Identity and Emergency Intervention 1 Running head: IDENTITY AND EMERGENCY INTERVENTION Identity and Emergency Intervention: How Social Group Membership and Inclusiveness of Group Boundaries Shapes Helping Behavior

Two experiments exploring the effects of social category membership on ‘real life’ helping behavior are reported. In study one, intergroup rivalries between soccer fans are used to examine the role of identity in emergency helping. An injured stranger wearing an ingroup team shirt is more likely to be helped than when wearing a rival team shirt or an unbranded sports shirt. In study two, a more inclusive social categorization is made salient for potential helpers. Helping is extended to those who were previously identified as outgroup members, but not to those who do not display signs of group membership. Taken together, the studies show the importance of both shared identity between bystander and victim and the inclusiveness of salient identity for increasing the likelihood of emergency intervention. Identity and Emergency Intervention 3 Identity and Emergency Intervention: How Social Group Membership and Inclusiveness of Group Boundaries Shapes Helping Behavior. Over the past 40 years, social psychologists have identified a number of factors which shape the likelihood of help being offered in an emergency situation. These include the number of people present (Darley & Latané 1968, Latané & Darley 1970), the location of the incident (Milgram 1970; Levine, Martinez, Brase & Sorenson (1994) and the costs of helping (Piliavin, Dovidio, Gaertner & Clark 1981). For the most part, psychological explanations for emergency intervention have tended to be subsumed within general theories of helping behavior. For example, the arousal: cost-reward model (Piliavin, Rodin & Piliavin, 1969; Piliavin, Dovidio, Gaertner & Clark 1981; Dovidio, Piliavin, Schroeder & Clark, 1991) begins with the aversive arousal caused by the distress of others in need. In this model it is the balance of cost-reward calculations made by an individual (as a means to reduce aversive arousal) that explains helping behavior. Batson’s empathy-altruism model (Batson, 1987, 1991; Batson, Batson, Griffitt, Barrientos, Brandt, Sprengelmeyer & Bayly, 1989; Batson & Shaw, 1991) also focuses on the place of emotion in helping. The empathy-altruism model argues that helping is related to the empathetic concern an individual feels (defined as an emotional reaction characterized by feelings like compassion, tenderness, softheartedness and sympathy) for others. Batson’s primary aim is to argue that empathy based helping provides evidence for genuine altruism or selflessness in the motivation to help others. Recent work concerning both of these models has begun to move from a focus on individual and interpersonal factors to exploring the importance of group and intergroup processes in helping. What is at issue in these debates is the conceptual possibility of shared identities (Brewer & Gardner, 1996; Marcus & Kitayama, 1991; Identity and Emergency Intervention 4 Tajfel & Turner, 1985; Turner, Hogg, Oakes, Reicher & Wetherell, 1987). Arguments center on the degree to which the bystander and the victim can be said to share a common identity and the role this common identity plays in explaining helping behavior. For example, the most recent formulation of the arousal: cost-reward model (Piliavin et al., 1981, Dovidio et al., 1991) includes the concept of ‘we-ness’, described as “a sense of connectedness or a categorization of another person as a member of one’s own group” (Dovidio et al., 1991 p102). The model suggests that the categorization of others as members of the ingroup (“we”) leads to multiple and simultaneous effects. Such categorization leads to perceptions of similarity, feelings of greater closeness and increased feelings of responsibility for welfare of others. This in turn increases both arousal and the costs of not helping a victim whilst decreasing (through feelings of greater familiarity) the cost of helping. The combination of such effects is predicted to increase the likelihood of intervention. In their review of the arousal: cost reward model literature, Dovidio et al., 1991 suggests there is substantial data consistent with this aspect of the model. Similar questions about the importance of group processes emerge in recent debates over the claims made by Batson’s empathy altruism model. For example, Cialdini, Brown, Lewis, Luce and Neuberg (1997) propose a reinterpretation of the empathy-altruism model based on the insight that the self-concept can be located outside the individual and inside close related others. They introduce the concept of ‘oneness’ to describe this self-other overlap and present data which suggests that feelings of empathetic concern, rather than determining helping directly, are predicated on feelings of ‘oneness’. Although Cialdini et al (1997) distinguish the concept of ‘oneness’ from the concept of ‘we-ness’ on the grounds that the later refers to a sense of Identity and Emergency Intervention 5 merger with a collectivity, where the former refers to relationships with a specific other, they remark on the “striking similarity between the concepts we have found useful to account for our findings and those used to account for the