Affective Information in Context and Judgment of Facial Expression: Cultural Similarities and Variations in Context Effects Between North Americans and East Asians

Abstract

Previous research in cultural psychology suggests that North Americans are less likely than their East Asian counterparts to be sensitive to contextual information. By contrast, much evidence suggests that even North Americans’ judgments are highly influenced by affective priming information, the effect of which can be seen as another type of contextual cue. However, the magnitude of such a priming effect has not been comprehensively tested in a cross-cultural context. Taking advantage of the methodology of the affective priming paradigm, we conducted two studies, in which we manipulated (a) the timing of priming information (simultaneous vs. sequential) and (b) the type of affective information (background landscape vs. background human figures), in which European Canadians and Japanese judged target faces that showed either happy or sad facial expressions in the focal area of the scene. The results in general indicate that a similar degree of contextual effect occurs in members of both cultures. The issue of generalization of cross-cultural findings and the necessity of overarching more than one research paradigm are discussed.

Full text

http://jcc.sagepub.com/
Journal of Cross-Cultural Psychology
http://jcc.sagepub.com/content/43/3/429
The online version of this article can be found at:
DOI: 10.1177/0022022110395139
2012 43: 429 originally published online 17 January 2011Journal of Cross-Cultural Psychology
Kenichi Ito, Takahiko Masuda and Koichi Hioki
Asians
Similarities and Variations in Context Effects Between North Americans and East
Affective Information in Context and Judgment of Facial Expression : Cultural
Published by:
http://www.sagepublications.com
On behalf of:
International Association for Cross-Cultural Psychology
can be found at:Journal of Cross-Cultural PsychologyAdditional services and information for
http://jcc.sagepub.com/cgi/alertsEmail Alerts:
http://jcc.sagepub.com/subscriptionsSubscriptions:
http://www.sagepub.com/journalsReprints.navReprints:
http://www.sagepub.com/journalsPermissions.navPermissions:
http://jcc.sagepub.com/content/43/3/429.refs.htmlCitations:
What is This?
- Jan 17, 2011 OnlineFirst Version of Record
- Mar 12, 2012Version of Record >>
by guest on February 19, 2013jcc.sagepub.comDownloaded from

Journal of Cross-Cultural Psychology
43(3) 429 –445
© The Author(s) 2012
Reprints and permission:
sagepub.com/journalsPermissions.nav
DOI: 10.1177/0022022110395139
jccp.sagepub.com
Affective Information in
Context and Judgment of
Facial Expression: Cultural
Similarities and Variations
in Context Effects
Between North Americans
and East Asians
Kenichi Ito
1
, Takahiko Masuda
1
, and Koichi Hioki
2
Abstract
Previous research in cultural psychology suggests that North Americans are less likely than their
East Asian counterparts to be sensitive to contextual information. By contrast, much evidence
suggests that even North Americans’ judgments are highly influenced by affective priming
information, the effect of which can be seen as another type of contextual cue. However, the
magnitude of such a priming effect has not been comprehensively tested in a cross-cultural
context. Taking advantage of the methodology of the affective priming paradigm, we conducted
two studies, in which we manipulated (a) the timing of priming information (simultaneous vs.
sequential) and (b) the type of affective information (background landscape vs. background
human figures), in which European Canadians and Japanese judged target faces that showed
either happy or sad facial expressions in the focal area of the scene. The results in general
indicate that a similar degree of contextual effect occurs in members of both cultures. The
issue of generalization of cross-cultural findings and the necessity of overarching more than one
research paradigm are discussed.
Keywords
cross-cultural comparison, facial expressions, contextual influence, cultural similarity, European
Canadian, Japanese
Research on culture and psychology has demonstrated systematic cross-cultural variations in
psychological processes. Notably, much evidence suggests that North Americans are less likely
than their East Asian counterparts to be sensitive to contextual information (e.g., Masuda & Nisbett,
2001, 2006; Nisbett & Masuda, 2003). However, how much these findings can be applicable and
1
University of Alberta, Edmonton, Alberta, Canada
2
Kobe University, Japan
Corresponding Author:
Kenichi Ito, Department of Psychology, University of Alberta, P-357, Biological Sciences Building,
Edmonton, Alberta, T6G 2E9, Canada.
Email: kenichi@ualberta.ca
by guest on February 19, 2013jcc.sagepub.comDownloaded from

