Evolved Disease-Avoidance Mechanisms and Contemporary ...

G
PGroup Processes &

Intergroup Relations
I2004 Vol 7(4) 333–353
R
Evolved Disease-Avoidance

Mechanisms and

Contemporary

Xenophobic Attitudes

Jason Faulkner, Mark Schaller, Justin H. Park
and Lesley A. Duncan

University of British Columbia

From evolutionary psychological reasoning, we derived the hypothesis that chronic and
contextually aroused feelings of vulnerability to disease motivate negative reactions to foreign
peoples. The hypothesis was tested and supported across four correlational studies: chronic
disease worries predicted implicit cognitions associating foreign outgroups with danger, and
also predicted less positive attitudes toward foreign (but not familiar) immigrant groups. The
hypothesis also received support in two experiments in which the salience of contagious disease
was manipulated: participants under high disease-salience conditions expressed less positive
attitudes toward foreign (but not familiar) immigrants and were more likely to endorse policies
that would favor the immigration of familiar rather than foreign peoples. These results reveal a
previously under-explored influence on xenophobic attitudes, and suggest interesting linkages
between evolved disease-avoidance mechanisms and contemporary social cognition.

keywords disease, immigration, intergroup attitudes, threat

History teaches us that society has no shortage of disease. This association is evident in xenopho-
means available to dehumanize ‘undesirable’ bic propaganda, in which ethnic outgroups are
groups. The grave risks of this process are magni- likened to non-human vectors of disease, such
fied when combined with the threat of infectious as rats, flies, and lice (Suedfeld & Schaller,
disease. (Markel, 1999) 2002). The associative link between foreign
peoples and disease shows up consistently in
HISTORICAL analyses of stereotypes, prejudice the social science literature on immigration.
and exclusionary intergroup behavior reveal a
tendency to associate foreign peoples with Author’s note
disease. Ethnic outgroups are often blamed for Address correspondence to Jason Faulkner,
outbreaks of epidemic diseases, and these out- Department of Psychology, University of
breaks can inspire violently xenophobic reac- British Columbia, 2136 West Mall, Vancouver,
tions to outsiders (Goldhagen, 1996; Markel, BC V6T 1Z4, Canada
1999; Oldstone, 1998). Foreigners are also [email: [email protected]]
associated with semantic concepts that connote

Copyright © 2004 SAGE Publications
(London, Thousand Oaks, CA and New Delhi)
7:4; 333–353; DOI: 10.1177/1368430204046142

Group Processes & Intergroup Relations 7(4)

Anti-immigrant literature in ancient Rome human mind evolved in response to adaptive
likened outsiders to detritus and scum (Noy, problems imposed by the environments in
2000). Anti-immigrant attitudes in England which ancestral populations lived; (b) for the
have been linked to beliefs that foreigners eat past several million years of this evolutionary
disturbingly unclean foods (roots, frogs, haggis history, ancestral populations lived in social
and other foods ‘fit only for dogs’; Stat, 1995, groups; (c) specific adaptive problems associ-
p. 192). And in the United States, Markel and ated with group life may have given rise to
Stern (2002, p. 757) write, ‘Despite the specific psychological mechanisms that facili-
dramatic changes in demography, the meaning tate negative perceptions of and reactions to
of citizenship, and the ability to treat and cure individuals associated with specific sorts of
acute and chronic diseases, foreigners were con- groups. In other words, certain prejudicial atti-
sistently associated with germs and contagion.’ tudes may have conferred adaptive benefits
within ancestral environments, and now—even
Remarkably, this topic has received virtually though contemporary environments are very
no empirical attention in the psychological different in very many ways—the processes
literature on intergroup stereotypes and preju- underlying those prejudicial attitudes persist.
dice. There are, of course, many additional
psychological processes implicated in inter- Within this broad evolutionary framework,
group relations, and so there are many other different theoretical models identify links
unrelated personality factors and contextual between specific adaptive problems of the past
variables that may contribute to xenophobic and specific psychological processes in the
attitudes (see Fiske, 1998, for a review). In fact, present. One model links the adaptive utility of
there are many other kinds of disease-irrelevant cooperative coalitions with contemporary ten-
threats that can contribute to intergroup anxi- dencies to exclude and exploit members of out-
eties and prejudices (e.g. Stangor & Crandall, groups (Kurzban & Leary, 2001). Another
2000; Stephan, Ybarra, & Bachman, 1999). Still, model links the adaptive utility of status hierar-
given the apparent historical importance of chies with contemporary tendencies to express
disease in shaping societal-level reactions to domination over outgroups (Sidanius & Pratto,
foreign peoples, it seems fruitful to consider 1999). Another model links the adaptive utility
closely its analogous role in the psychology of of avoiding potentially dangerous intergroup
individuals. That is the purpose of this article. encounters with contemporary tendencies to
We employ an evolutionary framework to perceive outgroups in derogatory ways
derive—and then empirically test—the hypoth- (Schaller, 2003). Yet another model links the
esis that people who feel most vulnerable to adaptive utility of avoiding harmful pathogens
contagious diseases are likely to react especially and parasites with contemporary prejudices
negatively toward subjectively foreign (but not against individuals who are perceived, at some
familiar) outgroups. heuristic level, to be potential carriers of
pathogens or parasites (Kurzban & Leary, 2001;
Disease-avoidance mechanisms and Park, Faulkner, & Schaller, 2003). It is this
intergroup attitudes disease-avoidance model specifically that
informs the current investigation.
A number of theoretical models have suggested
ways in which contemporary intergroup cogni- Disease-avoidance and prejudicial
tion and behavior may reflect psychological person perception
mechanisms that evolved a long time ago (e.g. People typically experience feelings of uneasi-
Kurzban & Leary, 2001; Neuberg, Smith, & ness and even disgust when they encounter
Asher, 2000; Reynolds, Falger, & Vine, 1987). individuals perceived to be carriers of conta-
The common evolutionary logic underlying gious disease. These psychological experi-
these perspectives is summarized by Kurzban ences—and the behavioral avoidance that they
and Leary (2001): (a) The contemporary precipitate—may have evolutionary roots

334

Faulkner et al. evolved disease-avoidance mechanisms

(Kurzban & Leary, 2001; Rozin, Haidt, behavioral avoidance, it is likely that there are
McCauley, & Imada, 1997). The logic is straight- evolved mechanisms through which the per-
forward. In ancestral environments, many ception of disease-connoting cues elicits a
communicable parasites and diseases were haz- specific affective reaction, and activates cogni-
ardous to individuals’ health. Behavioral ten- tions in which the perceptual object is associa-
dencies that helped individuals to avoid tively linked to semantic concepts that connote
infection would have been adaptive. One disease and more generally motivate behavioral
adaptive tendency would have been the selective avoidance. The affective reaction is likely to be
avoidance of individuals who were especially that of disgust. Disgust probably evolved pri-
likely to be carriers of contagious diseases. marily as a means to inhibit the ingestion of
poisons and other harmful agents, but it plays
A prerequisite for behavioral avoidance is an important role in interpersonal avoidance as
some set of psychological mechanisms that well; disgust is elicited by subjectively ‘offensive’
facilitate learning and detection of cues that people and motivates physical distancing from
heuristically indicate the possible presence of those people (Rozin, Haidt, & McCauley,
contagious diseases. Thus, individuals may be 2000). Schiefenhövel (1997) notes that people
perceptually sensitive to superficial physical often display disgust reactions when speaking
features such as lesions, disfigurements, and about ethnic outgroups, and Rozin et al. (1997,
other morphological abnormalities. (Cues such p. 73) suggest that ‘disgust in humans serves as
as these are, of course, highly imperfect indi- an ethnic or outgroup marker’. Cognitive
cators of disease; however, evolution of a cue- responses are likely to be varied. They may
recognition mechanism requires only a modest include a wide variety of moralist judgments
correlation between the presence of the cue and beliefs (Haidt, 2001). They may also
and the actual presence of disease; see Kurzban include stereotypical beliefs in which disease-
& Leary, 2001.) In addition to these and other relevant characteristics such as ‘dirty’ or ‘unhy-
morphological cues that heuristically connote gienic’ are attributed to subjectively foreign
disease, it is also likely that individuals are groups. Most obviously, perhaps, these cogni-
especially sensitive to superficial cues indicating tive responses are likely to precipitate xeno-
that an individual is subjectively foreign to phobic attitudes—attitudes indicating that
one’s own group. There are several reasons why foreigners are to be avoided, excluded, and
subjectively foreign peoples might have become otherwise kept at a distance.
heuristically associated with disease. First,
contact with previously unknown populations Moderating influences on disease-avoidant
brings with it an increased risk of encountering reactions to foreign peoples
contagious diseases to which one has no Responses designed over evolutionary time to
acquired immunity (Diamond, 1999). Also, avoid recurrent threats not only confer poten-
individuals from foreign cultures are more tial functional benefits; they entail potential
likely to engage in subjectively strange customs costs as well (e.g., avoidant behavior consumes
(pertaining to hygiene, food preparation, etc.) calories and distracts resources from other
that may violate local customs that inhibit trans- functional activities). Therefore, these responses
mission of disease. Consequently, individuals are more functional in some situations than
are likely to be especially adept at learning to others. They are most functional under con-
detect a wide range of cues that—within their ditions in which the benefits most clearly
local social environment—discriminate between outweigh the costs. As with most other evolved
familiar and foreign peoples. psychological mechanisms, evolved disease-
avoidance mechanisms are likely to be engaged
Cue learning and perception is, by itself, flexibly, depending on additional information
insufficient to cause behavioral avoidance. bearing on the relative costs and benefits of
Another prerequisite is the activation of response.
psychological mechanisms that translate per-
ception into action. Thus, in order to facilitate