powerful ingroup favoritism effect in which individuals allocate greater resources to their own groups (see Brewer 1979; Tajfel & Turner 1985)” (p492). Emergency Intervention and the Social Identity Tradition Given the increasing importance of group and intergroup phenomena for emotion based theories of helping, the aim of this paper is to explore the contribution that a group level perspective can bring to the literature on emergency intervention. As we have seen, both Dovidio et al., (1991) and Cialdini et al., (1997) write about the relevance of the social identity tradition (SIT: Tajfel 1978, 1982) for research on helping. The aim of this paper therefore will be to use recent work in the social identity tradition to explore the consequences of shared social category membership for the way bystanders make decisions about emergency intervention. There is already some evidence which suggests that the perception of common group membership with a victim will increase the likelihood of helping in non-emergency settings. For example, Hornstein (1972, 1976) and colleagues have shown that people are more likely to help others believed to be members of the same community or to share similar attitudes. In early studies using the ‘lost letter’ paradigm (Hornstein, Masor, Sole & Heilman, 1971; Hodgson, Hornstein & LaKind 1972), Hornstein found that pedestrians from a predominantly Jewish section of Brooklyn, New York, were more likely to post an apparently lost questionnaire, if the questionnaire appeared to have been completed by someone with similar (pro Israeli) than dissimilar (pro Arab) sentiments to their own. Later work established that the degree of agreement or disagreement with others Identity and Emergency Intervention 6 opinions (Sole, Marton & Hornstein, 1975) and the degree of perceived threat from the outgroup (Flippen, Hornstein, Siegal & Weitzman, 1996) was also important in this kind of helping There is also some evidence from the literature on race and helping that people are more likely to offer help to ingroup members than they are to outgroup members (Wegner & Crano 1975). However, this picture is not straightforward. For example, Dovidio and Gaertner (1981) have found that, under certain conditions, whites are more likely to help African Americans than other whites. Gaertner and Dovidio (1986) use the concept of aversive racism to help explain this complex picture. They argue that, in order to understand the helping behavior of white bystanders we need to understand not only the norms and values of the specific intergroup context, but also whether failure to help can be constructed as being indifferent to race. Finally, there is evidence from the literature on the contact hypothesis that ingroup members may be favored over outgroup members. In a study manipulating intergroup contact to examine the role of common ingroup identity on prejudice reduction and prosocial behavior, Dovidio et al. (1997) found evidence of in-group bias in helping. Students were more likely to volunteer to help a student distribute questionnaires for a research project when the student was believed to be an in-group member and contact occurred under common identity conditions. However, what is important about more recent theoretical work in the social identity tradition is not simply that it allows for collective identification with others, but also that it proposes the dynamic responsiveness of the self concept to social interactive contexts. This is particularly useful for studies of emergency intervention as such situations tend to be interactions between people who have had little or no prior contact, Identity and Emergency Intervention 7 tend to occur in public spaces, and to require immediate evaluations (and reevaluations) of relationships with those present in the emergency situation. Self-categorization theory (SCT: Turner, 1982; Turner et al., 1987, Turner, Oakes, Haslam and McGarty 1994), a recent development of the social identity tradition, proposes that changes in self-concept are related to changes in immediate social context. In this SCT approach, the self, rather than being a single entity, is a complex system with a number of different levels. Using the notion of the meta-contrast ratio (Turner et. al., 1987), SCT suggests that social identity changes as other groups or individuals enter (or leave) the relevant social context. What becomes important for explaining both perception and action is the particular self-definition that is salient at any given time. In addition, there is evidence from an SCT perspective that the level of inclusiveness of shared identity affects the degree to which we think what happens to others may be personally relevant. For example, Stapel, Reicher and Spears (1994) show that categorical relations affect the degree to which we see things as personally threatening. When the victim of an accident is described as an in-group rather than an out-group member, individuals are more likely to consider that the same fate might befall them. In the same way, perceiving self and ‘victim’ as members of a shared category may increase levels of intra-group co-operation and thus increase the likelihood of emergency intervention. There is already some evidence for this position. Using an SCT influenced approach to