Archive for March, 2010

Please find some notes outlining Walther & Parks (2002) study entitled “Cues Filtered In, Cues Filtered Out.”

Main Idea: Nothing is empirically tested in this study.  Rather, the authors provide a review of several prominent computer mediated communication (CMC) theories.

Important Theories:

Cues Filtered Out – The greater the bandwidth (number of communication cues) a system affords, the greater the social presence of communicators.  Walther & Parks discredit this theory due to time restrictions.  Also, group identity and attitude are not taken into consideration.

Media Richness Theory – the more emotionally charged the subject matter, the more complex media is required.  Therefore, texting is not an adequate forum for emotional issues (533).

Social Information Processing (SIP) – Communicators exchange social information through the content, style and timing of verbal messages on-line.  This diverges from traditional CMC thinking, which suggests that the lack of nonverbal cues restricts the flow of social information (535).

Mixed-Mode Relationships – people meet on-line and migrate off-line (550).

SIDE (539) – This theory seems to draw the least amount of criticism from Walther & Parks although they do note its shortcomings in terms of its application to interpersonal relationships.

I’ve provided below an overview of Vaes and Paladino’s paper on “The uniquely human content of stereotypes”. Let me know if you have questions.

Vaes and Paladino’s paper underlying aim is to discern the linkages between infrahumanization and intergroup stereotyping and extend this to the affinity between uniquely human and competence factors. In essence, the question is whether stereotypes associated to a group entail a uniquely human dimension.

The paper discusses the following:

In-group stereotypes – more human as opposed to out-group stereotypes, which are formulated as a consequence of the denial of out-group humanity. This occurs when the out-groups lack warmth and competence.

Stereotype Content Model (SCM) – This model illustrates how the relative status of a group plays a crucial role in ascribing the level of “perceived warmth” through

i) degree of competence, and

ii) type of interdependence (if competitive or not)

From this, four inter-group stereotypes emerge:

1) Competent and warm (i.e. in-group)

2) Competent but not warm (High status/competitive groups),

3) Incompetent but warm (Low status/uncompetitive groups)

4) Neither competent nor warm (Low status, competitive groups).

It is within this context that the paper aims to demonstrate which of the groups derived from SCM are more susceptible to being infrahumanized.

One thing to not is that when the out-group is infrahumanized, “both positive and negative secondary emotions are denied”. 3 hypotheses derived from earlier research are tested:

1) Infrahumanization hypothesis – this target-based approach to infrahumanization theory claims that the infrahumanization bias should occur in all “inter-group comparisons independently of the out-group’s position according to the SCM”.

2) Low–low out-group hypothesis – Based on the findings of Harris and Fiske (2006), low–low out-groups are infrahumanized to a larger degree than other out-groups as they are denied “full humanness”.

3) Competence hypothesis – The third and most important hypothesis examines the connection between stereotype content and the varied forms of humanness. It predicts that high competence/low warmth outgroups will be perceived as most uniquely human.


By exploring the human content of stereotypes, the paper’s findings show that the in-group is always identified as uniquely human, even when seen differently in terms of competence and warmth (based on the SCM). In this regard, Vaes and Paladino extend the infrahumanization bias to intergroup stereotypes.

Cues filtered out perspectives

Social Presence Theory

  • Greater bandwidth allows for more cues
  • Nonverbal cues make communicators more salient to each other
  • Thus, the greater the bandwidth, the greater the social presence
  • CMC results in less impression formation

Reduced Social Context Cues

  • Focused on lack of nonverbal cues to express purpose, decorum, status and affect
  • Leads to focus on task and self, and hostile, disinhibited behavior
  • When you don’t have all the cues, people behave in more selfish ways (focus more on self)
  • CMC will lead to poorly developed ,negative impressions of people

Hancock and Dunham: Results

  • CMC Breadth < FTF Breadth (in support of Social Presence Theory)
  • CMC Intensity > FTF Intensity (somewhat against Reduced Social Context Theory.  Impressions were more intense but not necessarily more negative)

In our study we want to examine the Social Identity Model of Deindividuation Effects (SIDE). The SIDE theory was developed and first named in 1991 by Lea and Spears, and then later expanded on in 1992. This theory is important in understanding computer mediated technology and communication.  The SIDE model expands on the basic deindividuation theory that examines how in crowds people will act in ways that are often not perceived as rational. When somebody is in a crowd there is a certain amount of anonymity that can effect how they will act. For example, normally if a rational person did not agree with a controversial decision made by a company, they would not usually go up to the company’s building by themselves and throw a glass bottle at it. On the other hand if a person is in a crowd of one hundred people and everyone is throwing glass bottles, then the person may be inclined to act irrational and proceed to deface the building with glass bottles.

