Profile of the Victimized Aggressors in Child-to-Parent Violence: Differences According to the Type of Victimization.

Background/Objective: One of the main predictors of child-to-parent violence (CPV) is childhood victimization. Recent research indicates the need to study different types of CPV aggressors. However, the distinctive characteristics of the profile of the victimized aggressor and whether these characteristics differ according to the type of victimization have not been yet analyzed. Were examined differences between four types of CPV aggressors: with family victimization, with school victimization, with polyvictimization, and without victimization experiences. Method: A total of 1,559 Spanish adolescents aged between 12 and 18 years participated.
Results: Compared to nonvictimized aggressors, victimized aggressors generally exercise more reactive and instrumental CPV and show more insecure parental attachment and less emotional and coping competencies. Additionally, among the types of victimization, polyvictimized aggressors show worse adjustment compared to those with a unique type of victimization. There are also significant differences according to the gender of the aggressor; however, the interaction effect between the type of aggressor and gender is not significant. Conclusions: Considering the profile of the victimized aggressor and the type of victimization experienced in CPV can provide valuable empirical information for the approach of differential explanatory mechanisms and for the design of prevention and intervention strategies adapted to the needs of this profile.

Child-to-parent violence (CPV) is a type of family violence defined as “any act of a child that is intended to cause physical, psychological or financial damage to gain power and control over a parent" (Cottrell, 2001, p. 3) that is conscious, intentional, and repeated over time (Molla-Esparza & Aroca-Montolío, 2018; Pereira et al., 2017). This complex social problem involves numerous and serious consequences that lead to a significant deterioration of family health and well-being. In terms of its magnitude, the prevalence of psychological violence ranges between 45% and 92%, physical violence ranges between 5.5% and 21%, and financial violence is approximately 59% (Beckmann et al. al., 2017; Calvete et al., 2015; Cano-Lozano, León, & Contreras, 2021; del Hoyo-Bilbao et al., 2018; Margolin & Baucom, 2014).

Research on CPV has increased substantially in recent years (see review by Simmons et al., 2018), with childhood victimization being one of the risk factors with the greatest empirical support. A high percentage of minors who assault their parents have a history of family victimization. A study of adolescents with CPV offenses (Nowakowski-Sims & Rowe, 2017) found that 25% had experienced direct family victimization (or violence by parents) and 54% had experienced vicarious family victimization (or exposure to violence between parents). Numerous studies have found that both types of victimization are powerful predictors of CPV (e.g., Beckmann, 2019; Calvete et al., 2014; Cano-Lozano, Navas-Martínez, & Contreras, 2021; Contreras & Cano-Lozano, 2016a; Contreras, León, & Cano-Lozano, 2020; Lyons et al., 2015; Margolin & Baucom, 2014; Navas-Martínez & Cano-Lozano, in press). More specifically, it has been observed that the likelihood of CPV in children with these victimization experiences is increased by approximately 70% compared to nonvictimized children (Gallego et al., 2019). Another type of victimization, school victimization (or violence by peers), also significantly predicts CPV (Beckmann, 2019; Navas-Martínez & Cano-Lozano, in press). Studies that analyze both family and school victimization have found that both types predict CPV independently and that both types jointly contribute to explaining a greater proportion of CPV (Beckmann, 2019; Navas-Martínez & Cano-Lozano, in press). In other words, the impact of several types of victimization on the development of CPV is greater than that of any single type.

Other fields of studies on violence have analyzed different types of aggressors (e.g., Martínez-Monteagudo et al., 2019; Rodríguez-Franco et al., 2017). Recently, research on CPV has also indicated the need to study different types of aggressors (Grace-Moulds et al., 2019; Navas-Martínez & Cano-Lozano, in press). Based on the aforementioned literature, the typology of the victimized aggressor (or the aggressor with victimization experiences) stands out. The characteristics of this type of aggressor have been analyzed in studies on school violence (del Moral et al., 2014; Haynie et al., 2001; Ireland & Power, 2004; Martínez-Monteagudo et al., 2019; Povedano et al., 2012; Ragatz et al., 2011), which have found the same proportions of girls and boys in this type of profile (Povedano et al., 2012). Likewise, several typologies of aggressive victims in bullying situations have been proposed. An especially relevant typology is the displaced aggression victim. This typology refers to adolescents who experience victimization at school and displaced aggression toward their parents (del Moral et al., 2014). However, in CPV research, it has not been common to address this phenomenon from the perspective of victimization, except for the study by Nowakowski-Sims and Rowe (2017), and no studies have analyzed the distinctive characteristics of this type of victimized aggressor and whether these characteristics differ according to the type of victimization. The understanding of CPV could be improved by analyzing the characteristics of specific types of aggressors, which could provide valuable empirical information for the approach of differential explanatory mechanisms in the development of CPV and for the design of prevention and intervention strategies adapted to the needs of each type.