bystander behavior, Levine, Cassidy, Brazier and Reicher (2002) found evidence that when victims of violence were presented as ingroup rather than outgroup members, then participants were more likely to express an intention to intervene. Identity and Emergency Intervention 8 However, before we can make strong claims for the proposition that common social category membership between bystander and victim will lead to increased likelihood of intervention in emergency situations, it is important to recall some of the strengths of the traditional work in this area. For example, the strength and the success of Latané and Darley’s work on bystander behavior was founded on a collection of ingeniously choreographed experimental studies. By having subjects respond to smoke coming into a room (Latané & Darley, 1970) and to an epileptic seizure (Darley & Latané, 1968) they created what Latané and Nida (1981) describe as an ‘experimental realism’ for the paradigm. Participants were engaged in meaningful, involving and realistic situations. Moreover, they were unaware that the emergency unfolding before them was part of a psychology experiment. By using non-reactive measures, Latané and Darley were able to ensure that participants’ behavior was not shaped by a reflexive consideration of what the experimenters might be expecting to see. This combination of experimental realism and behavioral data allowed Latané and Darley to develop a powerful case for the bystander effect. A strong case for the importance of shared social category membership for intervention in emergency settings must also be able to demonstrate the effects of shared identity on actual helping behavior. The experiments described in this paper attempt to emulate the spirit of those early studies and in doing so demonstrate the potential importance of social category relationships for emergency intervention. In order to do so, the experiments take advantage of real life social group memberships and associated intergroup relations. In particular, the experiments take advantage of the powerful sense of group membership felt by those people who support English football (soccer) teams. This strong sense of sharing group membership with thousands of other ‘strangers’ is also bound up with Identity and Emergency Intervention 9 intergroup rivalries with fans of other teams. Moreover, the current fashion in Britain of wearing team football shirts as casual wear means that group affiliations and intergroup relations are readily apparent even amongst strangers. By taking advantage of these ‘everyday’ demonstrations of group memberships, studies of emergency helping under conditions of ‘experimental realism’ can be carried out. To that end, the experiments focus on fans of Manchester United, the wealthiest, most popular and currently most successful team in England. They utilize the rivalry between Manchester United and Liverpool FC; near neighbors, arch rivals, and most successful team of the previous decade. Background to the Experiments Before the first experiment can be described in detail, it is important to know more about the place of football (soccer) in an English social context and importance of these particular social categories (and the nature of their intergroup relationship). In terms of popularity and appeal, Manchester United could be considered to be the New York Yankees of English football. In addition to their widespread popularity they have, in the past 10 years, been the dominant team in English football. At the same time, and perhaps as a result of this success, fans of all other football clubs treat Manchester United fans as potentially inauthentic (not ‘real’ fans) who are just attracted by glamour and success. This means that Manchester United fans are in the unusual position of supporting a high status team (in terms of on-field success) but belonging to one of the more negatively valenced football supporter identities in England. The particular rivalry with Liverpool Football Club is longstanding and not just confined to competition on the football field. Liverpool and Manchester are competing regional cities in the North 1 In fact, the Yankees and Manchester United have just agreed a commercial partnership in which each club will promote the merchandising of the other in their respective countries. Identity and Emergency Intervention 10 West of England. Although the cities are only 30 miles apart, people born in the respective cities have different accents, different self images, and draw on different historical and industrial traditions. Moreover, where Manchester United currently dominate English football, Liverpool FC were the dominant team of the 1980’s. The history of animosity between some of the fans of the teams is played out in communal chants which are directed at opposition fans inside football grounds as well as occasional violent skirmishes outside football grounds. Attachment to group identities and the intergroup rivalry between Manchester United and Liverpool fans is therefore deeply held and extremely meaningful. At the same time, over the past 10 years, football has come to play an increasingly visible part in English cultural life in general. It has become common for fans of all teams to wear the shirt of their chosen club as casual wear (as opposed