Why People Break Away From the Groups They Would Die For

Some people feel so deeply bound to a group that they describe it as an extension of themselves. Psychologists call this state identity fusion: a visceral sense of oneness in which the personal self and the group identity merge into a single, synergistic whole. Fusion is among the strongest known predictors of willingness to fight, die, or make other costly sacrifices for a collective, and it has been documented in contexts ranging from terrorist organizations and street gangs to sports ultras and nation-states. Yet a crucial question has remained largely unexplored: what makes fused people loosen that bond? A new five-study investigation published in Frontiers in Social Psychology provides the most systematic answer to date, showing that the reasons people imagine for breaking away differ sharply from the reasons that actually drive them out.

The research team, led by Ángel Gómez of the Universidad Nacional de Educación a Distancia in Madrid together with colleagues in Spain, the United States, and Sri Lanka, distinguished between two kinds of defusion. Hypothetical defusion refers to reasons people believe could, in the future, dissolve their feeling of unity with a group. Real defusion refers to the motives that genuinely caused such a rupture in the past. The team also asked whether the mechanisms observed in ordinary groups extend to groups that have exercised collective violence, examining both current members of a Spanish hooligan firm and former members of the Liberation Tigers of Tamil Eelam, the insurgent organization that fought for a separate Tamil homeland in Sri Lanka until its defeat in 2009.

The methodological approach was deliberately open-ended. Rather than presenting participants with a fixed menu of options, the researchers used a qualitative strategy grounded in interpretative phenomenology, asking people to spontaneously generate reasons and then coding those responses through a three-cycle process with independent judges. This design allowed unexpected categories to emerge. Across the studies, three main reasons surfaced: degrading relational ties, meaning deteriorating sentiments toward individual group members; degrading collective ties, meaning deteriorating sentiments toward the group as a whole; and external factors, circumstances beyond the individual’s control. A fourth, unplanned category also appeared: the unfeasibility of defusion, in which participants denied that any complete detachment from the group was even possible.

Study 1 recruited 223 Spanish adults and asked them to think of a group with which they felt maximally fused, either a local group such as family or friends, where members know one another directly, or an extended group such as their city or country, where most members never meet. Participants confirmed strong fusion on a verbal scale, then wrote three reasons that could break that connection. Relational tie degradation dominated, cited by roughly 51 percent of participants, far ahead of external factors at about 27 percent, collective ties at about 20 percent, and unfeasibility at under 3 percent. Statistical comparisons using McNemar’s tests confirmed that relational ties were mentioned significantly more often than every other category.

The group’s nature mattered, but only in specific ways. People fused with local groups cited relational problems—lack of unity, bad atmosphere, selfishness, insincerity—more often than those fused with extended groups, while those thinking about extended groups more often pointed to clashes with the group’s values, principles, or objectives. This pattern fits the theoretical logic: in small face-to-face groups, interpersonal bonds are the glue, whereas in large, abstract collectives, shared ideas carry more weight. Within relational degradation itself, problems among group members were reported more frequently than problems between the individual and the group, such as feeling betrayed or personally devalued.

Study 2 scaled the design up to 568 participants and compared ascribed groups, which people belong to by birth or assignment, such as family or nationality, with chosen groups, such as friends or a sports club. The hierarchy of reasons held steady: relational ties first, external factors second, collective ties third. Notably, participants fused with ascribed groups were significantly more likely to say defusion was unfeasible, a finding consistent with the material and psychological difficulty of abandoning a group one never chose to join. Overall fusion levels were also higher for ascribed groups than for chosen ones.

Study 3 shifted from imagination to lived experience. Of 432 participants, 58 percent could not recall any group from which they had ever defused—striking evidence of fusion’s resilience. The remaining 181 participants described real defusions, overwhelmingly from local groups like friendships, families, and workplaces. Here the picture changed in an important way: relational tie degradation was cited by 66 percent, but external factors rose to nearly 65 percent, statistically indistinguishable from relational ties. Jobs ending, geographic distance, lack of time, and other uncontrollable circumstances proved far more powerful in reality than participants had anticipated when imagining hypothetical futures.

The two exploratory studies then tested whether these mechanisms extend to violent collectives. Study 4 surveyed 30 members of the Riazor Blues, a well-known hooligan group attached to a Spanish soccer club. All participants were fused with the group, and relational tie degradation again emerged as the leading hypothetical reason for defusion, mostly framed as internal conflict and lack of brotherhood among members. But a striking 33 percent of hooligans left the reason fields blank or denied that any reason existed, treating defusion as unfeasible—a far higher rate than in the general population samples. The authors suggest that for members of violent groups, merely imagining detachment may be difficult, painful, or perceived as risky, given the possibility of retaliation.

Study 5 broke new ground by interviewing 66 former LTTE members in community reintegration centers in Sri Lanka. To reduce social desirability bias, interviewers used third-person language, asking what key reasons had encouraged other members to abandon the organization. The results inverted the pattern seen everywhere else: external factors dominated, cited by 54 percent of respondents, with the end of the war itself the single most common reason, followed by family problems and the sheer unsustainability of the situation. Relational and collective tie degradation trailed at roughly 24 and 21 percent respectively. Notably, within relational degradation, problems between the individual and the organization outweighed problems among members, plausibly reflecting the LTTE’s harsh training practices, coercive control, and severe punishments for those suspected of disloyalty.

The overall lesson is a tension between the hypothetical and the real. When people imagine breaking away, they foreground the quality of their relationships; when they actually break away, forces beyond their control often do the work, especially in violent contexts where exit costs are high and ideology is enforced. The authors caution that their findings, drawn largely from Spanish samples and exploratory fieldwork, should not be overgeneralized, and that mention frequency does not necessarily reflect causal weight. Still, the practical implications are concrete. Interventions aimed at disengaging people from harmful groups might target relational dynamics in tightly knit local collectives, while questioning shared belief systems may work better in ideologically structured organizations. And for former insurgents, the study suggests that external openings—war’s end, family pressure, unbearable conditions—may matter more than any internal disillusionment, a nuance that could reshape how deradicalization programs are designed.

Subject of Research: Reasons for identity defusion from ordinary and violent groups

Article Title: Breaking social bonds: from hypothetical to real reasons for defusion

Article References: Gómez, Á., Martínez, M., López-Rodríguez, L., Vázquez, A., & Hettiarachchi, M. (2026). Breaking social bonds: from hypothetical to real reasons for defusion. Frontiers in Social Psychology, 4, Article 1787919. https://doi.org/10.3389/frsps.2026.1787919

Image Credits: AI Generated

DOI: 10.3389/frsps.2026.1787919

Keywords: identity fusion, defusion, social psychology, collective violence, relational ties, collective ties, deradicalization, hooligans, LTTE, group identity, extremism, disengagement