335

Group Processes & Intergroup Relations 7(4)

The actual benefits of a xenophobic response individual differences in perceived vulnerabil-
to foreigners were likely to be very low (and ity to disease may predict xenophobic reactions
outweighed by costs) under conditions in which to members of subjectively foreign ethnic
foreign peoples are far away, or under con- groups. We tested this specific hypothesis in
ditions in which perceivers are invulnerable to one set of four studies reported here.
disease. The benefits were likely to be higher
under conditions in which perceivers were Contextually aroused perceptions of vulnera-
actually vulnerable to the interpersonal trans- bility to disease Just as evolved mechanisms
mission of infectious agents. Therefore, just as are responsive to chronic individual differ-
other evolved psychological mechanisms are ences, they are similarly responsive to input
moderated by a variety of functionally-relevant from temporary contextual variables (Schaller,
local variables (Conway & Schaller, 2002; 2003). Any contextual information that implies
Gangestad & Simpson, 2000), the disease- increased vulnerability to disease might also
avoidance process that inspires xenophobic amplify negative reactions to categories of
responses to foreign peoples is likely to have people heuristically associated with disease.
evolved in such a way that it is sensitive to, and Disease-relevant stimuli in contemporary
moderated by, variables influencing the extent environments are abundant and include, for
to which perceivers feel personally vulnerable example, news stories about the latest disease
to disease. epidemic, advertisements for hygienic products
that make salient the omnipresence of germs
Chronic perceptions of vulnerability to disease and bacteria, and disgust-eliciting objects like
Evolved threat-avoidance mechanisms may be sewers and odorous garbage cans. Even when
pan-human, but there is considerable variation these contextual stimuli do not explicitly impli-
across individuals in the tendency for these cate foreign outgroups as a source of disease,
mechanisms to be triggered. For example, they may nevertheless trigger xenophobic
disgust is a universal human emotion, but some responses to foreign peoples. This specific
people tend to be more sensitive to disgust hypothesis—that temporarily-heightened vul-
than others (Haidt, McCauley, & Rozin, 1994). nerability to disease amplifies negative reac-
Similarly, for a variety of reasons bearing both tions to foreigners—was tested in an additional
on genetic heritage and social-learning pro- set of two studies reported here.
cesses, some people are likely to feel more
vulnerable to disease than others. Regardless Overview of studies
of the origins of individual differences in
perceived vulnerability to disease, these indi- We conducted six studies to examine the effects
vidual differences are likely to moderate the of chronic and contextually-aroused percep-
activation of evolved disease-avoidance tions of vulnerability to disease on xenophobic
mechanisms in the domain of person percep- reactions to foreign peoples. Four studies were
tion. Previous research has, in various ways, correlational: we examined the extent to which
revealed relations between individuals’ self-per- xenophobic attitudes were predicted by self-
ceived vulnerability and negative reactions to reported chronic perceptions of vulnerability to
others who actually are diseased, as well as to disease. In addition, we conducted two experi-
others who are stigmatized by characteristics— ments, which included procedures designed to
such as physical disability—that heuristically experimentally manipulate perceptions of vul-
connote disease (Bishop, Alva, Cantu, & nerability to disease. Across all studies, partici-
Rittiman, 1991; Crandall, Glor, & Britt, 1997; pants’ attitudes toward subjectively foreign or
Herek & Capitaniato, 1998; Park et al., 2003; familiar outgroups were assessed. Based on the
Rozin, Markwith, & Nemeroff, 1992). This reasoning outlined above, we expected that
analysis suggests a testable hypothesis that has participants who felt (either chronically or
not previously received empirical attention: temporarily) more vulnerable to disease would

336

Faulkner et al. evolved disease-avoidance mechanisms

react more negatively toward subjectively Studies 1–4: perceived vulnerability to disease
foreign outgroups, but not toward outgroups correlates with xenophobic attitudes
who were subjectively familiar. Assessment of perceived vulnerability to disease
In order to test the hypothesis, we developed a
Our conceptual analysis implies that the pre- questionnaire measure to assess individual
dicted effects should be observed primarily on differences in chronic perceptions of vulnera-
measures that assess specific attitudes toward bility to disease. This ‘perceived vulnerability to
target groups marked by characteristics con- disease’ (PVD) measure included 14 items that
noting foreign-ness. These effects are less assessed general beliefs about personal suscep-
likely to occur on measures that assess non- tibility to illness (e.g. ‘I have a history of sus-
target-specific attitudes. All studies included ceptibility to infectious disease’, ‘My immune
target-specific attitude measures to test the pre- system protects me from most illnesses that
dictions. In addition, several studies included other people get’), and discomfort with behav-
additional measures that were not target- iors that connote a higher risk of germ trans-
specific (these measures assessed attitudes mission (e.g. ‘I prefer to wash my hands pretty
toward immigration in general); this allowed us soon after shaking someone’s hand’, ‘I’m com-
to examine the extent that the predicted effects fortable sharing a water bottle with a friend’).
were restricted to specific target groups. Each item was followed by a 7-point response
scale with the endpoints labeled ‘Strongly
The conceptual analysis also implies that the Disagree’ and ‘Strongly Agree.’ All items were
predicted effects are likely to be observed on scored such that higher values indicated
measures that assess attitudes more directly greater perceived vulnerability (6 items were
relevant to functional behavior, such as physical reverse scored). An overall PVD score for each
avoidance or social exclusion. Consequently, all participant was computed by averaging
studies included measures that seem directly responses across all 14 items.
relevant (e.g. implicit associations between the
target group and danger, attitudes about the Preliminary psychometric evaluation indi-
immigration of target groups to participants’ cated that the PVD scale had adequate internal
local area). These measures were the primary validity (Cronbach’s alphas were typically in
focus of our analyses. the range from .70 to .80 across multiple
samples). Indicators of discriminant validity
It is less clear how strongly the underlying were also encouraging: It was not meaningfully
process might compel perceivers to attribute correlated with conceptually irrelevant con-
very specific traits to target groups. It seems structs (e.g. social desirability, self-esteem, locus
plausible that, even in the absence of direct of control), and was correlated—modestly—
supporting information about unknown with only one of the Big Five personality dimen-
peoples, perceivers might be more likely to sions, neuroticism ( John & Srivastava, 1999; r s
judge these peoples more negatively on of .20 and .22 across two samples). Convergent
specific traits that seem logically relevant to validity was indicated by positive correlations
disease (e.g. unclean). However, the affective between PVD and a measure of ‘belief in a
system (disgust) that presumably underlies the dangerous world’ (Altemeyer, 1988; r s typically
disease-avoidance process tends to compel a ranged between .30 and .40 across samples),
rather diffuse set of negative associations and with sensitivity to disgust (Haidt et al.,
(Haidt, 2001). Moreover, in the absence of 1994; r s ranged between .22 and .53 across sub-
actual information supporting specific attribu- scales assessing disgust in different domains).
tions, participants may be reluctant to offer
specific stereotypical judgments. To explore The PVD measure was administered to
the possibility of effects on trait attributions, participants in Studies 1–4. In each study, we
several of the studies included measures assess- examined correlations between overall PVD
ing participants’ judgments of target groups on score and the various measures assessing
a variety of traits both relevant and irrelevant intergroup attitudes. Studies 1 and 2 focused
to disease.

