Abstract
Ostracism is an aversive situation that occurs frequently in everyday life; however, few empirical studies have investigated multiple experiences of inclusion or ostracism from the same group. The prior work in this area has also not evaluated the influence of subsequent inclusion and ostracism on identification with the group, perceptions of the group, or group member behaviors. Across three experiments, the current study investigated the impact of subsequent inclusion and ostracism on an individual’s fundamental needs, identification with the group, perceptions of the group, and risk taking to benefit the group. It was hypothesized that participants who were ostracized and then included would show significant increases in group identification, positive group perceptions, and risk taking to benefit the group. Support for these hypotheses was found. Results are discussed regarding the impact of subsequent experiences of ostracism and inclusion on the individual and his or her relationship with the group.
Humans have a need to affiliate, connect, and belong with groups (Baumeister & Leary, 1995). Unfortunately, threats to feelings of belonging lead to both physical and psychological consequences (see Bernstein, 2016). Ostracism is the experience of being ignored and rejected, is one common way that the need to belong is threatened (Williams, 2007b), and leads to feelings of anger and sadness (Williams, 2007a, 2009). Although ostracism is a popular research topic (e.g., Williams, 1997; Wirth & Williams, 2009), few empirical studies have examined what happens when a person is ostracized and then included by the same group; this is the goal of the current research.
Ostracism
Ostracism affects what Williams (2009) refers to as four fundamental needs: belonging, self-esteem, control, and meaningful existence. Research details the widespread impact ostracism has on an individual: for example, virtual ostracism shows the same effects as physical ostracism (Choi, 2000), is like physical pain (Eisenberger, Lieberman, & Williams, 2003), and ostracism is still impactful even when participants are aware that the other Cyberball “players” are a computer (Zadro, Williams, & Richardson, 2004). Ostracism is most commonly studied using the virtual ball-toss game called Cyberball (Williams & Jarvis, 2006), where participants are ostracized by two other group members without any clear reason in an online bass tossing game, and do not interact with these group members again. Although an experience such as Cyberball may happen in the real world, more commonly individuals who are ostracized often end up interacting with the ostracizing group again. For example, college students pledging for a fraternity or sorority may be completely ignored by the current group members at the first event during their initial application or “rush” phase. These prospective members are very likely to interact with the current members again at future events where the current members may either accept or ostracize them again. The specific example of college Greek life highlights situations that happen all the time in human daily life, individuals are willing to interact with an ostracizing group because they are motivated to become members of that group.
Few studies have examined multiple experiences of ostracism. The first was a study conducted by Tang and Richardson (2013), who examined the impact of ostracism on Williams’ (2009) fundamental needs, hurt feelings, affect, and anger. The source of ostracism was also manipulated: Participants were ostracized by the same group, a different group at Game 1 and Game 2, or by humans at Game 1 and computers at Game 2. They found that being included after an experience of ostracism eliminated the negative impact of ostracism, that initial inclusion did not buffer or protect participants from later ostracism, and that being ostracized twice by the same group increased anger and negative mood but only when ostracized by the same group of humans. Inclusion reduced the impact of ostracism if it came after an experience of ostracism but inclusion did not provide any sort of buffering effect to subsequent ostracism.
In another study, Zwolinski (2014) also incorporated a repeated ostracism design. Zwolinski was interested in the consequences and recovery of positive and negative affects following ostracism, specifically if subsequent inclusion led to greater affective and needs recovery in comparison with recovery over time. Participants were either included or ostracized during the first game and then assigned to be either included in a second game or to wait for a period of time. Those who were included were told that one of the other players was the same as Game 1, whereas the other was a newcomer. Results of this study indicated that inclusion after ostracism increased positive affect and decreased negative affect more than just waiting.
Both the Tang and Richardson (2013) and Zwolinski (2014) studies offer initial evidence of the responses and recovery after two subsequent experiences of ostracism. However, these two studies did not consider how participants viewed the group that ostracized or included them or how these perceptions could affect the person’s future behavior. Particularly, these studies did not consider the impact that ostracism or inclusion has on people’s identification with the group or desires to identify with the group in the future.
Group Identification
“Fitting in” is an important part of the human experience. The knowledge that one’s position within a group is secure generally provides a positive source of social identification (Hogg, Fielding, & Darley, 2005) as well as a structured social categorization from which to draw knowledge about one’s self in a group context (e.g., Hogg, 2007). The alternative experience—being ostracized by a group—brings with it identity insecurity (Jetten, Branscombe, & Spears, 2002) and feeling like you do not belong (e.g., Williams, 2007a). Identity insecurity should coincide with the feeling that you do not fit in with the group and that you are not a true member—meaning that you should identify less with the group (e.g., see, Hohman, Gaffney, & Hogg, 2017). Because group identification requires feeling like you belong to the group and that the other members of the group accept you as a member (Turner, Hogg, Oakes, Reicher, & Wetherell, 1987), identification should be thwarted following ostracism (because ostracism signals that the other members do not accept you as a group member). Conversely, if the group includes the individual, identification should increase. The current research will test how group identification changes after multiple rounds of Cyberball.
Following previous findings (Tang & Richardson, 2013; Zwolinski, 2014), if an individual is first ostracized, their thwarted group identification should be eliminated if they are later included. However, enhanced group identification after inclusion should not buffer against the negative consequences from being ostracized. The experience of ostracism appears to be so aversive that prior inclusion does little to mitigate the negative effects (see Tang & Richardson, 2013).
Although group identification seems similar to the fundamental need of belonging (Williams, 2009), belonging is fundamentally different from identification with a group. The measure of the need to belong is centered on the presence or absence of a connection to a group (e.g., “I felt ‘disconnected,’” “I felt rejected,” “I felt like an outsider”). Group identification, however, is a measurement of how much the individual feels like he or she is a true member of the group and fits in with the group (e.g., “How much do you feel like a group member as a whole?” “How well do you feel you fit in as a group member?”), and contains a cognitive, emotional, and behavior component (Tajfel & Turner, 1979). Identification with the group reflects the extent to which the person uses the group as a reference category to explain the current situation and the extent to which the person is persuaded by the group’s norms (Turner et al., 1987). An individual may feel that he or she belongs in a group without really identifying with that group. For example, a university student is aware he or she belongs in the category of students from university X, but that student may not actually feel as though he or she fits in with the group or adheres to the norms of the group. In this way, the feeling of belonging is necessary but not sufficient to foster a sense of group identification. Identification represents a more detailed association with the group where the individual uses the group to define himself or herself and explain the world around him or her (Turner et al., 1987), rather than simply belonging with the group.
