R U There? Ostracism by Cell Phone Text Messages

Anita Smith Kipling D. Williams Macquarie University Purdue University

Ostracism has a powerful negative effect on individuals. Face-to-face (i.e., social) ostracism is not necessary for these effects to emerge; they occur also in Internet ball toss games and within chat rooms. In previous research, ostracized individuals observed the interaction between other members of a group. In this experiment, the authors tested whether imagined ostracism is sufﬁcient to inﬂict psychological pain. They used a triadic cell phone text-messaging method such that after initial inclusion in a conversation, participants either continued to be included or received no further messages from the others (and saw no messages between the others). Ostracized participants reported worse mood; reported lower state levels of belonging, control, self-esteem, and meaningful existence; and wrote more provoking messages.

Considerable research in the areas of ostra- (1995) belongingness theory, social exclusion cism (i.e., being ignored and excluded; Eisen- causes anxiety because it signals a potential or berger, Lieberman, & Williams, 2003; Wil- actual loss of belonging. And according to Wil- liams, 2001; Williams, Cheung, & Choi, 2000; liams’s (1997, 2001) need–threat model of os- Zadro, Williams, & Richardson, 2004), social tracism, being ignored and excluded can simul- exclusion (Twenge, Baumeister, Tice, & taneously thwart desires of belonging, self-es- Stucke, 2001; Twenge & Campbell, 2003; teem, control, and meaningful existence. Twenge, Catanese, & Baumeister, 2003), and The research on ostracism, for example, has rejection (Leary, 1990, 2001; Leary, Kowalski, demonstrated that simply being ignored and ex- Smith, & Phillips, 2003) has demonstrated that cluded is enough to produce lower self-reported being ignored, excluded, or rejected is a uni- satisfaction levels of belonging, control, self- formly unpleasant, even painful (Eisenberger et esteem, and meaningful existence regardless of al., 2003) experience. According to Leary’s so- how irrational this reaction is. For instance, ciometer hypothesis (Leary & Baumeister, Eisenberger et al. (2003) found that participants 2000; Leary, Tambor, Terdal, & Downs, 1995), ostracized in a virtual ball toss game, Cyberball these aversive interpersonal events signal rela- (Williams et al., 2000), reported lower levels on tional devaluation, which results in hurt feelings these needs even when they were told that their and lowered levels of belonging and self-es- computer was not yet hooked up to the other teem. According to Baumeister and Leary’s two people’s computers; thus, the others could not throw the ball to them even if they wanted to. Zadro et al. (2004) found that being ostra- Anita Smith, Department of Psychology, Macquarie Uni- cized on Cyberball by the computer was just as versity, Sydney, New South Wales, Australia; Kipling D. unpleasant as being ostracized by other human Williams, Department of Psychological Sciences, Purdue University. beings. And more recently, Gonsalkorale and An earlier version of this article was presented at the Williams (2004) found that Cyberball partici- meeting of the Society for Australasian Social Psycholo- pants were just as negatively affected when they gists, Sydney, New South Wales, Australia, April 2003. We were ostracized by in-group members (of the thank the Australian Research Council for funding Kipling D. Williams for this research. We are indebted to Hausmann same political party), out-group members (of Communications, who supplied three Nokia mobile phones the rival political party), or even despised out- for this research. Thanks also go to Geoff MacDonald for group members (the Ku Klux Klan). Clearly, his comments on a draft. the research indicates that being ostracized sets Correspondence concerning this article should be addressed to Kipling D. Williams, Department of Psycholog- off pain signals to the brain (as shown in func- ical Sciences, Purdue University, West Lafayette, IN 47907. tional magnetic resonance imaging [fMRI] E-mail: [email protected] brain scans by Eisenberger et al., 2003) to pre-

291 292 SMITH AND WILLIAMS pare the individual to deal immediately with the tional aggression involves harming a person threat of exclusion, and that this sequence of psychologically by manipulating relationships events is so powerful that it appears to occur or threatening to hurt a person or that person’s precognitively. That is, cognitive factors that property (Crick, Casa, & Nelson, 2002). A sur- ought to diminish the negative impact of ostra- vey conducted by the National Children’s cism (nondeliberate, by hated out-groups or Home found that 16% of children between the computers) do not. ages of 11 and 19 were harassed by text mes- To date, research on these variously con- saging, whereas only 7% were harassed in In- nected concepts has been conducted in face-to- ternet chat rooms and 4% were harassed via face settings, in chat rooms, or through the use e-mail. In Scotland and New Zealand, text mes- of Cyberball (for reviews, see Leary, in press; sage bullying has become such a problem that Twenge, in press; Williams & Zadro, in press). some schools have banned mobile phones The one common element in all of these para- (“Text message bullies,” 2001; “Text message digms is that participants are excluded, ignored, bullying,” 2001). However, it is not only school or rejected while they see or hear the others in the group engaging in social intercourse. Thus, children that are taking advantage of the latest it is not clear whether being ignored and ex- way to terrorize others. After a horrific gang cluded is sufficient to cause the negative feel- rape in Sydney, Australia in 2002, one of the ings or whether being ignored and excluded perpetrators sent disturbing text messages to his must be coupled with seeing or knowing that the victims, and when his phone was seized, mes- others in the group are enjoying social inclusion sages such as “When you are feeling down . . . with each other. One purpose of the present bash a Christian or Catholic to lift up” were study was to devise a paradigm in which the found stored on his phone (Wockner, 2002, ignored and excluded participant does not see or p. 3). hear the others engaging in social interaction to There is evidence that text message commu- determine whether ostracism alone is sufficient nication can have a profound effect on an indi- to thwart the desired needs postulated by Wil- vidual’s emotions (Taylor & Harper, 2003). In liams (1997, 2001). One relatively new method severe cases of text message harassment, ag- of social communication, the cell phone, offers gressive text message campaigns have appar- such a paradigm. ently contributed to suicide attempts. Gail The invention of the cell phone has had a Jones, aged 15, took an overdose of drugs after profound effect on the way in which people receiving a series of aversive text messages and communicate and organize their lives (Geser, silent calls (McVeigh, 2001). 2002). Cell phones have made it easier for peo- According to Crick et al. (2002), a common ple to communicate both in verbal and in writ- approach of perpetrators of relational aggres- ten form by way of cell phone text message sion is to threaten withdrawal of affection or to communication, otherwise known as SMS exclude their target from important social gath- (short message service). The development of erings. Taylor and Harper (2003) found that if SMS communication has surpassed all expecta- an individual did not receive the text messages tions (Rautiainen & Kasesniemi, 2000). Ac- they expected, this breached underlying agree- cording to Vodafone, in the year 2000 the number of messages sent jumped fivefold, to 200 ments between friends and left the individual billion messages, and in 2002 the number of feeling excluded from social networks. There text messages rose to 366 billion. In the year is qualitative evidence that individuals who do 2000 a Danish medical clinic admitted its first not receive text messages when they expect to text-messaging addict, who sent up to 200 mes- feel left out and dejected (Taylor & Harper, sages a day (Underhill et al., 2001). 2003). However, so far this issue has not been New technologies such as text messaging are empirically addressed. One aim of this study making it easier for people to maintain contact. was to examine whether short episodes of text However, these technologies bring with them message ostracism lead to feelings of being new means to perpetrate existing social prob- ignored and excluded. If they do, the individuals lems such as harassment and relational aggres- are expected to suffer a cascade of negative sion. In contrast to physical aggression, rela- effects. OSTRACISM BY CELL PHONE TEXT MESSAGES 293

Ostracism lowing ostracism on a subsequent task (with a new set of people). The predictions made in this study are based To date, the accumulated literature on the on the theoretical predictions derived from Wil- immediate effects of ostracism demonstrates liams’s (2001) model of ostracism. The crux of that regardless of individual differences in the the theory proposes that ostracism has the targets of ostracism and irrespective of the pres- unique potential to threaten state levels of four ence of factors that ought to enable the target to fundamental human needs: belonging, control, discount the importance of the ostracism, ostra- self-esteem, and meaningful existence. When cism hurts. In fact, despite several attempts, ostracized, the individual perceives a severing there have been no crosscutting variables that of the bond between him- or herself and the moderate its immediate impact on self-reports others, is frustrated by having no control over of mood or state experiential levels of the four eliciting a response, assumes that he or she has needs. For example, Zadro et al. (2004) showed done something wrong or bad to deserve such that participants react just as negatively to Cy- treatment, and has the unusual experience of berball ostracism if they are told that they are feeling nonexistent. Each of these needs has playing with a computer rather than with other received extensive attention in the psychology people. What would seem to be easily dismissed literature as being fundamental to human moti- as not meaningful is nevertheless painful. vation and well-being (e.g., Baumeister & Eisenberger et al. (2003) had participants play Leary, 1995, for belonging; Seligman, 1975, for Cyberball within an MRI chamber and con- control; Tesser, 1988, for self-esteem; and ducted fMRI scans on them during the game. At Greenberg et al., 1992, for meaningful exis- first, participants were told their computer was tence). Consistent with the literatures on the not yet linked with the other two people’s com- four needs, the model predicts that targets of puters so they were simply to watch the other ostracism will attempt to cope with the thwarted two people play until their computer was at- needs by trying to refortify them. For example, tached. Even with this clearly rational explana- when targets’ belonging is threatened, they may tion for not being included, participants re- attempt to strengthen bonds with other individ- ported feeling badly (having lower levels of the uals or groups. four needs) and displayed activation in the an- Williams and Sommer (1997) first examined terior cingulate cortex (ACC). This activation ostracism with a triadic ball toss paradigm in also occurs when people experience physical which half of the participants were not thrown pain (Panksepp, 2003). When they were told the the ball after the first initial set of throws (for computer was attached and they were included, other studies using this paradigm, see Williams, they felt better and showed less ACC activation. 