People have a deeply-rooted tendency to dislike social or ethnic groups of which they themselves are not a member 
. Often referred to as prejudice
, this tendency is considered to be a central cause of various societal problems. Accordingly, researchers have extensively studied these negative attitudes in order to delineate their causes and consequences. An important finding from this field of research is that the experience of group-related threats—regardless of their exact source—boosts prejudicial attitudes and behaviors. Whether threat stems from competition with other groups for scarce resources, from conflicting values, or from potentially being harmed or devaluated by an out-group
, it non-specifically increases prejudice 
As previous research on this topic has mainly relied on self-report measures of threat, the neurophysiological basis of the threat–prejudice relationship has remained rather obscure. In the present article, we propose that the extent to which the hypothalamic–pituitary–adrenal (HPA) axis is activated during intergroup situations is predictive of expressions of prejudice. This proposal is supported by the idea that, although they have been examined in different fields of research, cortisol (i.e., the end-point of the HPA axis) and prejudice have the same antecedents and serve similar functions. That is, both elevations of cortisol and increases in prejudice have been proposed to be caused by social threats. Furthermore, both cortisol and prejudice have been proposed to have restorative functions, in that they help people to (re)gain control over the situation after threats have been encountered. We will now address these ideas in greater detail.
The secretion of cortisol is known to occur in response to stressors that are (a) uncontrollable and (b) involve social evaluation 
. For example, a procedure that has both of these components is the Trier Social Stress Test, in which people are required to give a speech and perform a cognitive task in front of an evaluating audience 
. This procedure reliably leads to the secretion of cortisol. Similarly, social-group-related threats also seem to have the potential to cause elevations in cortisol. For example, in a recent study, female subjects interacted with a male confederate who was introduced as endorsing sexist attitudes. Results
indicated that subjects who were predisposed towards detecting sexism in society showed elevated cortisol due to this interaction 
. So, the experience of social threat—e.g., due to inter-group situations that are perceived as highly stressful—can cause cortisol elevations 
. Importantly, many studies have shown that out-group threats also
increase out-group prejudice 
. So, social threat turns out to be a cause of both cortisol elevations and expressions of prejudice, raising the possibility that prejudice is underpinned by (or at least correlated with) HPA-axis activation.
In addition, elevations of cortisol and expressions of prejudice also overlap in their functionality, in that both have been proposed to have restorative functions. For cortisol, studies have shown that while social stressors normally decrease people's mood, administration of cortisol ameliorates this effect 
. Similar effects have been found in patients who suffer from post-traumatic stress disorder, who benefit from cortisol administration because this inhibits the retrieval of traumatic memories 
. These findings support the perspective that cortisol helps people to adaptively regain control over the situation after stressors are experienced, via the modulation of cognitive and metabolic processes 
. Intriguingly, similar functions have often been ascribed to prejudice. For example, expressing prejudice has been found to restore people's self-esteem after this has been harmed 
. Similarly, expressing prejudice towards other groups increases the extent to which people identify with and value their own group, especially after their own group is threatened in any way 
. Thus, there is also overlap in the functionality of cortisol and prejudice, in that both broadly restore well-being and serve to regain control after threats have been encountered.
The lines of reasoning addressed above indicate that cortisol and prejudice have shared antecedents (i.e., social threat) and compatible functions (i.e., restoring well-being, regaining control). It is thus an interesting possibility that they are intertwined, and that cortisol responses (due to intergroup situations) are predictive of expressions of prejudice. Indeed, some research suggests that HPA functioning is directly related to prejudice-related phenomena. Specifically, one study showed that HPA-axis dysregulation correlates with internalized racism (i.e., the extent to which people endorse negative stereotypes about their own group) 
. Yet, while this finding is generally in line with the ideas put forward in this paper, this previous study did not address prejudice directed at other groups (i.e., out-groups). Moreover, it is currently less clear whether acute
cortisol responses to intergroup stressors are linked to prejudice. Accordingly, it is the main purpose of the present experiment to establish whether a relation between acute cortisol elevations and self-reported prejudice towards an out-group exists. In our study, participants first anticipate an intergroup interaction, while their cortisol response to this anticipation is measured. Next, participants engage in a (bogus) inter-group interaction, after which self-report measures of prejudice are administered.
While this procedure allows us to test our main hypothesis, we also aim to explore two boundary conditions of the anticipated cortisol–prejudice relation. First, on the basis of previous research on threats in intergroup situations, we expect that the cortisol–prejudice relation is strongest for forms of prejudice that are difficult to suppress (i.e., subtle rather than blatant forms of prejudice) 
. Second, the cortisol–prejudice relation is expected be weaker when people feel similar to the out-group member with whom they will interact. That is, in such contexts of interpersonal similarity, people perceive social interactions—and any threats that are accompanied with it—to be about individuals, rather than about groups 
. Thus, if people experience threats and/or have elevated cortisol in such a context, these do not necessarily relate to the out-group as a whole. To test this additional idea, half of the participants were led to believe they were interpersonally similar (vs. dissimilar) to the bogus out-group member with whom they are about to interact.