Abstract
Understanding the relationship between social exclusion and behavioral inhibition is essential for mitigating its detrimental effects and developing targeted interventions. This study examines the impact of short- and long-term social exclusion on reactive and intentional inhibition through three experiments. Experiment 1 utilized the Cyberball game to induce a short-term experience of exclusion, Experiment 2 employed the Future Life Alone paradigm to evoke imagined long-term exclusion, and Experiment 3 used a questionnaire to identify individuals who had experienced actual long-term exclusion. Reactive and intentional inhibition were assessed using the Free Two-Choice Oddball Task. Three key findings emerged: First, both short- and long-term social exclusion significantly impaired intentional inhibition. Second, diffusion model analyses revealed distinct mechanisms underlying these impairments: short-term social exclusion increased the tendency toward standard responses during free-choice trials, while long-term exclusion led to more conservative response criteria. Third, reactive inhibition was unaffected by short-term or imagined long-term exclusion but was impaired by actual long-term exclusion. Collectively, these findings suggest that although short- and long-term exclusion share some similarities in their effects on behavioral inhibition, they also exhibit notable differences. This study provides novel insights into the influence of social exclusion on behavioral inhibition, with theoretical, methodological, and practical implications.
Challenging the fundamental human need for stable and enduring social connections, social exclusion represents one of the most pervasive and distressing social phenomena encountered in daily life (Williams & Nida, 2022). Extensive research has demonstrated that social exclusion impairs self-control. Individuals experiencing social exclusion are more likely to engage in unhealthy eating behaviors (Vartanian & Porter, 2016), develop both substance-related and non-substance-related addictions (Taş, 2021; Wesselmann & Parris, 2021), exhibit aggressive tendencies (Poon & Wong, 2019), and even attempt suicide (Chen et al., 2020). Given the significant negative impacts of social exclusion, investigating the mechanisms underlying its detrimental effects on self-control holds significant theoretical and practical importance.
A prevailing hypothesis posits that social exclusion undermines the capacity for behavioral inhibition, thereby resulting in deficiencies in self-control (White et al., 2015). This hypothesis seems reasonable, as behavioral inhibition denotes the ability to restrain inappropriate behaviors and is considered a fundamental component of self-control (Diamond, 2013). However, investigations into this hypothesis reveal that only a few studies provide support for it (Jamieson et al., 2010; Wang & Sha, 2018), while a substantial body of research contradicts it (Ernst et al., 2018; King et al., 2018; Otten & Jonas, 2013; Xu et al., 2016; Zhang et al., 2021). For instance, Otten and Jonas (2013) conducted an experiment where participants, following exposure to social exclusion induced by the Cyberball virtual ball-tossing game, were instructed to perform a go/no-go task. This task required responding to a high-probability “go” stimulus while inhibiting responses to a low-probability “no-go” stimulus. The results indicated that both excluded and included individuals exhibited comparable accuracy in no-go trials, suggesting that social exclusion did not significantly influence behavioral inhibition. Furthermore, Ernst et al. (2018) employed similar methods of Cyberball-induced social exclusion and the go/no-go task, revealing that the excluded individuals exhibited higher levels of accuracy in no-go trials than did included individuals, indicating that social exclusion enhanced behavioral inhibition. Notably, in studies observing the impairment effect of social exclusion, nonclassic tasks have been employed to assess behavioral inhibition, such as the antisaccade task utilized by Jamieson et al. (2010) and the flanker task employed by Wang and Sha (2018). Therefore, the “behavioral inhibition” examined in these studies may not accurately represent genuine behavioral inhibition (e.g., the flanker task could potentially serve as a suitable measure for assessing interference control; Diamond, 2013). Taken together, the current body of research does not provide evidence to support the hypothesis that deficits in behavioral inhibition underlie the negative impact of social exclusion on self-control.
Nevertheless, before reaching the definitive conclusion that social exclusion has no adverse effects on behavioral inhibition, two critical factors must be considered. Firstly, existing studies on behavioral inhibition have predominantly focused on externally driven inhibition (i.e., reactive inhibition), in which individuals must cease their responses based on external stop signals, such as the no-go stimulus in a go/no-go task. However, this narrow focus fails to capture the full spectrum of behavioral inhibition in real-life situations, where excluded individuals often need to restrain inappropriate impulsive behaviors without external cues (Filevich et al., 2012; Lynn et al., 2014). For instance, there is no red light in social contexts signaling one to suppress aggressive behaviors in response to exclusion, implying that the inhibition of such tendencies must be internally generated. This internally driven inhibition, termed intentional inhibition, is considered to be more closely associated with self-control and has been shown to exhibit distinct temporal dynamics (Parkinson & Haggard, 2015; Xu et al., 2020, 2025) and neural substrates (Ficarella & Battelli, 2017; Kühn et al., 2009) compared to reactive inhibition. This distinction is typically examined using “free-choice” paradigms, which allow individuals to freely choose whether to execute or inhibit a prepotent response (for detailed task descriptions, see Filevich et al., 2012; Lynn et al., 2014; Xu, Wen, et al., 2024). Electrophysiological evidence indicates that reactive inhibition is typically indexed by frontal N2 and P3 components, whereas intentional inhibition involves more complex temporal features, including N2, early and late P3 components, which reflect behavioral updating, inhibitory execution, and outcome evaluation processes (Xu, Xu, et al., 2024). Furthermore, resting-state electroencephalographic (EEG) analyses reveal that intentional inhibition uniquely correlates with specific EEG power bands and microstate dynamics, suggesting a distinct neural architecture (Xu et al., 2025). Neuroimaging studies further highlight the selective involvement of the dorsal fronto-median cortex in intentional inhibition, distinguishing it from reactive inhibition (Brass & Haggard, 2007; Ficarella & Battelli, 2017; Kühn et al., 2009). Therefore, to achieve a more comprehensive understanding of the relationship between social exclusion and behavioral inhibition, it is essential to investigate how social exclusion influences intentional inhibition.
Despite the lack of direct research on this issue, indirect studies have provided valuable insights. On the one hand, Filevich et al. (2012) proposed that intentional inhibition is a dynamic and resource-intensive process that demands heightened self-engagement, as participants must continuously estimate the distal consequences of their current commands, generate an intentional inhibition signal when the forthcoming action conflicts with their long-term interests, and ultimately execute a final decision. On the other hand, social exclusion has been found to have a range of detrimental effects on cognitive processes, including reducing working memory capacity for storage and updating (which is crucial for dynamic monitoring and continuous estimation; Ma & Pang, 2023; Xu et al., 2018), decreasing self-awareness (e.g., increased self-avoidance, Twenge et al., 2003; lower levels of self-continuity, Jiang et al., 2021; and reduced sense of agency, Malik & Obhi, 2019), and leading to low-level construal (i.e., concrete mental representation), which is characterized by focusing on detailed aspects of a target from a short-term or proximal perspective (due to prevention focus, Lee et al., 2010; Park & Baumeister, 2015; and ego-depletion, Bahrami & Borhani, 2023; Baumeister et al., 2002; Bruyneel & Dewitte, 2012). Based on these findings, it can be inferred that social exclusion is likely to impede intentional inhibition. Hence, the first aim of this study was to examine and substantiate this hypothesis.