In school violence, victimized aggressors are more aggressive and have a greater tendency to engage in criminal activities than nonvictimized aggressors (Haynie et al., 2001; Martínez-Monteagudo et al., 2019; Ragatz et al., 2011). One explanation is that being a victim of violence can perpetuate aggressive behavior as a way to relieve tension (Brezina, 1999) and protect oneself from threats (Martínez-Monteagudo et al., 2019). These aspects have not yet been examined with regard to CPV. Regarding differences in the gender of the aggressor, analyzed as a whole, CPV is more frequent in girls than in boys (Calvete et al., 2015; Cano-Lozano, Navas-Martínez, & Contreras, 2021). Analyzed by type of violence, most studies find that psychological violence is more frequent in girls than in boys (Calvete et al., 2015; Cano-Lozano, León, & Contreras, 2021; Cano-Lozano, Navas-Martínez, & Contreras, 2021; Contreras, Rodríguez-Díaz, & Cano-Lozano, 2020), and that physical violence, while some studies show that is more frequent in boys (Cano-Lozano, León, & Contreras, 2021; Cano-Lozano, Navas-Martínez, & Contreras, 2021) others find no differences (Calvete et al., 2015; Contreras, Rodríguez-Díaz, & Cano-Lozano, 2020).

Under circumstances of victimization, aggression seems to be motivated by both reactive reasons (in response to a threat) and instrumental reasons (to benefit oneself), with the former more common than the latter (Ford et al., 2012). In school violence, it has been found that compared to nonvictimized aggressors, aggressors who are victimized by their peers, exercise violence that is motivated by both reactive and instrumental reasons (Ragatz et al., 2011). Concerning CPV, family victimization has been found to predict both reactive and instrumental CPV, although it explains reactive CPV to a greater extent (Navas-Martínez & Cano-Lozano, 2020). Therefore, although victimization is related to violence motivated by reactive and instrumental reasons, it may occur through different mechanisms. In this line, Contreras, León, and Cano-Lozano (2020) found that family victimization is related to CPV motivated by reactive reasons through dysfunctional components of sociocognitive processing (e.g., anger) and to CPV motivated by instrumental reasons through other components (e.g., justification of violence). Regarding gender differences, while girls exercise CPV for reactive reasons more than boys do, both girls and boys also exercise CPV for instrumental reasons (Calvete & Orue, 2016; Contreras, Rodríguez-Díaz, & Cano-Lozano, 2020).

Parental attachment, defined as the emotional bond in terms of security and insecurity established from the interaction between children and their parents, seems to be an important variable in cases of victimized aggressors in school violence; these types of aggressors are characterized by more insecure attachment compared to nonvictimized aggressors (Ireland & Power, 2004). Regarding CPV, Nowakowski-Sims and Rowe (2017) found that the greater the extent of family victimization is, the lower the parental attachment, and family victimization is a significant predictor that contributes to explaining up to 12% of parental attachment. Together, these findings suggest the need to examine this variable in cases of victimized aggressors.

It has also been noted that aggressors victimized by their peers have less emotion regulation and empathy competencies (Povedano et al., 2012; Ragatz et al., 2011) and more difficulty resolving conflicts and exerting self-control (Haynie et al., 2001; Povedano et al., 2012) than nonvictimized aggressors, with no gender differences in these results (Povedano et al., 2012). Concerning CPV, some studies have related this type of violence to deficits in emotional and social competencies (Contreras & Cano-Lozano, 2016b; López-Martínez et al., 2019). More specifically, it has been found that boys are less able to identify their own emotions than girls, while there are no gender differences in the ability to understand and regulate their own emotions (López-Martínez et al., 2019). However, no studies have specifically analyzed the role of emotional competencies in victimized aggressors.

This study is the first to analyze the profile of victimized aggressors in a sample of adolescents who have shown aggressive behavior toward their parents. The first objective is to examine differences according to the type of aggressor (with family victimization, with school victimization, with polyvictimization and without victimization experienced) in the pattern of CPV and its reasons, in parental attachment, in emotional and coping competencies and in the gender variable. Although no studies on CPV have analyzed these aspects, the results are expected to be similar to those found in other types of violence. It is hypothesized that victimized aggressors, compared to nonvictimized aggressors, are characterized by higher levels of CPV (Haynie et al., 2001; Martínez-Monteagudo et al., 2019; Ragatz et al., 2011), more reactive and instrumental reasons (Ford et al., 2012; Ragatz et al., 2011), more insecure parental attachment (Ireland & Power, 2004), and less emotional and coping competencies (Haynie et al., 2001; Povedano et al., 2012; Ragatz et al., 2011). Likewise, worse adjustment is expected in aggressors with several types of victimization than those with a unique type of victimization (Beckmann, 2019; Navas-Martínez & Cano-Lozano, in press). Regarding gender, no significant differences are expected in the proportion of girls and boys in the different types of victimized aggressors (Povedano et al., 2012).