337

Group Processes & Intergroup Relations 7(4)

on attitudes toward African outgroups. Africans was designed to assess associations with
were expected to be perceived as culturally ‘danger’. The IAT procedures are described in
foreign to participants (all participants were detail elsewhere (Greenwald et al., 1998;
students at the University of British Columbia, Schaller et al., 2003); therefore, these methods
which is a campus populated predominantly by are described in abbreviated form here.
peoples of East Asian and European heritage,
but with very few students of African heritage). Both IAT tasks were presented on IBM-com-
Studies 3 and 4 examined attitudes toward patible computers. Participants were presented
several other outgroups that a separate sample with a series of forced-choice categorization
of UBC students rated as either subjectively trials, and were instructed to respond as quickly
foreign or familiar. as possible. On each trial, a stimulus (i.e. a word
or a picture) was presented in the center of the
The hypothesized relation between PVD and computer screen, along with two response
intergroup attitudes is clearly directional; indi- categories (e.g. ‘unpleasant’ and ‘pleasant’)
viduals who perceive themselves to be more presented on the upper left and upper right
vulnerable to disease are expected to react more sides of the screen. Participants responded by
negatively toward subjectively foreign ethnic pressing one of two keyboard keys: either the
groups. Given this clear directional prediction, E key, corresponding to the response category
when correlations consistent with this hypothe- on the upper left side of the screen, or the
sis are observed, we report p values associated I key, corresponding to the response category
with directional (i.e. ‘one-tailed’) tests of the on the upper right side of the screen.
statistical null hypothesis. When correlations are
observed that are not obviously predicted by the On an ‘unpleasant/pleasant’ IAT, partici-
hypothesis, we report p values associated with pants judged whether words were unpleasant or
more conservative nondirectional tests of the pleasant, and on a ‘danger/safety’ IAT, partici-
null hypothesis, and explicitly identify them pants judged whether words connoted danger
as ‘two-tailed’. Exact p values are reported or safety. For both IAT tasks, participants also
throughout, so that readers can employ alterna- judged the ethnicity of faces: participants of
tive strategies of statistical inference if desired. Asian backgrounds judged whether faces were
African or Asian and all other participants
Study 1 judged whether faces were African or European.
An initial test of the hypothesis emerged from Each IAT task consisted of distinct blocks of
a study that was designed primarily to test trials; in two critical blocks, participants com-
hypotheses distinct from those under consider- pleted both word and face categorization trials.
ation here (additional results are reported in In the first critical block, the same keyboard key
Schaller, Park, & Mueller, 2003; Study 2). The was used to indicate that faces were African and
study employed the ‘implicit association test’ words were unpleasant (in the ‘unpleasant/
(IAT; Greenwald, McGhee, & Schwartz, 1998; pleasant’ IAT) or connoted danger (in the
Rudman, Greenwald, Mellott, & Schwartz, ‘danger/safety’ IAT). Responses on trials in this
1999) to assess implicitly activated semantic block were expected to be relatively fast if
associations with the social category ‘African’. participants implicitly associated African faces
with ‘unpleasant’ or ‘danger’. In the second
Participants were 52 undergraduate students critical block, keyboard responses were
at UBC. All participants first completed a set of changed, so that different keys were used to
questionnaires, one of which was the 14-item indicate ‘African’ and ‘unpleasant’/’danger’.
PVD measure. Participants then completed two Responses on trials in this block were expected
IAT tasks, presented in counterbalanced order, to be relatively slow if participants implicitly
that were developed to assess implicit associ- associated African faces with ‘unpleasant’ or
ations between Africans and specific evaluative ‘danger’. The difference in average response
characteristics. One IAT was designed to assess times within each of the two critical blocks of
associations with ‘unpleasant’; the other IAT trials (mean rt in the 2nd critical block minus

338

Faulkner et al. evolved disease-avoidance mechanisms

mean rt in the 1st critical block) served as an Open-ended and rating scale techniques
indicator of the extent to which Africans were were then used to assess attitudes toward Kras-
differentially associated with ‘unpleasant’ or neeans. Participants were first asked to ‘Write 5
‘danger’. words that you feel describe Krasneeans’, and
were also asked to draw a picture representing
Results reported elsewhere (Schaller et al., ‘what you would imagine a typical Krasneean to
2003) revealed that there was a general look like’. One coder, working from a master
tendency among participants to differentially list, counted the number of words indicated by
associate Africans with both ‘unpleasant’ and each participant that connoted some sort of
with ‘danger’. The key question here is whether danger or threat. Four additional coders looked
variability in these responses was associated with at each drawing and rated on a 10-point scale
individual differences in perceived vulnerability the demeanor of the Krasneean(s) depicted in
to disease. Correlations revealed that PVD had the drawing (1 = most unfriendly; 10 = most
a negligible relation with the unpleasant/ friendly). The four sets of ratings were stan-
pleasant IAT index (r = .05, p = .351), but posi- dardized and then averaged to create an index
tively predicted the danger/safety IAT index of the perceived friendliness of the typical
(r = .24, p = .045). Krasneean (Cronbach’s alpha = .92).

These results provide initial evidence that Additional measures of participants’ atti-
individual differences in PVD predict negative tudes toward Krasneeans employed rating-scale
implicit attitudes about people of African techniques. Participants were asked to rate, on
heritage. Interestingly, PVD did not predict the six separate 9-point scales, the extent to which
extent to which Africans were associated with Krasneeans were likeable, hostile, trustworthy, open-
unpleasant characteristics in general; but it did minded, ignorant, and poor. To assess attitudes
predict the extent to which Africans were toward immigration, participants were also
associated with danger. This pattern of results asked to rate, on five point scales, the extent to
suggests that chronic feelings of vulnerability to which they agreed with various statements. One
disease may be associated specifically with item specifically assessed attitudes toward Kras-
threat-relevant beliefs, which are likely to stimu- neean immigration, and read, ‘None of the
late avoidance behaviors, but are not as strongly Krasneeans applying for citizenship should be
associated with other negative beliefs that are allowed to immigrate to Canada’, and four
less functionally relevant to threat-avoidance. additional items assessed attitudes toward
Canadian immigration policies in general (e.g.
Study 2 agreement with the statement, ‘Canada’s immi-
The predictive effects of PVD were also gration policies are too strict’). Participants’
examined in a second study that was originally scores on these four items were combined to
designed for a different primary purpose. create a composite index of general anti-
Participants in this study were 82 undergradu- immigration attitudes.
ate students at UBC. All participants completed
several questionnaires, including the PVD Of primary interest as an indicator of xeno-
measure. They were then presented with a one- phobia were responses to the item ‘None of the
page written passage that began, ‘There is a Krasneeans applying for citizenship should be
country in Central Africa, which for the allowed to immigrate to Canada.’ Results
purposes of this study we will refer to as revealed a modest positive correlation between
Krasnee, that has been experiencing a great PVD and endorsement of this statement (r = .21,
deal of civil unrest in recent years. A number of p = .032): individuals with higher PVD scores
people from this country are trying to leave.’ more strongly opposed Krasneean immigration.
The passage went on to describe a group of
Krasneeans who, for various reasons, were There was little relation between PVD and
seeking to immigrate to Vancouver, Canada, anti-immigration attitudes in general (r = .13,
where participants in the study lived. p = .117). On the measures assessing specific
beliefs about the characteristics of Krasneeans,

339

Group Processes & Intergroup Relations 7(4)

PVD scores correlated weakly with the number tudes only toward an immigrant group from
of danger-relevant traits participants used to Africa—in this study, the geographical origin of
describe Krasneeans (r = .14, p = .099), but the immigrant group was experimentally
correlated more strongly with the demeanor of manipulated: the group was described as being
the Krasneeans represented in participants’ either from eastern Africa, eastern Asia, or
drawings (r = –.31, p = .002): Those with eastern Europe. A separate sample of 23 UBC
higher PVD scores drew less friendly looking students rated immigrants from eastern Africa
Krasneeans. PVD correlated with ratings of as subjectively foreign in general, and rated
Krasneeans on two trait characteristics: likable them as likely to engage in unusual practices in
(r = –.33, p = .002), and trustworthy (r = –.22, disease-relevant domains (e.g. hygiene and
p = .025). Individuals with higher PVD scores food preparation). In contrast, immigrants
rated Krasneeans to be less likable and less from eastern Asia and eastern Europe were
trustworthy. There were negligible relations rated as subjectively familiar, and were per-
between PVD and ratings of hostile, open-minded, ceived to have hygiene and food preparation
ignorant, or poor (–.16 < r s < .06). practices similar to those of participants them-
selves. Therefore, PVD is expected to predict
These results complement those of Study 1 negative attitudes toward African immigrants,
and reveal PVD predicted exclusionary atti- but not toward East Asian and European immi-
tudes toward a specific African immigrant grants.
group. There was some additional evidence
that PVD predicted some more specific trait- Participants were 86 undergraduate students
based beliefs about the immigrant group as at UBC. All participants completed several
well. One could argue that the perceived questionnaires including the PVD scale. Partici-
characteristics predicted most strongly by PVD pants then listened to a voice on audiotape,
(unfriendliness, unlikability, and untrustwor- briefly describing an ethnic population
thiness) are traits that, when observed in identified as ‘Saznians’. As in the previous
others, typically motivate attempts to avoid study, the description stated that Saznians were
close contact. Characteristics that less clearly seeking to immigrate to Vancouver, Canada.
motivate avoidance (e.g. ‘ignorant’ or ‘poor’) Participants were randomly assigned to hear, as
were not predicted by PVD. Nor did PVD mean- part of the audiotaped message, that the
ingfully predict the perception of hostility; Saznian homeland was either in ‘eastern
although perceived hostility certainly connotes Africa’, ‘eastern Asia’, or ‘eastern Europe’.
a danger to be avoided, the danger it connotes
is not functionally relevant to disease avoid- Participants then rated the extent that they
ance. It is interesting too that, although PVD agreed with various statements using 5-point
predicted opposition to immigration from the scales. The first item was of primary interest in
specific target group, it did not as strongly that it assessed attitudes specific to Saznian
predict anti-immigration attitudes in general. immigration: participants rated the extent to
This perhaps provides some initial evidence which they disagreed or agreed with the state-
that chronic feelings of vulnerability to disease ment, ‘The Canadian government should allow
underlie attitudes reflecting disease-avoidance the Saznians from [Africa/Asia/Europe] to
concerns with subjectively foreign, but not enter Canada.’ On an additional six items,
familiar outgroups. We conducted additional participants rated the extent that they agreed
studies to more directly explore this possibility. with other general pro- or anti-immigration
statements. A composite index of general pro-
Study 3 immigration attitudes was created by reverse
In a third study, we used methods similar to the scoring responses to anti-immigration state-
previous study to test the relation between PVD ments and then averaging responses to all
and attitudes toward potential immigrant items.
groups. Unlike Study 2—which assessed atti-
As expected, the relation between PVD and
340 endorsement of Saznian immigration depended