Secondly, while there is extensive research on the impact of social exclusion on behavioral inhibition, most studies have primarily focused on short-term exclusion (e.g., induced by Cyberball; Wirth, 2016), leaving the long-term effects largely unexplored (Rudert et al., 2020; Wesselmann et al., 2023). It is essential to recognize that the impact of long-term social exclusion on behavioral inhibition may differ significantly from that of short-term exclusion. For one thing, the intensity and duration of long-term exclusion surpass those of short-term exclusion, which can result in more severe consequences (Bernstein & Claypool, 2012a, 2012b; Williams, 2007). For instance, Bernstein and Claypool (2012a, 2012b) found that imagining long-term social exclusion (manipulated by the Future Life Alone paradigm; Wirth, 2016) increased individuals’ sensory and decision thresholds compared to experiencing short-term social exclusion (manipulated by Cyberball; Wirth, 2016), leading to desensitization of emotional and physical pain. For another, researchers have proposed that individuals may display distinct reaction tendencies to short- versus long-term exclusions. According to the goal-driven resource redistribution theory (Shilling & Brown, 2016) and the multimotive model (Smart Richman & Leary, 2009), excluded individuals allocate attentional resources based on specific goal priorities. Meanwhile, in the temporal need-threat model, Williams (2009) proposes that individuals’ reactions to social exclusion can be divided into three sequential stages: reflexive (immediate), reflective (coping), and resignation (long-term). During the reflexive stage, individuals quickly detect exclusion and experience social pain because of threats to their fundamental psychological need for belonging, self-esteem, meaningful existence, and control. After this initial painful reaction, individuals enter the reflective stage, where they attempt to cope with exclusion to recover the satisfaction of their needs. If all coping strategies fail and social exclusion persists, individuals may enter the resignation stage, where they accept their inevitable condition of exclusion. As the resources necessary for fortifying threatened needs are depleted during this stage, individuals no longer seek fulfillment of these psychological needs but instead try to avoid further exclusion. Accordingly, due to the likely divergent goal priorities between individuals excluded in the short term (reflexive stage: restoring social inclusion) and those excluded in the long term (resignation stage: avoiding future social exclusion), they may adopt different approaches to allocating resources during tasks, resulting in differential performance in behavioral inhibition. Based on this rationale, it can be inferred that long-term exclusion may pose a more substantial challenge to behavioral inhibition compared to short-term exclusion. Moreover, the cognitive mechanisms underlying the detrimental effects of short- versus long-term exclusion are likely to differ significantly. Consequently, the second and third aims of this study were to investigate and validate these hypotheses.
In summary, this study was designed to systematically investigate the impacts of both short- and long-term social exclusion on reactive and intentional inhibition. Specifically, it aims to achieve three primary objectives: (a) to evaluate the effects of social exclusion (particularly, long-term exclusion) on reactive inhibition; (b) to examine the influence of social exclusion on intentional inhibition; and (c) to compare and elucidate the cognitive mechanisms underlying impairments in behavioral inhibition resulting from short- and long-term exclusion, as induced by different experimental paradigms. 1 These aims were addressed through a series of three experiments. Experiment 1 utilized the Cyberball game to induce a short-term experience of exclusion, while Experiment 2 employed the Future Life Alone paradigm to evoke long-term imagining of exclusion (Wirth, 2016). However, it is important to note that although the Future Life Alone paradigm is widely employed in laboratory settings to simulate long-term exclusion (Ghandchi et al., 2024; Wirth, 2016), imagining future exclusion may not fully capture the lived experience of enduring prolonged exclusion. Moreover, differences in procedural design between the Future Life Alone paradigm and Cyberball introduce additional complexity. Specifically, in Cyberball, exclusion is inferred by participants based on the lack of ball reception, whereas the Future Life Alone paradigm involves explicit notification of exclusion. Consequently, it remains a matter of debate whether the differing effects of these paradigms on behavioral inhibition genuinely reflect the distinction between short- and long-term exclusion. Therefore, Experiment 3 used the Social Exclusion Questionnaire for Undergraduates (Wu et al., 2013) to identify individuals who had experienced actual long-term exclusion. Furthermore, to assess reactive and intentional inhibition, we utilized the recently developed Free Two-Choice Oddball Task, which integrates free-choice trials into a conventional two-choice oddball task consisting of standard and deviant trials (Xu, Wen, et al., 2024; Xu et al., 2025). For standard and deviant trials, participants are instructed to press buttons “1” and “2,” respectively. However, for free-choice trials, participants have the autonomy to decide whether to make a “1” or a “2” response. Importantly, to induce participants’ response tendencies in this task, the proportion of standard, deviant, and free-choice trials is set at 4:1:2. Given that there is a higher probability of encountering standard trials, it leads to the formation of a dominant response preference where participants are more inclined towards making a “1” response. Therefore, when low-probability free-choice trials occur, and if participants choose to make a “2” response instead of their already formed tendency for making a “1” response, they may need to inhibit their preexisting inclination for making a “1” response. Conversely, if they choose to make a “1” response for these low-probability free-choice trials, there would be no need to inhibit their dominant tendency to make such responses. Consequently, longer response times (RTs) observed for “2” responses compared with “1” responses, both in the reactive condition (i.e., standard and deviant trials) and the intentional condition (i.e., free-choice trials), can indicate behavioral inhibition presence. Moreover, impaired reactive inhibition can be inferred from lower accuracy in deviant trials, larger accuracy cost (i.e., the difference in accuracy between standard and deviant trials), and larger RT cost (i.e., the difference in RTs between deviant and standard trials). Impaired intentional inhibition, on the other hand, can be indicated by lower inhibition rates (i.e., the proportion of “2” responses) and larger RT costs (i.e., the difference in RTs between “2” and “1” responses) in free-choice trials (Xu, Xu, et al., 2024; Xu et al., 2025).
Building on previous studies, we hypothesized that long-term exclusion, as opposed to short-term exclusion, would impair reactive inhibition. We therefore anticipated that participants experiencing short-term exclusion would show similar levels of accuracy and RT costs in the reactive condition as those who did not experience exclusion (Experiment 1). In contrast, we expected participants subjected to long-term exclusion to exhibit significantly larger accuracy costs or RT costs (Experiments 2 and 3). Additionally, we hypothesized that both short- and long-term exclusion would impair intentional inhibition. Consequently, we anticipated that participants experiencing either short- or long-term exclusion would demonstrate reduced inhibition rates or increased RT costs in the intentional condition (Experiments 1–3). Finally, we hypothesized that the cognitive mechanisms underlying the detrimental effects of short- versus long-term exclusion on intentional inhibition, while both impaired, will be likely to differ significantly. Therefore, we conducted a diffusion model analysis to elucidate the decision-making process in free-choice trials (Ratcliff et al., 2016; White & Kitchen, 2022). Given the limited existing research, we refrained from formulating specific hypotheses regarding the anticipated outcome patterns. The experimental procedures described above, along with the research hypotheses and analysis plans, were approved by the Institutional Review Board (IRB) of the School of Psychology at Shaanxi Normal University. This approval was granted on September 5, 2022, under reference number HR 2022-09-001. Written informed consent was obtained from all participants prior to the commencement of the experiments.
