Help Less or Help More — Perceived Intergroup Threat and Out-Group Helping

The present research investigated the relationship between intergroup threat and prosocial actions toward out-group members. We proposed that intergroup threat not only results in direct negative intergroup attitudes and behavior, but also promotes an indirect reduction in out-group helping. Study 1 (N = 54) was conducted in a realistic setting, whereas Study 2 (N = 92) employed an experimental context. Willingness to help (help intent) and monetary donation (help behavior) was compared between the threatening out-group and the non-threatening out-group. The results showed that when intergroup threat was experienced, more help was offered to the non-threatening out-group than the threatening out-group. The findings suggest that perceived intergroup threat attenuates prosocial behavior toward out-group members.

The central hypothesis of the present research is that, in addition to the promotion of negative attitudes and negative behavior, perceived intergroup threat may also result in the reduction of helping behavior toward out-group members.Specifically, we propose that when intergroup threat is perceived, people may offer less help to the threatening out-group than the non-threatening out-group.

Intergroup Threat and Out-group Helping
Intergroup threat is a well-established group-level phenomenon in social psychology.According to intergroup threat theory (Stephan, Ybarra & Morrison, 2009), intergroup threat is experienced when members of one group perceive that another group is in a position to cause them harm.In the original version of intergroup threat theory, labeled integrated threat theory (Stephan & Stephan, 2000), four types of threat were included, but this number has been reduced to two basic types-realistic and symbolic threats (Stephan & Renfro, 2002).Realistic threats include threats to in-group members' political or economic power, general welfare and other tangible resources.Symbolic threats include threats to in-group members' belief system, the loss of face or honor, and the undermining of in-group members' self-identity or self-esteem (Stephan, Renfro, & Davis, 2002;Stephan & Stephan, 1985;Stephan, et al., 2009).
Previous studies have revealed that intergroup threat has a major effect on social attitudes and social behavior.The consequences of threat are related to negative responses within an intergroup context.Perceived threat promotes negative attitudes toward out-group members (Riek, Mania, & Gaertner, 2006;Stephan, Renfro, Esses, Stephan, & Martin, 2005).When American students perceived that immigrants posed a threat to both their worldviews (symbolic threats) and their economic, political, and physical well-being (realistic threats), they responded with negative attitudes that reflected negative feelings (e.g., dislike, resentment, disapproval, rejection) (Stephan, et al., 2005).Intergroup threat also promotes negative behavior, ranging from aggression, hostility, and discrimination, to harassment, warfare and other forms of open conflict (Stephan, et al., 2009).For example, psychology students whose status was threatened by medical students subsequently discriminated against social work students who had a lower status (Cadinu & Reggiori, 2002).Male participants who were exposed to threat were more likely to sexually harass a female confederate than those who did not experience any threat (Maass, Cadinu, Guarnieri, & Grasselli, 2003).Stereotype disconfirmation bias, in which out-group stereotypes are thought to be more difficult to disconfirm than in-group stereotypes, is intensified under a threat condition (Ybarra, Stephan, Schaberg, & Lawrence, 2003).Moreover, violence against the out-group is more likely, and easier to justify, when intergroup threat is made salient (Stephan, et al., 2009).
Although the relationship between intergroup threat and negative responses has been extensively studied, only a small number of studies have focused on how intergroup threat influences positive behavior (Renfro, Duran, Stephan, & Clason, 2006;Sawires & Peacock, 2000).Logically, negative responses to threat not only entail the promotion of direct intergroup conflicts, such as aggression, discrimination and stereotyping, but also encompass an indirect reduction in positive behavior.For instance, when exposed to threat, individuals responded by opposing policies favoring the out-group (Renfro, et al., 2006;Sawires & Peacock, 2000).The present research focuses on how intergroup threat may affect out-group helping behavior, which is particularly important in social life.There would be destructive social consequences if (sufficient) help were not offered to out-group members who were in need.This is especially the case for victims at a grave disadvantage, whose life and existence may be threatened under extreme conditions, such as natural disasters or warfare.As E. O. Wilson (1978) once noted "If human beings [do] to a large extent … favor their own relatives and tribe, only a limited amount of global harmony is possible" (p.167).In other words, prosocial behavior toward an out-group is a clear manifestation of humanitarianism, and an important foundation for intergroup harmony.The present research therefore focuses on how intergroup threat affects out-group helping.