Faulkner et al. evolved disease-avoidance mechanisms

importantly on the geographical origin of vulnerable to disease to have stronger anti-
Saznians (see Figure 1). When Saznians were immigration attitudes in general. However,
described as African, there was a substantial when expressing attitudes toward specific
negative correlation between PVD and immigrant groups, PVD does not predict anti-
endorsement of Saznian immigration (r = –.38, immigrant attitudes equally. In this sample
p = .016): individuals with higher PVD scores (drawn from a population comprised primarily
were less keen on the immigration of African of people of either European or East Asian
Saznians. In contrast, correlations were slightly ethnicity), PVD did not predict more negative
positive under conditions in which Saznians reactions toward immigration of eastern
were described as either eastern Asian or European or eastern Asian outgroups, but it did
European (r s were .15 and .16, respectively; predict more negative reactions toward immi-
both ps > .40, two-tailed). Further analyses (z gration of an eastern African outgroup. These
tests comparing the magnitudes of these rs) results are consistent with the hypothesis that
substantiate the observation that the negative PVD predicts negative reactions toward subjec-
correlation observed in the Africa condition tively foreign, but not subjectively familiar
was distinct from the weak relations in the other ethnic groups.
two conditions, ps = .026 and .020, respectively.
Study 4
PVD was also negatively related to pro- The results of the first three studies reveal a
immigration attitudes in general (r = –.22, p = pattern of findings linking PVD with negative
.029). This relation was not meaningfully mod- reactions toward African outgroups. These
erated by geographical origin. results are consistent with a psychological
process in which ethnic foreign-ness serves as a
Overall, the results reveal some tendency for
individuals who perceive themselves to be

0.20
0.10
0.00
–0.10
–0.20
–0.30
–0.40

Eastern Africa Eastern Asia Eastern Europe

Origin of Immigrant Group

Figure 1. Correlations between PVD and pro-immigration attitudes as a function of the geographical origin of
potential immigrants (Study 3).

341

Group Processes & Intergroup Relations 7(4)

cue that heuristically signals a threat of inter- Sri Lanka] is unacceptably low and that its health
personal disease transmission. These results are and social systems do not meet the needs of
also consistent with a different process in which [Peruvian/Qatar/Sri Lankan] people. The majority
participants’ reactions might be influenced by of these refugees are applying to live in the greater
some rational assessment of the health risks Vancouver area.
associated specifically with Africa. Certain high-
profile infectious diseases, such as Ebola and To assess attitudes toward immigrants from
HIV, have been more notably epidemic in parts Peru, Qatar, or Sri Lanka, participants in each
of Africa than in other parts of the world. Con- condition rated the extent that they agreed with
sequently, it could be argued, PVD might three statements that assessed attitudes and
predict negative attitudes uniquely toward beliefs pertaining specifically to the refugee
African peoples. This possibility is not uninter- group described. As in Study 3, one statement
esting; but the implications—both conceptual assessed whether participants felt the Canadian
and practical—are more profound if PVD government should allow the refugees to enter
predicts negative attitudes toward a broader Canada. Two other statements assessed whether
range of subjectively foreign peoples. For this participants felt the refugees might bring
reason, in a fourth study, we assessed the health problems or criminal problems into
relation between PVD and attitudes toward Canada.
three non-African outgroups: immigrants from
Peru, Qatar, and Sri Lanka. A separate sample of Three additional statements assessed attitudes
23 UBC students (the same separate sample toward immigration in general. Responses to
who rated the foreign-ness of the immigrant these last three items were combined to form a
groups portrayed in Study 3) rated immigrants single index assessing pro-immigration atti-
from Peru, Qatar, and Sri Lanka to be subjec- tudes in general. Ratings on all statements were
tively foreign overall and in cultural practices made on 7-point scales.
pertaining to hygiene and food. Therefore,
PVD was expected to predict xenophobic reac- Participants also rated the target immigrant
tions to all three of these groups. group on a set of 13 specific characteristics. Five
of these characteristics pertained to hygiene
Participants were 47 undergraduate students (sanitary, filthy, hygienic, clean, and dirty); the
at the University of British Columbia, primarily rest tapped into a variety of other evaluative
of East Asian and European heritage (none traits (likeable, hostile, trustworthy, open-
were from Peru, Qatar, or Sri Lanka). Partici- minded, ignorant, poor, lazy, and unintelli-
pants were first presented with a written gent). Ratings were made on 9-point scales.
description of a refugee group seeking to immi- Following these procedures, participants
grate to Vancouver, Canada. The passage was completed various questionnaires, including
identical for all participants except for the the 14-item PVD measure.
apparent geographical origin of the refugees.
Participants were randomly assigned to one of A negative relation between PVD and
three conditions in which they read that the endorsement of immigration of all three
refugees were from Peru, Qatar, or Sri Lanka. refugee groups was expected, and so analyses
After first identifying the geographical origin of focused on the full set of data, collapsed across
the refugees, and describing their harsh living conditions. Results revealed a substantial
conditions resulting from arid climate with low negative correlation between PVD and attitudes
rainfall, the passage proceeded as follows: toward immigration of the specified group (r =
–.40, p = .003). (A negative correlation was
As a result of these living conditions, a large observed regardless of whether the refugee
number of refugees from the plains of [Peru/ group was described as originating in Peru,
Qatar/Sri Lanka] are seeking refugee status in Qatar, or Sri Lanka: The r s in each of these con-
Canada. They have told Canadian immigration ditions were –.30, –.36, and –.53, respectively.)
officials that the standard of living in [Peru/Qatar/ The higher individuals’ PVD scores were, the
less willing they were to allow the group to
immigrate.

342

Faulkner et al. evolved disease-avoidance mechanisms

PVD also predicted beliefs linking the refugee vidual differences in perceived vulnerability to
groups to health problems (r = .34, p = .012). disease. Although the results on some ancillary
There was a somewhat weaker correlation measures (e.g. attribution of personal traits and
between PVD and beliefs linking the refugee characteristics, anti-immigrant attitudes in
groups to crime (r = .25, p = .087, two-tailed). general) were somewhat inconsistent across
There was also a negative correlation between studies, there was a consistent pattern of results
PVD and the 3-item measure of pro-immigration on the primary measures assessing xenophobic
attitudes in general (r = –.38, p = .004). attitudes within each study. Canadian students
who perceived themselves chronically to be
In contrast to these relations to the attitude vulnerable to diseases were more likely to
measures, PVD did not strongly predict associate foreign peoples (Africans) with
responses on the 13-trait rating items. Only danger, and were less likely to endorse the
three correlations exceeded .2 in magnitude. immigration of foreigners (immigrants from
Two of these correlations were with traits per- Africa, Peru, Qatar, and Sri Lanka). Perceived
taining to hygiene: The correlation with ‘clean’ vulnerability to disease predicted xenophobic
was r = –.25 (p = .046) and the correlation with attitudes towards specific immigrant groups
‘dirty’ was r = .22 (p = .073). The other notable that were subjectively foreign, but not toward
correlation was that between PVD and ‘open- other groups (immigrants from eastern Asia
minded’ (r = –.29, p = .050, two-tailed). and eastern Europe) that were subjectively
familiar.
The results of this study reveal a strong
relation between PVD and anti-immigrant atti- These findings are consistent with an evolved
tudes toward three non-African foreign ethnic disease-avoidance process in which subjective
groups. Individuals with higher PVD scores foreign-ness heuristically triggers cognitions
were less willing to allow foreign outgroups to that motivate avoidance and exclusion,
immigrate to Canada. They also held stronger especially among individuals who personally
beliefs that outgroups might bring disease feel vulnerable to disease. However, these
problems to Canada and are less hygienic. To findings are correlational in nature, and so it is
some extent, PVD also predicted beliefs about necessary to consider the possibility that the
criminality—a result suggesting that the under- relation between PVD and xenophobic atti-
lying process may be associated with a variety of tudes may be the product of shared variance
attitudes that are not logically disease-relevant, with other individual difference variables that
but which nonetheless motivate exclusionary have nothing conceptually to do with disease.
action. PVD also clearly predicted anti-immi- Several additional results address some of these
grant attitudes in general. ‘third variable’ alternative explanations. For
instance, it might be argued that PVD is merely
It’s also worth noting that although PVD one indicator of a broader personality con-
clearly predicted xenophobic attitudes, it did struct, such as neuroticism, that more directly
not strongly predict the attribution of specific predicts xenophobic attitudes. This alternative
traits to the immigrant groups. The reason for seems unlikely. Not only is PVD only weakly cor-
this may lie partially in the nature of the affec- related with neuroticism (see results summar-
tive state—disgust—that is linked conceptually ized above), but neuroticism is not consistently
to the disease-avoidance process. Disgust typi- correlated with prejudicial attitudes. Some
cally elicits crude evaluative attitudes (distaste, studies have shown modest positive relations
dislike) and behavioral attempts to reduce between neuroticism and ethnic prejudice
contact, but it may not immediately connote (Hebron & Ridley, 1965), other studies show
specific trait information of the sort connoted little or no predictive effects of either neuroti-
by other emotions, such as fear. cism or anxiety (Maykovich, 1975; Orpen, 1973;
Prentice, 1962) and at least one study reveals a
Discussion of Studies 1–4 negative correlation between neuroticism and
These studies provide the first evidence that
xenophobic attitudes are moderated by indi-

