Infrahumanization in computer-mediated communication
Bryce T. Hoffman
334 Kennedy Hall
Ithaca, NY, 14850
This study examines whether computer-mediated communication (CMC) leads to greater evidence of infrahumanization than face-to-face communication. In line with the Social Identity Deindividuation Effect (SIDE), we find a stronger attribution of uniquely human emotions and human-associated words in face-to-face communication than in CMC. These findings are used to modify and extend current theoretical perspectives on infrahumanization within the context of intergroup relations.
Infrahumanization, intergroup relations, cmc, computer-mediated communication, instant messaging, group dynamics
ACM Classification Keywords
H.5.3 [Information Interfaces and Presentation]: Group and Organization Interfaces—Computer-supported cooperative work, Organizational design; K.4.3 [Computers and Society]: Organizational Impacts—Computer-supported cooperative work, Reengineering, Automation
It is well known that people on opposing sides of conflicts often have divergent ways of construing their social worlds. Groups serve the function of bringing like-minded individuals together but also set the stage for bias against outsiders, ranging from subtle derision to deadly violence. In a pluralistic society with innumerable groups and conflicts, it is valuable to identify factors that can foster intergroup understanding or at least minimize the detrimental effects of bias. The purpose of this study is to compare evidence of a particular form of bias, infrahumanization , when group members communicate via computer-mediated technology and when they communicate in a face-to-face setting. After a brief review of the relevant literature, we present the results of our study aimed at providing theoretical insight into our understanding of this subtle but potentially serious form of intergroup bias. In doing so, we will focus on how identity within a group may develop as a function of intergroup dynamics.
Infrahumanization refers to the tendency of group members to associate more uniquely human attributes to members of the ingroup than the outgroup. This has most often been studied through the attribution of emotions to ingroup and outgroup members. Leyens has argued that emotions can be categorized into two types: primary emotions that both human and animals can feel (such as anger, surprise, fear, and disgust) and secondary emotions, which are experienced only by humans. These include such complex emotions as admiration, fondness, disillusion, contempt, and conceit .
A key finding of infrahumanization research is that people tend to attribute secondary emotions more readily to ingroup members than to outgroup members . Aside from emotion words, evidence suggests people tend to attribute more human-related words to the ingroup members and more animal-related words to the outgroup members . While thought to occur on a subconscious level, studies have linked infrahumanization with behavioral effects. Denying the humanity of outgroup members can have negative effects on social interactions, such as nonintervention in cases of emergency .
Because communication is the point of intersection between all groups—and because of what is known about the factors that mediate infrahumanization—it seems especially valuable to study this phenomenon in a communication context. By comparing face-to-face communication to computer mediated communication, this study provides a unique scope to examine the effects of infrahumanization. The different settings and cues inherent in the face-to-face and CMC environments present an intriguing opportunity for further analysis.
The SIDE model
The SIDE model (Social Identity Deindividuation Effect) provides a useful platform for exploring infrahumanization in a communication context . The SIDE model suggests that anonymity can enhance social influence when a group identity is made salient. Social identity is salient when group members are identified as members of a group rather than as individuals. When social identity is salient, and members are made visually anonymous, group identification effects such as stereotyping and compliance with normative behavior are enhanced . In addition to exaggerated group identification effects, the SIDE model suggests that visual anonymous groups are more conducive to social influence than visually identifiable groups. In other words, visual anonymity increases group salience, which consequently increases social influence .
Thus, when comparing face-to-face with computer-mediated forms of communication, SIDE suggests that the visual anonymity of CMC will lead to overattributions and exaggerated group identification effects. From an infrahumanization perspective, we can expect to see similar effects. When group identity is made salient, and members are visually anonymous, we should expect to see exaggerated effects of infrahumanization. From the established SIDE and infrahumanization literature, we hypothesize the following:
H1: A greater degree of infrahumanization when describing the outgroup than the ingroup (for both FtF and CMC)
We predict that across all conditions we would find the classic infrahumanization effect of people attributing human characters preferentially to the ingroup.