Understanding how ostracism affects identification over time is important because ostracism is likely to thwart identification needs and motivate people to fulfill those needs (e.g., by identifying with a group), which could potentially affect group members’ perceptions and behavior. To date, we do not know how ostracism over time affects identification and how that affects group behavior.
Beyond social identity theory, there are other theoretical perspectives that would hypothesize specific patterns of results for successive experiences of ostracism. For instance, research on cognitive dissonance (Festinger, 1957) and the justification of unpleasant initiation (Aronson & Mills, 1959) would provide a rationale for how people would respond to multiple instances of ostracism. According to cognitive dissonance theory, when faced with an unpleasant experience during entry into a group, an individual will feel tension between his or her desire to join the group and the group’s negative response toward him or her. One way to reduce dissonance is to persuade oneself that the group is extremely desirable. Aronson and Mills (1959) tested this hypothesis in an experiment where female participants were randomly assigned to undergo a severe or mild initiation process to gain entry. Participants in the severe initiation condition read aloud several obscene sexual words; in the mild initiation condition, participants read aloud nonobscene sexual words. After being included by the group, severe initiation participants made more positive ratings of the group compared with the mild initiation participants. The authors reasoned that enduring an unpleasant initiation process causes dissonance, so participants exaggerate the positive characteristics of the group to justify their negative experience. This perspective would predict that participants who are first ostracized and then included by the group would show the strongest overall identification, where group members who experience ostracism followed by inclusion exaggerate the positive qualities of the group because entry was more difficult to obtain. An unpleasant initiation perspective would yield similar outcomes to the social identity approach, such that ostracism before inclusion should cause the highest levels of group identification and most positive group evaluations.
A perspective that would lead to different outcomes than the social identity approach and unpleasant initiation perspective is identity fusion. Identity fusion is the experience of being fused in a familial way to a group, where an individual has combined his or her personal and social self (Gómez, Morales, Hart, Vázquez, & Swann, 2011). Research on identity fusion and ostracism found that fused (vs. nonfused) individuals supported risky actions such as fighting and dying for their ingroup after experiences of ostracism (Gómez et al., 2011). Based on these findings, identity fusion research would predict that being included and then ostracized would lead to the strongest identification with the group. Initial inclusion with the group would cause high identification and fusion with the group, and then later ostracism would inspire compensatory behavior for the group to gain a sense of reaffiliation. This perspective contrasts with the social identity approach because it assumes that initial inclusion will buffer against later ostracism and lead to compensatory identification, which previous research does not support (e.g., Tang & Richardson, 2013; Zwolinski, 2014).
Another perspective that focuses on compensatory responses and reaffiliative behaviors is the desire to be prototypical. Research on prototypicality and peripheral group membership is important to consider when studying ostracism and group identification. Prototypicality refers to a set of expectations, norms, behaviors, and attitudes that represent the characteristics that most closely fit the definition of a group (Hogg, Cooper-Shaw, & Holzworth, 1993). A group prototype offers members a description of what the group is, how members of the group think, and a prescription of how to behave. Group members who are less prototypical (peripheral) are highly motivated to perform behaviors that will increase their standing in the group and acceptance by other members (Noel, Wann, & Branscombe, 1995), which often involves adhering to the group norms or matching the group prototype (e.g., Barreto & Ellemers, 2000). In the context of multiple experiences of ostracism, it is possible that individuals who are ostracized will feel peripheral and become motivated to change their peripheral group status by expressing compensatory group identification. Similar to identity fusion research, this perspective would predict that individuals who are included and then ostracized would show the strongest overall group identification.
Although identity fusion and peripheral membership research would suggest that being included and then ostracized would inspire compensatory reactions and, thus, higher identification, we predict that being included and then ostracized would, instead, damage group identification. Because previous research shows that prior inclusion does not buffer against the negative consequences of ostracism (Tang & Richardson, 2013; Zwolinski, 2014), we believe that prior inclusion will similarly fail to protect one’s group identification after being ostracized later. Being included after ostracism, however, eliminates the negative impact of ostracism (e.g., restores feelings of belonging, self-esteem, meaningful existence, control; Tang & Richardson, 2013; Zwolinski, 2014). This rebound may cause group members to exaggerate the positive aspects of the group such as in instances of unpleasant initiation and ingroup bias. For the first time, we will test the social identity effects of multiple rounds of ostracism and inclusion, while considering alternative explanations for group behavior.
The Current Study
This research expands on prior ostracism research (e.g., Tang & Richardson, 2013; Zwolinski, 2014) by investigating how two subsequent experiences of Cyberball impact group identification. Given the motivation to fit in and be part of a group, it is likely that ostracism will affect an individual’s desire to identify with the group, their perceptions of the group, and willingness to act to benefit the group. Across three experiments, we tested the impact of ostracism in a minimal group paradigm (Cyberball) on intragroup processes. The first experiment investigated the changes in identification across two rounds of Cyberball, the second experiment examined how perceptions of the group change, and the third experiment explored how people’s willingness to act to benefit the group changes.
The current research relied on a variant of the minimal group paradigm to test the relationship between multiple instances of ostracism and intragroup processes. Minimal group paradigms avoid confounding factors associated with real-world group membership by categorizing participants into groups based on an arbitrary criterion (Tajfel, 1970; for a review, see Diehl, 1990). Under minimal conditions, participants define themselves in terms of the minimal categorization (e.g., Chaiken & Maheswaran, 1994) causing self-categorization processes such as ingroup favoritism (Diehl, 1990; Tajfel, 1982) and ingroup identification (Mullin & Hogg, 1999). Being categorized into a minimal group can produce ingroup identification even without the presence of an outgroup (see Mullin & Hogg, 1999).
Experiment 1
The first experiment measured participant’s fundamental needs and group identification after two rounds of Cyberball. For this study, there were four conditions: Participants were included in Round 1 and ostracized in Round 2, ostracized in Round 1 and included in Round 2, included in both rounds, or ostracized in both rounds. We hypothesized that being ostracized in the first round but included in the second would lead to recovery of fundamental needs (replicating Tang & Richardson, 2013; Zwolinski, 2014) and an increase in identification (novel to the current study). Conversely, being included in the first round and ostracized in the second round should lead to a significant decrease in fundamental needs fulfillment and identification.