Another objective is to examine differences according to the gender of the aggressors in the pattern of CPV and its reasons, in parental attachment and in emotional and coping competencies. It is expected that CPV analyzed as a whole is greater in girls than in boys (Calvete et al., 2015; Cano-Lozano, Navas-Martínez, & Contreras, 2021). Girls exercise CPV for reactive reasons more than boys do, while there are no gender differences in instrumental reasons (Calvete & Orue, 2016; Contreras, Rodríguez-Díaz, & Cano-Lozano, 2020). Boys have less ability to identify their own emotions than girls (López-Martínez et al., 2019).

Finally, the present study analyzes differences according to the type of aggressor and the aggressor's gender in terms of the stated variables. It is expected that the interaction between the type of aggressor and gender is not significant (Povedano et al., 2012).

Method

Participants

We applied intentional nonprobabilistic sampling and selected a total of 3,142 adolescents who had exercised CPV behaviors repeatedly (any CPV behavior exercised 2 or more times) in the last year. The final sample included 1,559 adolescents (54.7% girls) aged between 12 and 18 years (Mage = 14.5, SD = 1.5) from educational centers (50.5% subsidized, 49.5% public) in two provinces in southern Spain.

Following the procedure used in similar research (e.g., Haynie et al., 2001; Ragatz et al., 2011; Rodríguez-Franco et al., 2017; Romera et al., 2021), four groups of aggressors classified by types of victimization were identified. The group with only family victimization (FV, 32.5%, n = 507) experienced direct violence by their parents or vicarious violence between their parents (any behavior 2 or more times) and did not experience school victimization (any behavior 1 time or none); the group with only school victimization (SV, 8.5%, n = 132) experienced violence by their peers directly or online (any behavior 2 or more times) and did not experience family victimization (any behavior 1 time or none); the polyvictimization group (PV, 22.4%, n = 349) experienced family (any behavior 2 or more times) and school victimization (any behavior 2 or more times); and the nonvictimized group (NV, 36.6%, n = 571) experienced neither family (any behavior 1 time or none) nor school victimization (any behavior 1 time or none).

Instruments

The Child-to-Parent Violence Questionnaire, Adolescent Version (CPV-Q-A; Contreras et al., 2019) evaluates how often violent behaviors were exercised toward the mother (α = .67) and the father (α = .66) in the last year through 14 parallel items on a Likert scale (0 = never; 4 = very often, six times or more) and the frequency of reactive (α = .79) and instrumental reasons (α = .82) for exercising CPV through 8 parallel items on a Likert scale (0 = never; 3 = always).

The Violence Exposure Scale (VES; Calvete et al., 2014) evaluates the frequency of violent behaviors by parents (Direct family victimization subscale) and the frequency of violent behaviors observed between parents (Vicarious family victimization subscale) through 6 items (α = .87) and 3 items (α = .73), respectively, on a Likert scale (0 = never; 4 = every day).

European Bullying/Cyberbullying Intervention Project Questionnaire (EBIP-Q and ECIP-Q; Brighi et al., 2012, Spanish validation; Ortega-Ruíz et al., 2016). These instruments evaluate the frequency of violent behaviors by peers in school (School victimization subscale) and online (School cybervictimization subscale) in the last two months through 7 items (α = .82) and 11 items (α = .75), respectively, on a Likert scale (0 = no; 4 = yes, more than once a week).

Attachment Representations Questionnaire, Short Version (CaMir-R; Pierrehumbert et al., 1996, Spanish validation; Balluerka et al., 2011). Past and present parental attachment experiences were evaluated to determine a secure or insecure attachment style (preoccupied, avoidant, and traumatized) through 32 items (α = .65) on a Likert scale (1 = strongly disagree; 5 = strongly agree).

Emotional competencies were evaluated with the Wong-Law Emotional Intelligence Scale (WLEIS; Wong & Law, 2002, Spanish validation; Extremera et al., 2019). This scale evaluates four dimensions of emotional intelligence (assessment and expression of one's own emotions, assessment and recognition of the emotions of other people, use or assimilation of one's own emotions for personal performance, and regulation of one's own emotions) through 16 items (α = .84) on a Likert scale (1 = completely disagree; 7 = completely agree).

Coping competencies were evaluated with the Connor-Davidson Resilience Scale, Short Version (CD-RISC-10; Connor & Davidson, 2003, Spanish validation; Notario-Pacheco et al., 2011), which evaluates the degree of resilience or coping with conflicts through 10 items (α = .79) on a Likert scale (degree of agreement: 0 = not at all; 4 = almost always).