Previous literature suggests that perceived threats reduce prosocial behavior toward out-group members.Jackson and Esses (1997) found that participants who perceived that their values were violated by an out-group were less likely to offer help to this group than those who did not perceive so.Economic competition between host populations and immigrants also reduced out-group helping (Jackson & Esses, 2000).After reading an editorial which aimed to elicit the perception of economic competition, university undergraduates were less likely to support immigrants than participants in the control condition.Group status instability also inhibited prosocial actions toward a lower status out-group (Cunningham & Platow, 2007).Participants from a relatively high status group (Australians), whose group status was perceived stable or unstable, were asked to rate their helping attitudes toward unemployed out-group members from a lower status group (immigrants).The results showed a significant decrease in helping attitudes toward out-group members in the unstable group status condition (compared with the stable condition), demonstrating that group status instability deterred members of a higher status group from helping a lower status group.In addition, harmful actions carried out by members of an out-group reduced helping behavior toward individuals from that group (Flippen, Hornstein, Siegal, & Weitzman, 1996).Participants who were informed that their health would suffer a negative impact due to the actions of members from a specific dissimilar community, offered less help to a person who came from the harmful community than a person who came from a harmless community.
Together, the existing literature suggests that perceived threats -including perceptions that one's values are being violated, economic competition, group status instability, and harmful actions to public health -reduce prosocial behavior toward the target out-group (Cunningham & Platow, 2007;Flippen, et al., 1996;Jackson & Esses, 1997, 2000).Therefore, we expect that intergroup threat not only increases negative intergroup attitudes and behavior, but also decreases positive intergroup behavior, specifically, out-group helping.To elucidate the relationship between perceived intergroup threat and out-group helping, a few issues need to be clarified.For example, one aim of Cunningham & Platow's (2007) study was to examine whether out-group helping was reduced due to perceived economic threats.However, the threat manipulation somehow was not effective; participants did not feel that any threat had been posed.Thus, the resulting helping pattern could not be attributed to perceived economic threat.In the present research, the manipulation of intergroup threat concerns a history of warfare between national groups, since evidence has shown that groups with a long history of conflict are the most prone to perceive realistic intergroup threat (Shamir & Sagiv-Schifter, 2006;Stephan et al., 2002).
The scenarios adopted in previous experiments are relatively homogeneous (most of which concerned unemployment problems) (Cunningham & Platow, 2007;Jackson & Esses, 1997, 2000).In the present research, the helping context concerns a natural disaster.We believe this scenario is extremely important, not only to victims, but to the world at large.The dependent variables of previous studies were restricted to helping attitudes (Cunningham & Platow, 2007;Jackson & Esses, 1997, 2000).As noted by Zagefka and colleagues (2011), there has been little research on intergroup prosocial behavior that utilizes anything other than self-report measures.In addition to prosocial intention, which is measured by assessing willingness to help, the current research examines prosocial behavior, which is measured in terms of monetary donation.More importantly, the current study focuses on the help given to large groups, instead of assessing the help provided to a specific individual from that group (Flippen, et al., 1996).The results will therefore be highly representative of group-level prosocial behavior, and will add to limited intergroup helping literature.
Employing real-world groups and scenarios will add an aspect of ecological validity to a research.However, real-world groups can sometimes involve confounding variables.In the study by Jackson and Esses (1997), help givers were Canadian college students, while help recipients were homosexuals and Native Canadians (Study 1), single mothers and students (Study 2).Interpretations of the findings could not exclude the possibility that Native Canadians and students, who received more help than homosexuals and single mothers, were perceived to be in-group members by Canadian students.In other words, the threat manipulation co-varied with group membership, which may result in increased in-group helping and decreased out-group helping, as people are likely to offer more help to in-group members than out-group members (Cuddy, Rock, & Norton, 2007;Levine, Prosser, Evans, & Reicher, 2005;Tumer, Hogg, Oakes, Reicher, & Wetherell, 1987).In the present research, we employed a realistic setting in Study 1, adding an aspect of ecological validity to the current research.In Study 2, by manipulating two laboratory out-groups, we attempted to attain greater control over extraneous variables that may be caused by a realistic setting