Experiment 1
Experiment 1 investigated the impact of experiencing short-term social exclusion on reactive and intentional inhibition, as well as the underlying cognitive mechanisms. To achieve this goal, we used the Cyberball game to manipulate the experience of short-term social exclusion, the Need Threat Scale and the Positive and Negative Affect Schedule (PANAS) to assess the effectiveness of the manipulation, and the Free Two-Choice Oddball Task to measure reactive and intentional inhibition.
Methods
Participants
To determine the appropriate sample size for this study, we conducted an a priori power analysis using G*Power (Faul et al., 2007), which indicated that approximately 34 participants were needed to capture a medium effect size (f = .25) for the within–between interaction (i.e., the Group × Task × Response interaction on RT, and the Group × Task interaction on accuracy) with 80% power and an alpha of .05. Anticipating exclusions, 70 female undergraduate students were recruited and randomly assigned to either the inclusion or the exclusion group. The decision to limit our sample to women was based on prior research indicating that women are more susceptible to the negative effects of social exclusion (Benenson et al., 2013). Two participants were excluded from the study due to their extreme response patterns in free-choice trials (i.e., 100% standard responses), and another participant was excluded because she disagreed with the social exclusion manipulation. This resulted in 34 participants in the exclusion group (M age = 19.65 years, SE age = 0.24), and 33 participants in the inclusion group (M age = 19.61 years, SE age = 0.21). A sensitivity power analysis showed that, with a final sample size of 67, the study was sufficient to detect a small- to medium-sized effect (f = .14–.17) for the within–between interaction.
Cyberball game
The Cyberball game was used to manipulate short-term social exclusion (Wirth, 2016). The participants engaged in a virtual toss game with two unfamiliar female players (i.e., three-player game) and did not anticipate future interactions. We manipulated the levels of social exclusion and inclusion by varying the frequency with which the participants received the ball from their fellow players (51 total throws). Participants assigned to the inclusion group were granted approximately one third of the total throws, whereas those in the exclusion group were limited to only two initial throws at the outset of the game.
Need Threat Scale
Following completion of the Cyberball game, participants were administered the 20-item Need Threat Scale (van Beest & Williams, 2006), which assesses self-reported levels of satisfaction with feelings of belonging (e.g., “I feel accepted by others”; Cronbach’s α = .93), self-esteem (e.g., “I have high self-esteem”; Cronbach’s α = .87), meaningful existence (e.g., “I am a useful person”; Cronbach’s α = .95), and control (e.g., “I feel that I can control the game”; Cronbach’s α = .84) during gameplay on a 7-point Likert scale (1 = strongly disagree, 7 = strongly agree). Lower scores indicate an increased perceived threat to social needs and suggest the effectiveness of the social exclusion manipulation (Cronbach’s α for each subscale are reported above, and Cronbach’s α for the entire scale was .97).
Positive and Negative Affect Schedule
Participants were administered the 20-item Positive and Negative Affect Schedule (PANAS; Watson et al., 1988), which comprises 10 items measuring positive emotions (e.g., interested; Cronbach’s α = .88) and 10 items measuring negative emotions (e.g., irritable; Cronbach’s α = .92). Participants were instructed to rate their current emotional state on a 5-point scale (1 = very slightly or not at all, 5 = extremely).
Free Two-Choice Oddball Task
The Free Two-Choice Oddball Task (Figure 1) was used to investigate reactive and intentional inhibition where free-choice trials were incorporated into the conventional two-choice oddball task (Xu, Wen, et al., 2024). To elicit strong behavioral impulses, the ratio of reactive standard, reactive deviant, and free-choice trials was set to 4:1:2. Because reactive standard trials are more frequent than reactive deviant trials, standard responses become dominant. Participants develop a strong impulse to make standard responses during the experiment and to perform standard responses in the free-choice trials.
Illustrations of the Free Two-Choice Oddball Task.
Each trial began with a centrally presented crosshair lasting for 300 ms, followed by an interstimulus interval of 500–1,000 ms. Next, a target stimulus was presented for 100 ms, during which the participants could provide responses. This was followed by another interstimulus interval of 700 ms and concluded with feedback delivery lasting 300–700 ms. The target stimuli consisted of a white square (visual angle of 0.47° × 0.47°), a white diamond (the same square rotated by 45°), and a white circle (0.47° × 0.47°).
Participants viewed the display from a fixed distance of 60 cm and were instructed to press button “1” for white square (i.e., standard responses for reactive standard trials), and button “2” for white diamond (i.e., deviant responses for reactive deviant trials). Importantly, for the white circle (i.e., free-choice trials), participants were explicitly instructed to deliberately select either button “1” or “2” (i.e., standard or deviant responses). The instructions provided for these trials were highly specific and comprehensive. Participants were informed that there was no right or wrong answer and were encouraged to choose quickly every time they saw the free-choice trials, without deciding in advance what that choice would be. They were encouraged to choose “1” and “2” responses “roughly equally” throughout the experiment, while avoiding strategic approaches such as alternating or counting response numbers. In the formal experiment, a total of 476 trials were divided into four blocks consisting of 119 trials each. Before commencing the main task, the participants completed an initial practice block that was not included in the analysis.
Statistical analyses
To assess the effectiveness of the social exclusion manipulation, separate independent sample t tests were conducted on the Need Threat Scale and PANAS scores of the exclusion and inclusion groups.
To examine the influence of social exclusion on reactive and intentional inhibition, mean accuracies in reactive standard and deviant trials, inhibition rates (i.e., the proportion of deviant responses to total responses) in free-choice trials, and mean RTs were separately analyzed. Specifically, mean accuracies were analyzed with a Group (exclusion, inclusion) × Response (standard, deviant) analysis of variance (ANOVA). Mean inhibition rates were analyzed with independent samples t tests between the exclusion and inclusion groups. And mean RTs were analyzed with a Group (exclusion, inclusion) × Action Source (reactive, intentional) × Response (standard, deviant) ANOVA. The RTs ANOVA revealed a significant three-way interaction, F(1, 65) = 5.73, p = .020, η2p = .08. To enhance clarity, we will present the accuracy and RT results for the reactive condition jointly, while presenting the inhibition rate and RT results for the intentional condition together. For RT analysis, trials with false responses and exceedingly short or long RTs (±2.5 SD from the mean RT for each condition and each participant) were removed.