Procedure

A cross-sectional descriptive survey study was designed (Montero & León, 2007). In accordance with the principles of the Declaration of Helsinki, the research was authorized by the Ethics Committee of the University of Jaén (grant number MAR.18/5.PRY) by the public administration in the field of education and by the educational centers. Signed informed consent was obtained from the parents and from the adolescent participants. Data collection was conducted in person and in groups in the classrooms of the educational centers. Participation consisted of completing a series of paper-and-pencil questionnaires administered by a single evaluator. The voluntary, confidential and anonymous nature of the responses was guaranteed.

Data analysis

Multivariate analyses of variance were performed to estimate the main effect of the type of aggressor (4 × 3 MANOVAs; with family victimization, with school victimization, with polyvictimization and nonvictimized) and the main effect of gender (2 × 3 MANOVAs; female and male) in six dependent variables grouped in a theoretically relevant way. In this way, the global scores of CPV and the reactive, and instrumental reasons for CPV were introduced in one MANOVA, and the global scores of parental attachment, emotional intelligence, and resilience were introduced in another. The interaction effect between the type of aggressor and gender on the dependent variables was also analyzed.

These analyses were followed by univariate analyses of variance (ANOVAs) to determine the dimensions in which the dependent variables in the groups differed and between which groups the differences were found with post hoc Games-Howell (heterogeneous variances) and Bonferroni (homogeneous variances) multiple comparisons. Gender differences were analyzed with the t test for independent samples. Finally, the effect size of the intergroup differences was calculated. Likewise, contingency analyses were performed with 2 comparisons to examine the differences in the proportion of girls and boys belonging to the different aggressor types. Additionally, to determine the possible relationship between the dependent variables of the study, a correlational analysis was performed.

Results

The MANOVAs showed statistically significant differences in the main effects of aggressor type on CPV and its reasons [λ = 0.86, F(9, 3675) = 25.7, p < .001, 2 = .05] and on parental attachment, emotional intelligence and resilience [λ = 0.89, F(9, 3748) = 20.7, p < .001, 2 = .04]. Statistically significant differences were also found in the main effects of gender on CPV and its reasons [λ = 0.98, F(3, 1510) = 12.6, p < .001, 2 = .02] and on parental attachment, emotional intelligence and resilience [λ = 0.99, F(3, 1540) = 4.5, p = .004, 2 = .01]. The significant variables and partial effects can be found in Table 1. No significant interaction effects were found between the type of aggressor and gender on CPV and its reasons [λ = 0.99, F(9, 3675) = 1.1, p = .322, 2 = .00] or on parental attachment, emotional intelligence and resilience [λ = 0.99, F(9, 3748) = 0.6, p = .740, 2 = .00].

Table 1

Main and interaction effects.

	Type of aggressor		Gender		Type of aggressor x Gender
	F	η2	F	η2	F	η2
	First MANOVA
Child-to-parent violence	48.7***	.09††	7.2**	.01†	1.2	.00
Reactive reasons	60.4***	.11††	37.1***	.02†	1.6	.00
Instrumental reasons	19.5***	.04†	0.2	.00	0.1	.00
	Second MANOVA
Parental attachment	51.1***	.09††	2.9	.00	0.5	.00
Emotional intelligence	23.2***	.04†	3.2	.00	0.9	.00
Resilience	16.2***	.03†	7.6**	.01†	0.9	.00

Note.

** p < .01; *** p < .001; 2 = .01 - .06 (small effect†); > .06 - .14 (medium effect††).

Regarding the type of aggressor, the results show significant differences in all the dimensions of the variables analyzed with the exception of interpersonal emotional perception and the gender variable. Regarding the CPV pattern (see Table 2), the three groups of victimized aggressors obtained higher CPV scores compared to the nonvictimized aggressors. By type of victimization, the PV group exercised more CPV than the FV group and the SV group. Concerning the reactive reasons for CPV, the PV and FV groups obtained higher scores than the NV group. By type of victimization, the PV group presented higher scores than the FV group and the SV group. Regarding the instrumental reasons for CPV toward the mother, all three groups of victimized aggressors obtained higher scores than the NV group. By type of victimization, the PV group only differed from the FV group and obtained higher scores than the FV group. Regarding the instrumental reasons for CPV toward the father, the PV and FV groups obtained higher scores than the NV group. By type of victimization, the PV group presented higher scores than the FV group and the SV group.

Table 2

Differences according to the type of aggressor in the study variables.