430 Journal of Cross-Cultural Psychology 43(3)
generalizable to the other domains of psychological phenomena has not been fully discussed.
Furthermore, such findings are sometimes used to form stereotypes about specific cultural or
ethnic groups and their patterns of behavior (Adams & Markus, 2004). More detailed analyses of
underlying processes that produce cultural differences and similarities in psychological pro-
cesses are highly expected in the field of cross-cultural research (Matsumoto & Yoo, 2006).
Given the evidence in cultural research that has accumulated over several decades, we maintain
that a close examination of the magnitude of context effect across other domains of behavioral
research will be of value. For this investigation, we targeted the affective priming effect on the
judgment of emotional facial expressions (e.g., Fazio, 2001, for review). Using this research
paradigm, we conducted two cross-cultural studies to reexamine the issue of culture and the
judgment of facial expressions in context.
Judgment of Facial Expressions in Context
Past research on the judgment of facial expressions in context suggests that North Americans and
Europeans are sensitive to the affective valence (positive vs. negative) of contextual information.
For example, Canadian participants’ judgment of a facial expression was influenced by preced-
ing emotional scenarios or faces (Carroll & Russell, 1996; Russell & Fehr, 1987). Several stud-
ies gave credence to this effect of the contextual valence on the judgment of facial expressions
(Russell, 1991; Tanaka-Matsumi, Attivissimo, Nelson, & D’Urso, 1995).
Of particular interest to the current study, North Americans’ and Europeans’ judgments are
also highly influenced by even a brief presentation of affective priming information, which is
also a type of contextual cue (Fazio, 2001, for review). Hietanen, Klemettilä, Kettunen, and
Korpela (2007), for example, had Finnish participants judge a happy face or a disgusted face
after briefly presenting them with various landscape images. The landscape images in one cate-
gory showed scenes consisting mainly of artificial materials (e.g., an urban apartment complex),
which activated negative mindsets. In the other category, the landscape images showed mainly
natural materials (e.g., a suburban park), which activated positive mindsets. In a stepwise man-
ner, Hietanen et al. (2007) manipulated the ratio of natural materials to artificial materials in the
scenes and created five degrees of combinations from mostly artificial materials to mostly natural
materials. The results showed that as the proportion of natural materials in the primes increased,
participants categorized happy faces faster and disgusted faces more slowly. That is, participants
judged a target’s facial emotion more quickly when the valence of the elicited mindset was con-
gruent with that of the target’s facial emotion (e.g., a suburban park with a happy face) than when
the valence of the elicited mindset was incongruent with the target’s facial emotion (e.g., an
urban apartment complex with a happy face). In addition to landscape images, North Americans’
and Europeans’ speed of categorizing facial emotions was also influenced by nonverbal sounds
(Carroll & Young, 2005), odors (Leppänen & Hietanen, 2003), or pictorial images that entailed
clear emotional valence (Carroll & Young, 2005).
Such sensitivity to contextual information among North Americans and Europeans, however,
appears incompatible with the theory and evidence of cross-cultural research (Nisbett, 2003;
Nisbett, Peng, Choi, & Norenzayan, 2001). Past research demonstrated that, relative to East
Asians, North Americans and Europeans pay less attention to contextual cues such as the frames
of abstract figures (Ji, Peng, & Nisbett, 2000; Kitayama, Duffy, Kawamura, & Larsen, 2003;
Masuda, Akase, Radford, & Wang, 2008) or the backgrounds of visual images (Chua, Boland, &
Nisbett, 2005; Masuda, Gonzalez, Kwan, & Nisbett, 2008; Masuda & Nisbett, 2006; Miyamoto,
Nisbett, & Masuda, 2006).
A series of studies conducted by Masuda and his colleagues (Masuda, Ellsworth et al., 2008;
Masuda, Wang, & Ishii, 2008) are particularly relevant to the current study. Masuda and his
by guest on February 19, 2013jcc.sagepub.comDownloaded from

Ito et al. 431
colleagues presented American and Japanese participants with lineups of five cartoon figures
(or facial pictures), in which one main character who showed a specific facial expression stood
against a background of four figures who showed their own facial expressions. The combination
of the facial expressions of the target figure and the background figures was manipulated so that
central figures’ emotions either matched or mismatched with four background figures’ emotions.
Intensity ratings of the central figures’ emotions showed that the discrepancy of the ratings between
the match and the mismatch lineups was significantly smaller for American participants than for
Japanese participants. The implication was that Americans excluded the background figures’
emotions from the evaluation of the target’s emotions, but the Japanese incorporated the background
figures’ emotions into the evaluation of the target’s emotions.
Although both paradigms investigate sensitivity to contextual information, the methodologies
of the affective priming study (e.g., Hietanen et al., 2007) and the lineup study (e.g., Masuda,
Ellsworth et al., 2008) are distinct from one another. First, in the priming study, participants were
shown contextual information and the target information sequentially, whereas participants in
the lineup study were shown the two pieces of information simultaneously. Second, the priming
study was conducted using a variety of priming stimuli, whereas the lineup study was conducted
using only background figures’ facial expressions as contextual information. Third, the priming
study focused only on a single culture, whereas the lineup study compared the effect across cul-
tures. These inconsistencies in methodology obscure the magnitude of the research findings.
Therefore, modification of experimental designs from the two paradigms is necessary to place
these two research paradigms on the same plane. However, little research has been conducted to
seamlessly blend these two paradigms. The current article addresses this issue by (a) manipulat-
ing the presentation (timing) of the contextual information, (b) manipulating the type of stimuli,
(c) and cross-culturally examining the magnitude of contextual effect between two cultural groups:
European Canadians and Japanese.
Overview of the Studies
We designed our experiments so that, first, the contextual stimuli were presented in two different
ways: simultaneous condition versus sequential condition. In the simultaneous condition (Study 1A
and Study 2A), we followed the lineup study (e.g., Masuda, Ellsworth et al., 2008) that the con-
textual information was presented as a background of the target stimuli. In the sequential condi-
tion (Study 1B and Study 2B), we followed the affective priming study: Participants were presented
with a fixation point at the center of the screen for 500 ms, followed by contextual priming
stimuli for 250 ms, followed by a blank for 50 ms.
Second, we used two kinds of contextual stimuli: background landscape (i.e., natural or industrial
landscape with either positive or negative valence) and background human figures (i.e., back-
ground figures expressing either happiness or sadness). For background landscape, we selected
landscape images as the contextual information (Study 1A and 1B), following Hietanen et al.’s
(2007) studies. However, unlike Hietanen et al., who assumed that landscapes containing nature
images entail positive valence and that city landscapes entail negative valence, we manipulated
the valence of landscape images more clearly by using stimuli from International Affective
Picture System (IAPS; Lang, Bradley, & Cuthbert, 2005), and by pretesting the equivalency of
evaluation between the two cultural groups. For the background human figures (Study 2A
and 2B), we used revised stimuli originally created by Masuda, Wang et al. (2008), in which two
background individuals showed either happy or sad facial expressions.
1
By synthesizing these two factors, we crossed the two kinds of contextual stimuli with the two
ways to present them. Study 1A (simultaneous condition) and Study 1B (sequential condition)
used landscape images with clear valences (landscape cues); Study 2A (simultaneous condition)
by guest on February 19, 2013jcc.sagepub.comDownloaded from