2001). The impact of ostracism is so immediate Then, when ostracized again (this time the os- and powerful that the social presence of the tracism could be inferred to be intentional), they ostracizers is unnecessary. Subsequently, it was once again showed higher activation of the shown that mood and state need levels would ACC. Clearly, it appears that even the slightest suffer even when individuals were ostracized by hint of ostracism is sufficient to sound a warn- the computer ball toss paradigm Cyberball.1 In ing alarm in humans, possibly because detection this game, participants are led to believe that of ostracism has evolved as a mechanism to they are simply engaging in an exercise that will ward off threats to survival (Panksepp, 2003). increase their mental visualization skills. Con- In all of the previous studies on social exclu- vinced that they are playing a game of Internet sion and ostracism, it is explicitly clear to the ball toss with two others (whom they do not participants that while they are being ostracized, know, see, or expect future meetings with), half the others continue to engage in interaction with are included in the game whereas the other half each other (whether it is ball tossing or convers- are ostracized after having been thrown the ball ing in a chat room; see Wheaton, 2001; Wil- once. Williams, Cheung, and Choi (2000) found that Cyberball-ostracized participants reported liams, 2001; Williams et al., 2000; Williams et worse mood and lower state levels of belonging, control, self-esteem, and meaningful existence. 1 This program can be freely downloaded from http:// They also found an increase in conformity fol- www.psy.mq.edu.au/staff/kip/Announce/Cyberball 294 SMITH AND WILLIAMS al., 2002; Williams & Sommer, 1997). This Two such moderating factors were chosen for type of ostracism is referred to by Williams this study. One involves an individual differ- (1997, 2001) as social ostracism—being ig- ence that may be related to how important os- nored and excluded while in the presence of tracism is to people’s sense of self: individual- others. Another type of ostracism is physical ism–collectivism (Yamaguchi, 1994). Because ostracism, which occurs when the ostracized individuals with collectivistic orientations place individual is physically removed from the oth- more importance on their connection to other ers, as in time-out disciplinary action in schools, people, they may be affected more negatively solitary confinement in prisons, or exile from a by ostracism. country. There has been speculation (Williams, Our second factor involves whether partici- 2001) that physical ostracism may not be as pants are interacting with in-group or out-group painful as social ostracism, because when phys- members. It seems reasonable to expect that ically removed, the target is not continuously being ostracized by in-group members would be reminded of being ignored. Immersed in a new more debilitating than being ostracized by out- environment, the individual can be distracted group members. It should be easier to dismiss from being socially outcast from the original the impact of ostracism by out-group members group by other stimuli. Furthermore, when in- because one could interpret the ostracism to be dividuals are physically ostracized, they are not based on group membership rather than the continuously exposed to the close social inter- result of personal characteristics and shortcom- actions between the others in the group. Never- ings. In this study, participants’ levels of indi- theless, there are no empirical examinations of vidualism and collectivism were assessed prior physical ostracism. to the experimental task, and they were led to In the present study, we used cell phone text believe that during the task they were interact- messaging as a means to achieve a form of ing with two others who were in-group mem- physical ostracism. Text messaging works well bers (i.e., they had similar smoking habits) or in this regard because the individual (a) is phys- out-group members (i.e., they had dissimilar ically separated from the others and (b) cannot smoking habits). see or hear the continuous interactions between Consistent with the literature that suggests the others. Additionally, because cell phones that even mild forms of ostracism are painful, require technical success for transmission, the our primary hypothesis was that despite possi- individual may be allowed to surmise that his or ble factors that could help participants minimize her ostracism is simply the result of technical the impact, participants who were ostracized difficulties that are affecting all members of the through cell phone text messaging would report group. Thus, we believed that using cell phone it as a negative experience (characterized by text messaging as the paradigm would provide a worse mood and lower state levels of needs), more conservative test of the impact of ostra- compared with the responses of participants cism. Individuals ostracized from text messages who were included. It was also predicted that do not have to endure the pain associated with ostracized participants would attempt to pro- being in the presence of others while being voke a response from the others when they were ignored and excluded; neither can they be sure left out of the social interaction (see Williams et that the others are closely interacting with each al., 2002). other. They are also likely aware of the possi- bility that they (and perhaps the others) are Method simply experiencing technical difficulties with their phones, and so they may not perceive Participants and Design deliberate ostracism. This paradigm, then, offers ostracized individuals several means by Forty-three psychology undergraduates from which to dismiss the importance of ostracism Macquarie University (27 women and 16 men; and, consequently, to show little negative im- mean age ϭ 19.76, SD ϭ 2.86) were randomly pact when they are subjected to it. assigned to a 2 (social interaction: ostracism or As in previous investigations, we once again inclusion) ϫ 2 (group type: in-group or out- attempted to discover possible moderating fac- group) between-subjects design. All partici- tors to the immediate experience of ostracism. pants were familiar with text messaging and