Violent behaviors in higher education are a focus of international research (Itani et al., 2017; Dobarro et al., 2018; Martínez-Martínez et al., 2018; Esquivel-Santovena et al., 2020; Alsawalqa, 2021). Versatility and adaptability of violence in its physical, psychological or social forms is a challenge to the educational community in general. Recently, it has been possible to see how higher education institutions have deteriorated in terms of coexistence, which has a negative impact on academic performance, engagement, and physical as well as psychological health, even causing students drop out of their studies (Bernardo et al., 2020; MacCann et al., 2020; Martínez-Monteagudo et al., 2020).

In recent years, violent behaviors have been adapted to new human interaction methodologies. Cyberbullying is one of the most widespread consequence in higher education (Watts et al., 2017; Alsawalqa, 2021) as well as mobile phone abuse (Polo et al., 2017; Meldrum et al., 2020). On the other hand, it is found that this kind of dissocial behavior is directly related to violence in all its variants and suicide (Washington, 2015; Vergel et al., 2016; Hinduja and Patchin, 2019; Martín-Baena et al., 2019).

One of the contexts where violent behaviors develop is the academic environment. The study conducted by Xu and Fang (2021) establishes that one of the reasons for these behaviors in the academic environment lies in the degree of achievement in different academic tasks. Such behavior is based on the theory of self-esteem orientation, stating that when subjects cannot obtain a sufficient degree of self-esteem support, they seek other ways to improve their self-esteem (Yiyi et al., 2022). When a person does not achieve the expected academic performance, a decrease in self-esteem is carried out, increasing self-esteem levels through violent behavior toward those who obtain better grades (Xiong et al., 2020).

Furthermore, these behaviors have a negative impact on both physical and mental health (Reisig and Golladay, 2019; Choe et al., 2021). In this situation, emotions play a key role in the appearance and development of violent behaviors (MacCann et al., 2020; Ubago-Jiménez et al., 2021) since they are a checkpoint for controlling one’s own impulses and the self-regulation of the person over himself (Bacon et al., 2014; Beltrán-Catalán et al., 2018; Sorina et al., 2019).

Regarding the importance of violent behaviors at the university setting, and given that little has been done on the interplay between emotional intelligence, self-concept, and academic performance as predictors of violent behaviors, the present study aims to identify and establish the relationship between academic performance, self-concept and violent behavior as a function of academic record average among Spanish university students, by (a) developing an explanatory model of violent behavior and its relationship with self-concept and emotional intelligence and (b) contrasting the structural equation model using a multi-group analysis according to academic record average.

In the model developed through variables assessed for the sample obtaining an average of pass marks (5–6.99), a good fit has been shown for the different indices. The Chi-square analysis showed a significant value of p (X² = 297.755; df = 32; pl = 0.000), but these data cannot be interpreted independently due to the influence of susceptibility and sample size (Tenenbaum and Eklund, 2007), so other standardized fit indices have been used. The comparative fit index (CFI) analysis obtained a value of 0.933, the normalized fit index (NFI) analysis obtained a value of 0.905, the incremental fit index (IFI) was 0.961 and the Tucker-Lewis index (TLI) obtained a value of 0.922, all of which were excellent. In addition, the root mean square error of approximation analysis (RMSEA) obtained a value of 0.032.

Figure 2 and Table 1 show the regression weights of the theoretical model, with statistically significant differences at p < 0.05 and p < 0.001. In relation to self-concept, positive relationships are observed with AS (r = 0.470), SS (p < 0.001; r = 0.781), FS (p < 0.001; r = 0.668), PS (p < 0.001; r = 0.471) and ES (p < 0.001; r = 0.276). Continuing with emotional intelligence, positive relationships are observed with ER (r = 0.474), EC (p < 0.001; r = 0.931) and EA (p < 0.001; r = 0.375). Looking at direct violence, it is observed that it shows a negative relationship with self-concept (r = −0.066) and emotional intelligence (r = −0.159). Finally, looking at indirect violence, positive relationships are observed with self-concept (p < 0.05; r = −0.096) and negative relationships with emotional intelligence (r = −0.051).

Furthermore, the model developed for students with an average of notable showed good scores for each of the different indices. The Chi-square analysis showed a significant value of p (X² = 231.490; df = 32; pl = 0.000). The comparative fit index (CFI) analysis obtained a value of 0.945, the normalized fit index (NFI) analysis obtained a value of 0.920, the incremental fit index (IFI) was 0.941 and the Tucker-Lewis index (TLI) obtained a value of 0.909, all of which were excellent. In addition, the root mean square error of approximation analysis (RMSEA) obtained a value of 0.039.

In Figure 3 and Table 2, positive relationships with AS (r = 0.069), SS (p < 0.001; r = 0.779), FS (p < 0.001; r = 0.646), PS (p < 0.001; r = 0.538) and ES (p < 0.001; r = 0.271) are observed. Following emotional intelligence, positive relationships were found with ER (r = 0.603), EC (p < 0.001; r = 0.739), EA (p < 0.001; r = 0.341) and self-concept (p < 0.05; r = 0.184). For the direct violence, positive relationships are observed with self-concept (r = 0.041) and emotional intelligence (r = 0.010). Finally, for indirect violence, negative relationships were found with self-concept (p < 0.001; r = −0.282) and emotional intelligence (p < 0.001; r = −0.187).