The SIDE model even more examines anonymity and how anonymity changes the salience of personal identity and social identity, thus having a profound effect on behavior. In our study we are asking people to choose a side on a subject in which they have strong feelings about as well as a concrete opinion, and then discuss these views with someone that has the opposite opinion. To keep the study valid, it is important that people are kept anonymous in order to have the full effect. We were thinking about just giving people standard instant messaging user names, so that for example user 1 would be having a conversation with user 2, and both users will have conflicting views on the subject matter.

The cognitive side of the SIDE model deals with group immersion and anonymity and the salience with personal and social identities. Anonymity in a group can enhance the salience of social identity and depersonalize the social perceptions of others and the self. This can also lead to people perceiving others in terms of stereotypes.

In some situations, making an individual identifiable can promote a stronger social categorization. SIDE describes the cognitive processes by which the salience of social identity is affected by making information individuated or by eliminating individuated information. On the strategic side of the use of SIDE, anonymity can have strategic consequences, and can affect the ability for people to express their personal and social identities. Strategic concerns come into play when an out-group has more power than the in-group, or when the norms of both groups are different. When this happens, the identifiably of the in-group towards the out-group will shift the power between groups. The identifiably towards a more powerful out-group will limit the degree to which the in-group’s identity can be expressed freely.

SIDE is used in order to explain the effects of anonymity and social isolation in various contexts. SIDE can be used to do research with online teams and electronic relationships.

Coding – Euphemistic Discourse

Email Coding

Euphemism Definition:

According to McGlone & Batchelor, the underlying definition of euphemism is “an expression referring to a stimulus that is perceived as more polite than the stimulus’ conventional literal label.” It is this linguistic substitution of a more agreeable expression in place of the original expression itself that this study examines. Such euphemistic circumlocutions are discerned through the coding of emails.

Coding Scheme:

1. Omit salutations and sign-offs, such as ‘sincerely’ and ‘regards’.
2. Separate out the ideas contained in each email: Disparate ideas are segmented into individual sentences.
3. Code each idea unit ( 1 = if it contains 1 or more euphemisms; 0 = no euphemism)


Inter-coder reliability: 89.8%
Percentage of euphemisms found (and agreed on): 0.88%

Confusion matrix:

Percentage of errors occurring because coder 1 coded euphemism and coder 2 did not: 100%
Vice versa percentage: 0%


Two euphemisms we agreed on:
i) “I want to eat you for breakfast, lunch, evening tea, supper, and midnight snack.”
ii) ” I want to devour you.”

Status Analysis

When you email a professor, do you call her “Mrs.”?  “Ms.”?  “Professor”?  What if she earned her doctorate degree?  This only complicates things further.  Alternatively, what if you are replying to your brother’s third email in the span of an hour?  Does your attitude change?  Do you even address him directly or does sharing the same family tree preclude him from a proper introduction?  The underlying presumption is that one’s tone, words and context of what is written changes depending upon the status of the person receiving the email.  Through a meticulous analysis of 25 emails sent to “low status” individuals and 25 emails sent to “high status” individuals, we discovered some intriguing differences in the parts of speech and semantic value in each corpus.

By examining the parts of speech applied in high and low status emails, our group focused exclusively on the individual words that were used without applying context to what was written.  One trend we discovered was that we used the preposition “of” more often in high status emails compared to low status emails. The reason for this could be when writing emails to high status emails, we may want to be very descriptive and carefully explain things; whereas, when we wrote emails to low status people, we were not as concerned with coming across crystal clear. On a similar note, words like “being” were used more in high status emails in order to be descriptive. Also, the preceding noun of a title was more evident in high status emails, which undoubtedly have an inherent formal and professional tone compared to low status emails

In terms of semantic analysis, the time period category manifested itself the most in both high and low status emails.  Although the frequency of time in high status emails edged out the frequency of time in low status emails 37-26, it is understandable that time plays a pivotal role in both high and low status emails.  These emails referred specifically to work and school deadlines or specific dates.  Time appears to be the only category for which the status of the email does not impact the frequency.  Education was one of the main areas of focus in high status emails compared to low status emails.  For example, every undergraduate in the group included at least one email to a professor as a high status email.  Money also served as a main discussion point in high status emails.  One may attribute this to the fact that our group is largely comprised of students, who are conversing with prospective employers and/or discussing funds devoted to a particular on campus group or association.  The topic of money may also seem prevalent since one of the group’s members owns his own business.

