Cognitive Psychology, 4, 55–81.
Chiroro, & Valentine, T. (1995). An investigation of the contact hypothesis of the own-race bias in
face recognition. Quarterly Journal of Experimental Psychology, 48A, 879–894.
564 C. A. Meissner et al.
Copyright #2005 John Wiley & Sons, Ltd. Appl. Cognit. Psychol. 19: 545–567 (2005)
Craik, F. I. M., & Lockhart, R. S. (1972). Levels of processing: a framework for memory research.
Journal of Verbal Learning & Verbal Behavior, 11, 671–684.
Cutler, B. L., Dexter, H. R., & Penrod, S. D. (1990). Nonadversarial methods for sensitizing jurors to
eyewitness evidence. Journal of Applied Social Psychology, 20, 1197–1207.
Dasgupta, N., McGhee, D. E., Greenwald, A. G., & Banaji, M. R. (2000). Automatic preference for
White Americans: eliminating the familiarity explanation. Journal of Experimental Social
Psychology, 36(3), 316–328.
Deffenbacher, K. A., Johanson, J., Vetter, T., & O’Toole, A. J. (2000). The face typicality-
recognizability relationship: encoding or retrieval locus? Memory & Cognition, 28, 1173–1182.
Devine, P. G., & Malpass, R. S. (1985). Orienting strategies in differential face recognition.
Personality and Social Psychology Bulletin, 11, 33–40.
Diamond, R., & Carey, S. (1986). Why faces are and are not special: an effect of expertise. Journal of
Experimental Psychology: General, 115, 107–117.
Elliott, E. S., Wills, E. J., & Goldstein, A. G. (1973). The effects of discrimination training on the
recognition of White and Oriental faces. Bulletin of the Psychonomic Society, 2, 71–73.
Fallshore, M., & Schooler, J. W. (1995). Verbal vulnerability of perceptual expertise. Journal of
Experimental Psychology: Learning, Memory, & Cognition, 21, 1608–1623.
Farah, M. J., Wilson, K. D., Drain, M., & Tanaka, J. N. (1998). What is ‘special’ about face
perception? Psychological Review, 105, 482–498.
Gardiner, J. M. (1988). Functional aspects of recollective experience. Memory & Cognition, 16,
309–313.
Gardiner, J. M., & Richardson-Klavehn, A. (2000). Remembering and knowing. In E. Tulving,
F. Craik, & I. M. Fergus (Eds.), The Oxford handbook of memory (pp. 229–244). London: Oxford
University Press.
Gauthier, I., Williams, P., Tarr, M. J., & Tanaka, J. (1998). Training ‘greeble’ experts: a framework
for studying expert object recognition processes. Vision Research, 38, 2401–2428.
Goldstein, A. G., & Chance, J. E. (1971). Visual recognition memory for complex configurations.
Perception & Psychophysics, 9, 237–241.
Goldstein, A. G., & Chance, J. E. (1985). Effects of training on Japanese faces recognition: reduction
of the other-race effect. Bulletin of the Psychonomic Society, 23, 211–214.
Jacoby, L. L. (1991). A process dissociation framework: separating automatic from intentional uses
of memory. Journal of Memory & Language, 30, 513–541.
Juslin, P., Olson, N., & Winman, A. (1996). Calibration and diagnosticity of confidence in
eyewitness identification: comments on what can and cannot be inferred from a low confi-
dence-accuracy correlation. Journal of Experimental Psychology: Learning, Memory, and
Cognition, 5, 1304–1316.
Kelley, C. M., & Jacoby, L. L. (2000). Recollection and familiarity: process-dissociation. In
E. Tulving, F. Craik, & I. M. Fergus (Eds.), The Oxford handbook of memory (pp. 215–228).
London: Oxford University Press.
Lavrakas, P. J., Buri, J. R., & Mayzner, M. S. (1976). A perspective on the recognition of other race
faces. Perception & Psychophysics, 20, 475–481.
Levin, D. T. (1996). Classifying faces by race: the structure of face categories. Journal of
Experimental Psychology: Learning, Memory, & Cognition, 22, 1364–1382.