The last model developed showed good scores for each of the different indices. The Chi-square analysis showed a significant value of p (X² = 229.431; df = 32; pl = 0.000). The comparative fit index (CFI) analysis obtained a value of 0.963, the normalized fit index (NFI) analysis obtained a value of 0.919, the incremental fit index (IFI) was 0.929 and the Tucker-Lewis index (TLI) obtained a value of 0.907, all of which were excellent. In addition, the root mean square error of approximation analysis (RMSEA) obtained a value of 0.043.

As shown in Figure 4 and Table 3, for self-concept, positive relationships are observed with AS (p < 0.001; r = 0.549), SS (p < 0.001; r = 0.762), FS (p < 0.001; r = 0.687), PS (p < 0.001; r = 0.445) and ES (p < 0.05; r = 0.313). In terms of emotional intelligence, positive relationships were obtained with ER (r = 0.264), EC (r = 0.901), EA (p < 0.05; r = 0.160), and self-concept (p < 0.001; r = 0.057). For direct violence, it shows negative relationships with self-concept (p < 0.001; r = −0.497) and emotional intelligence (p < 0.005; r = −0.104). Finally, for indirect violence, a negative relationship is obtained for self-concept (r = −0.371) and emotional intelligence (p < 0.05; r = −0.023).

This study showed the relationships between academic performance, self-concept, emotional intelligence, and violent behavior in undergraduates. The results obtained answered the aims stated in the research. Likewise, recent studies have been found that present similar characteristics and relate the research variables (Molero et al., 2014; Elipe et al., 2015; García-Martínez et al., 2022).

Based on a first approach to the comparative analysis, the data obtained show relationships between self-concept, emotional intelligence and violent behavior (Baroncelli and Ciucci, 2014; Estévez et al., 2019; Cañas et al., 2020; Skrzypiec et al., 2021). In fact, it can be observed that university students with an average of outstanding marks have a negative relationship between their self-concept and the two kinds of violence. Similar results are found in the meta-analysis of Samara et al. (2021) who analysed studies carried out in children and adolescents. They also found that students with lower school performance are those who tend to be victims of violent behavior. On the other hand, Ma et al. (2009) found that lower academic performance was associated with high levels of disruptive behavior in adolescents. Moreover, numerous studies confirm the relationship between school bullying and lower grades (Román and Murillo, 2011; Sinkkonen et al., 2014; Droppert et al., 2019; Pörhölä et al., 2019). The results showed that undergraduates with a pass average mark are more likely to engage in a type of indirect violence such as cyberbullying (Kowalski and Limber, 2013; Young-Jones et al., 2015; Martínez-Ramón et al., 2019).

The research carried out by Kowalski and Limber (2013) states that people who perform poorly academically tend to show behavior based on physical or virtual bullying toward those who perform better academically. Furthermore, the research carried out by Martínez-Ramón et al. (2019) states that the act of bullying within the university environment is undergoing a process of evolution, as it is moving from physical violence to covert violence, which is based on isolating and underestimating the victim. This type of academic bullying leads to a reduction in basic academic skills, producing a reduction in academic motivation levels and giving rise to an increase in anxiety levels (Young-Jones et al., 2015).

In addition, it is also observed that emotional intelligence correlates negatively with both forms of violence for students with a pass average mark and a notable average mark. Our findings give credence to Elipe et al. (2015) who confirm how students show a better emotional adaptation to different kinds of violence.

However, it has been found that undergraduates with a higher average mark have a negative correlation between their emotional condition and direct and indirect violence. Perhaps this could be caused by a higher stress exposure due to a higher academic performance (Lee, 2017). Also, their desire to have a higher average mark makes students more competitive, which causes them to present emotions such as anger, annoyance or despair (Elipe et al., 2015). The above data support previous investigations that have shown how suffering any type of school violence is associated with low academic performance (Ramírez-Granizo et al., 2019; Torres et al., 2019; Tekel and Karadag, 2020).

According to the above, it is worth highlighting that the present study shows the relationship between self-concept, emotional intelligence, violence behaviors and their impact on academic performance. Therefore, emotional education should be promoted in order to develop self-control in anxiety and stressful situations to prevent violent behavior. It is also suggested that teachers need to increase their teacher-student interpersonal behaviors, by drawing on the tenets of positive psychology and rhetorical/relational goal theory (Xie and Derakhshan, 2021). Likewise, it is also necessary to strengthen and reward efforts to obtain a better academic performance in order to prevent emotional problems and adverse emotions such as anger, boredom, and violence.

The original contributions presented in the study are included in the article/supplementary material, further inquiries can be directed to the corresponding author.

The studies involving human participants were reviewed and approved by Department of Corporal Expression and Ethics Committee of the Faculty of Education Sciences of University of Granada (Spain) with the code 1478/CEIH/2020. Written informed consent for participation was not required for this study in accordance with the national legislation and the institutional requirements.

JU-J, SC-R, JO-M, and EM-I: idea, state of the art, methodology, and data analysis, results, discussion and conclusions, and final revision. JU-J, SC-R, and JO-M: literature review (state of the art). JU-J and EM-I: original draft. All authors contributed to the article and approved the submitted version.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.