The Present Study
According to intergroup threat theory (Stephan, et al., 2009), the results of threat are primarily related to direct negative consequences.In the present research, we expect that intergroup threat not only engenders direct negative consequences, by promoting negative attitudes and negative behavior, but also brings about indirect negative consequences, by undermining positive behavior.Specifically, we propose that participants who perceive intergroup threat may offer less help to out-group members than participants who perceive no threat.
To test this assumption, Study 1 is conducted in a realistic setting in which threats are measured without any manipulation.Study 2 replicates Study 1, but in a controlled experimental context: laboratory groups are employed and intergroup threat is manipulated.In both studies, participants are informed that the target out-group is suffering a grave natural disaster, and are asked to offer help.Participants' help intention and monetary donation are assessed in order to compare the help received by the threatening out-group and the non-threatening out-group.Across the two studies, it is predicted that the perception of intergroup threat will result in the non-threatening out-group receiving more help than the threatening one.

Study 1
Study 1 was conducted in a realistic setting.Japan served as the threatening out-group, Kelu served as the non-threatening out-group.Japan is a potentially threatening country to Chinese participants, mainly because it was once at war with China.Previous intergroup threat literature has shown that prior conflicts can lead to intergroup threat (Shamir & Sagiv-Schifter, 2006;Stephan, et al., 2002); hence, it was expected that Japan would be perceived as a threatening out-group by Chinese participants.However, Kelu is a fictitious group, which was described as an island country in the Pacific Ocean.In fact, there are thousands of islands in the Pacific Ocean area and it is extremely difficult for a person to name them all.Participants would know nothing about Kelu except for the information presented.There have been no prior conflicts between Kelu and China.As a result, it would not be perceived as a threatening out-group by Chinese participants.In Study 1, we predicted that participants would provide more help to the non-threatening out-group (Kelu) than the threatening out-group (Japan).

Presentation
The presentation phrase served only as scenarios of natural disaster.Twelve photographs taken during the aftermath of the Japan earthquake in 2011 were selected from the Internet.They all portrayed ruins caused by the disaster.All of those photographs were highly consistent in style and contained no distinctive clues for participants to differentiate between different areas.The logo of the Xinhua News Agency (a news agency in China) was located in the bottom-right corner of each photograph to show the source and the veracity of the images.Participants in both conditions were presented with the same 12 photographs.One of the differences between conditions was the described location of the earthquake.Another difference was the slight variation in date.When the earthquake in Japan took place on March 11, 2011, no reports on other earthquakes in the Pacific Ocean could be seen.If the date of earthquake in Kelu was the same day or in the same period of time, participants could suspect about the news.So the date of the earthquake in Kelu was different from that of Japan.
The presentation phase consisted of 12 trials wherein the 12 photographs were randomly presented on the computer screen.It began with a brief introduction of the earthquake, including the date, the area affected and the source of the photographs.For example, the presentation phase for participants in the Kelu condition began with: What will be presented below are photos from the Kelu Island earthquake disaster area in the Pacific Ocean (March 21, 2011), all taken by journalists from the Xinhua News Agency.
Each photograph was presented on the screen for 1500ms, followed by a fixation point lasting for 500ms.

Help Giving
Participants were asked to rate their willingness to help the target group on a seven-point scale (1 = not at all, 7 = very much).They were then informed that they would receive ￥10 reward upon completion of the experiment and were asked to select a value (0 to 10) corresponding to the amount of money they would like to donate to the victims.

Threat Measurement
Participants' perceived threat of the target group was measured via ratings on a seven-point scale (1 = not at all, 7 = very much).For example, participants in the Kelu condition read: Is Kelu a threat to us? Please express your opinion on a scale of 1 to 7 (1 = not at all, 7 = very much).

Results
Table 1 presents the results.The mean scores of Study 1, including scores of participants' perceived threats, willingness to help, and monetary donation, were first submitted to a one-way ANOVA by gender, which yielded no sex difference in any of these dependent variables.