343

Group Processes & Intergroup Relations 7(4)

prejudice against ethnic immigrants (Ray & participants randomly assigned to one experi-
Kiefl, 1984). Potentially more plausible sources mental condition viewed a series of pictures
of spurious relations are other individual differ- that conveyed the ease with which bacteria and
ence constructs pertaining to other specific germs are transmitted in everyday life. The
threats and fears. For instance, PVD does cor- remaining participants (assigned to a control
relate substantially with ‘belief in a dangerous condition) viewed a series of pictures that
world’ (BDW), and prior research reveals that conveyed the ease with which accidents occur
BDW does predict prejudicial attitudes against in everyday life. All participants then com-
ethnic outgroups (Altemeyer, 1988). We were pleted measures assessing their attitudes toward
able to directly address this specific alternative one specific immigrant group, which—depend-
because participants in Studies 1–4 also com- ing on another experimental manipulation—
pleted Altemeyer’s (1988) measure of BDW. was either subjectively foreign or familiar.
For each study, we conducted regression Participants in Study 6 also completed an
analyses that included both PVD and BDW as additional measure in which they allocated
predictors of the primary measure of xenopho- government resources to attract immigrants
bic attitudes; results revealed that the predictive from a variety of both foreign and familiar loca-
effects of PVD were not substantially reduced. tions. In both studies, the conceptual hypothe-
We also created a composite dataset (N = 213), sis would be supported by evidence of
merging data from all four studies after first interactions between the disease-salience
standardizing responses on the primary depen- manipulation and the foreign-ness of target
dent measure (we reverse scored the values groups: compared to participants in the control
from Studies 3 and 4 so that, across all four condition, participants in the disease-salient
studies, higher standardized values reflected condition were expected to respond more
more negative attitudes). Within this composite negatively toward foreign—but not familiar—
dataset, PVD and BDW were entered simul- immigrant groups.
taneously as predictors. Results revealed a
modest effect of BDW on xenophobic attitudes Pretesting
(beta = .13, p = .065, two-tailed) and—more Prior to conducting these experiments, a
importantly—a more substantial and unique separate sample of 26 UBC students pre-rated a
effect of PVD (beta = .23, p < .001). set of eight immigrant groups representing
areas of the world with very different climatic
Although these analyses help rule out several and geographic characteristics. These groups
specific personality constructs as viable alterna- were immigrants from Taiwan, Peru, Poland,
tive explanations for the effects of PVD, they Nigeria, Scotland, Mongolia, Brazil, and
cannot address all possible third-variable expla- Iceland. Participants rated each immigrant
nations. Consequently, we conducted two group on scales assessing how different the
additional studies in which we attempted to group was from participants themselves (e.g.
experimentally manipulate the extent that how culturally and visually different they are).
people feel vulnerable to disease. This Participants also rated each group on several
approach allowed us to more rigorously test the scales assessing how different they were in
proposition that disease-avoidance processes specific disease-relevant domains (e.g. how
contribute to xenophobic attitudes. different their food preparation and hygienic
practices are). Ratings were made on 9-point
Studies 5 and 6: temporary salience of disease scales. On the basis of these responses, we
amplifies xenophobic attitudes computed an average foreign-ness score for
In two experiments, we manipulated partici- each group. Immigrants from Nigeria were
pants’ temporary perceptions of vulnerability rated as most foreign overall (M = 5.89),
to disease, and then assessed attitudes toward followed by immigrants from Mongolia (M =
the immigration of subjectively foreign or 5.48), Brazil and Peru (Ms = 5.38), Iceland (M
familiar immigrant groups. In each study,

344

Faulkner et al. evolved disease-avoidance mechanisms

= 4.62), Poland (M = 4.50), Taiwan (M = 4.33), Participants viewed their assigned slide show
and Scotland (M = 4.00). twice; the first time they rated the educational
informativeness of each slide on a 10-point
These pretest results informed the methods rating scale, and after a second viewing they
used in the two experiments. In one experi- wrote a short paragraph describing their overall
ment, participants indicated attitudes toward impressions of the slide show.
potential immigrants from either Nigeria
(foreign) or Scotland (familiar). In a second Participants were then told that the initial
experiment, participants indicated attitudes study was over and were asked to complete a
toward potential immigrants from either questionnaire for an ostensibly unrelated study.
Mongolia (foreign) or Taiwan (familiar), and The experimenter read aloud a brief descrip-
also completed an additional measure assessing tion of a group seeking to immigrate to Van-
relative preferences among all eight pre-rated couver. Participants were randomly assigned to
immigrant groups. hear that this group was from either Nigeria or
Scotland. They then completed a questionnaire
Study 5 consisting of rating scale items assessing their
Participants were 57 undergraduate students at attitudes toward the described immigrant
UBC. Participants first completed several ques- group. These items were similar to those in
tionnaires, including the PVD measure. They Study 4: three items assessed specific attitudes
were then randomly assigned to one of two toward the potential immigrant group, three
experimental conditions. In one condition, items assessed attitudes toward immigration in
participants viewed a slide show designed to general, and 13 items asked participants the
remind them of the various ways that diseases extent that various traits are characteristic of
are transmitted in daily life, thus making the the immigrant group.
threat of contagious diseases especially salient.
This ‘Disease’ slide show included 11 pictures In general, across all attitude and trait items,
that would ostensibly be used in a health edu- there was a general main effect for target
cation program (e.g. one slide pictures a group, in that participants tended to rate
woman unsuccessfully attempting to kill Nigerian immigrants less favorably than
cartoon germs in her kitchen, and is titled ‘The Scottish immigrants. On specific items, partici-
Horrors of the Kitchen Sponge and Family pants believed Nigerians were more likely to
Pets’; another slide depicts a strand of hair sur- bring costly health problems to Vancouver, and
rounded by bacteria and is labeled ‘Hair rated Nigerians to be less sanitary, less clean,
Bacteria. A microscopic view of a strand of hair less open-minded, and as poorer and dirtier
and some of the typical bacteria that surround than Scottish immigrants (all ps < .05).
it’). The rest of the participants were assigned
to a control condition, in which they viewed a For the purpose of testing the hypothesis
slide show designed to make the threat of about disease-salience and xenophobia, the
physical dangers especially salient. These slides attitude item of central interest was the item
were intended to make participants feel sus- assessing agreement with the statement: ‘The
ceptible to a threat that was irrelevant to Canadian government should allow [Nigerian/
disease. This ‘Accidents’ slide show included Scottish] immigrants to live in Vancouver.’ PVD
11 pictures that would ostensibly be used in scores, particularly for the average of PVD items
a safety education program. (Two examples are assessing discomfort with behaviors that
‘School Bus Hazards’ which depicts a girl connote a risk of germ transmission, predicted
bending over to pick up a book as a school responses on this item; therefore the average
bus is about to hit her, and ‘Electricity and score on this subset of PVD items was included
Water Don’t Mix’ which portrays a bathing as a covariate in a 2 ϫ 2 (slide show condition
woman surrounded by plugged-in electrical ϫ target group condition) analysis of covari-
appliances.) ance. This analysis revealed the predicted inter-
action, (F(1,52) = 6.00, p = .02). As depicted in
Figure 2, participants in the control condition

345

Group Processes & Intergroup Relations 7(4)

were approximately equally likely to support disease-avoidance processes. It is also possible
immigration of Scots (M = 5.85) and Nigerians that it reflects the operation of participants’
(M = 6.07), but those in the disease-salience self-presentational concerns: these students
condition were more likely to endorse the may be somewhat willing to explicitly express
immigration of Scots (M = 6.36) and less likely attitudes about governmental decisions bearing
to endorse the immigration of Nigerians (M = on desired and undesired immigrant groups,
5.53). but may be less willing to overtly express stereo-
typical judgments about those groups.
In contrast to the clear effects of the disease-
salience manipulation on this dependent The results depicted in Figure 2 provide
measure, there were weaker and inconsistent initial support for the hypothesis that tem-
effects on other items. This failure to find porarily-heightened worries about vulnerability
effects across a broad range of items is con- to disease can amplify xenophobic attitudes
sistent with the general nature of the PVD toward subjectively foreign (but not familiar)
effects obtained in Studies 1–4. The effects outgroups. Still, these results are limited. The
seem to show up primarily on attitude items purpose of the final study was to conceptually
that bear most directly on physical contact replicate and extend this finding by examining
and/or social exclusion. This may reflect the attitudes toward a broader range of potential
underlying behavioral function of evolved immigrant groups.