H2: A lesser degree of infrahumanization when describing the individual partner than the group (for both FtF and CMC).
We predict that across all conditions we would find that identifying the outgroup member as an individual rather than the group member would mitigate the infrahumanizaiton effect.
H3a: A greater degree of infrahumanization in CMC than FtF when describing outgroup members
H3b: A lesser degree of infrahumanization in CMC than FtF when describing ingroup members
We predict that visual anonymity will exaggerate the effects of infrahumanization. We expect a greater degree of infrahumanization when describing the outgroup and lesser degree of infrahumanization when describing the ingroup in both CMC and FtF, but to a greater degree (in both directions) in CMC.
Participants were recruited for a research study on “political negotiation” and told that they would have a conversation with someone whose views may or may not be different from their own before answering a few simple questions.
Upon signing an informed consent form and filling out a short survey about their political attitudes, political party membership, and strength of identification, participants were brought to a designated location in order to participate in either a face-to-face discussion or a computer-mediated discussion. Separate testing rooms for the instant messaging conversation were set up with a desk and computer with an instant messaging program. For the instant messaging conversation, subjects were prevented from seeing each other prior to their participation so as to preserve visual anonymity. For the face-to-face conversation the test subjects were brought together in one room where they participated in the study.
Participants were asked to come to agreement on what they believed to be a fair letter grade for the Obama administration. They were given a ten-minute time limit.
Group membership was made salient by informing each participant of their own and their partner’s political views prior to the exercise. Subsequent to the ten-minute discussion, all participants were separated and asked to complete a word selection task.
Based on pilot testing, we created a list of 32 descriptive terms in four categories: primary and secondary emotions, human-animal descriptors, morality judgments, and warmth-competence terms. Each category contained eight words and was balanced for valence. The emotion and human-animal categories were also balanced with four uniquely human words and four non-uniquely human words. All words were presented in a randomized order.
Participants were asked to choose eight words from the list they felt best described their partner. Using the same set of words, participants were then asked to choose sixteen words—the eight they felt best described members of their ingroup in general and the eight they felt best described members of the outgroup in general. For this second portion of the task, all 32 words were available but participants were asked not to ascribe the same word to more than one group, thus if a participant described Democrats as “folksy” s/he could not also describe Republicans with that term.
Data were obtained from 16 participants, 10 males and six females. Thirteen participants self-identified as Democrats, three as Republicans. This allowed for three intergroup pairings of Republicans with Democrats and five intragroup pairings, all Democrats. One set of responses was discarded because a participant did not follow the directions provided. Fifteen response sets are included in this analysis.
Based on pilot testing, eight of the 32 possible descriptors respondents could choose in their word selection task were identified as “uniquely human.” Four represented secondary emotions: hopeful, resentful, disenchanted, and optimistic. Four were characteristics that can describe humans but not animals: educated, citizen, folksy, and criminal.
For each participant, three scores were calculated based on the number of “uniquely human” attributes they ascribed to 1) their partner 2) outgroup members in general and 3) ingroup members in general. These scores were then averaged across all participants and according to each of the experimental conditions. The differences reported in this section, then, are differences in the average number of uniquely human words used to describe the relevant target individual or group.
Because of the small sample size, results will neither be analyzed nor discussed in terms of statistical significance. However, trends in the data relative to our hypotheses can be identified.
Our first hypothesis predicted the classic infrahumanization effect. We expected that participants across all conditions would preferentially attribute uniquely human qualities to the ingroup over the outgroup. Our second hypothesis predicted that there would be less evidence of infrahumanization of partners (i.e. the person they talked to) than of generic group members (i.e. Republicans or Democrats in general).
Our data appear to offer support for both hypotheses. Overall, participants described their partners with the greatest number of uniquely human words, using an average of 3.27 (SD=.8). Ingroup members in general were described with slightly fewer at 3 (SD=1.1). As predicted, there was a sharp decrease in the number of uniquely human words used to describe outgroup members in general, with an average of 1.73 (SD=1.1).