Method
Participants and design
Participants were 141 Mturk workers (n = 84 female, Mage = 30.65 years, SDage = 10.27 years, n = 112 White) who were randomly assigned to be included or ostracized by their group members across two rounds of Cyberball in a repeated-measures design. An initial sample of N = 186 was collected from Mturk, but 36 participants were excluded from analysis because they did not correctly answer two attention check questions (e.g., “If you are reading these questions thoroughly and are paying attention to their content, please select answer 5”). The removal of these 36 participants did not significantly affect the results of this experiment. An additional nine participants agreed to participate but quit before playing the first Cyberball game. This left us with 141 participants for analysis. We predicted a small to medium effect size (f = .25) and desired 95% power to detect our effect in this experiment, and, thus, we aimed to collect data from at least 132 participants. Participants were randomly assigned to be included or ostracized in Round 1 and randomly assigned again in Round 2. This led to four conditions over the two rounds: included–ostracized, ostracized–included, included–included, or ostracized–ostracized.
Materials and procedure
The experiment was hosted on the Qualtrics website and was introduced as a study about mental visualization. The manipulated predictor variable involved being included or ostracized by their Cyberball group members: Participants were randomly assigned to be included or ostracized. Dependent variables were measured after each round of Cyberball: fundamental needs, mood, and identification with one’s Cyberball group.
Cyberball Round 1: Participation manipulation
Participants were told that they would be connected with two other people online to play a ball-tossing game so that they could practice their mental visualization skills. They were told that once they were connected with their group members, to “mentally visualize the entire experience.” Participants were not connected to other players online, but were randomly assigned to be included or ostracized by computerized group members. In the inclusion condition, participants were thrown the ball 33% of the time. In the ostracism condition, participants were thrown the ball once at the beginning and then not thrown the ball again. Each participation condition lasted 2.5 min.
Fundamental needs
After the first round of Cyberball, participants reported how they were feeling during the game on four separate fundamental needs questionnaires (Williams, 1997) and one mood questionnaire. On scales from 1 (not at all) to 5 (extremely), participants reported their levels of belongingness (five items, α = .941; e.g., “I felt rejected”), self-esteem (five items, α = .914; e.g., “I felt good about myself”), meaningful existence (five items, α = .923; e.g., “I felt invisible”), control (five items, α = .823; e.g., “I felt powerful”), and mood (eight items, α = .952; e.g., “good,” “bad”). Items on the mood questionnaire were recoded such that higher scores indicated more positive mood. A composite score was created for each scale by averaging the items together.
Group identification
Next, participants completed a validated measure of group identification that had been adapted from previous research (Hohman & Hogg, 2011). Participants reported how strongly they identified with their Cyberball group on an eight-item scale from 1 (not very much) to 7 (very much; e.g., “How well do you feel you fit in as a group member?” Round 1 α = .953, Round 2 α = .958).
Self-uncertainty, an exploratory variable
After the identification questions, participants completed a self-uncertainty measure (Niedbala & Hohman, in press), for exploratory purposes, even though it was not part of the main analyses in this study. Participants reported how self-uncertain they felt on a six-item scale from 1 (strongly disagree) to 7 (strongly agree; e.g., “I am uncertain about what my future holds.” Round 1 α = .854, Round 2 α = .834), see online supplementary material for the self-uncertainty analysis and discussion.
Cyberball Round 2: Participation manipulation
After completing the dependent measure scales, participants were told that they would play a second round of Cyberball with the same group as in the first round. Participants were randomly assigned to be either included or ostracized in the second round. After the second round ended, participants completed the same dependent measure scales as before, reported demographic information, and were debriefed.
Results
Correlations between identification and fundamental needs scales are presented in Table 1. A factor analysis was first conducted to determine whether the measures of belonging and identification in this study could be considered distinct constructs. Results suggest that belonging and identification scales are related but separate factors. Additional details are provided in the online supplemental material.
Correlations for Fundamental Needs and Identification Across Time 1 and Time 2.
Note. All correlations significant at p < .01.
Main analyses
One-way repeated-measures ANOVAs were conducted to compare condition on fundamental needs and identification at Round 1 and Round 2. There was a significant effect of condition × time on belonging, Wilks’ lambda (Λ) = .268, F(3, 137) = 124.62, p < .001, ηp2 = .732; self-esteem, Λ = .512, F(3, 137) = 43.57, p < .001, ηp2 = .488; meaningful existence, Λ = .444, F(3, 137) = 52.22, p < .001, ηp2 = .556; control, Λ = .431, F(3, 137) = 60.30, p < .001, ηp2 = .569; mood, Λ = .476, F(3, 137) = 50.18, p < .001, ηp2 = .524; and identification, Λ = .434, F(3, 137) = 59.65, p < .001, ηp2 = .566. These analyses were also conducted while covarying the other fundamental needs and mood (e.g., for belonging, covarying Round 2 self-esteem, meaningful existence, control, and mood) to determine whether the effects were due to the simultaneous changes in these variables as well. Covarying these variables had no significant impact on the results.
For each dependent variable, simple effects of time were used in evaluating changes from Round 1 to Round 2. Means and standard deviations are presented in Table 2. From Round 1 to Round 2, participants in the ostracism–inclusion condition showed significant increases in belonging, Ms = 2.84 versus 5.60, Λ = .324, F(3, 137) = 285.61, p < .001, ηp2 = .676; self-esteem, Ms = 2.41 versus 3.96, Λ = .571, F(3, 137) = 103.01, p < .001, ηp2 = .429; meaningful existence, Ms = 2.27 versus 4.30, Λ = .531, F(3, 137) = 120.90, p < .001, ηp2 = .469; control, Ms = 1.65 versus 3.62, Λ = .475, F(3, 137) = 151.18, p < .001, ηp2 = .525; mood, Ms = 2.41 versus 4.06, Λ = .559, F(3, 137) = 107.97, p < .001, ηp2 = .441; and group identification, Ms = 2.11 versus 4.89, Λ = .497, F(3, 137) = 138.48, p < .001, ηp2 = .503; Figure 1.
Experiment 1 Means and Standard Deviations on Post Cyberball Dependent Measures.
Note. Averages were calculated across all items for each scale. Belonging, self-esteem, meaningful existence, control, and mood were measured using a 5-point scale. Identification was measured using a 7-point scale.
Differing letters indicate significant differences on identification between conditions at p < .05.
All changes from Round 1 to Round 2 were significant decreases at p < .001.
All changes from Round 1 to Round 2 were significant increases at p < .001.
Experiment 1 group identification at Round 1 and Round 2 as a function of condition.