6.50 Nigerian Immigrants
6.25 Scottish Immigrants
6.00
Endorsement 5.75
5.50
5.25
5.00

Accidents Diseases

Salient Threat

Figure 2. Interaction between disease salience and geographical origin of immigrants on endorsement of
potential immigration (Study 5).

346

Faulkner et al. evolved disease-avoidance mechanisms

Study 6 immigration advertising budget that should be
The methods of Study 6 were similar to those of spent in each of eight different countries:
Study 5, with two main changes. One change Taiwan, Peru, Poland, Nigeria, Scotland,
pertained to the specific target immigrant Mongolia, Brazil, and Iceland. These were the
groups that were described to participants: same eight countries that had been pre-tested
Mongolians and Taiwanese. In pre-testing for perceived foreign-ness. Therefore, budget
(described above), Mongolian and Taiwanese percentages that each participant allocated to
immigrants were rated, respectively, as the the four countries pre-rated to be most foreign
second-most and second-least foreign of the (Nigeria, Mongolia, Brazil, and Peru) were
eight pre-rated groups. The second change was summed to compute a single index; and per-
the inclusion of an additional measure, in centages allocated to the four countries pre-
which participants indicated the percentage of rated to be most familiar (Scotland, Taiwan,
a special government budget that should be Poland, and Iceland) were summed to compute
allocated to help attract immigrants from each a second index. This allowed for a within-
of the eight different pre-rated nations. This subjects comparison of budget allocations to
measure extends the analysis to include a attract foreign versus familiar immigrants.
broader range of subjectively foreign and Participants in the disease-salient condition,
familiar groups. In addition, it might also be compared to control participants, were
expected to be less reactive to participants’ self- expected to allocate relatively less of the budget
presentation concerns. to subjectively foreign areas.

Participants were 45 undergraduate students The results on the attitude and trait-rating
at UBC. Participants first completed several items indicated that, in contrast to Study 5,
questionnaires, including the PVD measure, participants did not generally rate the subjec-
and then were randomly assigned to view either tively foreign immigrant group (Mongolians)
the ‘Disease’ slide show or the ‘Accidents’ slide any less favorably than the familiar immigrant
show (this manipulation was identical to that in group (Taiwanese). A main effect for target
Study 5). Following this, participants heard a group emerged on only one specific item:
description about a specific immigrant group participants rated Mongolian immigrants to be
seeking to settle in Vancouver; they were poorer than Taiwanese (t (43) = 2.69, p = .01).
randomly assigned to hear that the immigrant Also in contrast to the results of Study 5, there
group was either from Mongolia (foreign) or emerged no interpretable effects of the disease-
Taiwan (familiar). Then, as in Study 5, they salience manipulation on these measures—
completed a set of rating-scale items assessing including no evidence of the predicted
attitudes and beliefs about the specific target interaction on the item assessing endorsement
immigrant group. of the target group’s immigration to Canada.
The predicted interaction did emerge on the
Participants then completed an additional budget allocation measure. A 2 ϫ 2 ϫ 2 (slide
dependent measure, labeled the ‘immigration show condition ϫ target group condition ϫ
advertising questionnaire’. At the beginning of budget allocation index) repeated measures
this questionnaire, a short paragraph explained analysis of variance revealed the anticipated
that the Canadian government was considering two-way interaction between slide show con-
‘advertising’ Canada as a suitable destination dition and budget allocation index (F(1, 41) =
for potential immigrants around the world. 4.58, p = .038). Whereas participants in the
This paragraph also explained that Canada had control condition allocated roughly equal
a finite amount of resources for this kind of amounts of advertising resources to foreign and
advertising, and so it sought input on the way familiar countries (47.48% vs. 52.52%, respec-
in which this advertising budget should be tively), participants in the disease-salience
allocated to different countries from which condition allocated less advertising resources to
potential immigrants might come. Participants foreign countries and more to familiar countries
were asked to indicate the percent of the total

347

Group Processes & Intergroup Relations 7(4)

(37.57% vs. 62.43%, respectively). Figure 3 them to appear especially welcoming to other
depicts this interaction. A deeper examination groups of immigrants—more familiar immi-
of the budget allocations revealed that alloca- grants, of course.
tions to three of the four foreign countries
(Nigeria, Mongolia, and Peru) were lower in An effect of PVD also emerged on the budget
the disease-salient condition. Thus, the effects allocation measure: participants with higher
of temporary disease salience extend beyond PVD scores allocated less of the advertising
reactions merely to Nigerians (documented in budget to foreign areas (r = –.24, p = .05).
Study 5) to other foreign peoples as well. These
results reveal also that, although participants in Discussion of Studies 5 and 6
the disease-salient condition did not express Across both experiments, there emerged some
greater xenophobic reactions to Mongolian evidence that temporarily-activated worries
immigrants when they simply rated Mongolians about disease led to more negative reactions to
alone (in the attitude rating scales), they did foreign, but not familiar, immigrant groups.
appear to do so when these reactions were Study 5 revealed evidence of this effect on an
embedded in a response context that allowed attitude item pertaining to support for the
immigration of specific foreign and familiar

Percent Allocation of Funds 65 Foreign Immigrants
60 Familiar Immigrants

55

50

45

40
35

30 Diseases
Accidents

Salient Threat

Figure 3. Interaction between disease salience and foreign-ness of geographical areas on allocation of
immigration advertising budget (Study 6).

348

Faulkner et al. evolved disease-avoidance mechanisms

target groups. Study 6 failed to replicate this associated with even more favorable reactions
effect exactly on the same kind of attitude item to familiar immigrants; see especially the results
rewritten to specify a different pair of target from Studies 3 and 5.) Thus, the process under-
groups, but did conceptually replicate the lying these effects appears to be conceptually
effect on a different measure (a budget allo- distinct from other psychological processes that
cation measure) that assessed interest in attract- precipitate prejudicial attitudes toward a less-
ing a greater variety of foreign and familiar specified range of outgroups and social
immigrant groups. It is possible that the pre- categories. This evidence of target-specificity
dicted effect emerged more clearly on this and domain-specificity is consistent with an evo-
latter measure because, in contrast to the more lutionarily-informed meta-theoretical approach
straightforward attitude measures, participants to prejudice processes in general. There are
may have been less likely to worry that their many different kinds of threat that may precip-
budget allocation choices were indicative of itate negative reactions to outgroups. These
anti-immigrant prejudice. In any case, although include symbolic threats at the level of self-
there was some inconsistency in the results concept and social identity, as well as perceived
across the two experiments, both experiments threats to political status, economic security,
yielded evidence consistent with the same con- and physical well-being (Schaller, 2003;
ceptual hypothesis. Sidanius & Pratto, 1999; Stephan, Diaz-Loving,
& Duran, 2000). Different outgroups—defined
It is worth noting also that the effects of by different features—precipitate different
PVD documented in Studies 1–4 were also repli- kinds of threat, and these different kinds of
cated on the budget allocation measure used in threat are associated with qualitatively different
Study 6. Thus, this experiment not only sub- kinds of negative responses (Neuberg &
stantiates the predicted effect of temporary Cottrell, 2002).
disease worries on xenophobia, it also further
documents the effects of chronic disease The results from several of these studies
worries on xenophobia as well. revealed some, albeit tentative, indication of
domain-specificity in beliefs about subjectively
General discussion foreign outgroups. For instance, in Study 1,
PVD predicted implicit cognitive associations
The six studies reported here provide the first linking Africans to danger, but not to unpleas-
empirical evidence that chronic and temporar- ant semantic concepts in general. In several
ily-aroused feelings of vulnerability to disease other studies, there was some evidence that
contribute to negative attitudes toward foreign PVD predicted beliefs linking foreign groups
peoples. These findings are consistent with an with specific kinds of traits that either connote
evolved disease-avoidance model of contem- uncleanliness or that typically motivate physical
porary stigma processes (Kurzban & Leary, avoidance. However, these results were weak
2001; Park et al., 2003), and they extend this and inconsistent across studies, and bear
model to a new domain of application. further investigation. Many other evolved pro-
cesses—such as those pertaining to fear—are
Subjective perceptions of familiarity and associated with fairly domain-specific judg-
foreign-ness appear to be fundamental to the ments and reactions. For instance, variables
activation of this underlying process. Individual connoting vulnerability to intentional physical
differences in perceived vulnerability to disease harm predict specific, functionally-relevant trait
predicted negative reactions only to subjectively judgments about ethnic outgroups (Schaller
foreign peoples, but not to subjectively familiar et al., 2003). It is possible that an evolved
peoples. Similarly, contextual information that disease-avoidance process (presumably linked
makes disease especially salient led to negative affectively to disgust rather than fear) precipi-
reactions to subjectively foreign immigrants, tates a less well-specified set of semantic
but not to subjectively familiar immigrants. (If associations. Given the interest that issues of
anything, greater worries about disease were