	PV (n = 349)	FV (n = 507)	SV (n = 132)	NV (n = 571)	F	df	Significant post hoc Games-Howella (Cohen's d)
Child-to-mother violence	8.4 (6.4)	6.9 (5.1)	6.3 (4.7)	4.6 (3.1)	52.2***	452.7	PV-FV (0.26); PV-SV (0.36); PV-NV (0.82); FV-NV (0.54); SV-NV (0.47)
Reactive reasons (M)	0.8 (0.7)	0.7 (0.6)	0.4 (0.4)	0.3 (0.4)	56.6***	486.3	PV-FV (0.20); PV-SV (0.53); PV-NV (0.82); FV-SV (0.37); FV-NV (0.61)
Instrumental reasons (M)	0.7 (0.5)	0.6 (0.5)	0.6 (0.4)	0.4 (0.4)	21.5***	477.2	PV-FV (0.20); PV-NV (0.54); FV-NV (0.33); SV-NV (0.33)
Child-to-father violence	7.3 (6.2)	6.1 (4.9)	5.3 (4.1)	4.1 (2.8)	44.8***	450.9	PV-FV (0.22); PV-SV (0.36); PV-NV (0.74); FV-NV (0.52); SV-NV (0.39)
Reactive reasons (F)	0.8 (0.8)	0.7 (0.6)	0.4 (0.4)	0.3 (0.4)	60.2***	481.4	PV-FV (0.20); PV-SV (0.54); PV-NV (0.84); FV-SV (0.38); FV-NV (0.63)
Instrumental reasons (F)	0.6 (0.5)	0.5 (0.5)	0.4 (0.4)	0.3 (0.4)	16.1***	470.6	PV-FV (0.22); PV-SV (0.29); PV-NV (0.49); FV-NV (0.26)
Secure attachment	23.9 (4.8)	24.8 (4.5)	26.6 (3.4)	27.0 (3.3)	53.9***	491.5	PV-FV (-0.19); PV-SV (-0.61); PV-NV (-0.79); FV-SV (-0.43); FV-NV (-0.57)
Preoccupied attachment	12.5 (3.5)	11.6 (3.5)	10.8 (3.4)	9.9 (3.1)	46.9***	485.0	PV-FV (0.27); PV-SV (0.50); PV-NV (0.78); FV-NV (0.48)
Avoidant attachment	13.3 (3.3)	12.5 (3.2)	11.7 (3.1)	11.1 (2.8)	41.8***	485.0	PV-FV (0.25); PV-SV (0.50); PV-NV (0.73); FV-SV (0.26); FV-NV (0.46)
Traumatized attachment	11.2 (4.5)	10.2 (3.9)	8.5 (2.8)	7.9 (2.8)	72.9***	494.2	PV-FV (0.24); PV-SV (0.66); PV-NV (0.92); FV-SV (0.47); FV-NV (0.68)
Emotional perception (A)	19.1 (4.9)	20.4 (4.6)	19.9 (4.9)	21.0 (4.3)	11.9***	478.3	PV-FV (-0.26); PV-NV (-0.41)
Emotional perception (E)	21.3 (3.9)	21.4 (3.9)	21.0 (4.8)	21.1 (3.8)	0.6	1547
Emotional assimilation	17.6 (5.6)	18.8 (5.1)	18.7 (5.4)	20.2 (5.0)	17.5***	482.5	PV-FV (-0.21); PV-NV (-0.48); FV-NV (-0.28); SV-NV (-0.28)
Emotional regulation	14.7 (5.6)	16.3 (5.5)	15.8 (5.4)	18.1 (5.2)	30.5***	1547	PV-FV (-0.28); PV-NV (-0.63); FV-NV (-0.35); SV-NV (-0.44)
Resilience	24.2 (7.2)	25.8 (6.5)	25.5 (6.7)	27.3 (5.9)	15.9***	480.0	PV-FV (-0.23); PV-NV (-0.47); FV-NV (-0.23); SV-NV (-0.30)

Note. The values indicate the mean and the (standard deviation). PV = polyvictimized aggressors; FV = aggressors with family victimization; SV = aggressors with school victimization; NV = nonvictimized aggressors; M = mother; F = father; A = intrapersonal; E = interpersonal.

aExcept in interpersonal emotional perception and emotional regulation (Bonferroni).

*** p < .001; d = 0.20 (small effect); 0.50 (medium effect); 0.80 (large effect).

The PV and FV groups obtained lower scores in secure parental attachment and higher scores in insecure attachment (preoccupied, avoidant, and traumatized) than the NV group. By type of victimization, the PV group presented lower scores in secure attachment and higher scores in the three insecure attachment styles than the FV group and the SV group. Finally, the three groups of victimized aggressors obtained lower scores in emotional assimilation and regulation and the capacity of resilience compared to the NV group. However, only the PV group differed from the NV group in intrapersonal emotional perception, with lower scores than the NV group. By type of victimization, the PV group differed only from the FV group, with lower scores in these variables than the FV group.

On the other hand, there were no significant differences in the proportion of girls and boys classified by type of aggressor (see Table 3) 2 (3, 1559) = 3.9, p = .270, φ = 0.05.