Our additional hypotheses predicted:
• H3a: That infrahumanization of the outgroup would be exaggerated among participants in the CMC condition. (see Figure 1).
Figure 1. Comparison of the number of uniquely human words used to describe target groups after computer chat or face-to-face discussion.
• H3b: That infrahumanization of the ingroup would be exaggerated among participants in the face-to-face condition.
Ingroups and outgroups alike were described with fewer uniquely human words by participants in the CMC condition, providing preliminary support for the former hypothesis but not the latter. As predicted, there was a sharp contrast in evaluations of the outgroup. Participants in the face-to-face condition used an average of 2.29 (SD=1.1) uniquely human words to describe the outgroup, while participants who spoke via instant messaging chose only 1.25 (SD=.9). Ingroup evaluations followed a similar but less dramatic pattern. Participants in the face-to-face condition chose 3.14 (SD=1.22) uniquely human words on average, while those in the CMC condition chose 2.87 (SD=.99).
Infrahumanization has far-reaching implications for the way we interact as a society. Whether or not we are becoming a more tolerant society that embraces different viewpoints or one that is increasingly polarized depends largely on the biases we display when communicating. With the ubiquity of CMC and the increasing globalization of this form of communication, the way we develop biases about “us” and “them” will significantly impact how we perceive, discriminate, and ultimately treat members of opposing groups.
Evidence from this study supports our hypothesis that the communication medium plays a significant role in infrahumanization. While CMC may foster bias and reinforce preconceived notions in intergroup dialogue, one solution may be to individuate people within CMC spaces. The effects of CMC on infrahumanization have demonstrated some consistency with SIDE and further integration of these theories should be explored.
We thank our participants, Ashley Downs for her logistical support, and all who have assisted in the development of this research.
1. Leyens, J-P., Demoulin, S., Vaes, J., Gaunt, R., Paladino, M. (2007). Infra-humanization: The wall of group differences. Social Issues and Policy Review, 1, 139-172.
2. Demoulin, S., Leyens, J.-P., Paladino, M., Rodriguez-Torres, R., Rodriguez-Perez, A., & Dovidio, J., (2004). Dimensions of ‘uniquely’ and ‘non-uniquely’ human emotions. Cognition and Emotion, 18(1), 71-96
3. Leyens,J-P., Paladino, P., Rodriguez-Torres, R., Vaes, J., Demoulin, S., Rodriguez-Perez, A., et al. (2000). The emotional side of prejudice: The attribution of secondary emotions to ingroups and outgroups. Personality & Social Psychology Review (Lawrence Erlbaum Associates), 4(2), 186-197. Retrieved from Academic Search Premier database.
4. Viki, G., Winchester, L., Titshall, L., Chisango, T., Pina, A., & Russell, R. (2006). Beyond secondary emotions: the infrahumanization of outgroups using human-related and animal-related words. Social Cognition,24(6), 753-775. Retrieved from Academic Search Premier database.
5. Vaes, J., Castelli, L., Paladino, M., Leyens, J-P., Giovanazzi, A. (2003). On the Behavioral Consequences of Infrahumanization: The Implicit Role of Uniquely Human Emotions in Intergroup Relations. Journal of Personality and Social Psychology, 85(6), 1016-1034.
6. Postmes, T., & Spears, R. (1998). Deindividuation and anti-normative behavior: A meta-analysis. Psychological Bulletin, 123, 238-259.
7. Reicher, S. D., Spears, R., & Postmes, T. (1995). A social identity model of deindividuation phenomena. In W. Stroebe & M. Hewstone (Eds.), European review of social psychology, Vol. 6 (pp. 161-198). Chichester, UK: Wiley.
8. Postmes, T., Spears, R., Sakhel, K., & de Groot, D. (2001). Social Influence in Computer-Mediated Communication: The Effects of Anonymity on Group Behavior. Personality and Social Psychology Bulletin, 27, 1242-1254.