432 Journal of Cross-Cultural Psychology 43(3)
and Study 2B (sequential condition) used background figures’ facial emotions as contextual stim-
uli (human figure cues). Crossing the two presentation styles with the two kinds of contextual
stimuli allowed us to examine whether the nature of the contextual stimuli interacted with the
manner of presentation. Using these experimental manipulations, we tested whether we could
observe cultural similarities or any differences in any specific experimental condition. If the con-
text effect was observed across all experimental conditions and across cultures, the result would
confirm the robustness of the context effect during category judgment of facial expressions. By
contrast, if any cultural variation in the context effect was observed, the assertion of cultural varia-
tion in context sensitivity would be generalizable beyond the original research paradigm.
Study 1A
Following previous works in cultural psychology (e.g., Masuda & Nisbett, 2001, 2006), Study 1A
examined whether European Canadians’ tendency to be influenced by context information was
similar to that of their Japanese counterparts. To investigate this question, we created a set of
stimuli by placing the contextual information on the same plane as the target information that the
participants were asked to judge.
Method
Participants. Thirty-eight European Canadians (20 females and 18 males) at the University of
Alberta and 36 Japanese (18 females and 18 males) at Kobe University participated in the experi-
ment. In exchange for their participation, European Canadian participants received a course
credit, and Japanese participants received a gift card worth about $5.
Stimuli. Selecting culturally equivalent facial expressions, we conducted a pilot study in which
European Canadian and Japanese participants evaluated the valence of facial expressions from
Masuda, Wang et al.’s (2008) stimuli. Eighty-two European Canadian and 42 Japanese partici-
pants were asked to rate the intensity of happiness and sadness expressed by either European
faces or Asian faces on a 10-point scale ranging from 0 (not at all intense) to 9 (extremely
intense). Each model showed both happy and sad facial emotions in accordance with criteria
suggested by Ekman and Friesen (1975). On the basis of the results, we selected 24 facial por-
traits that varied equally in terms of model’s ethnicity and gender.
2
For the selected 12 European
faces, a paired-samples t test confirmed that the mean happiness intensity and the mean sadness
intensity of 6 happy faces as well as those of 6 sad faces were significantly different from each
other among European Canadian participants, t
happy face
(40) = 30.85, p < .001 (M
happiness
= 6.81;
M
sadness
= .98), t
sad face
(40) = 13.00, p < .001 (M
happiness
= .70; M
sadness
= 4.20) and among Japanese
participants, t
happy face
(20) = 26.38, p < .001 (M
happiness
= 6.72; M
sadness
= .72), t
sad face
(20) = 10.28,
p < .001 (M
happiness
= .88; M
sadness
= 4.99). Similarly for the selected 12 Asian faces, a paired-
samples t test showed that happy faces and sad faces were clearly distinguishable from one
another among European Canadian participants, t
happy face
(40) = 30.52, p < .001 (M
happiness
= 6.55;
M
sadness
= .48), t
sad face
(40) = 25.53, p < .001 (M
happiness
= .55; M
sadness
= 6.24) and among Japanese
participants, t
happy face
(20) = 25.92, p < .001 (M
happiness
= 6.52; M
sadness
= .34), t
sad face
(20) = 14.18,
p < .001 (M
happiness
= .59; M
sadness
= 6.01).We conducted another pilot study in which 28 European
Canadians and 20 Japanese who had not participated in the first pilot study were asked to evalu-
ate the intensity of the valence of landscape images from IAPS (Lang et al., 2005). Following
Lang et al., we used Self-Assessment Manikin (SAM) as the scale, in which a schematic image
depicted values ranging from 1 (extremely positive) to 9 (extremely negative). On the basis of the
results, we selected 10 images (5 positive) that were either clearly positive or clearly negative.
A paired-samples t test confirmed that the mean of 5 positive landscapes was significantly different
by guest on February 19, 2013jcc.sagepub.comDownloaded from

Ito et al. 433
from the mean of 5 negative landscapes among European Canadian participants, t(27) = 22.10,
p < .001 (M
positive
= 2.41; M
negative
= 7.71), and among Japanese participants, t(19) = 9.86, p < .001
(M
positive
= 3.49; M
negative
= 7.00).
We then merged images of facial emotions with images of landscapes to make single pictures,
so that faces were presented with landscape backgrounds (see Figure 1). In total, we created
240 experimental stimuli by merging each of 24 facial images with five positive and five nega-
tive landscapes, using Adobe Photoshop CS2 Version 9.0.2. The size of the target’s facial images
was 566 × 453 pixels (11.4° × 14.2°). The size of the landscape images was 1,024 ×768 pixels
(25.3° × 18.9°) and covered the entire screen of the 15-inch monitor.
Procedure. Upon arrival, participants were asked to read instructions on a computer screen.
The instructions stated that the task was to judge as fast and accurate as possible whether the
person at the center of the picture expressed positive or negative emotions. Half of the partici-
pants were instructed to left-click for positive emotions and right-click for negative emotions; the
order was reversed for the other half of the participants. All participants used their right hand to
click. The instructions were presented in the participants’ native language. The first author trans-
lated the English version of the instructions into Japanese, and a bilingual research assistant
back-translated the instructions. Experimental stimuli were presented randomly using E-prime
version 1.2 (Schneider, Eschman, & Zuccolotto, 2002) on a 15-inch laptop computer at a
resolution of 1,024 × 768 pixels. Participants sat in front of the computer at a viewing dis-
tance of approximately 60 cm. They were given four practice trials and subsequently completed
Figure 1. Examples of the Stimuli Used in Study 1A.
Stimuli differed in terms of the valence of facial expression and of landscape. Four possible combinations were
(a) negative face–negative landscape, (b) positive face–negative landscape, (c) negative face–positive landscape, and
(d) positive face–positive landscape.
by guest on February 19, 2013jcc.sagepub.comDownloaded from