Table 3

Differences according to type of aggressor in the gender variable.

Type of aggressor	Girls n (%)	Boys n (%)
Polyvictimized	200 (23.5)	149 (21.1)
Family victimization	272 (32.0)	235 (33.2)
School victimization	63 (7.4)	69 (9.8)
Nonvictimized	317 (37.1)	254 (35.9)

Regarding gender, the results showed statistically significant differences between girls and boys in the pattern of CPV and its reasons (see Table 4). Specifically, girls obtain higher scores on CPV toward mothers than boys. On the other hand, girls exercise CPV towards mothers and fathers more for reactive reasons compared to boys.

Table 4

Differences according to gender in the study variables.

	Girls (n = 852)	Boys (n = 707)	t	df	Cohen's d
Child-to-mother violence	6.7 (5.1)	5.9 (4.7)	-2.7**	1550	0.16
Reactive reasons (M)	0.7 (0.6)	0.4 (0.5)	-7.1***	1548.7	0.53
Instrumental reasons (M)	0.5 (0.5)	0.5 (0.5)	-1.1	1537
Child-to-father violence	5.7 (4.6)	5.3 (4.7)	-1.4	1537
Reactive reasons (F)	0.6 (0.6)	0.4 (0.5)	-6.0***	1536.9	0.36
Instrumental reasons (F)	0.4 (0.4)	0.4 (0.4)	0.0	1524
Secure attachment	25.7 (4.1)	25.3 (4.4)	-1.9*	1557	0.09
Preoccupied attachment	10.9 (3.5)	11.4 (3.4)	2.5*	1557	-0.14
Avoidant attachment	12.3 (3.2)	11.9 (3.2)	-1.9	1557
Traumatized attachment	9.4 (4.0)	9.5 (3.7)	0.1	1557
Emotional perception (A)	19.5 (4.8)	21.2 (4.2)	7.2***	1550.8	-0.37
Emotional perception (E)	21.8 (3.8)	20.5 (4.1)	-6.5***	1557	0.33
Emotional assimilation	18.8 (5.2)	19.3 (5.3)	1.5	1557
Emotional regulation	15.9 (5.5)	17.3 (5.4)	5.0***	1557	-0.26
Resilience	25.3 (6.6)	26.7 (6.5)	4.3***	1557	-0.23

Note. The values indicate the mean and the (standard deviation). M = mother; F = father; A = intrapersonal; E = interpersonal.

* p < .05; ** p < .01; *** p < .001; d = 0.20 (small effect); 0.50 (medium effect); 0.80 (large effect).

Statistically significant differences were also found between girls and boys in parental attachment, emotional intelligence and resilience. Specifically, boys obtained lower scores in secure parental attachment and higher scores in preoccupied parental attachment compared to girls. On the other hand, girls obtained lower scores in intrapersonal emotional perception and emotional regulation, while boys obtained lower scores in interpersonal emotional perception. Finally, girls obtained lower scores in resilience than boys.

Finally, the results of Table 5 show significant relationships between all the study variables. Specifically, CPV and its reactive and instrumental reasons are positively related to each other and negatively related to parental attachment, emotional intelligence and resilience. The latter have significant and positive relationships with each other.

Table 5

Intercorrelations between the study variables.

	1	2	3	4	5	6
Child-to-parent violence	-
Reactive reasons	.43***	-
Instrumental reasons	.44***	.21***	-
Parental attachment	-.17***	-.22***	-.09***	-
Emotional intelligence	-.16***	-.23***	-.17***	.32***	-
Resilience	-.11***	-.10***	-.10***	.22***	.60***	-

Note. *** p < .001.

Discussion

This study finds that more than half of the adolescents who exercised CPV also experienced some type of victimization (63.4%), justifying the study of CPV from the perspective of victimization (Nowakowski-Sims & Rowe, 2017).

Regarding the analysis of differences in the variables examined according to the type of aggressor, the results show differences between victimized and nonvictimized aggressors that are not only significant but also have medium and large effect sizes in many cases.

Regarding the CPV pattern, all victimized aggressors exercised more CPV than nonvictimized aggressors. These results were similar to those found in studies on school violence (Haynie et al., 2001; Martínez-Monteagudo et al., 2019; Ragatz et al., 2011). This finding supports the idea that being a victim of violence can trigger more violence in response to the tension experienced (Brezina, 1999), resulting in displaced violence toward the parents (del Moral et al., 2014). By type of victimization, as expected, polyvictimized aggressors exercised more CPV than those with a unique type of victimization. This is consistent with a recent study concluding that family and school victimization together contribute to explaining a higher proportion of CPV than either type separately (Navas-Martínez & Cano-Lozano, in press). These results are relevant because they suggest that cumulative (repeated) and multifaceted (multiple types) victimization can trigger more aggressive responses (Ford et al., 2012), requiring more attention to this profile.