Conversely, participants in the inclusion–ostracism showed significant decreases in belonging, Ms = 4.99 versus 3.04, Λ = .564, F(3, 137) = 105.89, p < .001, ηp2 = .436; self-esteem, Ms = 3.43 versus 2.35, Λ = .785, F(3, 137) = 37.44, p < .001, ηp2 = .215; meaningful existence, Ms = 3.68 versus 2.05, Λ = .702, F(3, 137) = 58.24, p < .001, ηp2 = .298; control, Ms = 2.88 versus 1.74, Λ = .783, F(3, 137) = 38.00, p < .001, ηp2 = .217; mood, Ms = 3.84 versus 2.59, Λ = .746, F(3, 137) = 46.72, p < .001, ηp2 = .254; and group identification, Ms = 4.16 versus 2.29, Λ = .747, F(3, 137) = 46.34, p < .001, ηp2 = .253.
No significant differences in any dependent measures between Round 1 and Round 2 were found for the included–included participants or for ostracized–ostracized participants. An additional one-way ANOVA was conducted to evaluate conditional differences in identification with the group after Round 2. Results indicated that there were significant differences between conditions, F(3, 137) = 59.563, p < .001, ηp2 = .359, with Tukey post hoc comparisons showing the ostracism–inclusion condition being significantly higher in identification (M = 4.89, SD = 1.71, p = .020) than the inclusion–inclusion condition (M = 3.90, SD = 1.58), which was significantly higher (p < .001) than the inclusion–ostracism (M = 2.29, SD = 1.46) and the ostracism–ostracism (M = 2.15, SD = 1.24) conditions, which did not significantly differ from one another (p = .987).
As predicted, results from Experiment 1 demonstrated that being ostracized first and then later included by the same group significantly increased group identification. Not only do these participants show full recovery of fundamental needs after inclusion (replicating previous research, e.g., Tang & Richardson, 2013; Zwolinski, 2014) but they also identified more strongly with the group compared with any other condition at any time. Predicted decreases in fundamental needs, mood, and identification were also found for included–ostracized participants. Results supported the social identity perspective and unpleasant initiation perspective (see online supplemental material for a more detailed discussion about Experiment 1).
Experiment 2
Results from Experiment 1 demonstrate that different ostracism/inclusion experiences affect group identification. Because ostracism influences how one feels about the group, it is likely that perceptions of the group fluctuate depending on one’s interactions with the group. In Experiment 2, we explore how perceptions of warmth, competence, status, and entitativity change in response to multiple rounds of Cyberball.
Warmth and competence are considered two universal dimensions of group perception that have made significant contributions to social cognition research (e.g., Fiske, Cuddy, & Glick, 2007; Fiske, Cuddy, Glick, & Xu, 2002). Perceived warmth relates to the group’s intent—friendliness, helpfulness, and trustworthiness, whereas perceived competence relates to the group’s ability—skillfulness, efficacy, and intelligence (Fiske et al., 2007). Groups considered high in both warmth and competence are seen positively because they are perceived to have trustworthy motives and high ability (Fiske et al., 2002). Because group identification corresponds to more favorable impressions of the ingroup because of ingroup bias (Tajfel & Turner, 1979), we would expect perceptions of warmth and competence to fluctuate alongside group identification.
One variable related to perceived competence is group status (Fiske et al., 2002), which reflects the exclusivity or prestige of the group (Brewer, Manzi, & Shaw, 1993; Major et al., 2002). Compared with groups that include individuals all the time, groups that first ostracize an individual before including them may be perceived to be high in exclusivity and, therefore, high in status. Because of this, we expect the ostracism–inclusion condition to report the highest perceptions of status out of any condition.
Social identity theory literature also identifies perceived group entitativity as a characteristic strongly related to group identification (Castano, Yzerbyt, & Bourguignon, 2003). Entitativity refers to the following group properties: clarity of group boundaries, internal homogeneity, social interaction within the group, structure, common goals, and common fate (Campbell, 1958; Hamilton & Sherman, 1996). Highly entitative groups have a clear structure and set of rules associated with group membership, whereas groups that are low in entitativity are perceived as lacking structure and boundaries for membership (Hamilton & Sherman, 1996). Past research suggests that increases in identification are related to increases in entitativity (Castano et al., 2003). Gaertner and Schopler (1998) found that group members who had closer interaction with one another perceived the group to be more entitative because it was easier to conceptualize their group as an entity. Because inclusive groups will naturally interact more closely with their members, we would expect perceived entitativity to increase after inclusion and to decrease after ostracism.
Research from Knowles and Gardner (2008) also contributes to the idea that rejection can alter the perceptions one has of his or her group, although in the opposite direction. In a series of experiments, these authors found that rejection caused participants to have more active group constructs, to judge their group as more entitative, and that these active group constructs and entitativity were associated with better mood and higher self-esteem. Rejection then made the individuals think more about their group memberships and perceive their groups as higher in entitativity. Based on previous research (Castano et al., 2003; Gaertner & Schopler, 1998) and the results from Experiment 1, however, we predict that judgments of entitativity will fluctuate in the same direction as group identification.
Overall, we hypothesize that inclusion will cause perceptions of high warmth, high competence, high status, and high entitativity, which will correspond with increases in group identification. Second, because Experiment 1 showed the strongest group identification for the ostracized–included group after Round 2, we predict this condition will report the highest perceived warmth, competence, status, and entitativity out of any condition after Round 2. Our hypotheses for Experiment 2 are inspired by the prior research on group identification as well as justification of unpleasant initiation (Aronson & Mills, 1959; Festinger, 1957). After an initial experience of ostracism, participants are likely to experience dissonance between their desire to join the group and the negative response from the group. They will be low in identification at this point from feeling rejected and upset about being denied entry into the group. When they are later included, however, they should rebound with the strongest group identification and exaggerate the positive qualities of the group to justify their earlier experience of ostracism.
Method
Participants and design
Participants were 324 undergraduate students who participated in the study for partial course credit (n = 211 female, Mage = 19.07 years, SDage = 1.88 years, n = 229 White). We predicted a small to medium effect size (f = .25) and increased our desired power to 99% to detect our effect in this experiment, and, thus, we aimed to collect and collected data from 324 participants. These participants were randomly assigned to the same four conditions as in Experiment 1.