Levin, D. T. (2000). Race as a visual feature: using visual search and perceptual discrimination tasks
to understand face categories and the cross-race recognition deficit. Journal of Experimental
Psychology: General, 129, 559–574.
Lindsay, D. S., Jack, P. C., & Christian, M. A. (1991). Other-race face perception. Journal of Applied
Psychology, 76, 587–589.
Lindsay, D. S., Nilsen, E., & Read, J. D. (2000). Witnessing-condition heterogeneity and witnesses’
versus investigators’ confidence in the accuracy of witnesses’ identification decisions. Law and
Human Behavior, 24, 685–697.
MacLin, O. H., & Malpass, R. S. (2001). Racial categorization of faces: the ambiguous race face
effect. Psychology, Public Policy, & Law, 7, 98–118.
MacLin, O. H., & Malpass, R. S. (2003). The ambiguous-race face illusion. Perception, 32, 249–252.
Malpass, R. S. (1981). Training in face recognition. In G. Davies, H. Ellis, & J. Shepherd (Eds.),
Perceiving and remembering faces (pp. 271–285). London: Academic Press.
Cross-race effect 565
Copyright #2005 John Wiley & Sons, Ltd. Appl. Cognit. Psychol. 19: 545–567 (2005)
Malpass, R. S. (1990). An excursion into utilitarian analyses, with side trips. Behavioral Science
Research, 24, 1–15.
Malpass, R. S., Lavigueur, H., & Weldon, D. E. (1973). Verbal and visual training in face
recognition. Perception & Psychophysics, 14, 283–292.
Mandler, G. (1980). Recognizing: the judgment of previous occurrence. Psychological Review, 87,
252–271.
Mantyla, T. (1993). Knowing but not remembering: adult age differences in recollection experience.
Memory & Cognition, 21, 379–388.
Mantyla, T. (1997). Recollection of faces: remembering differences and knowing similarities.
Journal of Experimental Psychology: Learning, Memory, & Cognition, 23(5), 1203–1216.
Mantyla, T., & Cornoldi, C. (2002). Remembering changes: repetition effects in face recollection.
Acta Psychologica, 109(1), 95–105.
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, & Law, 7, 3–35.
Monteith, M. J., Ashburn-Nardo, L., Voils, C. I., & Czopp, A. (2002). Putting the brakes on
prejudice: on the development of operation cues for control. Journal of Personality & Social
Psychology, 83(5), 1029–1050.
Morris, P. E., & Wickham, L. H. V. (2001). Typicality and face recognition: a critical re-evaluation of
the two factor theory. Quarterly Journal of Experimental Psychology: Human Experimental
Psychology, 54A, 863–877.
Mullen, B., & Hu, L. (1989). Perceptions of ingroup and outgroup variability: a meta-analytic
integration. Basic & Applied Social Psychology, 10, 233–252.
Neil v. Biggers. (1972). 409 U.S. 188.
O’Toole, A. J., Deffenbacher, K. A., Valentin, D., & Abdi, H. (1994). Structural aspects of face
recognition and the other-race effect. Memory & Cognition, 22, 208–224.
Paller, K. A., Bozic, V. S., Ranganath, C., Grabowecky, M., & Yamada, S. (1999). Brain waves
following remembered faces index conscious recollection. Cognitive Brain Research, 7(4), 519–
534.
Parkin, A. J., Gardiner, J. M., & Rosser, R. (1995). Functional aspects of recollective experience in
face recognition. Consciousness & Cognition: An International Journal, 4(4), 387–398.
Payne, B. K. (2001). Prejudice and perception: the role of automatic and controlled processes in
misperceiving a weapon. Journal of Personality & Social Psychology, 81(2), 181–192.
Payne, B. K., Lambert, A. J., & Jacoby, L. L. (2002). Best laid plans: effects of goals on accessibility
bias and cognitive control in race-based misperceptions of weapons. Journal of Experimental
Social Psychology, 38(4), 384–396.
Rajaram, S. (1993). Remembering and knowing: two means of access to the personal past. Memory
& Cognition, 21, 89–102.