Manipulation Check
We conducted a one-way ANOVA on perceived threats, which showed that participants in the control (Kelu) condition perceived less threat (M = 3.30, SD = 1.44) than those in the threat (Japan) condition (M = 5.19, SD = 1.44;F (1, 53) = 26.17,p < 0.01).The results suggested that Kelu was perceived as a non-threatening out-group and Japan was perceived as a threatening out-group.

Help
The results show that the mean scores for help intention and monetary donation to the non-threatening out-group were higher than that for the threatening out-group.To test our hypothesis, we conducted one-way ANOVAs by group condition to assess the help patterns toward the two out-groups.We compared participants' willingness to help and their monetary donations for the non-threatening out-group and the threatening out-group.Participants in the control (Kelu) condition expressed more willingness to help (M = 6.04,SD = 0.85) than those in the threat (Japan) condition (M = 5.37, SD = 1.52;F (1, 53) = 3.94, p = 0.053).They also donated more money (M = 7.70, SD = 2.61) to the non-threatening out-group than to the threatening out-group (M = 5.22, SD = 2.78; F (1, 53) = 11.43,p < 0.01).

Discussion
The above findings suggest that when intergroup threat was perceived, participants offered less help to out-group members, compared to when no intergroup threat was perceived.These results support our hypothesis that intergroup threat attenuates out-group helping within a real-world context.One limitation, however, is that the non-threatening out-group (Kelu) is an unknown island country, whereas the threatening out-group (Japan) is a neighboring country to participants.It could be that, by contrast to the small island country of Kelu, powerful Japan was helped less because the Japanese were perceived as better able to cope with the disaster and so less help was needed from elsewhere.In this way, the group condition could be confounded with group status in Study 1.By employing two laboratory groups, Study 2 was designed to address this concern.

Study 2
Study 2 aimed to retest our hypothesis using laboratory groups.Participants in both the threat condition and the control condition read a passage, which described an anonymous out-group.The two anonymous out-groups, both called Group A, were described in the same manner, except for the mention of prior conflicts within the threat manipulation.As groups that have a history of conflicts are prone to perceive intergroup threat, we expected this manipulation to elicit intergroup threat (Shamir & Sagiv-Schifter, 2006;Stephan, et al., 2002).In Study 2, it was predicted that intergroup threat would undermine out-group helping.After being informed of prior conflicts, we hypothesized that participants would perceive intergroup threat and subsequently demonstrate less help intention and give less in terms of monetary donation to out-group members than participants who experienced no threat.

Participants and Design
Ninety-two Chinese students (28 males and 64 females, mean age = 21.27years, SD = 1.50) from Southwest University, China participated in Study 2. This study employed a single-factor between-subjects design.

Procedure
Participants were instructed that the study consisted of two tasks, a context-responding task and a recognition task.They were asked to respond to the context as if it were a real-life scenario and to carefully read the written materials that would be used for the recognition task.The recognition task served only to ensure that participants worked carefully on the threat manipulation materials.

Threat Manipulation
Seated in separate cubicles before computers, participants were presented with a passage that gave a brief introduction to Group A. Participants in both conditions read the same description as below: Group A live on an island.This island is mountainous and its forest cover rate is high.This island has a climate featuring distinct seasons and enjoys abundant rainfall.On this island, earthquakes and volcanism are frequent.
Participants in the threat condition read the following description: Group A have lived on the island for many years.They were once at war with our country and caused a great loss to our country.

Participants in the control condition read an alternative description:
Group A have lived on the island for many years.Their staple food is rice and fish.

Help Giving
Participants were subsequently told that a devastating earthquake had taken place on the island where Group A live.They were asked to rate their willingness to help Group A on a seven-point scale (1 = not at all, 7 = very much).Participants were then informed that they would receive ￥10 upon completion of the experiment.They could donate money to Group A by selecting a value from 0 to 10, corresponding to the amount of money they wished to give to the victims.