434 Journal of Cross-Cultural Psychology 43(3)
240 experimental trials. Participants’ reaction time and accuracy for each trial were recorded.
After answering a demographic information questionnaire at the end of the experiment, partici-
pants were fully debriefed.
Results
The focus of the current studies was to compare the effect of contexts on Canadian participants
with that on Japanese participants. Thus we combined the happy face–positive landscape condi-
tion with the sad face–negative landscape condition to create the congruent scores, and the happy
face–negative landscape condition with the sad face–positive landscape condition to create the
incongruent scores. In this way, discrepancies between congruent scores and incongruent scores
indicate the effect of contextual information on the categorization of target. We then conducted
2 (Culture: European Canadian vs. Japanese) × 2 (Congruency: Congruent vs. Incongruent)
ANOVA, with participants’ reaction time being the dependent variable. Culture was the between-
subjects variable, and congruency was the within-subject variable.
3
We found a significant main
effect of congruency, F(1, 72) = 33.66, p < .001, η
p
2
= .32. A main effect of culture was not sig-
nificant, F(1, 72) = 2.34, p = ns. The two-way interaction of participants’ cultural backgrounds
and congruency was not significant, F < 1, suggesting that the effect of contextual landscape was
similar across the two cultural groups. Within-cultural level, the results also showed that, similar
to Japanese participants, Canadian participants categorized the facial expression faster when the
landscapes and facial expressions had congruent valences (M
CND
= 830.69 and M
JPN
= 773.29)
than when the valences were incongruent (M
CND
= 866.25 and M
JPN
= 813.86; see Figure 2);
t(37) = 3.89, p < .001, t(35) = 4.31, p < .001, respectively.
4,5
Study 1B
In Study 1B, we isolated the affective component of contextual information from the target’s facial
expressions by using the affective priming paradigm. European Canadian and Japanese partici-
pants were briefly presented with pictorial images of positive or negative landscapes before the
Figure 2. Mean Reaction Times (+SE) as a Function of Affective Background–Foreground Congruence
and Culture (Study 1A).
by guest on February 19, 2013jcc.sagepub.comDownloaded from

Ito et al. 435
judgment of happy or sad facial expression. We examined whether the brief presentation of
salient affective contextual information influenced the two groups of participants in a similar
manner; that is, we tested whether both groups would categorize targets’ facial emotions that
were congruent with the valence of contextual information more quickly than facial emotions
that were incongruent with the valence of contextual information.
Method
Stimuli. In this experiment, we used the same facial expressions and landscape images as in
Study 1A. Instead of merging them as a single image, we used landscape images as primes and
facial expressions as targets. The sizes of the facial images and landscape images were identical
to those used in Study 1A.
Participants. Thirty-eight European Canadians (18 females and 20 males) at the University of
Alberta and 35 Japanese (18 females and 17 males) at Kobe University participated in the experi-
ment. These participants participated in neither the two pilot studies nor Study 1A. In exchange
for their participation, European Canadian participants received a course credit, and Japanese
participants received a gift card worth about $5.
Procedure. The experimental procedure was similar to the procedure used in Study 1A, except
that participants were presented with sets of two consecutive images. The task was to categorize
(by clicking mouse buttons) whether the second image, a facial portrait, appeared to depict a
positive or a negative emotion. Experimental trials followed the affective priming paradigm (see
Figure 3). Participants were presented with a fixation point at the center of the screen for 500 ms,
followed by primes for 250 ms, and a blank for 50 ms. Thus, the Stimulus Onset Asynchrony
(SOA) was set at 300 ms to separate affective priming effect from semantic priming effect.
6
Target stimuli were presented until participants responded or for 5 seconds. The intertrial inter-
val was 2 seconds.
Results
We conducted 2 (Culture: European Canadian vs. Japanese) × 2 (Congruency: Congruent vs.
Incongruent) ANOVA, with participants’ reaction time being the dependent variable. Culture
was the between-subjects variable, and congruency was the within-subject variable.
7
We found
a significant main effect of congruency, F(1, 70) = 14.59, p < .001, η
p
2
= .17. A main effect of
culture was not significant, F(1, 70) = 1.64, p = ns. The two-way interaction of participants’
cultural backgrounds and congruency was not significant, F < 1. Planned contrast revealed that,
similar to Japanese participants, Canadian participants categorized the target faster when the
pairs of prime and target had congruent valences (M
CND
= 727.33 and M
JPN
= 779.17) than when
the pairs had incongruent valences (M
CND
= 742.72 and M
JPN
= 792.10; see Figure 4); t(36) = 2.95,
p < .01, t(34) = 2.46, p < .05, respectively.
8,9
Study 2A
Following the procedure used in Study 1A, we placed the targets and the primes on the same
plane in Study 2A to examine whether European Canadians’ tendency to be influenced by the
valence of human figures was similar to that of their Japanese counterparts.
Method
Participants. Thirty-seven European Canadians (17 females and 20 males) at the University of
Alberta and 33 Japanese (16 females and 17 males) at Kobe University participated in the
by guest on February 19, 2013jcc.sagepub.comDownloaded from

436 Journal of Cross-Cultural Psychology 43(3)
experiment. These participants did not participate in any of the previous studies. In exchange for
their participation, European Canadian participants received a course credit, and Japanese par-
ticipants received a gift card worth about $5.
Stimuli. We conducted another pilot study to select happy or sad facial portraits to be used as
contexts. We prepared new facial portraits according to criteria used in the first pilot study. Eighteen
Canadian participants (8 females) and 20 Japanese participants (6 females) who had not partici-
pated in any of previous studies evaluated the intensity of valence. Participants evaluated each
portrait using the SAM, the same scale as the one we used to select landscape images in Study 1A.
On the basis of the result, we selected 16 facial portraits (8 happy and 8 sad) that varied equally
in terms of models’ ethnicity and gender. Models showed either a happy face or a sad face. A paired-
samples t test confirmed that the mean of 8 happy facial portraits were significantly different
from the mean of 8 sad facial portraits among European Canadian participants, t(17) = 17.26,
Figure 3. Illustration of the Affective Priming Paradigm Used in Study 1B.
After being presented with a fixation point, participants saw (a) positive or (b) negative landscape images for 250 ms
as a prime. Following the interval (50 ms), they were to categorize (c) positive or (d) negative facial expression.
Emotionally congruent prime–target pairs refer to the combination of (a) and (c) or (b) and (d), whereas emotionally
incongruent prime–target pairs refer to the combination of (a) and (d) or (b) and (c).
by guest on February 19, 2013jcc.sagepub.comDownloaded from