Rajaram, S. (1996). Perceptual effects on remembering: recollective processes in picture recognition
memory. Journal of Experimental Psychology: Learning, Memory, & Cognition, 22, 365–377.
Reder, L. M., Angstadt, P., Cary, M., Erickson, M. A., & Ayers, M. S. (2002). A reexamination of
stimulus-frequency effects in recognition: two mirrors for low- and high-frequency pseudowords.
Journal of Experimental Psychology: Learning, Memory, & Cognition, 28, 138–152.
Reingold, E. M., Charness, N., Pomplun, M., & Stampe, D. M. (2001). Visual span in expert chess
players: evidence from eye movements. Psychological Science, 12, 48–55.
Reinitz, M. T., Morrissey, J., & Demb, J. (1994). Role of attention in face encoding. Journal of
Experimental Psychology: Learning, Memory, & Cognition, 20(1), 161–168.
Rhodes, G., Brake, S., Taylor, K., & Tan, S. (1989). Expertise and configural coding in face
recognition. British Journal of Psychology, 80, 313–331.
Rodin, M. J. (1987). Who is memorable to whom: a study of cognitive disregard. Social Cognition, 5,
144–165.
Rossion, B., Gauthier, I., Goffaux, V., Tarr, M. J., & Crommelinck, M. (2002). Expertise training
with novel objects leads to left-lateralized facelike electrophysiological responses. Psychological
Science, 13, 250–257.
Sensening, J., Jones, R. A., & Varney, L. (1973). Inspection of faces of own and other race as a
function of subjects’ prejudice. Representative Research in Social Psychology, 4(2), 85–92.
Shepherd, J. W., Gibling, F., & Ellis, H. D. (1991). The effects of distinctiveness, presentation time
and delay on face recognition. European Journal of Cognitive Psychology, 3, 137–145.
566 C. A. Meissner et al.
Copyright #2005 John Wiley & Sons, Ltd. Appl. Cognit. Psychol. 19: 545–567 (2005)
Slone, A. E., Brigham, J. C., & Meissner, C. A. (2000). Social and cognitive factors affecting the
own-race bias in Whites. Basic & Applied Social Psychology, 22, 71–84.
Sporer, S. L. (1991). Encoding strategies and the recognition of human faces. Journal of Experi-
mental Psychology: Human Learning, Memory, & Cognition, 17, 323–333.
Sporer, S. L. (2001). Recognizing faces of other ethnic groups: an integration of theories.
Psychology, Public Policy, & Law, 7, 36–97.
Sporer, S. L., Penrod, S. D., Read, J. D., & Cutler, B. L. (1995). Choosing, confidence, and accuracy:
a meta-analysis of the confidence-accuracy relation in eyewitness identification studies. Psycho-
logical Bulletin, 118, 315–327.
Steblay, N. M. (1992). A meta-analytic review of the weapon focus effect. Law & Human Behavior,
16, 413–424.
Tulving, E. (1983). Elements of episodic memory. London: Oxford University Press.
Valentine, T. (1991). A unified account of the effects of distinctiveness, inversion and race on face
recognition. Quarterly Journal of Experimental Psychology, 43A, 161–204.
Valentine, T., & Bruce, V. (1986). The effect of race, inversion and encoding activity upon face
recognition. Acta Psychologica, 61, 259–273.
Valentine, T., & Endo, M. (1992). Towards an exemplar model of face processing: the effects of race
and distinctiveness. Quarterly Journal of Experimental Psychology, 44A, 671–703.
Vokey, J. R., & Read, J. D. (1988). Typicality, familiarity, and the recognition of male and female
faces. Canadian Journal of Psychology, 42, 489–495.
Vokey, J. R., & Read, J. D. (1992). Familiarity, memorability, and the effect of typicality on the
recognition of faces. Memory & Cognition, 20, 291–302.
Winograd, E. (1981). Elaboration and distinctiveness in memory for faces. Journal of Experimental
Psychology: Human Learning and Memory, 7, 181–190.
Wright, D. B., & Stroud, J. N. (2002). Age differences in lineup identification accuracy: people are
better with their own age. Law & Human Behavior, 26(6), 641–654.