Manipulation Check
To test if the threat manipulation was effective, participants' perceived threats were measured by asking the question "Does Group A pose a threat to us?" Ratings were made on a seven-point scale (1 = not at all, 7 = very much).

Memory Check
After all ratings had been made, participants were asked to confirm the presence/absence of a war-description and a food-description presented on the screen previously.Participants were required to press F if the description had been presented before and to press J if it had not.

Results
Table 2 presents the results.The mean scores of Study 2, including scores of participants' perceived threats, willingness to help, and monetary donation, were first submitted to a one-way ANOVA by gender.Although the overall gender distribution of participants was not even, it did not differ significantly on any of the variables.

Manipulation Check
Participants' perceived threat was submitted to a one-way ANOVA by the group condition.Participants in the control condition perceived less threat (M = 3.16, SD = 1.51) than those in the threat condition (M = 4.38, SD = 1.23;F (1, 91) = 18.06, p<0.01).These results confirmed that the threat manipulation in Study 2 was successful.

Help
Main effects for willingness to help and monetary donation also emerged.Participants in the control condition expressed greater willingness to help Group A (M = 5.89, SD = 1.04) than participants in the threat condition (M = 4.92, SD = 1.70;F (1, 91) = 10.67,p < 0.01).Participants in the control condition also donated more money to Group A (M = 6.68,SD = 2.82) than participants in the threat condition (M = 5.54, SD = 2.73; F (1, 91) = 3.88, p = 0.05).

Discussion
The above results, which conceptually replicated those of Study 1, but using two laboratory groups in an experimental condition, demonstrated that intergroup threat reduced out-group helping.As predicted, the threatening out-group received less help than the non-threatening out-group, as reflected by participants' willingness to help (help intention) and their monetary donation (help behavior).Study 2 retested our hypothesis, but provides more refined evidence.The only difference between the two conditions of Study 2 concerned the threat manipulation.Thus, the potentially confounding variable in Study 1 -group status (Kelu vs. Japan)could now be ruled out.Taken together, the results of the two studies strongly support our central hypothesis that intergroup threat attenuates prosocial behavior toward out-group members.

Helping Out-groups under Intergroup Threat
The present studies explored whether out-group helping would be attenuated when people perceived intergroup threat, which was induced by prior conflict between the in-group and an out-group.Previous studies have demonstrated that this manipulation can elicit intergroup threat (Shamir & Sagiv-Schifter, 2006;Stephan, et al., 2002).In Study 1, participants were exposed to photographs of earthquake ruins, and their help intention and monetary donation to those who were suffering from the disaster was measured.In the intergroup threat condition, victims of the earthquake were from Japan, a country that was at war with China in the last century.In the non-threatening condition, victims were from Kelu, a fictitious nation which have no previous history of conflict with China.As expected, we observed that threatened participants had a decreased help intention, and donated less money to out-group victims than non-threatened participants.Study 2 retested our hypothesis using two laboratory out-groups.Intergroup threat was manipulated by informing the participants of a history of warfare between an out-groups and the in-group.After being informed of prior warfare, participants perceived intergroup threat and expressed less willingness to help the target out-group, compared with participants who did not feel threatened.Threatened participants also donated less money to the out-group than participants who perceived no threats.
There was a stable pattern of findings across the two studies regarding the attenuation of prosocial behavior toward the threatening out-group associated with intergroup threat.The results support our reasoning that intergroup threat not only leads to direct negative consequences, by promoting negative attitudes and negative behavior, but also results in indirect negative behavior, by attenuating out-group prosocial behavior.These findings add to the growing literature on how intergroup prosocial behavior is reduced by perceived threats.In addition to harmful actions (Flippen, et al., 1996), violating values (Jackson & Esses, 1997), economic competition (Jackson & Esses, 2000), and group status instability (Cunningham & Platow, 2007), prior conflict between groups attenuates out-group helping behavior as well.Although the threatening out-groups received less help than the non-threatening out-groups, this does not mean that threatened in-group members refuse to help out-group members.Across the two studies, participants were helpful to the threatening out-groups.The money donated to the threatening out-group was slightly higher than the middle point (but not statistically significant, i.e., 5; M = 5.43, p > 0.1).In addition, participants' help intention toward the threatening out-group was significantly above the middle point (i.e., 4; M = 5.08, p < 0.01), which suggests that they were willing to help the out-group members despite feeling threatened.It can be assumed that when the threatening out-group is in desperate need of help because of natural forces, in-group members are still willing to help them.
The present research also expands on researches into group-level prosocial behavior.Group-level prosocial behavior is important, as individuals in a group may behave differently from how they behave when they are on their own (Brown & Gaertner, 2002;Haslam & Hasiam, 2004;Mackie, Devos, & Smith, 2000).A careful scrutiny of prosocial behavior in real life will reveal that helping is offered and received to a large extent, between people who belong to different social groups (van Leeuwen, 2007).Prior research suggests that studies of helping behavior should go beyond the investigation of individual and interpersonal processes, to explore group processes (Cialdini, Brown, Lewis, Luce, & Neuberg, 1997).The present research investigated prosocial actions from members of one group toward another large group, rather than addressing helping behavior toward an individual from that group (Flippen, et al., 1996).Moreover, instead of relying exclusively on ratings of attitudes, as in previous intergroup helping research (Cunningham & Platow, 2007;Jackson & Esses, 1997, 2000), prosocial behavior was explored using measurements of both help intention (willingness to help) and help behavior (monetary donation).It was assumed that this would enable the results to be more representative of group-level prosocial behavior.
Our research also provides insight into how charities and international relief organizations can increase donations, or promote other forms of help, to victims of disasters or war who would otherwise be overlooked.For instance, when planning relief appeals, it may be more effective to avoid focusing on factors that could elicit intergroup threat.Instead, it might be more useful to highlight factors that will attenuate perceived intergroup threat.