to just wearing the shirt to the stadium to watch the team). This means that it is not unusual to see people wearing football shirts on the streets. It also allows the possibility of reading a shared social category membership with a stranger (when they wear the shirt of the team you also support). It is this combination of strongly held identities, intergroup rivalry, and facility to read common category membership with a stranger, that will form the basis of the first experiment. Structure of the Experiment The structure of the experiment borrows several features of Darley and Batson’s (1973) ‘good Samaritan’ experiment. In particular, participants are exposed to an incident while on their way to a second location having begun an experiment in the psychology department. The experiment itself is centered around a choreographed ‘accident’ in which a confederate falls over and shouts out in pain. Participants in the experiment are selfidentified Manchester United fans who witness the ‘accident’ while Identity and Emergency Intervention 11 walking between buildings on the Lancaster University Campus. The participants have already taken part in the first stage of a study in which their identity as Manchester United fans has been made salient. They are walking between buildings at the request of the experimenters who have asked them to go to a separate location to watch a video. The social category membership of the victim is manipulated by the clothing the confederate is wearing. He either wears a shirt which designates him as an ingroup member (Manchester United shirt); or an outgroup member (Liverpool FC shirt); or wears a shirt which offers no social category information (plain, unbranded sports shirt). As the participants travel between buildings, they see a confederate come jogging into view. The confederate runs down a bank which leads to a car park across which the participants are walking. The confederate trips and falls in line of sight (although not directly in the path of) of the participant and about 15 feet away. The measures of intervention are taken from Darley and Batson (1973) and range on a five-point scale from not noticing the victim to physically assisting the victim out of the experimental context. The behavior of the participants is assessed by three independent observers and it is predicted that intervention levels will be highest when the victim is clearly identified as an ingroup member and least likely when the victim is clearly identified as an outgroup member. Identity and Emergency Intervention 12 STUDY 1 Method Participants 45 male students from Lancaster University, aged between 18 21 years. All participants were self-identified Manchester United Football Club supporters. Participants received no payment for taking part. Procedure. Recruitment. The recruitment procedure for participants in the study was designed to ensure that, while only Manchester United fans would take part in the study, the participants themselves were not aware of this fact. This was important for two reasons: firstly, if participants felt that Manchester United fans were being targeted in some way it might make them more suspicious of the experimental manipulation which involved the Manchester United football shirt; secondly, given the particular (negative) reputation of Manchester United fans amongst football fans in general, it might make participants feel that the experiment was an examination of their ‘character’. Thus, posters were displayed around the University campus asking for people who were supporters of all English Premier League football teams to take part in a study of football clubs and their fans. Participants who signed up to take part in the study were asked to give a contact number/ address and to identify which Premier League team they supported. Only those participants who identified themselves as Manchester United supporters were contacted to take part in the experiment (supporters of other teams were told that we had sufficient participants for the study). Raising the salience of identity. Having identified a pool of Manchester United supporters, participants were then invited to come to the Psychology Department. Identity and Emergency Intervention 13 Participants were invited separately and 30 minutes was allowed between each appointment. When a participant arrived at the psychology department (on the second floor of the building), they were met by an experimenter, who accompanied them to a research cubicle. Once seated in the cubicle, participants were reminded that they were being asked to take part in research on football teams and their fans and invited to fill in two questionnaires. The first questionnaire asked them to identify the team they supported and to answer (in an open ended format) questions about why they supported this team; how long they have supported them; how often they watch their team play; and how they feel about the success and failure of their team. The second questionnaire asked the participants to answer (on a five point Likert scale format) ten questions which explored the participants identification with other supporters of their team. This questionnaire was adapted from Brown, Condor, Mathews, Wade & Williams (1986) identification scale and included items on cognitive (‘I am a person who identifies with ... fans’), emotional (‘I am a person who is glad to be a .......fan’) and evaluative (‘I am a person who makes excuses for being a ........ fan’) aspects of identification. In order to ensure that participants remained unaware that Manchester United supporters were the focus of the experiment (as they might have done if they had been presented with scales which had been pre-printed with ‘Manchester United’) each person was asked to write ‘Manchester United’ in the relevant position in each item before choosing an appropriate number from the scale. Creating the ‘helping’ incident. After the participants had completed the two questionnaires they were told that, as part of the study, they would be asked to watch a short video about football teams, their supporters, and crowd behavior at football matches. They were also told that the cubicle was too small to show the video and that a Identity and Emergency Intervention 14 projection room with a large screen display had been booked to show the video (the location of the projection room was familiar to the participants in the study). As this projection room was in another building adjacent to the psychology department, the experimenter informed the participant that they would accompany them to the new room. The experimenter and participant walked down the two flights of stairs, and out onto the car park which separated the two buildings. Having walked a short distance across the car park the experimenter informed the participant that they needed to go back to meet the next person in the study, but that he should continue across the car park and round the back of the building where they would be met by a second experimenter who would show him the video. The participant was then left to walk on alone. The path around the back of the building was a quiet and secluded part of the campus which narrowed as it neared the corner of the building. At the corner of the building the path had a clear line of sight to a grassy area a short distance away. Hidden observers ensured that all other people were kept out of the area and the sight of the participant. As the participant approached the corner of the building, a confederate appeared, jogging across the grass and preparing to run down a grass bank a short distance in front of the participant. The confederate had mud splattered up his legs and was wearing shorts, socks and running shoes. In addition (depending on experimental condition) he was also wearing either a Manchester United team shirt; a Liverpool FC team shirt; or an ordinary, unbranded sportshirt. As the confederate ran down the grass bank he slipped and fell over, holding onto his ankle and shouting out in pain. Having crumpled to the ground he groaned and winced, indicating that he might need help. The confederate did not make eye contact with or ask the participant for help. As the key dependant measure was whether the participant, having noticed the accident, offered Identity and Emergency Intervention 15 help, the confederate was instructed to be non-directive in response to any first contact from the participant. If a participant asked the confederate if he ‘was alright’, or if he needed help, the confederate replied, ‘I think I might have done something to my ankle’. Any further solicitation from the participant was met with tentative reassurance that the ankle was going to be alright. As the ‘accident’ unfolded, the participant was observed and rated by three independent observers, all hidden at different vantage points around the accident site. The observers were aware in advance of the shirt condition the participant was to be exposed too. When the participant departed the scene of the accident (having intervened or not) they continued on to the projection room where they were met by another member of the research team. The researcher asked the participant if they noticed anything on the journey to the projection room, and if so, what it was and how serious it appeared to be. The participant was then fully debriefed. Dependent Variable. The degree of intervention offered by the participants towards the accident victim was assessed by three independent observers, hidden from the view of the participants, and who were required to score the behavior of the participants on a 5point scale. The scale was adapted from the one developed by Darley and Batson (1973) for their ‘Good Samaritan’ experiment: 1 The participant failed to notice the victim was in need of help at all 2 Perceived the victim to be in need of help (i.e. glanced in the victim’s direction) but did not offer any form of help at all 3 Stopped and asked the victim if they were in need of help 4 Stopped and asked the victim if they were in need of help and then directly helped the victim themselves. 5 After stopping to provide assistance, participant did not leave victim and escorted them out of the experimental context Identity and Emergency Intervention 16 The only item to be excluded from Darley and Batson’s original scale was the item referring to ‘failing to intervene directly, but helping indirectly by asking someone else to help’. This item was excluded as the experiment was designed to ensure that the participants encountered the victim in the absence of any other potential bystanders.