In addition to traditional analyses of inhibition rates and RTs, a diffusion model analysis was conducted using fast-dm software (version 30.2) to further investigate the underlying cognitive mechanisms of the impact of social exclusion on intentional inhibition (Voss et al., 2015). This analysis could provide crucial information for understanding decision-making behavior by incorporating all of the behavioral data, including the proportion of different responses and RT distributions for both standard and deviant responses (Ratcliff et al., 2016; White & Kitchen, 2022). The model proposes that the overall RT in a single trial can be decomposed into two distinct components: decisional and nondecisional. The decisional component pertains to the choice or response selection process, whereas the nondecisional component encompasses all cognitive processes that occur before and after response selection, such as stimulus encoding and motor processes. In its classical version, the decision-making process is characterized by an initial bias toward one of two responses (starting point bias), the average rate of evidence accumulation for each response alternative (drift rate), and the threshold level of evidence required to reach a decision (boundary separation). Thus, four basic parameters, namely the starting point bias (z), drift rate (v), boundary separation (a), and nondecisional RT (t), could be estimated for each condition and each participant, with the lower and upper thresholds set to reflect the standard and deviant responses for free-choice trials, respectively. Therefore, to investigate the cognitive mechanisms underlying the impact of social exclusion on intentional inhibition, the exclusion and inclusion groups were compared by analyzing each of these four parameters separately using independent samples t tests. In the diffusion model analysis, all other parameters implemented in fast-dm-30 were set to zero, and fast RTs below 100 ms were excluded. The Kolmogorov–Smirnov test was chosen as the optimization criterion, following the recommendations of Lerche et al. (2017) and Voss et al. (2013, 2015) for trial numbers ranging from approximately 100 to 500 (medium size). Regarding the quality of model fit, the Kolmogorov–Smirnov statistic provided by the fast-dm-30 did not reveal any significant deviations between empirical and predicted RT distributions (Experiment 1: ps ⩾ .437; Experiment 2: ps ⩾ .419; Experiment 3: ps ⩾ .420), suggesting that the model fitted the data reasonably well for all participants.
Results and Discussion
Manipulation check
For the need threat scores (average of the 20 items), results revealed significantly lower scores for the exclusion group (M = 3.08, SE = 0.18) than for the inclusion group (M = 5.25, SE = 0.16), t(65) = 9.06, p < .001, Cohen’s d = 2.22. Moreover, further analyses revealed that the exclusion group scored lower than the inclusion group on the self-esteem, t(65) = 7.24, p < .001, Cohen’s d = 1.77; sense of belonging, t(65) = 8.65, p < .001, Cohen’s d = 2.12; meaningful existence, t(65) = 8.77, p < .001, Cohen’s d = 2.14; and sense of control, t(65) = 8.09, p < .001, Cohen’s d = 1.98 dimensions. These results indicate that the excluded participants experienced greater threats to their needs than the included ones, confirming the effectiveness of the exclusion manipulation.
Regarding the PANAS scores, results revealed that the exclusion group had lower positive and higher negative emotions than the inclusion group: positive (M = 27.38, SE = 1.10 vs. M = 31.97, SE = 0.98), t(65) = 3.10, p = .003, Cohen’s d = 0.76; negative (M = 19.62, SE = 1.20 vs. M = 16.00, SE = 1.01), t(65) = −2.30, p = .025, Cohen’s d = 0.56.
Impact of experiencing short-term social exclusion on reactive inhibition
For accuracy, the ANOVA revealed a significant main effect of response, F(1, 65) = 105.54, p < .001, η2p = .62, with higher accuracy for the standard response (M = 0.98, SE = 0.01) than for the deviant one (M = 0.87, SE = 0.01). However, neither the main effect of group nor the interaction between group and response was significant, F(1, 65) < 0.01, p = .971, η2p < .01, F(1, 65) = 0.88, p = .351, η2p = .01 (Figure 2A).
Results of Experiment 1. (A) Accuracy (ACC) costs (i.e., the differences between standard and deviant responses) in the reactive condition and the intentional rates in the intentional condition; (B) RT costs in the reactive and intentional conditions; (C) Diffusion model analysis results in the intentional condition, for the exclusion and inclusion groups.
For RT, analyses revealed that both the exclusion and inclusion groups exhibited faster RTs for standard responses than for deviant ones, F(1, 65) = 220.42, p < .001, η2p = .77, F(1, 65) = 192.26, p < .001, η2p = .75, suggesting that reactive inhibition manifested across groups. Additionally, there was no significant difference in RT costs (i.e., deviant response minus standard response) between the two groups, F(1, 65) = 0.71, p = .401, η2p = .01 (Figure 2B).
The combined analyses of accuracy and RT suggest that experiencing short-term social exclusion does not affect reactive inhibition.
Impact of experiencing short-term social exclusion on intentional inhibition
For inhibition rate, the t test failed to reveal any significant differences between the exclusion and inclusion groups, t(65) = 0.42, p = .679, Cohen’s d = 0.10 (Figure 2A).
For RT, analyses revealed that both the exclusion and inclusion groups showed faster RTs for standard responses than for deviant ones, F(1, 65) = 50.63, p < .001, η2p = .44, F(1, 65) = 12.82, p = .001, η2p = .17, suggesting that intentional inhibition was evident across groups. Additionally, the exclusion group showed a greater RT cost than the inclusion group, F(1, 65) = 5.97, p = .017, η2p = .08 (Figure 2B), suggesting that social exclusion impairs intentional inhibition.
For the diffusion model analysis, t tests revealed that the exclusion group displayed a smaller starting point bias than the inclusion group, t(65) = 3.04, p = .003, Cohen’s d = 0.74. This suggests that the exclusion group had a stronger bias toward making standard responses in the free-choice trials (Figure 2C). However, no significant differences were found between groups in boundary separation, t(65) = 0.34, p = .736, Cohen’s d = 0.08; drift rate, t(65) = 0.96, p = .340, Cohen’s d = 0.24; and nondecisional RT, t(65) = 0.43, p = .669, Cohen’s d = 0.11.
The analyses of RT and starting point bias suggested that experiencing short-term social exclusion impairs intentional inhibition, and this impairment effect may be associated with a stronger tendency to make standard responses during free-choice trials.
Experiment 2
Experiment 2 investigated the impact of imagining long-term social exclusion on reactive and intentional inhibition, as well as the underlying cognitive mechanisms. For this, we used the Future Life Alone paradigm to manipulate the imagination of long-term social exclusion; the Need Threat Scale and PANAS to assess the effectiveness of the manipulation; and the Free Two-Choice Oddball Task to measure reactive and intentional inhibition.
Methods
Participants
Another 70 female undergraduate students were recruited and randomly assigned to the inclusion (M age = 18.42 years, SE age = 0.19) or exclusion group (M age = 18.11 years, SE age = 0.08). A sensitivity power analysis showed that this sample allowed us to detect a small- to medium-sized effect (f = 0.14–.17) for the within–between interaction, with 80% power and an alpha of .05.
Future Life Alone paradigm
The Future Life Alone paradigm is used to manipulate the imagining of long-term social exclusion (Wirth, 2016). Participants were informed that they would undergo a comprehensive battery of assessments and receive personalized feedback regarding their personality traits, along with potential implications for their future lives based on their responses (Twenge et al., 2001). They completed the Eysenck Personality Inventory and received accurate feedback on their level of extraversion, which served to enhance the credibility of the subsequent social experience manipulation. Subsequently, all participants were randomly assigned to receive one of two types of false feedback. Participants in the inclusion group (i.e., future belonging) were informed of an upcoming life filled with enriching social relationships, whereas those in the exclusion group (i.e., future alone) were presented with the prospect of lacking meaningful connections (for detailed scripts, see Twenge et al., 2001).