OSTRACISM BY CELL PHONE TEXT MESSAGES 295 received course credit for their participation. tions, the first a distracter question (“How often Three participants were excluded from the do you exercise?”). analyses because they guessed the purpose of In-group/out-group manipulation. The sec- the study, resulting in 10 people per condition ond question was used to create the in-group/ (N ϭ 40). out-group manipulation. The group was instructed to send this question: “Do you smoke?” This was also one of the personal habits ques- Procedure tions in the preexperimental questionnaire, and so the experimenter told the confederates After the participant and the confederates whether the participant was a smoker or a (one man and one woman) arrived at the labo- nonsmoker prior to the in-group/out-group ratory, they were taken to a room where there manipulation. were three seats arranged in a triangular forma- For the in-group condition, the confederates tion, 60 cm apart. After providing consent, par- would say that they had the same smoking hab- ticipants completed a preexperimental question- its as the participant. For example, if the partic- naire that assessed demographic and personal ipant was a smoker, then the confederates habit information and contained items from would say that they were also smokers. In the Yamaguchi’s (1994) Collectivism Scale. out-group condition, the confederates would say Predictor variable. The items from they did not share the same smoking habit with Yamaguchi’s Collectivism Scale that were in- the participant: If the participant was a smoker, cluded in the questionnaire were those that re- then the confederates would say that they were ceived a factor loading greater than .35 in nonsmokers. After the participant and the con- Kashima et al.’s (1995) factor analysis. Seven federates had sent and replied to the specified items measured collectivism, six measured questions, the group was told to begin their agency, and five measured assertiveness. All SMS conversation. three factors are components of collectivism. Ostracism–inclusion manipulation. Partici- Participants were asked to circle the number pants were then instructed to compose their own that best described them on a 10-point scale text messages. At this point participants were (where 0 ϭ not at all and 9 ϭ very much). After either ostracized or included for 8 min. If the the questionnaire was completed, the SMS in- participant was in the inclusion condition, the teraction commenced. confederates continued to interact with the par- SMS interaction. Participants were told ticipant for the entire 8 min. Alternatively, if the they would be communicating with each other participant was in the ostracism condition, the via SMS text messaging because we were in- confederates responded to none of the partici- terested in the nature of SMS communication. pant’s messages. In the inclusion condition, the They were told they would begin by answering confederates were told to be responsive to the two questions supplied by the experimenter participant’s messages, but the participants (one of which constituted the in-group/out- rarely deviated from the two questions they group manipulation), after which they were to were asked. maintain the interaction until the experimenter Postexperimental questionnaire. At the told them to stop. The two confederates were conclusion of the SMS conversation, partici- taken to another room, leaving the participant pants were asked to complete a questionnaire, alone. On the participant’s table was a Nokia which assessed the four fundamental needs (be- 3310 cell phone, instructions on how to use the longing, control, self-esteem, and meaningful mobile phone, and instructions on what to do in existence), assessed mood (i.e., happy or sad, the experiment. frustrated, angry, anxious), and contained other To check that participants knew how to send questions about their evaluation of the conver- and retrieve SMSs, we asked them to send the sation. The questionnaire consisted of self-de- letter “A” to each of the confederates and to scription items for which they were to respond retrieve the same message that had been sent to between 1 (not at all) and 10 (very much). There them by the confederate. If the participants were 3 questions on each of the four fundamen- could perform this task, the experiment pro- tal needs, 10 items measuring affect, 4 items on ceeded, with the group responding to two ques- the participants’ experience with the phones, 1 296 SMITH AND WILLIAMS question measuring the desire for group har- asking participants to rate themselves from 0 mony, and 2 manipulation checks for ostra- (not at all) to 9 (very much) on these two items: cism–inclusion manipulation (these items have “I was excluded” and “I was ignored.” Both been used in several previous studies, including items were highly correlated (Cronbach’s al- Williams et al., 2000, and Eisenberger et al., pha ϭ .96) and were averaged to form an os- 2003, and are listed verbatim in Zadro et al., tracism index score. Ostracized participants re- 2004). A SIM card reader was used to download ported feeling more excluded and ignored from the text messages into an Excel file so that they the text message communication than included could be content analyzed by the experimenter participants, F(1, 36) ϭ 116.8, p Ͻ .01. and one coder, both blind to condition. Participants were actively engaged in the in- After the experiment, the participants were group/out-group manipulation by sending text thoroughly debriefed as to the purpose of the messages saying whether they were smokers or study and thanked. Participants were then rein- nonsmokers, and in many instances their con- troduced to the confederates and asked to sign a versations continued on this topic. Inspection of postdebriefing form that gave us permission to their text messages seemed to indicate that par- use their data (all participants agreed). ticipants were fully aware of their group membership status (83% explicitly mentioned this). Results For example, one participant wrote