Concerning the reasons for CPV, the hypothesis is almost completely confirmed given that most victimized aggressors, specifically those with polyvictimization and family victimization, exercise CPV that is more motivated by both reactive and instrumental reasons compared to nonvictimized aggressors. These results are in line with previous findings regarding other types of violence (Ford et al., 2012; Ragatz et al., 2011). These differences are not found in aggressors with school victimization compared to nonvictimized aggressors. On the other hand, more relevant effect sizes have also been found in reactive reasons than in instrumental reasons. One possible explanation is that in cases of victimization, there is a greater predisposition to respond reactively with violence to defend oneself from the victimization experienced (Brezina, 1999; Ford et al., 2012), but it is also possible that victimization contributes to normalizing and internalizing violence as an acceptable instrument to gain benefits, which in this case would be the cessation of victimization. By types of victimization, as expected, polyvictimized aggressors exercise CPV that is more motivated by both reactive and instrumental reasons compared to those with a unique type of victimization.

The results almost completely confirm the hypothesis regarding parental attachment. Most victimized aggressors, specifically those who experienced polyvictimization and family victimization, have lower levels of secure parental attachment and higher levels of insecure attachment than nonvictimized aggressors. These results are consistent with studies on CPV (Nowakowski-Sims & Rowe, 2017) and are similar to studies on other types of violence (Ireland & Power, 2004). However, these differences are not found in aggressors with school victimization compared to nonvictimized aggressors. By type of victimization, polyvictimized aggressors have less secure parental attachment and more insecure attachment compared to those with a unique type of victimization, as expected. Therefore, victimized aggressors, especially those with several types of victimization, have parental-child relationships marked by anxiety, ambivalence, self-sufficiency, rejection, lack of affection, and even abuse or indifference. Recently, it has been found that a lack of parental warmth (characteristic of insecure attachment) generates emotional deficits that lead to reactive CPV, while parental criticism and rejection (also characteristic of insecure attachment) are associated with both reactive and instrumental CPV (Cano-Lozano et al., 2020). This and similar evidence previously discussed (Contreras, León, & Cano-Lozano, 2020) are examples of differential mechanisms in the development of CPV motivated by reactive and instrumental reasons, which, together with the mechanisms discussed in this study, suggest the benefit of continuing this line of research.

Although it has been noted that adolescents who exercise CPV present emotional and coping difficulties (Contreras & Cano-Lozano, 2016b; López-Martínez et al., 2019), this study deepens the finding that all victimized aggressors have lower abilities to regulate and use their own emotions and lower abilities to cope with conflicts than do nonvictimized aggressors. These results are in line with studies on typologies of aggressors (Haynie et al., 2001; Povedano et al., 2012; Ragatz et al., 2011) and studies that highlight the role of victimization in emotional regulation (Rey et al., 2020), confirming the proposed hypotheses. By type of victimization, the results partially coincide with expectations. Specifically, polyvictimized aggressors have more emotional and coping difficulties than aggressors with only family victimization. There are no differences in these variables between polyvictimized aggressors and aggressors with school victimization; they show similarly low levels of emotional intelligence and resilience.

As expected, the proportions of girls and boys classified into the three types of victimized aggressors were the same, in line with similar studies in which the proportion of victimized aggressors in school violence situations also did not differ according to gender (Povedano et al., 2012), contributing to the field of CPV.

In general, the results related to polyvictimized aggressors and those who experienced family victimization are in line with the hypotheses proposed. However, aggressors who experienced school victimization present, on the one hand, a profile more similar to that of nonvictimized aggressors in the reasons for CPV and parental attachment and, on the other hand, a profile more similar to that of polyvictimized aggressors in emotional and coping competencies. Likewise, although no hypotheses were proposed about the differences between aggressors with unique types of victimization, aggressors who experienced family victimization show more reactive CPV and more insecure parental attachment compared with those who experienced school victimization. These findings underscore the importance of separately analyzing several types of victimization in CPV (Beckmann, 2019; Navas-Martínez & Cano-Lozano, in press).

Regarding the analysis of gender-related differences in the variables examined among aggressors, this study finds that girls exercise more CPV toward mothers than boys do, which is consistent with previous research (Calvete et al., 2015; Cano-Lozano, Navas-Martínez, & Contreras, 2021). However, no differences were found between girls and boys in CPV toward the father. In this line, a recent study that analyzed violence by children toward parents and by parents toward children found that daughters exert more violence toward mothers and, in turn, mothers exert more violence toward daughters, which suggests a two-way relationship. In contrast, there are no differences between sons and daughters in the violence they exert or suffer from their fathers (Cano-Lozano, Navas-Martínez, & Contreras, 2021).