Need Threat Scale
The Cronbach’s α values for the subscales of belonging, self-esteem, control, and meaningful existence were .64, .69, .62, and.77, respectively. The Cronbach’s α for the entire scale was .87.
PANAS
The Cronbach’s α values for the subscales of positive and negative emotions were .87 and .82, respectively.
Free Two-Choice Oddball Task
The Free Two-Choice Oddball Task as in Experiment 1 was conducted.
Statistical analyses
Analyses were similar to those in Experiment 1. The ANOVA of RTs revealed a significant three-way interaction, F(1, 68) = 4.30, p = .042, η2p = .06. To enhance clarity, we present the accuracy and RT results for the reactive condition together, and the inhibition rate and RT results for the intentional condition together.
Results and Discussion
Manipulation check
For the need threat scores (average of the 20 items), results revealed no significant difference between the exclusion (M = 5.08, SE = 0.10) and inclusion groups (M = 5.28, SE = 0.12), t(68) = 1.24, p = .220, Cohen’s d = 0.30. However, further analyses revealed that the exclusion group scored lower than the inclusion one on the self-esteem dimension, t(68) = 2.28, p = .026, Cohen’s d = 0.54, but not on the sense of belonging, t(68) = 0.07, p = .944, Cohen’s d = 0.02; meaningful existence, t(68) = 0.25, p = .801, Cohen’s d = 0.06; or sense of control dimensions, t(68) = 1.38, p = .173, Cohen’s d = 0.33. This suggests that excluded participants’ self-esteem needs were more threatened than those of the included participants, confirming the effectiveness of the exclusion manipulation (Bernstein et al., 2013).
For the PANAS scores, results indicated that the exclusion group demonstrated similar positive but higher negative emotions than the inclusion group: positive (M = 31.20, SE = 1.11 vs. M = 31.83, SE = 1.12), t(68) = 0.40, p = .691, Cohen’s d = 0.10; negative (M = 23.11, SE = 0.97 vs. M = 18.94, SE = 0.97), t(68) = 3.04, p = .003, Cohen’s d = 0.73.
Impact of imagining long-term social exclusion on reactive inhibition
For accuracy, the ANOVA revealed only a main effect of response, F(1, 68) = 209.45, p < .001, η2p = .76, with higher accuracy for standard responses (M = 0.95, SE = 0.01) than for deviant ones (M = 0.78, SE = 0.02). However, neither the main effect of group nor the interaction between group and response was significant, F(1, 68) = 0.18, p = .673, η2p < .01, F(1, 68) = 0.02, p = .887, η2p < .01 (Figure 3A).
Results of Experiment 2. (A)Accuracy (ACC) costs (i.e., the differences between standard and deviant responses) in the reactive condition and the intentional rates in the intentional condition; (B) RT costs in the reactive and intentional conditions; (C) Diffusion model analysis results in the intentional condition, for the exclusion and inclusion groups.
For RT, analyses revealed that both the exclusion and inclusion groups showed faster RTs for standard responses than for deviant ones, F(1, 68) = 197.50, p < .001, η2p = .74, F(1, 68) = 203.51, p < .001, η2p = .75, suggesting that reactive inhibition was evident across groups. Additionally, both groups showed similar RT costs, F(1, 65) = 0.02, p = .881, η2p < .01 (Figure 3B).
The combined analyses of accuracy and RT suggest that imagining long-term social exclusion does not affect reactive inhibition.
Impact of imagining long-term social exclusion on intentional inhibition
For inhibition rate, the t test failed to reveal any significant differences between the exclusion and inclusion groups, t(68) = 0.24, p = .810, Cohen’s d = 0.06 (Figure 3A).
For RT, analyses revealed that both the exclusion and inclusion groups showed faster RTs for standard responses than for deviant ones, F(1, 68) = 76.42, p < .001, η2p = .53, F(1, 68) = 24.73, p < .001, η2p = .27, suggesting that intentional inhibition was evident across groups. Additionally, the exclusion group exhibited greater RT cost than the inclusion one, F(1, 68) = 7.10, p = .010, η2p = .10 (Figure 3B), suggesting that social exclusion impairs intentional inhibition.
For the diffusion model analysis, t tests revealed that the exclusion group had a larger boundary separation than the inclusion one, t(68) = 2.61, p = .011, Cohen’s d = 0.62. This suggests that the exclusion group exhibited more conservative response criteria when making decisions during the free-choice trials (Figure 3C). However, no significant group differences were found for starting point bias, t(68) = 1.53, p = .131, Cohen’s d = 0.37; drift rate, t(65) = 0.77, p = .443, Cohen’s d = 0.19; or nondecisional RT, t(68) = 0.09, p = .926, Cohen’s d = 0.02.
The analyses of RT and boundary separation suggest that imagining long-term social exclusion impairs intentional inhibition, and this impairment effect may be associated with more conservative response criteria in free-choice trials.
Experiment 3
Experiment 3 sought to identify individuals who had experienced long-term exclusion in real-life settings and to reexamine the impact of these experiences on both reactive and intentional inhibition, as well as the underlying cognitive mechanisms. To achieve this objective, we first employed the Social Exclusion Questionnaire for Undergraduates (Wu et al., 2013) to screen for individuals who had experienced prolonged social exclusion (defined as lasting more than 3 months, analogous to the duration criterion for chronic physical pain; Büttner et al., 2024; Riva et al., 2016). Subsequently, we used the Free Two-Choice Oddball Task to assess their reactive and intentional inhibition, and collected subjective assessments from participants regarding task difficulty and attentional resource allocation during this task. 2 Additionally, an intertemporal choice task was conducted (Rachlin & Jones, 2008); however, since this was not the primary focus of our study and no significant group differences were observed, these results are omitted from the main text but provided in the Supplemental Material.
Methods
Participants
Initially, 500 female undergraduate students were recruited to participate in a mass screening using the Social Exclusion Questionnaire for Undergraduates (Cronbach’s α = .92; Wu et al., 2013). This questionnaire comprises 19 items, each rated on a 5-point Likert scale. Participants were instructed to reflect on the frequency with which their classmates exhibited the behaviors described in the questions (e.g., “Whenever I participate in a group discussion, the conversation consistently becomes silent”) or the occurrence of such events/situations since they commenced their university studies (all participants had been enrolled for a minimum period of 6 months). A higher total score indicates a greater degree of perceived social exclusion in daily life. Based on the distribution of scores, individuals falling within the top and bottom 27% were selected for further analysis (Kelley, 1939; Xu, Xu, et al., 2024). Specifically, 38 participants who scored at or above the 73rd percentile—indicating high levels of long-term social exclusion—were invited to participate (M age = 18.63 years, SE age = 0.14), along with 37 participants who scored at or below the 27th percentile, indicating low levels of exclusion (M age = 18.68 years, SE age = 0.16). An independent sample t test conducted on exclusion experience scores revealed a significant difference between groups, t(73) = 26.23, p < .001, with the exclusion group (M = 49.66, SE = 1.05) reporting significantly higher exclusion experience scores compared to the nonexclusion (i.e., inclusion) group (M = 20.68, SE = 0.31). A sensitivity power analysis showed that this sample allowed us to detect a small- to medium-sized effect (f = .14–.16) for the within–between interaction with 80% power and an alpha of .05.