that she was ignored “because I responded that I didn’t Initially, for each dependent variable, regres- smoke, and they both did. I think that seemed to sion analyses were conducted with individual- put up an invisible barrier!” ism–collectivism as a continuous variable and group membership and ostracism–inclusion Belonging, Control, Self-Esteem, and dummy coded. In no analysis did individual- Meaningful Existence ism–collectivism yield significant effects, nor did it interact with any other variable. Thus, we The self-reported satisfaction levels for the excluded individualism–collectivism from the four needs were each assessed by three postex- analysis and conducted 2 ϫ 2 analyses of vari- perimental items that asked participants to rate ance. Table 1 contains the means and standard their current feelings from 0 (not at all) to 9 deviations of the dependent variables. (very much). All of the needs were highly reli- able except control. Cronbach’s alpha coeffi- Manipulation Checks cients for each need were as follows: belonging ϭ .84; self-esteem ϭ .83; meaningful exis- The ostracism–inclusion manipulation was tence ϭ .81; control ϭ .60. Consequently, each successful. We assessed this manipulation by set of items was averaged to form a single index

Table 1 Means (and Standard Deviations) of Variables in the Postexperimental Questionnaire Ostracism Inclusion In-group Out-group In-group Out-group Measure (n ϭ 10) (n ϭ 10) (n ϭ 10) (n ϭ 10) Manipulation check 7.05 (2.19) 6.95 (1.75) 1.20 (1.35) 0.75 (1.73) Fundamental needs Belonging 2.83 (1.74) 2.97 (1.84) 6.87 (1.48) 7.23 (1.12) Control 2.70 (1.17) 3.13 (1.18) 5.47 (1.92) 5.53 (1.94) Self-esteem 3.57 (1.53) 3.30 (1.53) 5.37 (1.86) 5.17 (2.01) Meaningful existence 3.53 (2.14) 3.10 (1.74) 6.83 (1.00) 7.77 (1.83) Affective measures Negative mood 3.45 (1.69) 4.73 (1.83) 1.43 (1.96) 1.38 (2.00) Anger 2.25 (1.68) 1.95 (1.73) 1.08 (1.18) 0.68 (1.09) Group harmony 5.40 (1.71) 5.50 (2.59) 3.80 (3.12) 3.10 (3.48) Note. Higher scores indicate that participants reported more of a particular quality. OSTRACISM BY CELL PHONE TEXT MESSAGES 297 for each need. Participants who were ostracized Content Analysis of Postexperimental had a lower sense of belonging, F(1, Comments 36) ϭ 71.55, p Ͻ .01, d ϭ 1.61; control, F(1, 36) ϭ 26.17, p Ͻ .01, d ϭ 1.28; self-esteem, Only a few participants filled out open-ended F(1, 36) ϭ 11.04, p Ͻ .01, d ϭ 0.96; and questions regarding their attributions for the meaningful existence, F(1, 36) ϭ 53.16, p Ͻ course of the social interactions, and so statis- .01, d ϭ 1.51. No main effects or interactions tical analyses were not possible to conduct. Ca- for group membership emerged. sual observation of the comments suggests personal attributions (i.e., ostracism was done to the person because of something the person Affect Measures him- or herself did; e.g., “Maybe because I responded that I didn’t smoke and they both Mood, anger, and desire for group harmony did”) were the most common attribution used to were assessed by the participants’ ratings on explain ostracism (n ϭ 7), followed by techni- postexperimental items, with possible scores cal difficulties (n ϭ 5), and then blaming the ranging from 0 (not at all) to 9 (very much). experimenter (n ϭ 3). Included participants at- We averaged six items measuring mood (such tributed their treatment to the mobile phones that high scores indicated negative moods) to (n ϭ 7) or the experimenter (n ϭ 3). form a single mood index (Cronbach’s ␣ ϭ .89), and the four items on anger were also averaged (Cronbach’s ␣ ϭ .83). Participants Content Analysis of Text Messages answered a single item that asked whether they “wanted things to be more harmonious.” The text messages were rated by the experi- Participants who were ostracized reported menter and a rater, who were both blind to having a more negative mood, F(1, condition. The interrater reliability was high 36) ϭ 20.47, p Ͻ .01, d ϭ 1.17; were angrier, (Cohen’s ␬ ϭ .93). The categories used to code F(1, 36) ϭ 7.13, p Ͻ .01, d ϭ 0.79; and the messages are shown in Table 2. The most wanted more harmony, F(1, 36) ϭ 5.08, p Ͻ notable finding is that ostracized participants .03, d ϭ 0.69, than those who were included. attempted to provoke responses more than did No significant main effects for group mem- included participants. As shown in Table 3, 12 bership or interactions with ostracism were participants in the ostracism condition typed in found for these factors. text that was coded as provocation, whereas

Table 2 Coding Categories and Descriptions for Text Messages Category Description Provocation Any comments that participants may make in an effort to provoke someone to text message them. For example, “Are you people not speaking to me. I am being oppressed” and “How come no 1 is messaging me?!!” Suspicion Expressions of suspicion that the aim of the experiment was different from what participants had been told. For example, “Is this part of the experiment u guys not msging me back??” Boring Expressions that participants were finding the experiment boring. For example, “Blah blah! Cmon, getting bored!” Smoking Any mention the participants make about smoking, beyond what was specified by the experimenter. For example, “u bOth smOke except 4 me!” Bad technology Any mention the participants make specifically about the adequacy of the phones and how they are functioning. For example, “I thought it would make a noise when a msg came in so i thought it Wasnt working.” 298 SMITH AND WILLIAMS