Wright, D. B., Boyd, C. E., & Tredoux, C. G. (2001). A field study of own-race bias in South Africa
and England. Psychology, Public Policy, & Law, 7, 119–133.
Wright, D. B., Boyd, C. E., & Tredoux, C. G. (2003). Inter-racial contact and the own-race bias for
face recognition in South Africa and England. Applied Cognitive Psychology, 17(3), 365–373.
DOI: 10.1002/acp.898
Yin, R. K. (1969). Looking at upside-down faces. Journal of Experimental Psychology, 81, 141–145.
Yonelinas, A. P. (2002). The nature of recollection and familiarity: a review of 30 years of research.
Journal of Memory & Language, 46, 441–517.
Yonelinas, A. P., & Jacoby, L. L. (1995). The relation between remembering and knowing as bases
for recognition: effects of size congruency. Journal of Memory & Language, 34(5), 622–643.
Yonelinas, A. P., Kroll, N. E. A., Dobbins, I. G., & Soltani, M. (1999). Recognition memory of faces:
when familiarity supports associative recognition judgments. Psychonomic Bulletin & Review, 6,
654–661.
Cross-race effect 567
Copyright #2005 John Wiley & Sons, Ltd. Appl. Cognit. Psychol. 19: 545–567 (2005)
... Research investigating the CRE has resulted in several key theoretical positions (for a detailed review, see Young et al., 2012): the perceptual expertise hypothesis (Meissner & Brigham, 2001), the social-cognitive hypothesis (Anthony et al.,1992;Levin, 2000), and integrative models Meissner, Brigham, & Butz, 2005;Sporer, 2001). ...
Preprint
Full-text available
The Cross-Race Effect (CRE) is a reliable and robust phenomenon, whereby individuals better recognize faces that belong to their race compared to another race. Our goal was to produce items for a self-report Inventory (i.e., CRE-I) that brings together known predictors of the CRE to improve the postdiction of cross-race eyewitness accuracy.We expected a CRE for White and Asian participants. We anticipated that developed CRE-I subscales would correlate positively with extant (some modified) scales and predict accuracy. Participants completed four trials (two White targets and two Asian targets). For each trial, they watched a mock crime video, performed a distractor task, made a sequential lineup decision (target-present or target-absent), and indicated confidence in their lineup decision. After all trials, participants completed the potential items for the CRE-I.We replicated prior findings of a CRE for White participants but did not find a CRE for Asian participants. Exploratory factor analysis produced internally reliable scales for the CRE-I to be used with White eyewitnesses: general face recognition ability, race-specific face recognition ability, racial attitudes, quantity of contact, quality of contact, motivated individuation, and cognitive disregard. Responses to several scales predicted identification accuracy. In particular, three CRE-I scales predicted identification accuracy beyond the predictiveness of confidence: race-specific face recognition ability, racial attitudes towards White people, and motivated individuation of White people. Variables suggested separately by the perceptual expertise hypothesis and the social cognitive hypothesis predicted identification accuracy, providing support for integrative models of the CRE. The CRE-I contributes to the CRE literature both in terms of theory—by showing which factors among many may best relate to recognition—and practice—by improving evaluations of eyewitness reliability.
... In contrast, social-cognitive accounts suggest that the CRD stems from factors associated with social categorization and motivation (e.g., Hugenberg and Sacco 2008;Hugenberg et al. 2010;Levin 2000;Meissner, Brigham, and Butz 2005;Rodin 1987). For example, according to a prominent model, when individuals encounter faces perceived as cross-race, their attention tends to gravitate towards features perceived to be associated with or signal race, causing them to overlook the identity-diagnostic information needed to distinguish between different cross-race faces (Levin 2000). ...
