Aggression — both in general and in its gendered manifestations — remains an underrecognized social and public health issue (see, for example, Gomez-Pulido et al., 2024; Vinagre-González et al., 2023). For this reason, it is important to analyze variables related not only to perpetrators (Bonilla-Algovia and Rivas-Rivero, 2022; Expósito-Álvarez et al., 2024) but also to victims (Galera et al., 2023; Marcos et al., 2024; Medinilla-Tena et al., 2024; Molero et al., 2023; Ríos Lechuga et al., 2024). Aggressive behaviors — regardless of their lethality or severity — are frequently reported in correctional settings (Reidy et al., 2016; Safranoff and Kaiser, 2020). Among the wide range of institutional misconduct, those involving self-harm (for a review, see Favril et al., 2020) or harm toward others (for a review, see Schenk and Fremouw, 2012) are particularly concerning for prison management. This problem is especially pronounced in correctional systems within lower-middle-income countries, where limited security infrastructure compounds the risks (Oficina de las Naciones Unidas contra la droga y el delito [UNODC], 2020). Thus, the high prevalence of aggression in these settings has prompted a growing body of research into its predictors and consequences, with the aim of informing security level classification, treatment planning, and recidivism prevention strategies for both male and female incarcerated populations (Gilbert et al., 2015b; Steiner and Wooldredge, 2013). Regardless of its origin or underlying purpose, aggression is best understood as a multidimensional construct encompassing a continuum of behaviors that vary in terms of planning, intent, and lethality. Although numerous psychological theories address the underpinnings of human aggression, the absence of a universally accepted definition of hostile or harmful behavior has led to the development of integrative frameworks. Among them, Riber and Treiber's (1989) psychometric model is particularly influential. This model conceptualizes aggression according to three dimensions: 1) expression of anger/hostility, 2) expression of verbal/adaptive anger, and 3) expression of maladaptive/physical anger. These dimensions are intended to capture the relative stability and magnitude of aggressive tendencies within the personality structure, serving classification, clinical, and predictive purposes. This model underpins the development of the multiscale self-report questionnaire known as the Personality Assessment Inventory (PAI; Morey, 2012; Ortiz-Tallo et al., 2012, 2013).

The PAI (PAI; Morey, 2012; Ortiz-Tallo et al., 2013) has emerged as a comprehensive and widely validated self-report instrument designed for use in diagnostic and treatment planning across general, clinical, forensic, and correctional populations (for a review, see Paulino et al., 2024). The instrument comprises 11 clinical scales (Somatic Complaints, SOM; Anxiety, ANX; Anxiety-Related Disorders, ARD; Depression, DEP; Mania, MAN; Paranoia, PAR; Schizophrenia, SCZ; Borderline Features, BOR; Antisocial Features, ANT; Alcohol Problems, ALC; and Drug Problems, DRG). Aligned with Riber and Treiber's (1989) psychometric model of aggression, the PAI also includes the Aggression scale (AGG) within its set of treatment consideration scales, which additionally encompass Suicidal Ideation (SUI), Stress (STR), Nonsupport (NON), and Treatment Rejection (RXR). The AGG scale assesses dispositional tendencies related to anger, assertiveness, hostility, and overt aggression, and is subdivided into three targeted subscales: Aggressive Attitude (AGG-A), Verbal Aggression (AGG-V), and Physical Aggression (AGG-P). Specifically, the AGG-A subscale assesses cognitive and affective dimensions such as chronic hostility, poor anger regulation, and beliefs about the instrumental use of aggression. The AGG-V measures the frequency and intensity of verbal expressions of anger, while the AGG-P measures the tendency to engage in physical aggression, including direct acts of violence and threats. In addition, the PAI includes interpersonal scales — Dominance (DOM) and Warmth (WRM) — bringing the total number of core scales to 18.