Table 3 Percentages (and Raw Scores) of Categories as a Function of Ostracism–Inclusion Categories Social interaction Boring Smoking Provocation Bad technology Ostracism (n ϭ 20) 18 (3.5) 60 (12.0) 60 (12.0) 18 (3.5) Inclusion (n ϭ 20) 2 (0.5) 50 (10.0) 10 (2.0) 8 (1.5) Note. The raw scores represent the number of individuals in a condition who sent one or more text messages in a category. only 2 were scored as doing this in the inclusion the effect sizes were large (for most, d Ͼ 1.00) condition. and comparable to the effect sizes observed for face-to-face and Cyberball ostracism. Discussion In the text message paradigm, participants were provided with little information to help The primary hypothesis of this study was that them make sense of their situation. But the pain individuals who were ostracized from a text associated with ostracism was not alleviated, message interaction would be adversely af- even for those who did have an explanation for fected by the experience. Although considerable their ostracism (i.e., being the odd person out research has shown ostracism to be painful, all with respect to smoking habits). As has been of the previous studies exposed ostracized par- shown in other studies (Eisenberger et al., 2003; ticipants to the other individuals interacting Zadro et al., 2004), factors that should logically with each other. It is plausible that an individ- diminish the impact of ostracism have no effect ual’s exclusion would be magnified when he or on the immediate experience. she saw that others continued to enjoy inclu- The uncertainty associated with this text mes- sion. We therefore wondered whether the ab- sage paradigm may actually add to the ostra- sence of text messages, without explicit knowl- cism effect, rather than subtracting from it. Sub- edge that the others were communicating with stantial research has demonstrated that individ- each other, would be enough to cause similar uals are uncomfortable with uncertainty and are painful experiences. The results of our study motivated to reduce it (Evans & Over, 1996; demonstrate that ostracism from a text message Guerin, 2001; Jetten, Hogg, & Mullin, 2000; interaction is sufficiently negative to lower self- Luck, Hillyard, Mouloua, & Hawkins, 1996). reported state levels of belonging, control, self- esteem, and meaningful existence and to pro- The ambiguity of the text message situation duce more negative affective ratings. may contribute to the negative effects of ostra- It is remarkable that text message ostracism cism because of the effort required for ostra- was found to be so aversive, considering the cized targets to reduce uncertainty. Research by following factors. First, the participants did not Kramer (1994) suggests that uncertainty can be know the confederates and did not anticipate a culture that breeds paranoid attributions, any further interaction with them. Second, text which could also magnify ostracism’s impact in messages lack nonverbal and paralanguage in- this medium. formation that is important when conveying in- There is evidence within this study, as there terpersonal and potentially threatening attitudes has been in others, that individuals who are (Short, Williams, & Christie, 1976). Third, the ostracized ruminate over possible explanations participants felt the effects of ostracism even for why they are not being included (Williams, though they had no direct evidence that they 2001). Content analysis revealed that people were being deliberately ostracized or that the often blame themselves for the ostracism, even others continued interacting with each other. though they have the option of blaming the Yet, not only were the state levels of the four others or the mobile phones. When people use needs significantly lower for those who were personal attributions to explain their plight, it ostracized by the text-messaging method, but magnifies the effects of ostracism (Fenigstein, OSTRACISM BY CELL PHONE TEXT MESSAGES 299

1979; Nezlek, Kowalski, Leary, Blevins, & logically meaningful to university students than Holgate, 1997). manipulations have been in typical minimal- group paradigms (e.g., being told they prefer Klee or Kandinsky, that they are over- or un- Provocation in Cyberostracism derestimators, being assigned to the heads or tails group as determined by a coin flip). This study also supports the contention that Collective and individualistic individuals did targets of cyberostracism (being ostracized not significantly differ in their reactions to os- through electronic media) display bravado that tracism. This finding suggests that one’s imme- is not typical in face-to-face (or social) ostra- diate reaction to ostracism is not moderated by cism (Williams, 2001; Williams et al., 2002). this particular cultural self-construct. Neverthe- This may be because social ostracism is more less, it is premature to form any definitive con- explicit and awkward than SMS ostracism. Tar- clusions. It is possible that collectivism would gets of social ostracism sit in the same room as interact with ostracism if the reactions of those the confederates, and so the ostracism is obvi- who were extremely and consistently individu- ously deliberate and must be endured in the alistic or collectivistic were compared (Bem & presence of others. However, targets of SMS Allen, 1974). It is also possible that this indi- ostracism (as with e-mail and chat room ostra- vidual difference may reveal its effects on dif- cism) cannot see or hear the confederates, and ferential coping responses after a certain period so they do not have any direct evidence that of time elapses. For example, in a recent exper- they are being deliberately ignored. This may iment in which ostracism was manipulated in lead the target to seek clarification that they are the Cyberball paradigm, Bowland, Richardson, actually being deliberately ostracized, and prov- and Zadro (2003) found that initially, normal ocation may be an attempt at achieving this. and social phobic participants reported similar Cyberostracized participants may also feel low levels of the four needs following ostracism bolder because being socially removed from the compared with those who had been included. others reduces a person’s inhibitions to confront After a delay of 45 min, however, the normal other people (Short et al., 1976). Seeking clar- sample’s self-reported levels were as high as ification and provocation can be used as a those of included participants, but according to means to regain control, which should reduce their self-reports, those participants who were the negative impact of ostracism (see Warbur- high in social phobia had only partially rein- ton, Cairns, & Williams, 2003). stated their need levels.