Likewise, it was found that CPV exercised by girls is more motivated by reactive reasons than in the case of boys, although there are no gender differences in the instrumental reasons for exercising CPV. These results confirm the proposed hypotheses in line with previous findings (Calvete & Orue, 2016; Contreras, Rodríguez-Díaz, & Cano-Lozano, 2020), providing further evidence of gender differences in terms of motivations for CPV.

Although no hypotheses were proposed regarding gender differences in parental attachment style given the absence of evidence in this regard, it has been found that boys have more insecure emotional ties with their parents than girls do and are characterized more by a preoccupied attachment style. There are no gender differences in the other insecure attachment styles. These data are novel and should be replicated in future studies given the importance that this infrequently studied variable seems to have in relation to CPV (Nowakowski-Sims & Rowe, 2017).

Regarding gender differences in emotional and coping competencies, while girls have more difficulties recognizing and regulating their own emotions, boys have more difficulties recognizing the emotions of others. These results do not coincide with the study by López-Martínez et al. (2019), who found that boys have a more difficult time identifying their own emotions than girls do. However, the aforementioned study did not analyze gender differences in dimensions of emotional intelligence, such as recognizing the emotions of others, while this study provides that complementary information. It is also found that girls have weaker resilience or poorer competencies in coping with conflicts than boys, which provides new data for CPV research.

On the other hand, this study did not find a significant interaction effect between the type of aggressor and gender in the variables analyzed. These results confirm the hypothesis by showing that girls and boys belonging to the different types of aggressors do not differ in the pattern of CPV, reactive or instrumental reasons, parental attachment and emotional and coping competencies. The findings are consistent with the results found in school violence, where girls and boys of the victimized aggressor typology no differ in the characteristics analyzed (Povedano et al., 2012). More studies on CPV are needed that analyze different types of aggressors and their gender while taking into account other relevant variables in this type of violence.

Finally, the dependent variables of this study present significant relationships with each other. Specifically, a positive relationship was found between CPV, reactive reasons and instrumental reasons for CPV, as well as a positive relationship between parental attachment, emotional intelligence and resilience. In turn, CPV and its reasons show a negative relationship with parental attachment style, emotional intelligence and resilience.

In summary, in CPV, the victimized aggressor generally presents a more violent reactive and instrumental-type profile, is characterized by more insecure parental attachment, and has less emotional and coping competencies than the nonvictimized aggressor. Additionally, aggressors with several types of victimization generally exercise more reactive and instrumental CPV, show more insecure parental attachment, and have less emotional and coping competencies than those with a unique type of victimization. On the other hand, girls exercise more CPV toward mothers and do it for more reactive reasons than boys do. Likewise, boys have more insecure emotional ties with their parents and more difficulty recognizing the emotions of others, while girls have more difficulty recognizing and regulating their own emotions as well as poorer coping competencies to adaptively face conflictive or difficult situations.

This study has several limitations. Given its cross-sectional nature, it is not possible to establish causal relationships. Future longitudinal studies could analyze the temporal sequence of the different types of victimization. On the other hand, group size is heterogeneous in some cases. Although this study started with a large sample, it would be recommended for future research to have even larger sample sizes to obtain groups of more equal size. Likewise, the sample is based on two specific provinces of southern Spain, limiting the generalization of the results to similar contexts. This study should be replicated in other provinces and countries. Finally, for a broader perspective, it would be wise to complement the self-reports of adolescents with those of parents and peers.

Despite the limitations, and although additional research is needed, this is the first study on CPV to address specific profiles of aggressors. The main contribution is the finding that the aggressors of CPV differ from each other in terms of not only experiencing or not experiencing victimization but also according to different experiences of victimization, findings that could have important implications for research and professional practice. Specifically, given that victimization and its different types seem to set in motion differential mechanisms in the development of reactive and instrumental CPV, it would be interesting for future research to deepen the analysis of the mediating and moderating characteristics of the relationship between the different types of childhood victimization and CPV motivated by reactive and instrumental reasons. On the other hand, these findings show the need to detect early cases of victimization, especially those of polyvictimization, and to apply treatments that take into account the specific difficulties that these adolescents face compared to those who have not experienced victimization. Finally, this finding also suggests the importance of working on parental attachment bonds and improving the emotional and coping competencies of these adolescents.

In conclusion, this study provides valuable empirical information for the approach of differential explanatory mechanisms in CPV development according to the typology of aggressors and for the design of prevention and intervention strategies adapted to the needs of this profile of victimized aggressors in CPV.

Funding

This study was funded by a grant from the Research Support Plan of the University of Jaén [grant number EI_SEJ8_2019] and by a Predoctoral Fellowship for Research Personnel Development of the University of Jaén [grant number R6/04/2018].

The authors declare there are no conflicts of interest.