Article
Full-text available
The other-ethnicity recognition deficit (OERD) involves poorer recognition of faces perceived as other-ethnicity compared to faces perceived as same-ethnicity. In this literature, research has examined social and perceptual encoding in the OERD separately. Recent research comparing these strategies shows that encoding faces based on social information (personality traits) enhances face recognition relative to encoding faces based on perceptual information (facial features), with a similar effect for both same-ethnicity and other-ethnicity faces. Expanding on this research, we conducted three experiments (one pre-registered), manipulating the perception of face ethnicity within participants and using different social and perceptual encoding questions to examine their impacts on the OERD. Results showed that social encoding equally facilitated the recognition of faces perceived as same-ethnicity and other-ethnicity, replicating previous research (Experiment 1). However, perceptual encoding divergently impacted face recognition, improving other-ethnicity but disrupting same-ethnicity (Experiment 2) or having comparable effects to standard encoding (Experiment 3). We discuss these findings and their implications for existing OERD theories. Keywords: face recognition, other-ethnicity recognition deficit, social encoding, perceptual encoding, face ethnicity or race
... After a 5-minute delay, participants saw 12 six-person simultaneous, photo lineups randomly intermixed during the test portion of the experiment with the stipulation that no more than three target-present or target-absent lineups were shown in consecutive order. These lineups showed front-facing photos of six White individuals with neutral expressions wearing identical maroon-colored clothing presented in a 2 X 3 array (Meissner, Brigham, & Butz, 2005). Gettleman et al. thoroughly evaluated the lineups for fairness. ...
Article
Full-text available
Is confidence most diagnostic of accuracy when expressed in numbers or when expressed in words? This question bears immense importance in many real-world contexts especially within the confines of eyewitness identification. In an eyewitness identification task, we compared the diagnostic value of numeric confidence across rating scales that varied in grain size (3-point vs. 6-point vs. 21-point vs. 101-point rating scales). We also compared the diagnostic value of numeric confidence to verbal confidence statements using several machine-learning algorithms. We found that fine-grain ratings are more diagnostic of identification accuracy than coarse-grain ratings, which suggests that the former provides a closer correspondence to memory strength than the latter. Moreover, we found that verbal confidence statements capture diagnostic information about the likely accuracy of an identification that numeric confidence ratings do not capture. This suggests that verbal confidence statements and numeric confidence ratings reflect partially independent, nonoverlapping sources of information. These results shed light on the processes that provide diagnostic value to confidence. From an applied standpoint, these results suggest that verbal confidence statements and numeric confidence ratings ought to be collected from eyewitnesses after an identification decision. Collecting both captures more diagnostic information than either can capture in isolation.
... The primary stimuli required for this study were two targets and several associated filler faces. We selected 26 photos of young Caucasian males with short brown hair (and no distinctive features) from the Meissner face database (Meissner et al., 2005). This database contains multiple views of each face so that we could utilize one version (smiling) at encoding, and another (neutral) in target-present lineups. ...
Article
Many criminals have distinctive facial features such as tattoos, yet the potential impact on eyewitness memory has received little research attention. Does such a feature harm memory for the face at encoding, and can police do anything about this when constructing the lineup? Does it matter whether the feature is on the interior (e.g., tattoo on face) or exterior (e.g., tattoo on neck)? These are the kinds of questions that we investigated by randomly assigning a large nationwide sample of online participants to conditions within an experimental design in which we manipulated target exposure time, presence (and location) of a distinctive feature, and whether it is replicated or removed from lineup members. Results indicate that a distinctive feature harms memory for the face regardless of location, but replicating the feature in the lineup may attenuate this effect. Fortunately, high confidence was indicative of high accuracy regardless of our manipulations.
... Materials: Eight Caucasian male faces with neutral expressions were selected for this experiment (from [25]). Each face was cropped in Photoshop and placed within an oval frame so that no hair was visible. ...
