Custom services
text in 
English (pdf)
Article in xml format
Article references
Automatic translation
Send this article by e-mail
Cited by SciELO 
Cited by Google 
Similars in
SciELO 
Similars in Google 
Permalink
Introduction
The existence of differences between violent men and women is a great source of debate and discussion in forensic psychology and criminology. Both instruments and intervention programs have been developed using male samples (Loinaz, 2016). Furthermore, women are a minority group in criminal contexts, leading to many of their needs being ignored (Loinaz, 2014, 2016). In Spain, less than 2 out of 10 juvenile offenses are committed by women, which is double the proportion of adult women who are incarcerated. Although the overall figures from the National Statistical Institute (INE) show a slight increase in female crime in recent years in Spain (17.7% in 2013; 19.2% in 2017), the more serious the crime the lower the proportion of women offenders (homicides 10.5%, violence with injuries 25%, for example) and almost no women commit sexual offences (0.3% in 2017).
Violence risk assessment research confirms that men and women share many risk factors in common, yet some key differences remain (Penney, Lee, & Moretti, 2010; Wong, Slotboom, & Bijleveld, 2010). Regarding youth-specific tools, a recent meta-analysis (Pusch & Holtfreter, 2018) concludes that there are no significant differences between girls and boys in the predictive validity of YLS/CMI (Hoge, Andrews, & Leschied, 2002). Although these results converge with those from a previous meta-analysis (Olver, Stockdale, & Wormith, 2009), findings between studies are not homogeneous. Prediction research shows with higher scores, recidivism is lower among girls compared to boys (Anderson et al., 2016). Another benchmark in the risk assessment of youths, the SAVRY (Borum, Bartel, & Forth, 2006), shows similar predictive validity for both sexes (Childs, Frick, & Gottlieb, 2016; Gammelgård, Koivisto, Eronen, & Kaltiala-Heino, 2008; Meyers & Schmidt, 2008; Penney et al., 2010), although different prevalence rates are also observed depending on the type of risk factor (more antisocial behavior in boys and more past violence and life stressors in girls) (Gammelgård, Weizmann-Henelius, Koivisto, Eronen, & Kaltiala-Heino, 2012). It has also been noted that correlations between SAVRY factors differ between sexes, therefore encouraging a gender-specific focus on risk factors (Hilterman, Bongers, Nicholls, & van Nieuwenhuizen, 2016). Doing so would aid practitioners to decide which elements are most relevant to female or male offenders as well as help interpret gender-specific differences and adjust risk management strategies accordingly. These contradictory results lead to the question of whether or not there is a need to differentiate between sex in criminology and risk prediction studies, or whether indicators should be taken as equally useful regardless of sex (Emeka & Sorensen, 2009). This issue is paramount as it has a major impact on the way treatments are carried out.
Gender and child-to-parent violence
Child-to-parent violence (CPV), consisting in repeated behaviors of physical, psychological (verbal or nonverbal) or financial abuse, directed toward the parents or caregivers (Pereira et al., 2017), is arguably the kind of violence in which there are purported to be the most gender similarities. First, regarding prevalence of the problem, the proportion of male and female aggressors is more balanced than for other kinds of crimes. However, prevalence rates depend on the samples used (with percentages being more equal in more normalized samples, such as students) and male offenders are the majority in almost all studies using judicial samples (Armstrong, Cain, Wylie, Muftić, & Bouffard, 2018; Boxer, Gullan, & Mahoney, 2009; Ibabe & Jaureguizar, 2010; Simmons, McEwan, Purcell, & Huynh, 2019; Walsh & Krienert, 2007). When analyzing community samples (students’ self-reports), the proportions of male and female assailants are similar, but in criminal samples the proportion of men ranges from 59% to 87% (Simmons, McEwan, Purcell, & Ogloff, 2018).