Free Two-Choice Oddball Task
The Free Two-Choice Oddball Task as in Experiment 1 was conducted.
Subjective assessments
Following the completion of the Free Two-Choice Oddball Task, participants were asked to rate their perceived task difficulty on a 5-point Likert scale (1 = not at all difficult, 5 = extremely difficult) for both the reactive and intentional conditions (i.e., reactive trials and free-choice trials). Specifically, they were asked to evaluate: (a) the difficulty in responding to the white square and white diamond, and (b) the difficulty in responding to the white circle. Additionally, participants were asked to report their attentional resource allocation between the reactive and intentional conditions by indicating which category of target stimuli garnered more of their attention during the key-pressing task. Responses were recorded on a 5-point scale (1 = all attention was exclusively focused on the white square and white diamond, 5 = all attention was directed towards the white circle).
Statistical analyses
These analyses were similar to those conducted in Experiment 1. The ANOVA of RTs did not yield a significant three-way interaction, F(1, 73) = 1.73, p = .193, η2p = .02. Despite the lack of statistical significance for this interaction, we conducted an exploratory analysis of simple effects. Consistent with Experiment 1 and to enhance clarity, we present the accuracy and RT results for the reactive condition together, and the inhibition rate and RT results for the intentional condition together. Furthermore, subjective assessments of task difficulty and attentional resource allocation were analyzed using independent samples t tests between the exclusion and inclusion groups.
Results and Discussion
Impact of experiencing long-term social exclusion on reactive inhibition
For accuracy, ANOVA revealed a significant interaction between group and response, F(1, 73) = 4.44, p = .039, η2p = .06. Further analyses indicated that both the exclusion and inclusion groups exhibited higher accuracy for standard responses compared to deviant ones, Fs > 72.07, ps < .001, all η2p > .50, suggesting that reactive inhibition was evident across groups. Furthermore, the exclusion group demonstrated lower accuracy for both standard and deviant responses relative to the inclusion group, Fs > 5.82, ps < .018, all η2p > .07. Additionally, the exclusion group showed greater accuracy cost than the inclusion group (Figure 4A), suggesting that social exclusion impairs reactive inhibition.
Results of Experiment 3. (A)Accuracy (ACC) costs (i.e., the differences between standard and deviant responses) in the reactive condition and the intentional rates in the intentional condition; (B) RT costs in the reactive and intentional conditions; (C) Diffusion model analysis results in the intentional condition, for the exclusion and inclusion groups.
For RT, analyses revealed that both the exclusion and inclusion groups showed faster RTs for standard responses than for deviant ones, F(1, 73) = 328.77, p < .001, η2p = .82, F(1, 73) = 209.89, p < .001, η2p = .74, suggesting that reactive inhibition was evident across groups. Additionally, the exclusion group exhibited greater RT cost than the inclusion group, F(1, 73) = 5.87, p = .018, η2p = .07 (Figure 4B), suggesting that social exclusion impairs reactive inhibition.
The combined analyses of accuracy and RT suggest that experiencing long-term social exclusion impairs reactive inhibition.
Impact of experiencing long-term social exclusion on intentional inhibition
For inhibition rate, the t test revealed a significant difference between groups, t(73) = 2.02, p = .048, Cohen’s d = 0.47, with the exclusion group exhibiting lower inhibition rate compared to the inclusion group (Figure 4A), suggesting that social exclusion impairs intentional inhibition.
For RT, analyses revealed that both the exclusion and inclusion groups showed faster RTs for standard responses than for deviant ones, F(1, 73) = 32.34, p < .001, η2p = .31, F(1, 73) = 24.08, p < .001, η2p = .25, suggesting that intentional inhibition was evident across groups. However, there was no significant difference in RT costs between the two groups, F(1, 73) = 0.25, p = .622, η2p < .01 (Figure 4B).
For the diffusion model analysis, t tests revealed that the exclusion group had a larger boundary separation compared to the inclusion group, t(73) = 2.03, p = .046, Cohen’s d = 0.47. There was also a trend toward a lower drift rate in the exclusion group, although this difference did not reach conventional levels of significance, t(73) = 1.73, p = .088, Cohen’s d = 0.40 (Figure 4C). These results suggest that the exclusion group demonstrated a slower rate of evidence accumulation and adopted more conservative decision criteria when making decisions during the free-choice trials. No significant differences were observed between groups for starting point bias, t(73) = 0.37, p = .713, Cohen’s d = 0.09, or nondecisional RT, t(73) = 1.56, p = .123, Cohen’s d = 0.36.
The analyses of inhibition rate, starting point bias, and drift rate suggest that experiencing long-term social exclusion impairs intentional inhibition, and this impairment effect may be associated with a slower rate of evidence accumulation and more conservative response criteria in free-choice trials.
Impact of experiencing long-term social exclusion on perceived task difficulty and attentional resource allocation
For the difficulty of reactive trials (i.e., white square and white diamond), the t test failed to reveal any significant differences between the exclusion (M = 2.34, SE = 0.13) and inclusion groups (M = 2.24, SE = 0.10), t(73) = 0.60, p = .551, Cohen’s d = 0.14.
For the difficulty of free-choice trials (i.e., white circle), the t test revealed a significant difference between groups, t(73) = 2.11, p = .038, Cohen’s d = 0.49, with the exclusion group (M = 2.76, SE = 0.14) reporting higher difficulty compared to the inclusion group (M = 2.32, SE = 0.16).
For attentional resource allocation between reactive trials and free-choice trials, the t test revealed a marginal significant difference between groups, t(73) = 1.96, p = .054, Cohen’s d = 0.45, with the exclusion group (M = 3.26, SE = 0.15) deploying more resources to the free-choice trials compared to the inclusion group (M = 2.84, SE = 0.16).
General Discussion
The present study conducted three experiments to investigate the impacts of short- and long-term social exclusion on both reactive and intentional inhibition. Experiment 1 employed the Cyberball game to induce a short-term experience of exclusion, Experiment 2 utilized the Future Life Alone paradigm to evoke long-term imagined exclusion, while Experiment 3 used the Social Exclusion Questionnaire for Undergraduates to identify individuals who had experienced long-term exclusion in real-life contexts. Reactive and intentional inhibition were assessed using the Free Two-Choice Oddball Task. Consistent with the majority of our hypotheses, the study revealed three significant findings. First, both short- and long-term exclusion negatively affected intentional inhibition. Second, diffusion model analyses revealed distinct mechanisms underlying these impairments: short-term social exclusion increased the tendency toward standard responses during free-choice trials, while long-term exclusion led to more conservative response criteria. Third, while neither experiencing short-term exclusion nor imagining long-term exclusion had a significant impact on reactive inhibition, actual long-term exclusion was found to impair this function. Collectively, these findings suggest that although short- and long-term exclusion share some similarities in their effects on behavioral inhibition, they also exhibit notable differences.