Although several differences were also demonstrated in other categories, the frequencies of time, education and money were the most prevalent.  Similarly, in the parts of speech analysis, prepositions and official titles accounted for the most prevalent occurrence.  While our suspicions that official titles, such as “Mr.” and “Dr.”, were applied more frequently in high status emails, this analysis also generated some unexpected results.  The fact that time, education and money frequently occurred in high status emails may not have been as easy to surmise prior to the analysis.  However, after a thorough examination of our group’s emails, it seems obvious that these categories would account for the largest frequencies.

Research Proposal

The InGroup Research Proposal

It is well known that people on opposing sides of conflicts often have  divergent ways of construing their social worlds. Where an insurgent sees a martyr, a Western soldier may see a terrorist. Where an environmentalist sees devastation, a mineworker may see economic opportunity. 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 destructive effects of bias.

Against this backdrop, we propose a study that will examine the  influence of different communication settings on a particular form of bias, infrahumanization (Leyens et al., 2000). Specifically, we wish to examine whether computer-mediated communication between members of opposing groups leads to greater evidence of infrahumanization than face-to-face communication. This investigation can provide theoretical insight into our understanding of this subtle-but-serious form of intergroup bias. It will also help to determine the usefulness of CMC as a tool for fruitful intergroup interaction.

Infrahumanization and Communication

Infrahumanization refers to the tendency of group members to reserve the full essence of humanity for members of their ingroup but not the outgroup (Leyens et al., 2007). This is most often studied by using 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 thought to be experienced only by humans. These include such complex emotions as admiration, fondness, disillusion, contempt, and conceit (Demoulin et al., 2004). A key finding of infrahumanization research is that people tend to attribute secondary emotions more readily to ingroup members than to outgroup members (Leyens et al., 2000). The valence (positivity or negativity) of the emotions, regardless of primary or secondary, had no bearing (Leyens et al., 2000).

Interestingly, Leyens found that the infrahumanization effect was significantly diminished when outgroup members were established as individuals, for example by giving them names (Leyens et al 2001). This suggests that infrahumanization is more likely to occur in a low-context environment where group-level attributes take precedence over nuanced, personal interaction.  Additionally, the degree of infrahumanization relied on the relevance of the outgroup to the ingroup (Cortes et al., 2005).  If members of the outgroup were not perceived as somehow relevant to the fate of the ingroup, the degree of infrahumanization was reduced (Cortes et al., 2005).

While thought to take place on a subconscious level (Gaunt, Leyens & Demoulin, 2002), 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 (Leyens et al 2001). Vaes et al. (2003) demonstrated that when secondary emotions were evoked , people tended to approach ingroup members and avoid outgroup members.  The result was discrimination of outgroup members on the basis of expression (or lack thereof) of secondary emotions. In other words, the bias operating below our conscious awareness may still affect our conscious actions toward others.

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.  Specifically, if individualization of outgroup members diminishes infrahumanization, then communications media that are rich in contextual information should provoke less bias in intergroup interactions than context-poor media. To date, such a link has not been studied by infrahumanization researchers despite its appeal as a means of fostering intergroup harmony or, at the least, avoiding unnecessary dischord.

Face-to-face communication and computer mediated communication are different. Whittaker (2003) explains the differences as a function of the “affordances” of various technologies.  For instance,  instant messaging may constrain an individual’s ability to communicate feedback compared to face-to-face communication, even though they share the qualities of being synchronous and bidirectional. Modes that different technologies support are considered one type of affordance (3).  Instant messaging provides only a linguistic mode while face to face communication includes an additional visual mode as well.  Interactivity, another affordance, measures the technology’s feedback speed (4).