Ito et al. 437
p < .001 (M
positive
= 2.35; M
negative
= 7.12), and among Japanese participants, t(19) = 25.20,
p < .001 (M
positive
= 2.09; M
negative
= 7.08). The selected facial portraits, the same size as the target
stimuli used in Study 1A and 1B, were then paired with portraits of the opposite gender within
the same ethnic group to form 8 context images (4 pairs of happy faces and 4 pairs of sad faces).
A pair of portraits was placed side by side, 3 cm apart (2.9°). We then merged facial portraits
used in Study 1A and 1B with the context images to make single images, so that target faces were
presented with backgrounds of two other faces (see Figure 5). In total, we created 192 experi-
mental stimuli by merging each of 24 target facial images with 8 contextual facial images. The
merged images were presented at the center of the screen (870 × 459 pixels) on a 15-inch monitor
(1,024 × 768 pixels).
Procedure. The experimental procedure was similar to that used in Study 1A, except that target
facial emotions were presented with background human figures. Participants’ task was to judge
whether the person at the center of the picture expressed positive or negative emotions as fast and
accurate as possible.
Results
We conducted 2 (Culture: European Canadian vs. Japanese) × 2 (Congruency: Congruent vs.
Incongruent) ANOVA, with participants’ reaction time being the dependent variable. Culture
was the between-subjects variable, and congruency was the within-subject variable.
10
We found
a significant main effect of congruency, F(1, 68) = 14.69, p < .001, η
p
2
= .18. Neither a main
effect of culture nor the two-way interaction of participants’ cultural backgrounds and congru-
ency were significant, Fs < 1. Planned contrasts revealed that, similar to Japanese participants,
Canadian participants categorized the targets’ facial emotions faster when the target and back-
ground figures had congruent valences (M
CND
= 791.91 and M
JPN
= 821.08) than when they had
incongruent valences (M
CND
= 818.18 and M
JPN
= 839.62; see Figure 6), t(36) = 2.99, p < .01,
t(32) = 2.48, p < .05, respectively.
11
Figure 4. Mean Reaction Times (+SE) as a Function of Affective Prime-Target Congruence and Culture
(Study 1B).
by guest on February 19, 2013jcc.sagepub.comDownloaded from

438 Journal of Cross-Cultural Psychology 43(3)
Study 2B
Following the methodology of Study 1B, we used an affective priming paradigm in which
European Canadian and Japanese participants were briefly presented with pictorial images of
Figure 5. Examples of the Stimuli Used in Study 2A.
Stimuli differed in terms of the valence of facial expression in the foreground and of facial expression in the
background. Four possible combinations were (a) negative foreground face–negative background faces, (b) positive
foreground face–negative background faces, (c) negative foreground face–positive background faces, and (d) positive
foreground face–positive background faces.
Figure 6. Mean Reaction Times (+SE) as a Function of Affective Background–Foreground Congruence
and Culture (Study 2A).
by guest on February 19, 2013jcc.sagepub.comDownloaded from

Ito et al. 439
happy or sad facial expressions before viewing and categorizing happy or sad facial emotions in
the target images.
Method
Stimuli. In this experiment, we used the same facial portraits as in Study 2A. Background faces
used as contexts were used as primes, and target faces were the same as in the previous three
studies. The size of background faces and target faces was identical to those used in Study 2A.
Participants. Forty-one European Canadians (21 females and 20 males) at the University of
Alberta and 36 Japanese (17 females and 19 males) at Kobe University participated in the experi-
ment. These participants had not participated in any of the previous studies. In exchange for their
participation, European Canadian participants received a course credit, and Japanese participants
received a gift card worth about $5.
Procedure. Experimental procedure was the same as in Study 1B, except that there were 192 trials
in total, and primes were facial emotions rather than landscape images (see Figure 7). Partici-
pants’ task was to categorize as fast and accurate as possible whether the person in the second
picture depicted positive or negative emotions.
Figure 7. Illustration of the Affective Priming Paradigm Used in Study 2B.
After being presented with a fixation point, participants saw (a) positive or (b) negative facial expressions for 250
ms as a prime. Following the interval (50 ms), they were to categorize (c) positive or (d) negative facial expression.
Emotionally congruent prime–target pairs refer to the combination of (a) and (c) or (b) and (d), whereas emotionally
incongruent prime–target pairs refer to the combination of (a) and (d) or (b) and (c).
by guest on February 19, 2013jcc.sagepub.comDownloaded from

440 Journal of Cross-Cultural Psychology 43(3)
Results
We then conducted 2 (Culture: European Canadian vs. Japanese) × 2 (Congruency: Congruent vs.
Incongruent) ANOVA, with participants’ reaction time as the dependent variable. Culture was the
between-subjects variable, and congruency was the within-subject variable.
12
We found a signifi-
cant main effect of congruency, F(1, 75) = 11.16, p < .001, η
p
2
= .13. Neither a main effect of
culture nor the two-way interaction of participants’ cultural backgrounds and congruency was not
significant, Fs < 1. However, planned contrast showed that Canadian participants’ response latency
in the congruent condition (M
CND
= 836.09) was not significantly faster than their response latency
in the incongruent condition (M
CND
= 848.86), t(40) = 1.54, p = .13, whereas Japanese participants
categorized targets significantly faster in the congruent condition (M
JPN
= 788.91) than in the
incongruent condition (M
JPN
= 813.56; see Figure 8), t(35) = 3.35, p = .002.
13,14
General Discussion
We reported two cross-cultural studies that investigated the effect of contextual affective priming
on the judgment of facial expressions. Study 1A placed the positive or negative landscape images
in the background of the target stimuli. Study 1B used a classic affective priming paradigm, in
which the positive or negative landscape images were briefly presented before the target stimulus
appeared on the screen. Study 2 used exactly the same experimental design as Study 1 except that,
instead of landscape images, facial expressions were used as the contextual information.
Is there any cultural similarity in findings across all experimental conditions? To verify the
effect of presentation methods and types of stimuli, we combined data from two studies and con-
ducted 2 (Presentation: Sequential vs. Simultaneous) × 2 (Stimuli: Landscape vs. Face) × 2 (Culture:
European Canadian vs. Japanese) between-subject ANOVA with the effect of context (the response
latency of the congruent condition subtracted from that of the incongruent condition) being the
dependent variable. We found a significant main effect of Presentation, F(1, 285) = 6.18, p < .02,
η
p
2
= .02. The context stimuli in Study 1A and Study 2A (i.e., simultaneous presentation) were
Figure 8. Mean Reaction Times (+SE) as a Function of Affective Prime-Target Congruence and Culture
(Study 2B)
by guest on February 19, 2013jcc.sagepub.comDownloaded from