Measures and procedure
The experiment was hosted on the Qualtrics website and participants completed the study within a laboratory. All other procedures directly replicated that of Experiment 1. Dependent variables were measured after each round of Cyberball: fundamental needs, mood, identification with one’s Cyberball group, and the perceived status, entitativity, warmth, and competence of his or her group. After the second Cyberball round ended, participants completed the same dependent measure scales as after Round 1, reported demographic information, and were debriefed.
Status
Participants completed a measure of their perceived status (created for this study) of the Cyberball group that they had just interacted with. A total of four items (α = .965; e.g., “How much status do you feel your Cyberball group has?”) were used and participants indicated their responses on a scale from 1 (low status; not prominent) to 7 (high status; very prominent).
Entitativity
The measure of entitativity for this study involved a total of eight items (α = .961; e.g., “How cohesive is the group?” Spencer-Rodgers, Hamilton, & Sherman, 2007) and was measured on a scale from 1 (not at all) to 9 (extremely).
Competence
Participants evaluated the competence of the group based on five characteristics (α = .844; e.g., “competent, intelligent”; Fiske et al., 2002) on a scale from 1 (not at all) to 5 (extremely).
Warmth
Participants evaluated the warmth of the group based on four characteristics (α = .978; e.g., “sincere, good natured”; Fiske et al., 2002) on a scale from 1 (not at all) to 5 (extremely).
Results
One-way repeated-measures ANOVAs were conducted to compare condition on fundamental needs, mood, identification, perceived status, entitativity, warmth, and competence at Round 1 and Round 2. There was a significant effect of condition × time on belonging, Λ = .279, F(3, 320) = 275.99, p < .001, ηp2 = .721; self-esteem, Λ = .422, F(3, 320) = 146.25, p < .001, ηp2 = .578; meaningful existence, Λ = .324, F(3, 320) = 222.83, p < .001, ηp2 = .676; control, Λ = .389, F(3, 320) = 167.52, p < .001, ηp2 = .611; mood, Λ = .356, F(3, 320) = 193.23, p < .001, ηp2 = .644; identification, Λ = .365, F(3, 320) = 185.24, p < .001, ηp2 = .635; status, Λ= .610, F(3, 320) = 68.15, p < .001, ηp2 = .390; entitativity, Λ = .478, F(3, 320) = 116.56, p < .001, ηp2 = .522; warmth, Λ = .392, F(3, 320) = 165.28, p < .001, ηp2 = .608; competence, Λ = .362, F(3, 320) = 38.64, p < .001, ηp2 = .266. These analyses were also conducted while covarying the other fundamental needs and mood (e.g., for belonging, covarying Round 2 self-esteem, meaningful existence, control, and mood) to determine whether the effects were due to the simultaneous changes in these variables as well. Covarying these variables had no significant impact on the results.
For each dependent variable, simple effects of time were used in evaluating changes from Round 1 to Round 2. Participants in the ostracism–inclusion condition showed significant increases in all fundamental needs, mood, and identification from Round 1 to Round 2. Participants in the inclusion–ostracism condition showed significant decreases in all fundamental needs, mood, and identification from Round 1 to Round 2. Novel to Experiment 2, participants in the inclusion–inclusion condition showed significant increases in belonging, meaningful existence, control, and identification from Round 1 to Round 2, as well as participants in the ostracism–ostracism condition showing a significant increase in control from Round 1 to Round 2. These results are presented in Table 3.
Experiment 2 Simple Effects.
Note. All analyses had degrees of freedom within = 3 and degrees of freedom between = 320 (df = 3, 320). Averages were calculated across all items for each scale. Belonging, self-esteem, meaningful existence, control, mood, warmth, and competence were measured using a 5-point scale. Identification and status were measured using a 7-point scale. Entitativity was measured using a 9-point scale.
As predicted, from Round 1 to Round 2, participants in the ostracism–inclusion condition showed significant increases in perceived status, Ms = 2.55 versus 4.97, Λ = .645, F(1, 320) = 176.34, p < .001, ηp2 = .355; perceived entitativity, Ms = 3.48 versus 6.80, Λ = .562, F(1, 320) = 249.26, p < .001, ηp2 = .438; perceived warmth, Ms = 1.72 versus 3.87, Λ = .502, F(1, 320) = 317.37, p < .001, ηp2 = .498; and perceived competence, Ms = 1.72 versus 3.87, Λ = .502, F(1, 320) = 317.37, p < .001, ηp2 = .498.
Also in line with predictions, participants in the included–ostracism condition showed significant decreases in perceived status, Ms = 3.36 versus 2.21, Λ = .889, F(1, 320) = 39.90, p < .001, ηp2 = .111; perceived entitativity, Ms = 5.11 versus 2.95, Λ = .751, F(1, 320) = 106.30, p < .001, ηp2 = .249; perceived warmth, Ms = 3.06 versus 1.44, Λ = .637, F(1, 320) = 182.08, p < .001, ηp2 = .363; and perceived competence, Ms = 3.13 versus 2.27, Λ = .851, F(1, 320) = 55.99, p < .001, ηp2 = .149.
Analyses for the inclusion–inclusion and ostracism–ostracism conditions are found in the online supplemental material. To test conditional differences in group perceptions and identification after Round 2, a multivariate ANOVA was conducted. To be more stringent with our test because of possible issues with heterogeneity of covariance matrices, we used Pillai’s Trace for our significance test. The omnibus test was significant, Pillai’s Trace = .709, F(15, 954) = 19.67, p < .001, ηp2 = .236. Tukey post hoc comparisons indicated that for identification, status, entitativity, and warmth, the ostracism–inclusion condition was significantly higher (p < .001) than the inclusion–inclusion condition, which was significantly higher (p < .001) than both the inclusion–ostracism and ostracism–ostracism conditions, which did not differ from one another (p > .05). See Table 3 for comparisons.
Findings from Experiment 2 confirm that successive experiences of inclusion and ostracism significantly influence group perceptions. Ostracism–inclusion participants showed significant increases in positive group perceptions, whereas inclusion–ostracism participants showed significant decreases. These results extend the findings of Experiment 1 and support the social identity and unpleasant initiation perspectives (see online supplemental material for a more detailed discussion about Experiment 2).
Experiment 3
Experiment 2 showed that experiences of ostracism and inclusion affect group perceptions. Experiment 3 will test how ostracism and inclusion change an individual’s behavior. Specifically, the current study will test whether changes in group identification drive individuals to act in risky ways that benefit the group.