Scientific evidence supporting the psychometric properties of the PAI aggression-related measures in prison populations has primarily focused on male inmates, since most studies evaluating the ability of these scales to discriminate or predict aggressive or misconduct-related behaviors have utilized male samples (Newberry and Shuker, 2012). For instance, in a large sample of incarcerated men convicted of common offenses such as robbery, murder, drug possession, and rape, BOR and ANT emerged as robust and consistent predictors of all three aggression subscales—AGG-A, AGG-V, and AGG-P (Burneo-Garcés et al., 2021). In contrast, a meta-analysis by Gardner et al. (2015) reported that, in female inmate samples, ANT demonstrated stronger and more moderate predictive validity for institutional misconduct than AGG, a pattern opposite to that observed in males. A study conducted with a mixed-gender sample further supports the incremental and predictive validity of both ANT and AGG in relation to general and aggressive disciplinary infractions, surpassing the predictive capacity of other scales such as PAR, BOR, and MAN (Reidy et al., 2016). Similarly, Davidson et al. (2016) found that, in a female prison sample, AGG, ANT, and PAR were most strongly associated with both general and assaultive disciplinary infractions. For male inmates, higher scores on AGG-A, AGG-V, and AGG-P were linked to increased infractions and high scores on ANT. Among female inmates, however, AGG and DOM were linked to the greater presence of aggressive and violent behaviors, along with elevated scores on BOR and PAR (Edens et al., 2005; Reidy et al., 2016).

While aggression is a phenomenon observed across genders, research in this field has historically concentrated on male prison populations. Even when female participants are included, they are often underrepresented or comprise a non-representative portion of the sample (Giordano and Copp, 2018; Messina and Calhoun, 2021). This imbalance is even more pronounced in Spanish-speaking prison environments, where empirical data remain scarce (Burneo-Garcés and Pérez-García, 2018). As a result, there is a lack of conclusive evidence regarding gender-based similarities and differences in psychological, psychopathological, and criminogenic variables within incarcerated populations (Guerino et al., 2011; Steiner and Wooldredge, 2013). In the case of female inmates, the literature has primarily focused on the underlying causes of aggression, such as prior trauma or psychosocial vulnerabilities, rather than systematically examining their aggression profiles or patterns of recidivism, despite the interconnected nature of these outcomes. Indeed, most studies have emphasized explanatory factors (Bouffard et al., 2022; Gover et al., 2008), including personal needs, institutional victimization, levels of social support, emotional and familial dynamics, and histories of childhood trauma and abuse (for a review, see Fedock and Murray, 2024).

In any case, generalizing findings from male samples to the female prison population — or making decisions in the absence of gender-specific and methodologically sound evidence (e.g., regarding correlates of disruptive behavior, prior prison terms, classification, or risk levels) is not advisable (Gover et al., 2008). The current state of the literature clearly highlights the need for more research on aggression in both the male and female prison populations to generate conclusive data. Such evidence is essential for the design of psychological assessment protocols, the classification of security levels, and the implementation of comprehensive rehabilitation strategies tailored to the specific characteristics and needs of each group (Gover et al., 2008; Prieto-Macías et al., 2020; Steiner and Wooldredge, 2013). Evidence-based strategies will likely have even greater social impact in countries with under-resourced correctional systems, substandard living conditions, and insufficient institutional safeguards (Oficina de las Naciones Unidas contra la droga y el delito UNODC, 2020). Consequently, the present study aimed to analyze aggression-related features (AGG, AGG-A, AGG-V, and AGG-P) by gender and security level in a sample of incarcerated men and women from three correctional centers in Guayaquil(Ecuador). The Spanish version of the PAI (Ortiz-Tallo et al., 2013) was employed for this purpose. Although the literature on aggression in female prison populations remains limited from a psychometric standpoint, prior findings support the hypothesis that women may score higher than men on AGG-A and AGG-V (Camlibel et al., 2021; Rueda et al., 2006; Smith et al., 2020; Steiner and Wooldredge, 2013).