Ito et al. 441
more salient than those in Study 1B and Study 2B (i.e., sequential presentation), suggesting that the
context were easily connected with the target’s facial expression when the two pieces of informa-
tion appear simultaneously. By contrast, if the context was presented before the target facial expres-
sion, such a time lag attenuated the context effect on the reaction speeds of the judgment of the
target’s facial expression.
15
In general, the magnitude of the context effect was similar for European
Canadians and Japanese participants that neither Culture × Presentation nor Culture × Stimuli were
significant, Fs < 1. The results successfully replicated Hietanen et al.’s (2007), who tested only the
sequential condition. Furthermore, we extended findings of their study by identifying a stronger
context effect in simultaneous condition compared with the sequential condition.
These cultural similarities across all experimental conditions lead us to conclude that when
people categorize a target’s facial expression, there might be a universal tendency for reaction
speed to be influenced by contextual cues. Just as emotional expressions do not occur in a vac-
uum in real life, judgment of facial expressions always occurs in context. So it is reasonable to
assume that people generally take into account the face-context linkage and that they are influenced
by the congruency and incongruency of this linkage (Matsumoto & Hwang, 2010; Mastumoto,
Hwang, & Yamada, 2012). Our findings regarding reaction times provide proof that the face-
congruent contexts influence judgment speed of the emotion on the target’s face equally for both
European Canadians and Japanese and that there are similar patterns between the two cultural
groups in terms of the magnitude of contextual effect. The results of the current study therefore
suggest that Masuda, Ellsworth et al.’s (2008) findings cannot be generalized to this experimen-
tal paradigm.
What were the differences in the Masuda, Ellsworth et al.’s (2008) study and the current
study, which could have demarcated cultural similarity from cultural variation? Because the
stimuli used in Study 2A or 2B and in Masuda, Ellsworth et al.’s study are qualitatively similar
to the extent that both stimuli consist of the target’s facial expressions as well as the background
people’s facial expressions, we speculate that the discrepancy in findings can be attributable to
the differences in measurement. The objective of the current studies was to investigate whether
or not the previously found cultural differences in context sensitivity were observable in the
information processing speed of affectively congruent or incongruent information; thus, we used
categorization tasks, in which we measured the reaction speed of categorizing the target’s facial
emotions. By contrast, Masuda, Ellsworth et al. asked participants to rate the intensity of targets’
facial emotions, targeting the interpretation of the facial emotions surrounded by congruent or
incongruent facial emotions. Therefore, it is possible that even in Masuda, Ellsworth et al.’s inten-
sity judgment task, both East Asians’ and North Americans’ reaction speed in identifying the
category of the target’s emotion is equally influenced by the contextual information. But in the
intensity judgment, which is believed to take longer than the category judgment, North Americans
succeeded in discounting the contextual influence to consistently judge the target’s facial expres-
sion across different background figures, whereas East Asians, because of their cultural belief
regarding facial expressions, did not make such an adjustment. Although experimental manipu-
lation of differences in measurement is beyond the scope of the current article, this issue is wor-
thy of future research with well-controlled experimentation.
Conclusion
The comparison of different paradigms reveals the interconnectedness of constructs in a theory
and provides new insights (Aronson, 1997). By focusing on competing research findings from
two research paradigms, the current studies investigated the magnitude of context effects in a
variety of experimental conditions, and the findings confirmed the robustness of the context effect
during the category judgment of facial expressions. Furthermore, conducting a series of systematic
by guest on February 19, 2013jcc.sagepub.comDownloaded from