Effects of Social Exclusion on Intentional Inhibition
Among the various potential mechanisms underlying self-control deficiencies in excluded individuals, a prominent hypothesis suggests that social exclusion undermines the capacity for behavioral inhibition (White et al., 2015). In line with this hypothesis, our findings reveal a significant impairment in intentional inhibition resulting from social exclusion. This effect is evident irrespective of the severity or duration of exclusion (whether short- or long-term) and the method of exclusion manipulation (whether experienced or imagined), as demonstrated by greater RT costs in both Experiments 1 and 2, and lower inhibition rate in Experiment 3, observed in the exclusion group compared to the inclusion one. The impairment effect of social exclusion, as previously mentioned in the introduction, may be attributed to the characteristics of intentional inhibition and the extensive effects of social exclusion on cognitive processing. In brief, excluded individuals tend to display greater self-avoidance behaviors (Malik & Obhi, 2019; Twenge et al., 2003). This is often accompanied by a lack of sufficient attentional resources and the utilization of inappropriate target representations (i.e., concrete mental representation due to low-level mindset; Bahrami & Borhani, 2023; Baumeister et al., 2002; Bruyneel & Dewitte, 2012; Lee et al., 2010; Park & Baumeister, 2015). Consequently, their performance in intentional inhibition, which is a dynamic and resource-intensive process that requires heightened self-engagement and abstract mental representation (i.e., high-level mindset; Filevich et al., 2012; Xu, Xu, et al., 2024), becomes impaired.
Importantly, diffusion model analyses revealed that the detrimental effects of exclusion induced by various methods—short-term exclusion experiences, imagined long-term exclusion, and actual long-term exclusion—are underpinned by distinct cognitive mechanisms. In Experiment 1, participants who experienced short-term exclusion via Cyberball exhibited a reduced starting point bias relative to those who were included, indicating an increased tendency toward generating standard responses during free-choice trials. In Experiment 2, following the imagining of long-term exclusion evoked by the Future Life Alone paradigm, the exclusion group demonstrated greater boundary separation compared to the inclusion group, suggesting the adoption of more conservative response criteria in free-choice trials. Moreover, as identified by the Social Exclusion Questionnaire for Undergraduates in Experiment 3, individuals who had experienced real-life long-term exclusion showed a marginally lower drift rate and significantly greater boundary separation compared to those who had not experienced exclusion, indicating a slower rate of evidence accumulation and more conservative response criteria in free-choice trials.
As far as we know, this study represents the first attempt to employ diffusion model analysis in investigating the impact of different types of exclusion on intentional inhibition. Given the lack of prior analogous findings, we propose a potential explanation for the distinct effects observed in short- versus long-term exclusion scenarios. Specifically, this difference may be attributed to individuals’ varying goal priorities in response to different types and levels of exclusion. Considering that long-term exclusion is more severe than short-term one, we infer that individuals who had experienced short-term exclusion may have entered the reflective stage and prioritized restoring social inclusion as a primary goal (Williams, 2009). Therefore, they likely complied with experimenter requirements (which can serve as potential sources of acceptance) and allocated more attentional resources to enhance task performance in reactive trials (i.e., reactive standard and deviant trials), where there was a clear right or wrong answer compared to the free-choice task (Shilling & Brown, 2016; Smart Richman & Leary, 2009). As a result, excluded individuals may have developed a habit of relying on external instructions, which could have reduced their flexibility in transitioning from relying on external stimuli to internal decision-making when faced with a task that allows for free choice (particularly if their capacity for updating and cognitive flexibility were diminished; Ma & Pang, 2023; Mortezazadeh et al., 2021). Consequently, they tended to exhibit a similar and strong tendency toward generating standard responses during the free-choice task as during the reactive task. Conversely, we infer that individuals who imagined and experienced long-term exclusion may entered the resignation stage, wherein they no longer actively sought social inclusion and prioritized avoiding painful future exclusion as their primary goal (Büttner et al., 2024; Williams, 2007). Additionally, given the inherent ambiguity in the response criteria for free-choice trials (where there was no definitively correct response but rather a requirement for “roughly equal” distribution of different responses), and considering that excluded individuals tend to perceive ambiguous stimuli as potential threats (DeWall et al., 2009; Restrepo et al., 2024), these individuals may have processed such stimuli with heightened caution. This hypothesis was partially supported by the subjective assessments of excluded individuals in Experiment 3, which indicated higher perceived task difficulty and increased allocation of attentional resources during the free-choice task. Consequently, to avoid reexperiencing social exclusion due to potential errors, these individuals may have adopted a more cautious and conservative approach to tasks, thereby raising their decision-making threshold. Moreover, actual experiences of long-term exclusion may have depleted attentional resources to a greater extent than imagined scenarios, leading to slower rate of evidence accumulation during decision-making in free-choice trials.
Effects of Social Exclusion on Reactive Inhibition
In addition to the primary findings on intentional inhibition, the results from three experiments suggested that only prolonged and intense experiences of social exclusion significantly impair reactive inhibition. This conclusion is supported by the greater accuracy and RT costs observed in the exclusion group compared to the inclusion group exclusively in Experiment 3, which utilized an exclusion questionnaire to identify individuals who had experienced long-term exclusion. In contrast, no significant differences were found between the exclusion and inclusion groups in Experiment 1, where a Cyberball game was used to induce short-term exclusion, or in Experiment 2, where the Future Life Alone paradigm was employed to evoke imagined long-term exclusion. These findings are partially consistent with previous studies (Otten & Jonas, 2013; Xu et al., 2016), which demonstrated that short-term exclusion does not impact reactive inhibition. The effects of different forms of exclusion appear to be correlated with the severity and duration of the exclusion manipulated by various methods. Specifically, the Cyberball game induced only mild and transient short-term exclusion, while individuals identified by the exclusion questionnaire had endured prolonged and intense real-life social exclusion. Furthermore, despite the utilization of the Future Life Alone paradigm to manipulate participants’ imagining of long-term social exclusion (Ghandchi et al., 2024; Wirth, 2016), imagining such exclusion may not fully replicate the actual experience of enduring prolonged exclusion, and its severity might only be mild to moderate. Therefore, although self-regulation of exclusion-related negative feelings consumes certain attentional resources (Chester & DeWall, 2014; Riva et al., 2015), individuals who experienced short-term exclusion or imagined long-term exclusion may still have retained sufficient attentional resources (Otten & Jonas, 2013; Xu et al., 2016). However, after experiencing prolonged exclusion, the available resources for subsequent tasks may have become depleted (Williams, 2009), leading to impaired reactive inhibition.