There has been considerable research conducted on risk taking in response to rejection. Research in this area illustrates that exclusion increases risk taking (e.g., Mead, Baumeister, Stillman, Rawn, & Vohs, 2011). Examples include spending money on unappealing food and being willing to try drugs (Mead et al., 2011), behaviors that are performed by the excluded individuals to reaffiliate with a group. Group members fused with their group (vs. nonfused) also endorse risky behaviors such as fighting or dying for the ingroup after experiences of ostracism (Gómez et al., 2011). From the identity fusion perspective, taking risky behaviors for the group will only increase after ostracism because excluded group members will react with compensatory behavior to reaffiliate with the group. In our mixed ostracism–inclusion design that would mean individuals would be most risky if they are included first and then ostracized.
However, our hypothesis is generated based on the research on group identification. Because being ostracized and then included increases group identification, we reason that these group members will be the most willing to engage in group-related behavior, even if that behavior is risky. Past research demonstrates that group members identify with groups even if they are extreme or radical (Hogg, 2007; Hogg & Adelman, 2013) or if they are risky to one’s health, safety, and well-being (Hogg, Siegel, & Hohman, 2011). Highly identified group members show ingroup favoritism (Perreault & Bourhis, 1999) and want to see their group in a positive light (Tajfel & Turner, 1979), so they should be willing to engage in risky behaviors if the behaviors benefit their group. Based on the findings from group identification literature, we predict that those who are first ostracized and then included will be the most willing to take risks to benefit their group. Conversely, those who are first included and then ostracized should be less willing to take risks to benefit the group.
The impact of ostracism on risk taking for the group will also be further analyzed with a mediation analysis to determine whether identification is the driving factor of changes in risk taking. Fundamental needs and mood will be covaried to evaluate whether changes in identity uniquely predict changes in risk. It is expected that as identification increases, so too will risk taking, and as identification decreases, risk taking will also decrease. Analyses will evaluate the indirect effects of condition on risk taking through identification and specifically focus on the ostracized–included and included–ostracized groups.
Method
Participants and design
Participants were 194 undergraduate students who participated in the study for partial course credit (n = 121 female, Mage = 18.86 years, SDage = 1.53 years, n = 113 White). We predicted a small to medium effect size (f = .25) and desired 95% power to detect our effect in this experiment, and, thus, we aimed to collect data from 200 participants (due to a computer issue with the Cyberball program, data from six participants were not recorded, which left us with 194 participants). These participants were randomly assigned to the same four conditions as in Experiment 1.
Materials and procedure
Materials and procedures for Experiment 3 replicated Experiment 1, with the addition of the risk-taking measure, the Balloon Analog Risk Task (BART; Lejuez et al., 2002). The BART is a validated measure of risk taking that has been used as an alternative to self-report personality measures of risk taking and temporary risk preference (e.g., Benjamin & Robbins, 2007). The BART was originally developed as a measure of individual risk taking but has since been used to measure differences in risk taking for a friend or for a local charity (Benjamin & Robbins, 2007). Because previous research shows that the BART can successfully measure risk taking on behalf of others, we decided to include it as a measure of risk taking for the benefit of one’s group. After participants completed the first round of Cyberball and dependent measures, participants were told that they would play a balloon challenge game to earn money for their Cyberball group. They were told that the online Cyberball groups currently playing on campus can earn money for their team, and that the winning groups are listed on a Leader Board for everyone to see. Participants were then shown a fake Leader Board with other groups listed (e.g., “Group 32,” “Group 19”) alongside the amount of money earned for their own groups. Participants were instructed that they would play two rounds of the balloon challenge game to earn money for their own group.
Participants then completed the BART. In the BART, participants pump up a virtual balloon while receiving money for each pump. Participants can push a “Collect $$$” button at any point to collect the money from that balloon. However, the balloon may burst on any given pump, and any money not “collected” before the burst is lost. Participants had 30 balloons and were informed of all game instructions. In the BART, more pumps per balloon indicate higher risk taking. For this study, participants were aware that the money they earned would benefit their whole group, so this measure was intended to reflect risk taking for one’s group. After the first round of the BART, participants were told that their team did not earn enough money to be on the Leader Board, but that they were not far behind. They were then reminded that they had one more round of the balloon challenge game to play later. Participants then played a second round of Cyberball with their same group, completed the same dependent measures as before, completed a second round of the BART, and finally were debriefed.
Results
One-way repeated-measures ANOVAs were conducted to compare the ostracism conditions on fundamental needs, mood, identification, and risk taking at Round 1 and Round 2. There was a significant effect of condition × time on belonging, Λ = .422, F(3, 190) = 86.64, p < .001, ηp2 = .578; self-esteem, Λ = .555, F(3, 190) = 50.75, p < .001, ηp2 = .445; meaningful existence, Λ = .406, F(3, 190) = 92.526, p < .001, ηp2 = .594; control, Λ = .525, F(3, 190) = 57.33, p < .001, ηp2 = .475; mood, Λ = .473, F(3, 190) = 70.67, p < .001, ηp2 = .527; and identification, Λ = .438, F(3, 190) = 81.27, p < .001, ηp2 = .562. These analyses were also conducted while covarying the other fundamental needs and mood (e.g., for belonging, covarying Round 2 self-esteem, meaningful existence, control, and mood) to determine whether the effects were due to the simultaneous changes in these variables as well. Covarying these variables had no significant impact on the results.
Risk was quantified in this study using participant’s adjusted average pump count, or the average number of pumps per balloon when controlling for the number of exploded balloons. It is recommended (Lejuez, Aklin, Zvolensky, & Pedulla, 2003) that adjusted average pump count across all 30 balloon trails be used as the measure of risk for the BART, as higher adjusted average pump count corresponds to higher risk taking. In a slight modification of the original recommendation, 1 analyses were conducted using the adjusted average pump count for the middle third balloons (balloons 11-20), we modified the original measure to control for participants getting acquainted to the task during the first 10 balloons or for participants potentially becoming fatigued during the final 10 balloons (see online supplemental material). The overall effect of condition on risk was not significant, Λ = .969, F(3, 190) = 2.02, p = .113, ηp2 = .031, but planned pairwise comparisons revealed the hypothesized significant effects between conditions.
For each dependent variable, simple effects of time were used in evaluating changes from Round 1 to Round 2. Results replicated those of Experiments 1 and 2, see Table 4.
Experiment 3 Simple Effects.
Note. All analyses had degrees of freedom within = 3 and degrees of freedom between = 190 (df = 3, 190). Averages were calculated across all items for each scale. Belonging, self-esteem, meaningful existence, control, and mood were measured using a 5-point scale. Identification was measured using a 7-point scale. Risk was measured by the average adjusted pumps for balloons 11 to 20.