Group Membership and Individualism– Summary Collectivism Factors This research breaks new ground by extending Individuals who were ostracized by their in- the consistent ﬁndings that ostracism is painful. It group were not more adversely affected by os- appears that ostracism is painful whether it occurs tracism than those who were ostracized by their face-to-face with the relatively rich array of non- out-group. Similar ﬁndings (or a lack thereof) verbal communication or in an Internet ball toss have occurred in other ostracism research (Gon- game (being played by other people or by the salkorale & Williams, 2004; Williams et al., computer, or even when the computer is not at- 2000; Williams et al., 2002). Because in-group/ tached to the others’ computers), where commu- out-group manipulations are typically quite ef- nication of any sort is vastly diminished. This fective in modifying people’s responses to the research extends this list to instances in which the situation (see Hogg & Abrams, 2003), we think ostracized individual cannot see or hear the inclu- our results combined with those of the previous sive behavior of the other group members and studies suggest that reactions to ostracism are so must simply deduce that he or she has been ex- powerful and deep (Brewer, 2003) that it over- cluded. The accumulated evidence points to the whelms factors such as group membership that conclusion that other potentially mitigating fac- require more cognitive processing. Addition- tors, like in-group/out-group membership, indi- ally, our manipulation of in-group/out-group vidual differences, ability to attribute ostracism to (smoking habits) was probably more psycho- nonpunitive causes, and lack of knowledge that 300 SMITH AND WILLIAMS others in the group are communicating with each Greenberg, J., Solomon, S., Pyszczynski, T., Rosen- other, are overwhelmed by merely being ignored blatt, A., Veeder, M., Kirkland, S., & Lyon, D. and excluded. Apparently, the higher level cogni- (1992). Why do people need self-esteem? Con- tive processes necessary to appreciate why these verging evidence that self-esteem serves an anxiety-buffering function. Journal of Personality and factors should diminish the importance of ostra- Social Psychology, 63, 913–922. cism are bypassed when we immediately confront Guerin, B. (2001). Replacing catharsis and uncer- ostracism. Immediate reactions to ostracism, we tainty reduction theories with descriptions of his- argue, are precognitive, because humans have torical and social context. Review of General Psy- evolved to detect even the slightest hint of ostra- chology, 5, 44–61. cism and to experience it negatively, to warn that Hogg, M. A., & Abrams, D. (2003). Intergroup be- something must be done in order to be reincluded havior and social identity. In M. A. Hogg & J. and prevent a threat to survival. Cooper (Eds.), The Sage handbook of social psychology (pp. 407–431). London: Sage. Jetten, J., Hogg, M. A., & Mullin, B. (2000). In- group variability and motivation to reduce subjec- References tive uncertainty. Group Dynamics: Theory, Re- search, and Practice, 4, 184–198. Baumeister, R. F., & Leary, M. R. (1995). The need Kashima, Y., Kim, U., Gelfand, M., Yamaguchi, S., to belong: Desire for interpersonal attachments as Choi, S., Yuki, M., et al. (1995). Culture, gender, a fundamental human motivation. Psychological and self: A perspective from individualism-collec- Bulletin, 117, 497–529. tivism research. Journal of Personality and Social Bem, D. J., & Allen, A. (1974). On predicting some Psychology, 69, 925–937. of the people some of the time: The search for Kramer, R. M. (1994). The sinister attribution error: cross-situational consistencies in behavior. Psy- Paranoid cognition and collective distrust in chological Review, 81, 506–520. groups and organizations. Motivation and Emo- Bowland, C., Richardson, R., & Zadro, L. (2003). Im- tion, 53, 199–229. mediate and delayed effects of ostracism on normals Leary, M. R. (1990). Responses to social exclusion: and social phobes. Unpublished manuscript, Univer- Social anxiety, jealousy, loneliness, depression, sity of New South Wales, Sydney, Australia. and low self-esteem. Journal of Social and Clini- Brewer, M. B. (2003). Implicit and explicit processes in cal Psychology, 9, 221–229. social judgment: Deep and high. In J. P. Forgas, Leary, M. R. (2001). Toward a conceptualization of K. D. Williams, & W. von Hippel (Eds.), Social interpersonal rejection. In M. R. Leary (Ed.), In- judgment: Implicit and explicit processes (pp. 387– terpersonal rejection (pp. 3–20). New York: Ox- 396). New York: Cambridge University Press. ford University Press. Crick, N. R., Casa, J. F., & Nelson, D. A. (2002). Leary, M. R. (in press). Varieties of interpersonal Toward a more comprehensive understanding of rejection. In K. D. Williams, J. P. Forgas, & W. peer maltreatment: Studies of relationship victim- von Hippel (Eds.), The social outcast: Effects of ization. Current Directions in Psychological Sci- ostracism, social exclusion, rejection, and bully- ence, 11, 98–101. ing. New York: Psychology Press. Eisenberger, N. I., Lieberman, M. D., & Williams, K. D. Leary, M. R., & Baumeister, R. F. (2000). The nature (2003, October 10). Does rejection hurt? An fMRI and function of self-esteem: Sociometer theory. In study of social exclusion. Science, 302, 290–292. M. P. Zanna (Ed.), Advances in experimental so- Evans, J., & Over, D. E. (1996). Rationality in the cial psychology (Vol. 32, pp. 1–62). San Diego, selection task: Epistemic utility versus uncertainty CA: Academic Press. reduction. Psychological Review, 103, 356–363. Leary, M. R., Kowalski, R. M., Smith, L., & Phillips, Fenigstein, A. (1979). Self-consciousness, self-atten- S. (2003). Teasing, rejection, and violence: Case tion, and social interaction. Journal of Personality studies of the school shootings. Aggressive Behav- and Social Psychology, 37, 75–86. ior, 29, 202–214. Geser, H. (2002, June). Towards a sociological the- Leary, M. R., Tambor, E. S., Terdal, S. K., & Downs, ory of the mobile phone. Retrieved July 1, 2002, D. L. (1995). Self-esteem as an interpersonal mon- from the Sociology in Switzerland Web site: http:// itor: The sociometer hypothesis. Journal of Per- socio.ch/mobile/t_geser1.htm sonality and Social Psychology, 68, 518–530. Gonsalkorale, K., & Williams, K. D. (2004, April). Luck, S. J., Hillyard, S. A., Mouloua, M., & Hawkins, The KKK won’t let me play: Ostracism by despised H. L. (1996). Mechanisms of visual–spatial attention: outgroups still hurts. Paper presented at the 76th Resource allocation or uncertainty reduction? Jour- Annual Midwestern Psychological Association nal of Experimental Psychology: Human Perception Conference, Chicago. and Performance, 22, 725–737. OSTRACISM BY CELL PHONE TEXT MESSAGES 301