Short- and Long-Term Social Exclusion Differently Impact Intentional and Reactive Inhibition
The subsequent question that merits further exploration is the mechanism through which short- and long-term social exclusion differentially affect intentional and reactive inhibition. Specifically, only prolonged and intense experiences of exclusion hindered reactive inhibition, while both short- and long-term exclusion impeded intentional inhibition. These differential effects can be attributed to two key factors. First, intentional inhibition requires greater self-engagement compared to reactive inhibition, making it more susceptible to resource limitations (Baumeister et al., 2002) and the detrimental impacts of social exclusion on one’s sense of self (Malik & Obhi, 2019; Twenge et al., 2003). Second, social exclusion induces a low-level construal mindset, characterized by a focus on detailed aspects from a short-term or proximal perspective (Bruyneel & Dewitte, 2012; Lee et al., 2010; Park & Baumeister, 2015). This low-level mindset impairs intentional inhibition, which necessitates a long-term or distal perspective (i.e., evaluating whether forthcoming actions align with long-term interests), but may enhance reactive inhibition, which requires close attention to immediate environmental features (e.g., low-probability deviant trials; Schmeichel et al., 2010; Xu, Xu, et al., 2024). Consequently, under mild to moderate exclusion conditions, as observed in Experiments 1 and 2, the impairment effect due to resource limitations and the improvement effect from low-level construal offset each other, resulting in null influences on reactive inhibition. However, under intense exclusion conditions, as demonstrated in Experiment 3, the impairment effect outweighs the improvement one, leading to impaired reactive inhibition.
Theoretical, Methodological, and Practical Implications
To the best of our knowledge, this study represents the first systematic investigation into the effects of both short- and long-term social exclusion on reactive and intentional inhibition. Our three experiments revealed that while short- and long-term exclusion share certain similarities in their impact on behavioral inhibition, they also exhibit distinct differences. These findings hold significant implications from theoretical, methodological, and practical perspectives. At the theoretical level, by elucidating the general impairment effect of both forms of exclusion on intentional inhibition, and the specific impairment effect of actual long-term exclusion on reactive inhibition, this research resolves previous controversies regarding the impact of social exclusion on behavioral inhibition, and partially validates the hypothesis that deficits in behavioral inhibition underlie the negative impact of social exclusion on self-control (White et al., 2015). Moreover, by demonstrating the similar yet distinct effects of different forms of social exclusion on reactive and intentional inhibition (i.e., short-term exclusion experiences induced by the Cyberball game, imagined long-term exclusion evoked by the Future Life Alone paradigm, and actual long-term exclusion experienced in real-life settings), the current findings enhance our understanding of the nuances among various exclusion paradigms and, in conjunction with recent studies, indicate the necessity for further investigation to compare diverse types of exclusion (Rajchert et al., 2023; Rudert et al., 2019; Wesselmann et al., 2023). At the methodological level, the findings lend support to Xu, Wen, et al.’s (2024) proposal, indicating that the Free Two-Choice Oddball Task offers distinct advantages in examining intentional inhibition because of its capacity to provide both an index of the inhibition rate and a more sensitive index of response time. Moreover, the findings also suggest that diffusion model analysis could yield more comprehensive insights than traditional analyses, thereby advocating its adoption within the realm of social exclusion studies (White & Kitchen, 2022). Finally, at the practical level, they offer a novel direction for designing targeted interventions aimed at alleviating social exclusion’s adverse consequences. A potential avenue for future research is to specifically target intentional inhibition and explore the effectiveness of cognitive training or brain stimulation as an intervention to mitigate the detrimental effects of social exclusion (Ma & Pang, 2023; Smith et al., 2019).
Limitations and Future Research
Despite its strengths, some limitations of this study should be acknowledged. First, although the Free Two-Choice Oddball Task was designed to assess intentional inhibition, the intentional inhibition demonstrated by participants in this task was, to some extent, a response to the experimenter’s request. Therefore, it may differ from intentional inhibition observed in real-life scenarios (e.g., voluntarily resisting inappropriate impulses to eat when having the goal of losing weight). Future studies should design more sophisticated and nuanced experiments to address these issues. Second, due to the inclusion of only female participants in this study, our findings may not be directly generalizable to male participants. While our decision was informed by previous research indicating that women may experience social exclusion more acutely (Benenson et al., 2013), future studies should aim to include both male and female participants to facilitate comparative analyses. Third, despite observing impairments in both intentional and reactive inhibition due to social exclusion, our study did not concurrently assess other self-control indicators such as aggressive behavior. Although Experiment 3 included an intertemporal choice task (Rachlin & Jones, 2008), we did not detect a significant impairment effect of exclusion on this indicator (see Supplemental Material). Consequently, we were unable to comprehensively evaluate the mediating role of behavioral inhibition, which would have provided robust support for the hypothesis that deficits in behavioral inhibition underlie the adverse effects of social exclusion on self-control. Therefore, we recommend that future research conduct comprehensive and direct tests of this hypothesis. Fourth, although one of the primary objectives of this study was to investigate the differential effects of short- and long-term exclusion on behavioral inhibition, these two forms of exclusion were not manipulated simultaneously within a single experiment. As a result, the conclusions regarding their distinct impacts should be interpreted with caution. Future research is encouraged to develop novel exclusion paradigms that allow for the simultaneous manipulation of both short- and long-term exclusion within the same experimental design, enabling a more direct and rigorous comparison of their effects (Bernstein & Claypool, 2012a; Ghandchi et al., 2024). Fifth, although this study frames long-term exclusion as more severe than short-term exclusion to explain differences across experiments, participants reported greater threats to their basic needs in Experiment 1 (Cyberball) than in Experiment 2 (Future Life Alone). This pattern, consistent with previous findings, may reflect numbness from prolonged exclusion or differing trajectories of emotional distress over time (see Supplemental Material; Bernstein & Claypool, 2012a; Ghandchi et al., 2024). Nonetheless, we encourage future research to directly manipulate exclusion intensity and assess responses at multiple time points to more accurately evaluate the explanatory role of severity.
Supplemental Material
sj-docx-1-gpi-10.1177_13684302251372904 – Supplemental material for Similar yet distinct effects of short-term and long-term social exclusion on behavioral inhibition
Supplemental material, sj-docx-1-gpi-10.1177_13684302251372904 for Similar yet distinct effects of short-term and long-term social exclusion on behavioral inhibition by Mengsi Xu, Jiayu Wen and Zhiai Li in Group Processes & Intergroup Relations
Footnotes
Data Availability Statement
The data that support the findings of this study are available from the corresponding authors upon reasonable request.
Ethical Approval and Informed Consent Statements
All studies were conducted with institutional IRB (School of Psychology at Shaanxi Normal University) approval, granted on September 5, 2022 under Reference No. HR 2022-09-001, and written informed consent was obtained before conducting the experiment.
Funding
The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This work was supported by the Science and Technology Innovation (STI) 2030 Major Projects (2021ZD0200500), the National Natural Science Foundation of China (32100871), the Social Science Foundation of Shaanxi Province (2021P008), the China Postdoctoral Science Foundation (2021M702064), the 2024 Shaanxi Provincial Education Science “14th Five-Year Plan” Project (SGH24Y2212), and the Shaanxi Normal University Graduate Education and Teaching Reform Research Project (GERP-24-26).
Supplemental Material
Supplemental material for this article is available online.
Notes
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.