Future Direction
Based on the intergroup threat theory (Stephan, et al., 2009), there are two types of intergroup threat.Realistic threats are threats to general welfare and the tangible resources of a group, whereas symbolic threats are threats to a group's belief system.In the present research, we limited our analysis to realistic intergroup threat (i.e., prior conflict).Future research could expand on the current study by investigating the relationship between symbolic intergroup threat and out-group helping.In the study by Stephan et al. (2005), realistic and symbolic intergroup threat resulted in different consequences (Stephan, et al., 2005); people responded to realistic and symbolic threat by having different attitudes toward immigrants.Future research could explore whether symbolic threat results in a different helping pattern to that of realistic threat.To generalize the present findings, the effects of intergroup threat on out-group helping could be further examined in different helping contexts and with diverse measures of prosocial behavior.
More importantly, to promote out-group helping behavior, researchers could further explore ways of decreasing perceived intergroup threat.In the present study, although the threatening out-groups were helped by the in-group members, they received significantly less help than the non-threatening out-groups.By reducing perceived intergroup threat, intergroup prosocial behavior could be promoted, and intergroup relations between adversarial groups could be improved.This idea is illustrated by the real-world example of the relations between Greece and Turkey.The humanitarian assistance toward Turkey, provided by Greek government in the aftermath of the 1999 Turkey earthquake, was proclaimed to promote the reciprocal actions of the Turkish government, when Greece was struck by an earthquake less than a month afterwards.This example illustrates how it is possible to alleviate tensions in intergroup relations.By reducing intergroup threat, mutual help can be expected, in addition to improved intergroup relations.Future investigations could test the assumption that reducing perceived intergroup threat will increase helping behavior toward the threatening out-group.Such investigations could also develop means of intervention in situations where intergroup threat is perceived.

Conclusion
The current research focused on the effects of intergroup threat on group-level prosocial behavior.It provides direct evidence for the relationship between intergroup threat and out-group prosocial behavior.The results suggest that intergroup threat not only promotes direct negative consequences in terms of negative attitudes and negative behavior, but also engenders indirect negative responses by attenuating prosocial actions toward out-group members.It also complements previous studies, which similarly found that threat reduces prosocial behavior, and it sheds light on how to promote intergroup helping.Future research should explore ways of reducing intergroup threat to increase out-group helping and improve intergroup relations.

Table 1 .
The mean scores (SD) for study 1

Table 2 .
The mean scores (SD) for study 2