Whittaker suggests without some type of feedback a “shared understanding and establishing common ground” is virtually impossible (35).  Unlike most theories of face-to-face interaction that do not attempt to explain the impact of visible behaviors on communication, Whittaker confirmed that factors such as “gaze and gesture” are integral to communicating interpersonal information (36).  As a result, technologies that do not transmit visual information concerning facial expressions “lead to interactions that are more impersonal and likely to end in deadlock” (35).  The underlying assumption is as the number of visual cues increase, so does the quality of the conversation on a personal and deeper level. This perspective dovetails nicely with the notion in infrahumanization research that the amount of interpersonal context may affect the level of bias generated during intergroup interactions. From this theoretical background, we derive the following hypothesis:

H1: Infrahumanization will be more pronounced when group members communicate via computer-mediated technology than when they communicate in a face-to-face setting.


The proposed study will examine the direct interactions of dyads, using mixed pairs composed of members of opposing groups.  Because of the established positive correlation between ingroup-outgroup relevance and strength of identification on infrahumanization, we propose to maximize those factors in our design so as to diminish the possiblity of mitigating infrahumanization through a confounding factor.

First, we will identify two topics of current relevance within our available sample pool, students at a large northeastern university. Participants will be asked to indicate their personal level of interest in those topics and also to indicate the strength and valence of their opinions. This measurement will be the basis of identifying individuals who are high-identifying members of opposing groups. At the time of the experiment, participants will be told they are part of a study on negotiation and that they will have a specified period of time to reach a creative solution concerning their area of disagreement. Each participant will be informed of the other participant’s group membership prior to the beginning of the conversation.

Conversations will take place either face-to-face or using an instant messaging program. Face-to-face conversations will be digitally recorded, and IM conversations will be logged as text.  While the participants in the instant messaging condition will not be permitted to see one another, they will be permitted the use of “emoticons” as a means of expressing emotion. No other nontextual information will be available to participants in the IM condition, but they will otherwise be free to express any information they feel is helpful and relevant. Individuals in the face-to-face condition will be permitted to speak freely, to touch, and to exchange any information they feel is relevant to their task.

When their time has expired, participants will be asked to turn in their solution (if they have one) and will then be given a paper-and-pencil evaluation form. It will ask them to list all of the emotions they felt and expressed during the conversation and all of the emotions their partner felt and expressed. A second measure will contain a matrix of animal-related words, neutral words, and uniquely human words. Participants will be asked to circle all of the words they felt applied to the communication style of their partner. Together, these two measures will evaluate the level of infrahumanization present in the interaction. Finally, participants will be asked to assess, on a seven-point scale, how typical they felt their partner was of their group. This measure will evaluate how individualized the outgroup member became during the course of the interaction, with the expectation that individualization would reduce typicality. While this study does not propose doing so, analysis of the voice and text data and of the solutions generated by the dyads may also prove fruitful.


Infrahumanization has far-reaching implications in the way we interact as a society.  Whether or not we are moving more towards a society that embraces different viewpoints or one that is increasingly polarized depends largely on the biases we display when communicating. With the ubiquity of computer-mediated communication and the increasing globalization of this communication, the way we form biases about “us” and “them” will significantly impact how we perceive, discriminate, and ultimately treat members of opposing groups.


Cortes, B., Demoulin, S., Rodriguez, R., Rodriguez, A., Leyens, J-P. (2005). Infrahumanization or Familiarity? Attribution of Uniquely Human Emotions to the Self, the Ingroup, and the Outgroup. Society for Personality and Social Psychology, 13(2), 243-252.

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

Gaunt, R., Leyens, J.-P., &Demoulin, S. (2002). Intergroup relations and the attribution of emotions: Control over memory for secondary emotions associated with ingroup or outgroup. Journal of Experimental Social Psychology38, 508-514.

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.

Leyens, J-P., Rodriguez-Perez, A., Rodriguez-Torres, R., Gaunt, R., Paladino, M-P., Vaes, J., Demoulin, S. (2001). Psychological essentialism and the differential attribution

of uniquely human emotions to ingroups and outgroups. European Journal of Psychology, 31, 395-411.

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.

Whittaker, S. (2003). Theories and methods in mediated communication. In A. C. Graesser, M. A. Gernsbacher, & S. R. Goldman (Eds.), Handbook of discourse processes (pp. 243-286). Mahwah, NJ: Erlbaum.