349

Group Processes & Intergroup Relations 7(4)

domain-specificity inspire among evolutionary vulnerable to disease. It might be argued, for
psychologists (Kenrick, Sadalla, & Keefe, 1998), instance, that the effects documented here may
it will be useful for future research to investi- result from a rational assessment of the disease-
gate more thoroughly the domain-specificity of relevant risks associated with particular immi-
disease-based xenophobic reactions. grant groups (e.g. those from developing
countries), rather than from the heuristic acti-
Empirical support for a disease-avoidance vation of evolved mechanisms. A rational risk-
mechanism does not preclude the operation of assessment process, however, implies that
additional adaptive mechanisms promoting exclusionary attitudes should be based on attri-
xenophobic reactions to foreign peoples. The butions of disease-relevant traits to foreign
violation of local cultural norms not only peoples. While our findings occasionally indi-
signals the threat of contagion, but may also cated that PVD predicted such trait attribu-
signal a threat to social efficiency and effective tions, these effects were weak when obtained
group functioning (Gil-White, 2001). Just as (see Study 4) and were inconsistent across
specific psychological reactions (such as studies; it seems unlikely that trait attributions
disgust) may be associated with the perceived mediate the relation between PVD and exclu-
threat of disease, other reactions (such as sionary attitudes. People may occasionally
contempt and moral disapproval) may be engage in a rational analysis when there are
associated with the perception of norm salient reasons (e.g. the recent outbreak of
violation (Boyd, Gintis, Bowles, & Richerson, SARS) to associate specific groups with specific
2003; Neuberg & Cottrell, 2002). A multi- disease threats. However, the present findings
faceted evolutionary approach can help predict suggest that there are additional non-rational
the circumstances under which subjectively processes that contribute to similar xenophobic
foreign ethnic outgroups might most strongly attitudes.
elicit disease-avoidant reactions rather than
reactions designed to maintain effective group Additional evidence for a non-rational
functioning. If an outgroup violates norms that process also emerges from statistical analyses
are particularly germane to disease (e.g. that attend to the two kinds of items from which
customs pertaining to personal hygiene, food the PVD measure is comprised. Recall that
preparation, or sex), then disease-avoidance some items measure beliefs about personal
mechanisms are especially likely to be engaged. vulnerability to disease, while other items
However, if an outgroup violates other kinds of measure discomfort with situations in which
norms that are particularly pertinent to social germs are likely to be transmitted. In a sense,
efficiency (e.g. language use, reciprocity), then the first set of items assesses thoughtful con-
xenophobia might be predicted primarily by siderations about disease susceptibility, while
concerns with effective group functioning. In the second set assesses more emotion-based
the present studies, we did not assess the extent avoidant reactions. Across Studies 1–4, we
to which target groups were perceived as found that the emotion-based ‘germ-aversion’
foreign in these other kinds of normative subscale typically exerted a stronger effect on
domains; further research is needed to test pre- xenophobic attitudes (mean r = .28 for the
dictions about the relations between specific germ-aversion subscale, compared to a mean
domains of foreign-ness, the threats they imply, r = .16 for the subscale measuring more
and the psychological reactions they elicit. thoughtful beliefs about disease susceptibility).
These results are consistent with other findings
While the empirical results presented here on PVD and prejudice (Park et al., 2003), and
are consistent with the underlying evolutionary provide some indirect evidence that the xeno-
model from which these hypotheses were phobia results are due less to a rational assess-
derived, these results alone cannot rule out ment of disease risk than they are to an
alternative accounts for the contemporary automatically-activated aversion to heuristically
tendency of xenophobic attitudes to be pro- threatening social situations.
nounced in individuals who feel particularly

350

Faulkner et al. evolved disease-avoidance mechanisms

Indirect evidence for an automated non- such as disgust (Heath, Bell, & Sternberg,
rational mechanism does not, of course, 2001), and more generally are those that help
provide verification for the hypothesized role of individuals solve fundamental adaptive problems
evolutionary pressures in sculpting such a (Krebs & Janicki, 2004).
mechanism. No such verification can ever be
produced by merely psychological research Concluding remarks
methods (Conway & Schaller, 2002). The value
of this evolutionary level of analysis lies not in The research presented here provides a specific
its explanatory necessity, but in its meta- example of the value of evolutionary logic in
theoretical utility in generating new discoveries psychological inquiry: reasoning about the
and connecting different scholarly literatures— recurrent adaptive problems imposed by ances-
including qualitative analyses of anti-immigrant tral human environments, and the cognitive
attitudes throughout human history (e.g. mechanisms that arose to solve them, can yield
Markel & Stern, 2002), zoological studies of novel hypotheses about psychological processes
xenophobia in other primate species (e.g. in the here-and-now (Conway & Schaller,
Holloway, 1974; Southwick, Siddiqi, Farooqui, 2002). We used this approach to generate the
& Pal, 1974), and the extensive psychological previously-untested hypothesis that individuals’
literature on intergroup prejudice. feelings of vulnerability to disease have impli-
cations for the seemingly unrelated domain of
The utility of an evolutionary approach to intergroup cognition. Our findings provide
contemporary psychological processes is maxi- support for this hypothesis and indicate that
mized when it is considered in conjunction with chronic and contextually aroused feelings of
more ‘traditional’ meta-theoretical approaches vulnerability to disease should be added to the
in the social sciences. Consider the processes of list of psychological factors that contribute to
socialization and cultural change. One might xenophobic attitudes.
argue that a process of cultural—not bio-
logical—evolution over the past several millen- References
nia has contributed to beliefs about foreigners
that are functionally related to disease-avoid- Altemeyer, B. (1988). Enemies of freedom.
ance concerns. As a result of contact with San Francisco: Jossey-Bass.
foreign populations and the inevitable co-
occurence of disease outbreaks, cultural beliefs Bishop, G. D., Alva, A. L., Cantu, L., & Rittiman,
may have arisen that associate foreigners with T. K. (1991). Responses to persons with AIDS:
disease and promote behavioral avoidance or Fear of contagion or stigma? Journal of Applied
social exclusion. These beliefs may be passed Social Psychology, 21, 1877–1888.
on inter-generationally, and so may become a
permanent part of cultural knowledge. Of Boyd, R., Gintis, H., Bowles, S., & Richerson, P. J.
course, processes of cultural evolution are (2003). The evolution of altruistic punishment.
entirely compatible—and often necessarily Proceedings of the National Academy of Sciences (USA),
inter-related—with processes of biological evol- 100, 3531–3535.
ution (Boyd & Richerson, 1985; Kameda,
Takezawa, & Hastie, 2003). For instance, the Boyd, R., & Richerson, P. J. (1985). Culture and the
ancient Greek belief that foreigners possess an evolutionary process. Chicago: University of Chicago
‘evil eye’, which by itself was thought to cause Press.
illness (Nutton, 2000), may have emerged and
persisted because it motivates avoidance or Conway, L. G., III, & Schaller, M. (2002). On the
exclusionary behaviors in the presence of for- verifiability of evolutionary psychological theories:
eigners. Beliefs that are especially likely to An analysis of the psychology of scientific
emerge and persist within cultures are those persuasion. Personality and Social Psychology Review,
that exploit fundamental emotional systems 6, 152–166.

Crandall, C. S., Glor, J., & Britt, T. W. (1997).
Aids-related stigmatization: Instrumental and
symbolic attitudes. Journal of Applied Social
Psychology, 27, 95–123.

Diamond, J. (1999). Guns, germs, and steel. New York:
Norton.