For risk taking, planned comparisons (Rosenthal & Rosnow, 1984) for time were used in evaluating changes in risk from Round 1 to Round 2 (Figure 2). As predicted, participants who were first ostracized and then included were significantly riskier from Round 1 to Round 2, Ms = 27.02 versus 32.59, Λ = .968, F(1, 190) = 6.30, p = .013, ηp2 = .032. Contrary to predictions, participants in the inclusion–ostracism condition did not differ in risk taking between rounds, Ms = 33.39 versus 36.34, Λ = .989, F(1, 190) = 2.16, p = .143, ηp2 = .011. There were no significant differences for included–included participants, Ms = 30.45 versus 29.40, Λ = .999, F(1, 190) = .252, p = .617, ηp2 = .001, but ostracism–ostracism participants showed a significant increase in risk taking, Ms = 30.72 versus 35.95, Λ = .972, F(1, 190) = 5.44, p = .021, ηp2 = .028.
Experiment 3 risk taking for the group at Round 1 and Round 2 as a function of condition.
To confirm that group identification was driving increases in risk taking for ostracized–included individuals, a mediation analysis was conducted using Model 4 of the PROCESS macro (Hayes, 2012) within SPSS Version 22. Our planned mediation analysis confirmed that from Round 1 to Round 2, ostracized–included individuals were significantly more willing to take risks for their group, and stronger group identification explained this increase. Because the included–ostracized group did not decrease in risk taking as we predicted, we could not test the mediating effect of group identification on their willingness to take risks for the group. Also, identification did not mediate the relationship between ostracism–ostracism and risk taking. See online supplemental material for additional details.
Experiment 3 demonstrated that group behavior changes depending on different combinations of ostracism and inclusion over time. Ostracized–included participants were significantly riskier than other participants and this increase was explained by group identification. Results from Experiment 3 support both the social identity and unpleasant initiation perspectives (see online supplemental material for a more detailed discussion).
General Discussion
The current research examined how multiple experiences of ostracism/inclusion affect a group member. Prior ostracism research focuses on singular instances of being ignored and excluded by a group (e.g., Williams & Govan, 2005; Wirth & Williams, 2009). However, it is not difficult to imagine situations where one may be ostracized by the same group on multiple occasions. The experience of being ostracized and then included by a group has only been studied a few times previously (Tang & Richardson, 2013; Zwolinski, 2014), but has not investigated how group identification, group perceptions, and risk taking are affected. Across three experiments, we showed that individuals who are ostracized and then included significantly increase in their identification with, positive perceptions of, and the amount of risk they are willing to take to benefit the group.
Experiment 1 demonstrated that identification significantly increased for participants who were ostracized and then included and significantly decreased for included and then ostracized participants. Experiment 1 also replicated prior work (e.g., Tang & Richardson, 2013; Zwolinski, 2014) showing that inclusion after ostracism led to recovery of an individual’s thwarted fundamental needs and mood. Experiment 2 demonstrated that ostracism and then inclusion also affects perceptions of the group. Specifically, being ostracized and then included led participants to perceive the group as being higher in status, entitativity, warmth, and competence. Finally, Experiment 3 measured behavioral risk taking for the group. Participants who were ostracized and then included took significantly more risks for their group, and this increase was explained by identification. The findings from all three experiments align with prior work on group identification as well as the justification of unpleasant initiation (Aronson & Mills, 1959; Festinger, 1957), advancing our understanding of what conditions most strongly motivate group perceptions and behavior. The findings of this research suggest that subsequent experiences of ostracism affect far more than an individual’s fundamental needs; they also affect how much the individual embraces the group as part of his or her self-concept, how positively he or she perceives the group, and how willing he or she is to perform group-serving behaviors.
The current research extends prior findings by showing that ostracism also damages a group member’s identification with the group. This is a valuable finding, because identification is an important aspect of how much an individual takes on the group’s norms as part of his or her self-concept (Turner et al., 1987). After ostracism, identification was quite low, but increased to the highest levels of any other condition upon inclusion. Interestingly, the current study found these changes in identity within a minimal groups paradigm (Experiments 1 and 2) and with a minimally labeled group (Experiment 3). It could be hypothesized that these findings may be even more pronounced if the identities were real social groups. Future work involving identities such as workplace and gender identities (Zimmerman, Carter-Sowell, & Xu, 2016) or racial identities (Carter-Sowell & Zimmerman, 2015) should investigate how multiple experiences of ostracism from an existing social identity affects individual’s identification.
It is noteworthy that people change their perceptions of the group even when no information about the group is provided. It appears that even when very little information is offered, participants will use whatever information they can obtain to make decisions about the group. Groups that withhold inclusion from individuals by ostracizing them first improve the perception of their group. Although the current study found increases in positive group perceptions for minimal groups where no group information was known, it is possible that this tactic could be used to improve the perceptions of groups that are known to be lower in characteristics such as status and entitativity. Future studies should manipulate group characteristics such as status to determine whether participants who experience ostracism and then inclusion from a low-status group would overestimate the prestige of that group.
Groups may also benefit from the findings of Experiment 3, where the experience affected ostracized–included members’ behaviors. Our results suggest that ostracism and inclusion could potentially be harmful for group members who are asked to perform dangerous and risky behaviors, such as pranking a rival fraternity. As the mediation analysis in Experiment 3 shows, changes in risk taking for the group are not due to changes in their needs or mood, but specifically connected to identification with the group. The member who now highly identifies with his or her group will want to display his or her strong identification by helping the group succeed in their goals, and the current research suggests that this will happen even if the behavior involves taking risks. Future research should investigate how risky group members are willing to act for their group following ostracism. For instance, does the boost in group identification drive group members to take risks that would be physically or financially harmful to themselves (e.g., using drugs, drinking underage, gambling, stealing from a rival group)?
In addition to those who were ostracized and then included by the group, participants who were included and then ostracized were also significantly affected. Included–ostracized participants showed significant declines in fundamental needs, mood, group identification, perceived entitativity, status, warmth, and competence. The results for participants in this condition illustrate how quickly an individual can go from feeling good about himself or herself and his or her group to feeling very bad and wanting to distance himself or herself from the group.