Article
Full-text available
While our direct observations of the features or behaviours of the stimuli around us tell us much about them (e.g., should they be feared?), the origin of much of our knowledge is often untethered from directly observable properties (e.g., through what we have learned or have been told about them, or “semantic knowledge”). Here, we ask whether otherwise neutral visual stimuli that participants learn to associate with emotional qualities in the lab cause the stimuli to be attended in a similar way as stimuli whose emotional qualities can be discerned through their visual properties. In Experiment 1, participants learned to associate negative or neutral characteristics with neutral faces, which then served as valid or invalid spatial cues to targets in an attentional disengagement paradigm. The performance of participants higher in trait anxiety was consistent with attentional avoidance of faces with learned negative associations, while participants lower in trait anxiety showed a general response slowing in trials with these stimuli, compared to those with neutral associations. In contrast, in Experiment 2, using (visually) expressive (angry) faces, the performance of participants higher in trait anxiety was consistent with difficulty disengaging from visually threatening faces, while the performance of those with lower trait anxiety appeared unaffected by the valence of the stimuli. These findings suggest that (1) emotionality acquired indirectly via learned semantic knowledge impacts how attention is allocated to face stimuli, and this impact is influenced by trait anxiety, and (2) there are differences in the effects of stimulus emotionality depending on whether it is acquired indirectly or directly via the perceptual features of the stimulus. These differences are discussed in the context of the variability of attention bias effects reported in the literature and the time course of impacts of emotionality on stimulus processing.
... Various, not mutually exclusive, explanations have been offered for the own-race advantage, including the theory that greater exposure to people of one's own race may lead to greater perceptual expertise for processing and remembering own-race faces (Brigham & Malpass, 1985;Carroo, 1986;Correll et al., 2021;Fallshore & Schooler, 1995;Golby et al., 2001;Lindsay et al., 1991;Tanaka et al., 2004). Another possibility is that people are more motivated to individuate own-race faces compared to other race faces, which again leads to better recognition and memory of own-race faces (Hugenberg et al., 2010;Levin & Beale, 2000;Meissner et al., 2005;Simon et al., 2023). While a number of factors might influence when the own-race advantage is likely to be observed, the basic pattern has been demonstrated across multiple methodologies and populations including looking time and/or eye tracking in young infants (Gaither et al., 2012;Sugden & Marquis, 2017), using functional magnetic resonance imaging methodology in Chinese adults (Feng et al., 2011), and categorization tasks in Chinese and Israeli college students (Zhao & Bentin, 2008) among others (for reviews, see Hugenberg et al., 2010;Lee et al., 2011;Meissner & Brigham, 2001). ...
Article
Full-text available
Across six preregistered studies (N = 1,292; recruited from university subject pools and Prolific Academic), we investigate how face perception along the dimensions of gender/sex and race can vary based on immediate contextual information as well as personal experience. In Studies 1a and 1b, we find that when placing stimuli along a continuum from male to female, cisgender participants sort prototypical gender/sex faces in a bimodal fashion and show less consensus and greater error when placing faces of intermediate gender/sex. We replicate and extend these findings to race in Study 2. In Study 3, we test whether sorting patterns can be influenced by preexisting experiences, and find evidence that transgender/nonbinary participants show less error than cisgender heterosexual participants when sorting intermediary faces. Finally, in Studies 4 and 5, we test whether cisgender participants’ judgments of intermediary faces along the continuum are influenced by the specific circumstances under which they are asked to sort. Here, we find that changing the sorting framework to include a third category resulted in less error when placing intermediary faces along the continuum than when participants were provided with only two category labels or two categories and a line at the midpoint, suggesting that new perceptual categories introduced with minimal training can be adopted quickly and successfully in a perceptual task. These data suggest that both long-term life experiences and quick experimental interventions can shape how we think about gender/sex and race.
Book
Full-text available
When Face Recognition Goes Wrong explores the myriad ways that humans and machines make mistakes in facial recognition. Adopting a critical stance throughout, the book explores why and how humans and machines make mistakes, covering topics including racial and gender biases, neuropsychological disorders, and widespread algorithm problems. The book features personal anecdotes alongside real-world examples to showcase the often life-changing consequences of facial recognition going wrong. These range from problems with everyday social interactions through to eyewitness identification leading to miscarriages of justice and border control passport verification. Concluding with a look to the future of facial recognition, the author asks the world’s leading experts what are the big questions that still need to be answered, and can we train humans and machines to be super recognisers? This book is a must-read for anyone interested in facial recognition, or in psychology, criminal justice and law.