351

Group Processes & Intergroup Relations 7(4)

Fiske, S. T. (1998). Stereotyping, prejudice, and & D. L. Krebs (Eds.), Handbook of evolutionary
discrimination. In D. T. Gilbert, S. T. Fiske, & psychology (pp. 485–514). Mahwah, NJ: Erlbaum.
G. Lindzey (Eds.), The handbook of social psychology Krebs, D., & Janicki, M. (2004). Biological
(4th edn, pp. 357–411). New York: McGraw-Hill. foundations of moral norms. In M. Schaller &
C. S. Crandall (Eds.), The psychological foundations
Gangestad, S. W., & Simpson, J. A. (2000). The of culture (pp. 125–148). Mahwah, NJ: Erlbaum.
evolution of human mating: Trade-offs and Kurzban, R., & Leary, M. R. (2001). Evolutionary
strategic pluralism. Behavioral and Brain Sciences, origins of stigmatization: The functions of social
23, 573–644. exclusion. Psychological Bulletin, 127, 187–208.
Markel, H. (1999). When germs travel. The American
Gil-White, F. J. (2001). Are ethnic groups biological Scholar, 68, 61–69.
‘species’ to the human brain? Essentialism in our Markel, H., & Stern, A. M. (2002). The foreignness
cognition of some social categories. Current of germs: The persistent association of immigrants
Anthropology, 42, 515–554. and disease in American society. The Milbank
Quarterly, 80, 757–788.
Goldhagen, D. J. (1996). Hitler’s willing executioners. Maykovich, M. K. (1975). Correlates of racial
New York: Random House. prejudice. Journal of Personality and Social Psychology,
32, 1014–1020.
Greenwald, A. G., McGhee, D. E., & Schwartz, J. L. Neuberg, S. L., & Cottrell, C. A. (2002). Intergroup
K. (1998). Measuring individual differences in emotions: A biocultural approach. In
implicit cognition: The Implicit Association Test. D. M. Mackie & E. R. Smith (Eds.), From prejudice
Journal of Personality and Social Psychology, 74, to intergroup relations: Differentiated reactions to social
1464–1480. groups (pp. 265–283). New York: Psychology Press.
Neuberg, S. L., Smith, D. M., & Asher, T. (2000).
Haidt, J. (2001). The emotional dog and its rational Why people stigmatize: Toward a biocultural
tail: A social intuitionist approach to moral framework. In T. Heatherton, R. Kleck, J. G. Hull,
judgment. Psychological Review, 108, 814–834. & M. Hebl (Eds.), The social psychology of stigma
(pp. 31–61). New York: Guilford.
Haidt, J., McCauley, C., & Rozin, P. (1994). Noy, D. (2000). Foreigners at Rome: Citizens and
Individual differences in sensitivity to disgust: strangers. London: Duckworth.
A scale sampling seven domains of disgust elicitors. Nutton, V. (2000). Did the Greeks have a word for
Personality and Individual Differences, 16, 701–713. it? In L. I. Conrad & D. Wujastyk (Eds.), Contagion:
Perspectives from pre-modern societies (pp. 137–163).
Heath, C., Bell, C., & Sternberg, E. (2001). Burlington, VT: Ashgate.
Emotional selection in memes: the case of urban Oldstone, M. B. A. (1998). Viruses, plagues, and
legends. Journal of Personality and Social Psychology, history. New York: Oxford University Press.
81, 1028–1041. Orpen, C. (1973). The cross-cultural validity of the
Eysenck Personality Inventory: A test of
Hebron, M. E., & Ridley, F. (1965). Characteristics Afrikaans-speaking South Africa. British Journal of
associated with racial prejudice in adolescent boys. Social and Clinical Psychology, 11, 244–247.
British Journal of Social and Clinical Psychology, 4, Park, J. H., Faulkner, J., & Schaller, M. (2003).
92–97. Evolved disease-avoidance processes and
contemporary anti-social behavior: Prejudicial
Herek, G. M., & Capitaniato, J. P. (1998). Symbolic attitudes and avoidance of people with physical
prejudice or fear of infection? A functional disabilities. Journal of Nonverbal Behavior, 27, 65–87.
analysis of AIDS-related stigma among Prentice, N. M. (1962). Ethnic attitudes,
heterosexual adults. Basic and Applied Social neuroticism, and culture. Journal of Social
Psychology, 20, 230–241. Psychology, 54, 75–82.
Ray, J. J., & Kiefl, W. (1984). Authoritarianism and
Holloway, R. (Ed.). (1974). Primate aggression, achievement motivation in contemporary West
territoriality, and xenophobia. New York: Academic Germany. Journal of Social Psychology, 122, 3–19.
Press. Reynolds, V., Falger, V., & Vine, I. (Eds.). (1987).
The sociobiology of ethnocentrism. London: Croom
John, O. P., & Srivastava, S. (1999). The big five trait Helm.
taxonomy: History, measurement and theoretical
perspectives. In L. A. Pervin & O. P. John (Eds.),
Handbook of personality: Theory and research (2nd
edn, pp. 102–138). New York: Guilford Press.

Kameda, T., Takezawa, M., & Hastie, R. (2003). The
logic of social sharing: An evolutionary game
analysis of adaptive norm development. Personality
and Social Psychology Review, 7, 2–19.

Kenrick, D. T., Sadalla, E. K., & Keefe, R. C. (1998).
Evolutionary cognitive psychology: The missing
heart of modern cognitive science. In C. Crawford

352

Faulkner et al. evolved disease-avoidance mechanisms

Rozin, P., Haidt, J., & McCauley, C.R. (2000). theory. Journal of Applied Social Psychology, 29,
Disgust. In M. Lewis & J. M. Haviland-Jones (Eds.), 2221–2237.
Handbook of emotions (2nd edn, pp. 637–653). New Suedfeld, P., & Schaller, M. (2002).
York: Guilford Press. Authoritarianism and the Holocaust: Some
cognitive and affective implications. In
Rozin, P., Haidt, J., McCauley, C., & Imada, S. L. S. Newman & R. Erber (Eds.), What social
(1997). Disgust: Preadaptation and the cultural psychology can tell us about the Holocaust:
evolution of a food-based emotion. In H. MacBeth Understanding perpetrator behavior (pp. 68–90).
(Ed.), Food preferences and taste (pp. 65–82). Oxford, UK: Oxford University Press.
Providence, RI: Berghahn.
Acknowledgements
Rozin, P., Markwith, M., & Nemeroff, C. (1992).
Magical contagion beliefs and fear of AIDS. This research was supported by a Natural Sciences
Journal of Applied Social Psychology, 22, 1081–1092. and Engineering Research Council of Canada
fellowship awarded to Jason Faulkner and a
Rudman, L. A., Greenwald, A. G., Mellott, D. S., & Social Sciences and Humanities Research Council
Schwartz, J. L. K. (1999). Measuring the automatic of Canada grant awarded to Mark Schaller.
components of prejudice: Flexibility and We are grateful to Chris Crandall, Steve Heine,
generality of the implicit association test. Social Doug Kenrick, and Ara Norenzayan for their
Cognition, 17, 437–465. helpful comments on drafts of this article. We
also thank Gurbreena Dhanoa, Chloe Kelly,
Schaller, M. (2003). Ancestral environments and David Loran, and Jennifer Scott for their assistance
motivated social perception: Goal-like blasts from in data collection.
the evolutionary past. In S. J. Spencer, S. Fein,
M. P. Zanna, & J. M. Olson (Eds.), Motivated Paper received 29 June 2003; revision accepted 12 March
social perception: The Ontario Symposium 2004.
(Vol. 9, pp. 215–231). Mahwah, NJ: Erlbaum.
Biographical notes
Schaller, M., Park, J. H., & Mueller, A. (2003). Fear of
the dark: Interactive effects of beliefs about danger J A S O N FA U L K N E R is a PhD student at the
and ambient darkness on ethnic stereotypes. University of British Columbia. His research
Personality and Social Psychology Bulletin, 29, 637–649. examines stereotyping, prejudice, and
discrimination from an evolutionary
Schiefenhövel, W. (1997). Good tastes and bad psychological perspective.
tastes: Preferences and aversions as biological
principles. In H. MacBeth (Ed.), Food preferences M A R K S C H A L L E R recently co-edited the book
and taste (pp. 55–64). Providence, RI: Berghahn. The Psychological Foundations of Culture (2004,
Lawrence Erlbaum Associates), and is currently
Sidanius, J., & Pratto, F. (1999). Social dominance: An co-editing the book Evolution and Social Psychology
intergroup theory of social hierarchy and oppression. (forthcoming, Psychology Press).
New York: Cambridge University Press.
J U S T I N H . PA R K is currently a PhD student at the
Southwick, C. H., Siddiqi, M. F., Farooqui, M. Y., & University of British Columbia. He investigates
Pal, B. C. (1974). Xenophobia among free-ranging psychological mechanisms that contribute to
rhesus groups in India. In R. Holloway (Ed.), people’s construction of their social reality, in
Primate aggression, territoriality, and xenophobia particular, mechanisms that have been shaped
(pp. 185–209). New York: Academic Press. by evolution.

Stangor, C., & Crandall, C. S. (2000). Threat and the L E S L E Y A . D U N C A N is an MA student at the
social construction of stigma. In T. F. Heatherton, University of British Columbia. Her research
R. E. Kleck, M. R. Hebl, & J. G. Hull (Eds.), The investigates disease-avoidance processes and
social psychology of stigma (pp. 62–87). New York, stigmatization.
Guilford.

Stat, D. (1995). Foreigners and Englishmen: The
controversy over immigration and population,
1660–1760. Newark: University of Delaware Press.

Stephan, W. G., Diaz-Loving, R., & Duran, A. (2000).
Integrated threat theory and intercultural
attitudes: Mexico and the United States. Journal of
Cross-Cultural Psychology, 31, 240–249.

Stephan, W. G., Ybarra, O., & Bachman, G. (1999).
Prejudice toward immigrants: An integrated threat

353