442 Journal of Cross-Cultural Psychology 43(3)
cross-cultural studies allowed us to examine the cultural similarity and differences in the context
effect between European Canadians and Japanese. Consistent with findings of research in the field
of emotion, the current study suggests that contextual cues indeed play an important role in the
judgment of facial expressions, even in the case of North American observers. Taken together, our
two studies provide evidence that systematic cross-cultural comparison is an insightful methodol-
ogy not only for solving conflicts between two competing findings but also to articulate the mag-
nitude of effects while avoiding any overgeneralization due to the use of a single paradigm. To
date, little research has comprehensively attempted to synthesize competing paradigms by using
a cross-cultural comparison methodology. We maintain that the current article provides a good
model for further advances in the field of cross-cultural psychology.
Acknowledgments
We thank Keiko Ishii for her feedback in the early part of the research project. We also thank research
assistants at the Culture and Cognition Lab at the University of Alberta.
Declaration of Conflicting Interests
The authors declared that they had no conflicts of interests with respect to their authorship or the publica-
tion of this article.
Financial Disclosure/Funding
The project is supported by the standard grant of the Social Science and Humanities Research Council of
Canada (G124130642).
Notes
1. Methodologically, emotionally neutral contextual stimuli are difficult to obtain, because, if stimuli are
not emotionally salient, people’s impression for the stimuli is likely to depend on meanings associated
with the stimuli. Controlling for the meaning of the stimuli between two cultural groups or even within
one cultural group is difficult. Therefore, we decided to focus on emotionally positive or negative
stimuli.
2. We intentionally varied the ethnicity and the gender of faces to control for possible biases in judging
facial emotions (Elfenbein & Ambady, 2003; Merten, 2005).
3. Prior to the analysis, we excluded nonresponse cases, incorrect responses, and responses that were 3 SD
away from each participant’s condition means (5.0%).
4. The condition means indicated that Canadian participants seemed to be slower at evaluating facial
emotions than Japanese participants. Thus, we conducted an independent sample t test to compare
reaction times for the congruent condition and the incongruent condition between cultures. Neither
the congruent condition, t(72) = 1.46, p =.15, nor the incongruent condition, t(72) = 1.23, p = .22, were
significantly different across cultures.
5. We also conducted 2 (Culture: European Canadian vs. Japanese) × 2 (Congruency: Congruent vs.
Incongruent) ANOVA, with participants’ accuracy rate as the dependent variable. We found a sig-
nificant main effect of congruency, F(1, 72) = 17.15, p < .001, η
p
2
= .19. Both Canadian and Japanese
participants categorized the facial emotions more accurately in the condition (M
CND
= 98.6% and M
JPN
=
98.3%) than in the incongruent condition (M
CND
= 95.9% for Canadian and M
JPN
= 95.9% for Japanese).
However, the result should be evaluated with caution, because the significant effect might be due to the
ceiling effect that all condition means were above 95%.
6. Affective priming research (Hermans, Spruyt, & Eelen, 2003) and neurological evidence (Adolphs,
2002) suggest that affective priming occurs within 300 ms SOA.
7. Prior to the analysis, we excluded nonresponse cases, incorrect responses, and responses that were 3 SD
away from each participant’s condition means (4.6%).
by guest on February 19, 2013jcc.sagepub.comDownloaded from

Ito et al. 443
8. The condition means indicated that Canadian participants seemed to be faster at categorizing facial
emotions than Japanese participants. Thus, we conducted an independent sample t test to compare
reaction times for the congruent condition and the incongruent condition between cultures. Neither
the congruent condition, t(70) = 1.36, p =.19, nor the incongruent condition, t(70) = 1.20, p = .23, were
significantly different between cultures.
9. We also conducted 2 (Culture: European Canadian vs. Japanese) × 2 (Congruency: Congruent vs.
Incongruent) ANOVA, with participants’ accuracy rate as the dependent variable. There was neither
significant main effect nor interaction effect of culture and congruency, Fs < 1. The lack of significant
effect of congruency, which was significant with reaction time as the dependent variable, may be due
to the ceiling effect, because all condition means were above 97%.
10. Prior to the analysis, we excluded nonresponse cases, incorrect responses, and responses that were 3 SD
away from each participant’s condition means (4.8%).
11. We also conducted 2 (Culture: European Canadian vs. Japanese) × 2 (Congruency: Congruent vs.
Incongruent) ANOVA, with participants’ accuracy rate as the dependent variable. We found a signifi-
cant main effect of the congruency, F(1, 68) = 8.16, p = .006, η
p
2
= .11. Both Canadian and Japanese
participants categorized the facial emotions more accurately in the congruent condition (M
CND
= 97.2%
and M
JPN
= 98.7%) than in the incongruent condition (M
CND
= 96.5% and M
JPN
= 96.6%). However, the
result should be evaluated with caution, because the significant effect might be due to the ceiling effect
that all condition means were above 96%.
12. Prior to the analysis, we excluded nonresponse cases, incorrect responses, and responses that were 3 SD
away from each participant’s condition means (4.4%).
13. We also conducted 2 (Culture: European Canadian vs. Japanese) × 2 (Congruency: Congruent
vs. Incongruent) ANOVA, with participants’ accuracy rate as the dependent variable. We found
a marginally significant main effect of the congruency, F(1, 75) = 3.91, p = .52, η
p
2
= .05. Both
Canadian and Japanese participants categorized the person more accurately when the background–
target pair had congruent valences (M
CND
= 97.4% and M
JPN
= 98.4%) than when the pair had
incongruent valences (M
CND
= 97.1% and M
JPN
= 97.5%). However, the result should be evaluated
with caution, because the significant effect might be due to the ceiling effect that all condition
means were above 97%.
14. The results support previous cross-cultural findings (e.g., Masuda, Ellsworth et al., 2008); that is,
although Japanese participants remained sensitive to context information, Canadian participants did
not show evidence of sensitivity to contextual information when the valence of context was modest.
However, the results must be interpreted with caution, because they may be due to the lack of power
given that the pattern of results is similar to the rest of the study.
15. Stimuli × Presentation interaction was marginally significant, F(1, 285) = 3.32, p < .07, η
p
2
= .01.
Simple effect analysis revealed a significant effect of presentation controlling for the type of stimuli.
That is, the simultaneously presented landscape images influenced participants’ judgment of facial
expression more than sequentially presented landscape images, F(1, 144) = 9.96, p < .01, η
p
2
= .07, but
such effect was not significant when facial expressions were used as contexts, F(1, 145) < 1.
References
Adams, G., & Markus, H. R. (2004). Toward a conception of culture suitable for a social psychology of
culture. In M. Schaller & C. S. Crandall (Eds.), The psychological foundations of culture (pp. 335-360).
Mahwah, NJ: Lawrence Erlbaum.
Adolphs, R. (2002). Recognizing emotion from facial expressions: Psychological and neurological mecha-
nisms. Behavioral and Cognitive Neuroscience Reviews, 1, 21-62.
Aronson, E. (1997). The theory of cognitive dissonance: The evolution and vicissitudes of an idea. In
C. McGarty & S. A. Haslam (Eds.), The message of social psychology (pp. 20-35). Cambridge, MA:
Blackwell Publishers Inc.
by guest on February 19, 2013jcc.sagepub.comDownloaded from