Article
Full-text available
Perceivers typically exhibit better recognition memory for same-race faces than for cross-race faces, a phenomenon known as the cross-race effect (CRE). Despite its ubiquity, it is yet unclear whether people are metacognitively aware of the CRE. This research thoroughly investigates perceivers’ metacognitive awareness of the CRE across five experiments. Experiments 1 and 2 demonstrated that both prospective (judgments of learning) and retrospective (confidence) metamemory judgments are sensitive to variations in the racial category and prototypicality of faces. Experiment 3 showed that participants’ item-level prospective judgments are informed by beliefs about the impact of face race on memory performance. Experiment 4 revealed that global predictions are influenced by face race in the absence of direct stimulus experience, emphasizing the role of preexisting beliefs. Experiment 5 extended these findings by showing large crossover interactions between face race and participant race in both global predictions and item-level prospective judgments, indicating that both White and Black participants have higher metamemory estimates for ingroup faces. This experiment further showed that preexisting beliefs intensify the impact of face race on metamemory judgments yet do not fully account for it. Collectively, these experiments provide robust evidence of good metamemory accuracy for faces varying in racial categories and prototypicality among White participants and demonstrate that beliefs underlie the effect of face race on metamemory judgments among both White and Black participants, though this may not be the only mechanism involved.
Article
https://authors.elsevier.com/a/1j1hJ51f8shvg People typically remember faces of their own race better than the faces of other races, which is known as the own-race bias. Social-cognitive theories on the own-race bias debate if motivation plays a role in racial face recognition under the “Asian-Caucasian” context. We investigated the effect of social-identity-threat based motivation on own-race bias. In Experiment 1, Chinese participants were primed with a negative image of own-race by associating Chinese (i.e., own-race) faces with negative personality traits. Results showed that such social identity threat marginally affect the own-race bias in face recognition sensitivity (Experiment 1a), and it did not influence participants' implicit positive attitudes toward their ethnicity (Experiment 1b). In Experiments 2, 3 and 4, the social identity threat was manipulated by reading an article before the face recognition task. Experiment 2 found that the social identity threat significantly affected the own-race bias in face recognition false alarm. Although we did not find the modulation effect of group on ORB with ANOVA in Experiments 3 and 4, we observed the significant indirect effect of group on ORB through the mediator of generalized threat. Together, the current study provides evidence of social motivation in the own-race bias in the “Asian-Caucasian” setting.
Article
Full-text available
Whether the format of a recognition memory task influences the contribution of recollection and familiarity to performance is a matter of debate. The authors investigated this issue by comparing the performance of 64 young (mean age = 21.7 years; mean education = 14.5 years) and 62 older participants (mean age = 64.4 years; mean education = 14.2 years) on a yes-no and a forced-choice recognition task for unfamiliar faces using the remember-know-guess procedure. Familiarity contributed more to forced-choice than to yes-no performance. Moreover, older participants, who showed a decrease in recollection together with an increase in familiarity, performed better on the forced-choice task than on the yes-no task, whereas younger participants showed the opposite pattern.
Article
Full-text available
Accusations of discriminatory treatment of minority persons in the criminal justice system create a need for policy and procedure development to create real and perceived equal treatment. A facial recognition deficit among law enforcement officers and witnesses for persons of another “race” contributes to unequal treatment of minority group members. This article demonstrates the other-race effect in an unusual context, reveals theoretical weaknesses, reveals the role of categorical processes in the phenomenon, and discusses policy implications. Experiment 1, based on feature and trait ratings, demonstrates that identical and racially ambiguous faces with different racial markers (hair) are perceived according to the marker. Experiment 2 demonstrates an other-race recognition effect using these faces. A feature acting as a racial marker can cause a face to be perceived and remembered differently. Other-race faces are perceived categorically, which drives the recognition process.
Article
Full-text available
The other-race effect (or own-race bias or cross-racial identification effect) refers to the finding that recognition memory tends to be better for faces of members of subjects’ own race than for faces of members of other races. The current study was designed to test the hypothesis that perceptual skills specific to identifying faces of particular racial groups contribute to this effect. On each of 50 trials, a photograph of a face was tachistoscopically presented for 120 ms, followed by a pattern mask and then a plain-view test pair composed of the previously presented face and a matched foil. As predicted, an other-race effect was obtained on this perceptual task: White subjects performed significantly more poorly on trials involving African American faces than on trials involving White faces, whereas no such difference was obtained among African American subjects.