McVeigh, T. (2001, January 8). Bullies prey on half lack of emotion, and self-awareness. Journal of of all children. Retrieved May 2, 2002, from http:// Personality and Social Psychology, 85, 409–423. society.guardian.co.uk/socialcare/news/0,8372, Underhill, W., Theil, S., Chiou, F., Itori, K., Nadeau, 419274,00.html B., Ferro, C., & Vitug, M. (2001, April 26). The Nezlek, J. B., Kowalski, R. M., Leary, M. R., text generation (short message service). Newsweek Blevins, T., & Holgate, S. (1997). Personality International, Article A72519301. Retrieved July 3, moderators of reactions to interpersonal rejection: 2002, from http://80-web5.infotrac.galegroup Depression and trait self-esteem. Personality and .com.simsrad.net.ocs.mq.edu.au/itw/infomark/ Social Psychology Bulletin, 23, 1235–1244. 0/1/1/purlϭrc6_EAIM?sw_aepϭMacquarie Panksepp, J. (2003, October 10). Feeling the pain of Warburton, W., Cairns, D., & Williams, K. D. (2003, social loss. Science, 302, 237–239. April). Effects of ostracism and loss of control on Rautiainen, P., & Kasesniemi, E. (2000). Mobile aggression. Paper presented at the 32nd Annual communication of children and teenagers: Case Meeting of the Society of Australasian Social Psy- Finland 1997–2000. Retrieved July 1, 2002, from chology, Bondi Beach, Australia. the Sociology in Switzerland Web site: http://www. Wheaton, A. (2001). Ostracism and susceptibility to telenor.no/fou/prosjekter/Fremtidens_Brukere/ the overtures of socially deviant groups and indi- seminarer/mobilpresentasjoner/Proceedings%20_ viduals. Unpublished master’s thesis, Macquarie FoU%20notat_.pdf University, Sydney, Australia. Seligman, M. E. P. (1975). Helplessness: On depression, Williams, K. D. (1997). Social ostracism: The causes development, and death. San Francisco: Freeman. and consequences of “the silent treatment.” In R. Short, J., Williams, E., & Christie, B. (1976). The social Kowalski (Ed.), Aversive interpersonal behaviors psychology of telecommunications. London: Wiley. (pp. 133–170). New York: Plenum Publishing. Taylor, A. S., & Harper, R. (2003). The gift of the Williams, K. D. (2001). Ostracism: The power of gab: A design oriented sociology of young peo- silence. New York: Guilford Press. ple’s use of mobiles. Journal of Computer Sup- Williams, K. D., Cheung, C. K. T., & Choi, W. (2000). Cyberostracism: Effects of being ignored ported Cooperative Work (CSCW), 12, 267–296. over the Internet. Journal of Personality and So- Tesser, A. (1988). Toward a self-evaluation model of cial Psychology, 79, 748–762. social behavior. In L. Berkowitz (Ed.), Advances Williams, K. D., Govan, C. L., Croker, V., Tynan, D., in experimental social psychology (Vol. 21, pp. Cruickshank, M., & Lam, A. (2002). Investiga- 181–227). San Diego, CA: Academic Press. tions into differences between social- and cyberos- Text message bullies action call. (2001, March 16). tracism. Group Dynamics, 6, 65–77. Retrieved May 9, 2002, from the BBC News Web Williams, K. D., & Sommer, K. L. (1997). Social site: http://news.bbc.co.uk/hi/english/uk/scotland/ ostracism by one’s coworkers: Does rejection lead newsid_1225000/1225223.stm to loaﬁng or compensation? Personality and Social Text message bullying ‘suited to girls.’ (2001, April Psychology Bulletin, 23, 693–706. 11). Retrieved May 2, 2002, from http://www. Williams, K. D., & Zadro, L. (in press). Ostracism: ϭ ananova.com/news/story/sm_236060.html?menu The indiscriminate early detection system. In K. D. news.technology Williams, J. P. Forgas, & W. von Hippel (Eds.), Twenge, J. M. (in press). When does social rejection The social outcast: Effects of ostracism, social lead to aggression? The inﬂuences of situations, exclusion, rejection, and bullying. New York: Psy- narcissism, emotion, and replenishing connections. chology Press. In K. D. Williams, J. P. Forgas, & W. von Hippel Wockner, C. (2002, July 17). Bash a Christian: Rapists’ (Eds.), The social outcast: Effects of ostracism, hi-tech message of hate. The Daily Telegraph, p. 3. social exclusion, rejection, and bullying. New Yamaguchi, S. (1994). Collectivism among the Jap- York: Psychology Press. anese: A perspective from the self. In U. Kim, Twenge, J. M., Baumeister, R. F., Tice, D. M., & H. C. Triandis, C. Kagitcibasi, S. C. Choi, & G. Stucke, T. S. (2001). If you can’t join them, beat Yoon (Ed.), Individualism and collectivism (pp. them: Effects of social exclusion on aggressive 175–188). Newbury Park, CA: Sage. behavior. Journal of Personality and Social Psy- Zadro, L., Williams, K. D., & Richardson, R. (2004). chology, 81, 1058–1069. How low can you go? Ostracism by a computer Twenge, J. M., & Campbell, W. K. (2003). “Isn’t it lowers belonging, control, self-esteem, and mean- fun to get the respect that we’re going to deserve?” ingful existence. Journal of Experimental Social Narcissism, social rejection, and aggression. Person- Psychology, 40, 560–567. ality and Social Psychology Bulletin, 29, 261–272. Twenge, J. M., Catanese, K. R., & Baumeister, R. F. Received November 20, 2003 (2003). Social exclusion and the deconstructed Revision received June 25, 2004 state: Time perception, meaninglessness, lethargy, Accepted June 25, 2004 Ⅲ