444 Journal of Cross-Cultural Psychology 43(3)
Carroll, J. M., & Russell, J. A. (1996). Do facial expressions signal specific emotions? Judging emotion
from the face in context. Journal of Personality and Social Psychology, 70, 205-218.
Carroll, N. C., & Young, A. W. (2005). Priming of emotion recognition. Quarterly Journal of Experimental
Psychology A: Human Experimental Psychology, 58A, 1173-1197.
Chua, H. F., Boland, J. E., & Nisbett, R. E. (2005). Cultural variation in eye movements during scene
perception. Proceedings of the National Academy of Sciences of the United States of America, 102,
12629-12633.
Ekman, P., & Friesen, W. V. (1975). Unmasking the face: A guide to recognizing emotions from facial
clues. Englewood Cliffs, NJ: Prentice-Hall.
Elfenbein, H. A., & Ambady, N. (2003). When familiarity breeds accuracy: Cultural exposure and facial
emotion recognition. Journal of Personality and Social Psychology, 85, 276-290.
Fazio, R. H. (2001). On the automatic activation of associated evaluations: An overview. Cognition &
Emotion, 15, 115-141.
Hermans, D., Spruyt, A., & Eelen, P. (2003). Automatic affective priming of recently acquired stimulus
valence: Priming at SOA 300 but not at SOA 1000. Cognition & Emotion, 17, 83-99.
Hietanen, J. K., Klemettilä, T., Kettunen, J. E., & Korpela, K. M. (2007). What is a nice smile like that
doing in a place like this? Automatic affective responses to environments influence the recognition of
facial expressions. Psychological Research, 71, 539-552.
Ji, L., Peng, K., & Nisbett, R. E. (2000). Culture, control, and perception of relationships in the environ-
ment. Journal of Personality and Social Psychology, 78, 943-955.
Kitayama, S., Duffy, S., Kawamura, T., & Larsen, J. T. (2003). Perceiving an object and its context in dif-
ferent cultures: A cultural look at New Look. Psychological Science, 14, 201-206.
Lang, P. J., Bradley, M. M., & Cuthbert, B. N. (2005). International affective picture system (IAPS): Affec-
tive ratings of pictures and instruction manual (Tech. Rep. No. A-6). Gainesville, FL: University of
Florida, Center for Research in Psychophysiology.
Leppänen, J. M., & Hietanen, J. K. (2003). Affect and face perception: Odors modulate the recognition
advantage of happy faces. Emotion, 3, 315-326.
Masuda, T., Akase, M., Radford, M. H. B., & Wang, H. (2008). Jokyo yoin ga gankyu undo pattern ni
oyobosu eikyo: Nihonjin to Seiyojin no syuken jyoho heno bunkasa no kikaku kenkyu [Cross-cultural
research on the pattern of eye-movement: Comparing the level of concentration between Japanese and
Western participants]. Japanese Journal of Psychology, 79, 35-43.
Masuda, T., Ellsworth, P. C., Mesquita, B., Leu, J., Tanida, S., & Van de Veerdonk, E. (2008). Placing
the face in context: Cultural differences in the perception of facial emotion. Journal of Personality and
Social Psychology, 94, 365-381.
Masuda, T., Gonzalez, R., Kwan, L., & Nisbett, R. E. (2008). Culture and aesthetic preference: Comparing
the attention to context of East Asians and Americans. Personality and Social Psychology Bulletin, 34,
1260-1275.
Masuda, T., & Nisbett, R. E. (2001). Attending holistically versus analytically: Comparing the context
sensitivity of Japanese and Americans. Journal of Personality and Social Psychology, 81, 922-934.
Masuda, T., & Nisbett, R. E. (2006). Culture and change blindness. Cognitive Science: A Multidisciplinary
Journal, 30, 381-399.
Masuda, T., Wang, H., & Ishii, K. (2008, February). Comparing patterns of attention between Canadian
and East Asian international students. Poster session presented at the annual meeting of the Society of
Personality and Social Psychology, Albuquerque, New Mexico.
Matsumoto, D., & Hwang, H. S. (2010). Judging faces in context. Social and Personality Psychology
Compass, 3, 1-10.
Matsumoto, D., Hwang, H. S., & Yamada, H. (2012). Cultural differences in the relative contributions of
face and context to judgments of emotion. Journal of Cross-Cultural Psychology, 43(2), 198–218.
by guest on February 19, 2013jcc.sagepub.comDownloaded from

Ito et al. 445
Matsumoto, D., & Yoo, S. H. (2006). Toward a new generation of cross-cultural research. Perspectives on
Psychological Science, 1, 234-250.
Merten, J. (2005). Culture, gender and the recognition of the basic emotions. Psychologia: An International
Journal of Psychology in the Orient, 48, 306-316.
Miyamoto, Y., Nisbett, R. E., & Masuda, T. (2006). Culture and the physical environment: Holistic versus
analytic perceptual affordances. Psychological Science, 17, 113-119.
Nisbett, R. E. (2003). The geography of thought: How Asians and Westerners think differently . . . and why.
New York: Free Press.
Nisbett, R. E., & Masuda, T. (2003). Culture and point of view. Proceedings of the National Academy of
Sciences of the United States of America, 100, 11163-11175.
Nisbett, R. E., Peng, K., Choi, I., & Norenzayan, A. (2001). Culture and systems of thought: Holistic versus
analytic cognition. Psychological Review, 108, 291-310.
Russell, J. A. (1991). The contempt expression and the relativity thesis. Motivation and Emotion, 15, 149-168.
Russell, J. A., & Fehr, B. (1987). Relativity in the perception of emotion in facial expressions. Journal of
Experimental Psychology: General, 116, 223-237.
Schneider, W., Eschman, A., & Zuccolotto, A. (2002). E-Prime user’s guide. Pittsburgh, PA: Psychology
Software Tools.
Tanaka-Matsumi, J., Attivissimo, D., Nelson, S., & D’Urso, T. (1995). Context effects on the judgment of
basic emotions in the face. Motivation & Emotion, 19, 139-155.
by guest on February 19, 2013jcc.sagepub.comDownloaded from