Article
Full-text available
Recognition memory for faces is hampered much more by inverted presentation than is memory for any other material so far examined. The present study demonstrates that faces are not unique with regard to this vulnerability to inversion. The experiments also attempt to isolate the source of the inversion effect. In one experiment, use of stimuli (landscapes) in which spatial relations among elements are potentially important distinguishing features is shown not to guarantee a large inversion effect. Two additional experiments show that for dog experts sufficiently knowledgeable to individuate dogs of the same breed, memory for photographs of dogs of that breed is as disrupted by inversion as is face recognition. A final experiment indicates that the effect of orientation on memory for faces does not depend on inability to identify single features of these stimuli upside down. These experiments are consistent with the view that experts represent items in memory in terms of distinguishing features of a different kind than do novices. Speculations as to the type of feature used and neuropsychological and developmental implications of this accomplishment are offered.
Article
Full-text available
The current article reviews the own-race bias (ORB) phenomenon in memory for human faces, the finding that own-race faces are better remembered when compared with memory for faces of another, less familiar race. Data were analyzed from 39 research articles, involving 91 independent samples and nearly 5,000 participants. Measures of hit and false alarm rates, and aggregate measures of discrimination accuracy and response criterion were examined, including an analysis of 8 study moderators. Several theoretical relationships were also assessed (i.e., the influence of racial attitudes and interracial contact). Overall, results indicated a "mirror effect" pattern in which own-race faces yielded a higher proportion of hits and a lower proportion of false alarms compared with other-race faces. Consistent with this effect, a significant ORB was also found in aggregate measures of discrimination accuracy and response criterion. The influence of perceptual learning and differentiation processes in the ORB are discussed, in addition to the practical implications of this phenomenon.
Chapter
Due to the advent of neuropsychology, it has become clear that there is a multiplicity of memory systems or, at the very least, of dissociably different modes of processing memory in the brain. As the Oxford Handbook of Memory demonstrates, the frontier of memory research has been enriched by breakthroughs of the last decades, with lines of continuity and important departures, and it will continue to be enriched by changes in technology that will propel future research. In turn, such changes are beginning to impact the legal and professional therapeutic professions and will have considerable future significance in realms outside of psychology and memory research. Endel Tulving and Fergus Craik, two world-class experts on memory, provide this handbook as a roadmap to the huge and unwieldy field of memory research. By enlisting an eminent group of researchers, they are able to offer insight into breakthroughs for the work that lies ahead. The outline is comprehensive and covers such topics as the development of memory, the contents of memory, memory in the laboratory and in everyday use, memory in decline, the organization of memory, and theories of memory.
Article
Two experiments used a priming paradigm to investigate the influence of racial cues on the perceptual identification of weapons. In Experiment 1, participants identified guns faster when primed with Black faces compared with White faces. In Experiment 2, participants were required to respond quickly, causing the racial bias to shift from reaction time to accuracy. Participants misidentified tools as guns more often when primed with a Black face than with a White face. L. L. Jacoby's (1991) process dissociation procedure was applied to demonstrate that racial primes influenced automatic (A) processing, but not controlled (C) processing. The response deadline reduced the C estimate but not the A estimate. The motivation to control prejudice moderated the relationship between explicit prejudice and automatic bias. Implications are discussed on applied and theoretical levels.
Article
In shopping centers in South Africa and England, Black and White people were approached by either a Black or a White confederate. Later they were asked to identify the confederate from both a sequential lineup, where photographs are seen one at a time, and a forced-choice recognition test, where photographs are seen simultaneously. In both countries and with both identification tasks the confederates were better recognized by people of their own race. By observing this own-race bias in two countries with different demographics, using methods similar to those used in some criminal investigations of eyewitness memory, the authors demonstrated the robust nature of the own-race bias. In addition, witness confidence and accuracy were found to be correlated but only when the participant was the same race as the confederate. Implications for legal procedure are discussed.