The ostracism–inclusion condition in this study could be interpreted as an unpleasant initiation process. From this perspective, our findings parallel research on cognitive dissonance (Festinger, 1957) and the justification of unpleasant initiation (Aronson & Mills, 1959). Participants’ initial experience of ostracism was quite unpleasant, which then changed their perceptions of the group (Experiment 2). When these participants were later included in Round 2, they exaggerated the positive aspects of the group possibly to justify their earlier experience of ostracism. Follow-up studies should consider cognitive dissonance as one mechanism enhancing group perceptions. The amount of dissonance felt in response to ostracism should be directly related to the strength of identification with the group, meaning that the more one believes he or she fits in with the group, the more tension he or she will feel in response to rejection. Ostracism, in turn, should cause highly identifying individuals to exaggerate the positive qualities of the group even more.
The current work also adds to the literature on the rejection–identification hypothesis. The rejection–identification model (RIM, Branscombe, Schmitt, & Harvey, 1999) focuses on the impact of rejection by an outgroup on ingroup identification, hypothesizing that outgroup rejection leads to minority group members increasing in their identification with their ingroup. Considering the results of the current study, it would seem that both ingroup rejection and outgroup rejection increase ingroup identification. Future studies should investigate both outgroup rejection and ingroup rejection simultaneously to determine the effectiveness of both in fostering stronger ingroup identification.
Results of the current study were mixed regarding the perspective of identity fusion. Participants identified less with the group after ostracism, even when included in Round 1 and then ostracized in Round 2. Research on identity fusion (e.g., Gómez et al., 2011) would hypothesize that individuals in the included–ostracized group would increase identification with the group at Round 2 due to their initial fusion at Round 1. Results did not support the identity fusion hypothesis. However, identity fusion would predict that individuals who are fused with their group would endorse risk behaviors for the group. This was supported by the current data as the increase in risk taking for the group was shown to increase through increases in identification with the group. The discrepancy between results could be due to the use of minimal groups. It is possible that individuals do not become as fused with minimal groups and that the use of preexisting groups would yield different results. Future research should examine the impact of preexisting groups versus minimal groups in situations of repeated ostracism.
The results of this study do not support the hypotheses outlined by prototypicality and peripheral group membership on multiple experiences of inclusion and ostracism. Although the current study did not measure perceptions of prototypicality of participants, ostracized individuals did not increase in their identification. Increasing identification could have been a way to attempt to become more prototypical of the group, which past research demonstrates (Noel et al., 1995). However, in the current research, ostracism led to decreases in identification with the group even after inclusion in Round 1. Future research should consider measuring perceived prototypicality of participants to the group to determine the role of prototypicality in the situation of simultaneous experiences of ostracism.
Limitations and Future Directions
The dependent variables used in this study should be highlighted as a necessary area for future investigation (specifically, fundamental needs and group identification). As shown in Table 1, these measures are strongly correlated with one another. Prior ostracism research has used the four needs as independent constructs (e.g., Williams, 2009) but more recent research has suggested that the four needs may fit better into a two-factor structure (Gerber, Chang, & Reimel, 2017). Despite our exploratory factor analysis showing belonging and identification loading on two separate factors, the correlations from this study and study results (e.g., similar increases from Round 1 to Round 2 for ostracized–included participants) suggest that, perhaps, belonging and group identification (as well as other fundamental needs such as control) have significant overlap as constructs, which the current study cannot disentangle. Future research should investigate scenarios where differences are expected between belonging and identification to disentangle the unique impact of each construct. An example could be a recent immigrant to the United States. This individual might feel extremely welcomed in his or her new neighborhood and workplace and feel a high sense of belongingness. However, he or she may still feel strong ties to his or her national origins and not yet identify himself or herself as an American. Future research should investigate specific groups and situations to determine whether using these measures as distinct constructs is warranted.
Another limitation of the study is the risk-taking findings in Experiment 3. Although there were significant increases in risk taking for individuals in the ostracized–included condition, there were also significant increases for participants in the ostracized–ostracized condition and increases (though not significant) for participants in the included–ostracized group. It could be that all three of these conditions were acting in a way that the participants anticipated would help them gain entry into the group, thereby fulfilling their need to belong. Future work should continue to investigate the behavioral component of risk taking following situations of ostracism by using more varied risk-taking behaviors (e.g., risky spending of personal funds, physical harm) that represent more real-life behaviors.
It is important to highlight the constraints of generality for the current data to describe the scope of our conclusions. Participants in this study included both college students and Mturk workers. Because results were consistent across both populations, we can conclude these phenomena do generalize; however, the extent of the generalizability to other populations (beyond college students and Mturk workers) remains to be tested by future research. The stimuli and operations of this study included placing participants into a minimal groups context and exposing them to the Cyberball ostracism paradigm, suggesting that these results will generalize to other small groups research on ostracism. However, these findings may be limited to small groups research (or minimal groups), so future work should investigate the procedures used here in other intragroup contexts (e.g., preexisting identities or large-scale social identities). Reproduction of these findings should not be constrained by sample selection and should be possible using other small groups contexts and experiences of ostracism. Finally, the findings of this study should broadly generalize to contexts where individuals experience ostracism from the same group. The current study does not make any claims regarding ostracism and inclusion experiences involving different groups, and future research would be required to draw conclusions about these experiences. We have no reason to believe that the results depend on other characteristics of the participants, materials, or context.
Conclusion
Three experiments presented in this study illustrate how subsequent experiences of inclusion or ostracism from the same group affects an individual’s fundamental needs fulfillment, and, more important, his or her identification with, perceptions of, and willingness to take risks for the group. Specifically, being ostracized and then included causes individuals to identify very highly with the group; perceive the group as higher in status, entitativity, warmth, and competence; and to take more risks for the group. Situations where an individual is first denied inclusion to a group and then later accepted leads to strong feelings of group membership and commitment. We all have gone through experiences of ostracism and inclusion, but it appears that first being denied inclusion leads to us feeling even more like a member of a group than if we had just been included from the beginning.
Supplemental Material
PSBP_Final_Online_Supplemental_Materials – Supplemental material for Loving the Group That Denies You First: Social Identity Effects of Ostracism Before Inclusion
Supplemental material, PSBP_Final_Online_Supplemental_Materials for Loving the Group That Denies You First: Social Identity Effects of Ostracism Before Inclusion by Ethan Dahl, Elizabeth M. Niedbala and Zachary P. Hohman in Personality and Social Psychology Bulletin
Footnotes
Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The author(s) received no financial support for the research, authorship, and/or publication of this article.
Notes
Supplemental Material
Supplemental material is available online with this article.
References
Supplementary Material
Please find the following supplemental material available below.
For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.
For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.