Regarding the type of violence committed, most research in community or general population samples find no differences between boys and girls in the perpetration of psychological and physical violence (Agnew & Huguley, 1989; Browne & Hamilton, 1998; Calvete, Orue, Gámez-Guadix, & Bushman, 2015; Elliott, Cunningham, Colangelo, & Gelles, 2011; Gámez-Guadix & Calvete, 2012; Ibabe & Bentler, 2016; McCloskey & Lichter, 2003; Pagani, Larocque, Vitaro, & Tremblay, 2003; Pagani et al., 2004, 2009). Some studies in these same samples describe more psychological abuse by girls (Beckmann, Bergmann, Fischer, & Mößle, 2017; Calvete & Orue, 2016; Elliott et al., 2011; Jaureguizar, Ibabe, & Straus, 2013; Lyons, Bell, Fréchette, & Romano, 2015; Margolin & Baucom, 2014; Rosado, Rico, & Cantón-Cortés, 2017). In clinical or judicial samples more physical violence from boys has been observed (Boxer et al., 2009; Evans & Warren-Sohlberg, 1988; Kuay, Tiffin, Boothroyd, Towl, & Centifanti, 2017; Nock & Kazdin, 2002; O’Hara, Duchschere, Beck, & Lawrence, 2017; Routt & Anderson, 2011; Walsh & Krienert, 2009), but it has also been argued that detained or convicted girls use violence of the same nature and severity as boys (Condry & Miles, 2014; Simmons et al., 2018; Strom, Warner, Tichavsky, & Zahn, 2014).
Finally, there also appear to be differences with respect to the parent who has been assaulted. Studies conclude that mothers are the most frequent target of violence, although the most serious physical assaults are committed against fathers and by male children (Condry & Miles, 2014; Lyons et al., 2015; Simmons et al., 2019). Other research has shown that daughters target their mothers more, especially when there is physical violence (Calvete & Orue, 2016; Del Hoyo-Bilbao, Gámez-Guadix, Orue, & Calvete, 2018; Walsh & Krienert, 2007). These interactions between type of violence, sex of the assaulted parent, sex of the offender and even age of the aggressor have been argued to be related to developmental changes and psychological and physical maturation of the children, which facilitates one form of violence or another (Walsh & Krienert, 2007). Yet, it is important to keep in mind that for all these findings, there are studies with contradictory results.
The parental, judicial, and police response also differs depending on the sex of the offender. Males receive most legal complaints given that parents are often more intimidated by young men than by young women (Armstrong et al., 2018; Miles & Condry, 2016; Simmons et al., 2018). This may partly explain the fact that in official samples the proportion of boys is higher (Walsh & Krienert, 2007). However, the pattern is also changing, as shown in, for example, the study of Strom et al. (2014) on police response to CVP. Regardless of the law, police response to incidents of CPV with female perpetrators is changing more than the response to male perpetrators. This is connected to the reduction of tolerance for female violence towards parents.
Characteristics of the abusive youth
There is comprehensive literature on risk factors or variables involved in CPV, examining individual traits and psychosocial adaptation of aggressors, characteristics of parents, family functioning, and so on. Although samples vary greatly from study to study, (student self-reports, cases in treatment, cases with judicial measures) and the establishment of clear patterns has proved difficult, what is known has allowed the development of the first violence risk assessment tool for CPV, the Child-to-Parent Violence Risk guide -CPVR- (Loinaz, Andrés-Pueyo, & Pereira, 2017; Loinaz & Ma de Sousa, 2020).
Along with parenting style (see Ruiz-Hernández, Moral-Zafra, Llor-Esteban, & Jiménez-Barbero, 2019, for a review of its influence on externalized behaviors), the most analyzed variable at the family level is parental violence, which increases the likelihood of CPV by around 70% (Gallego, Novo, Fariña, & Arce, 2019). The use of multivariate models (Del Hoyo-Bilbao, Orue, Gámez-Guadix, & Calvete, 2020) with individual, family and contextual factors, emphasizes the importance of paying attention to family variables, such as how parents implement disciplinary strategies and individual factors such as impulsivity and substance abuse. The influence of peers on boys is greater than on girls, and emotional instability is more common among girls. Another recent research line shows that clinical and judicial samples, for example, also differ in their risk factors, with especially striking differences in family characteristics, which are more complex and conflictive in judicial contexts (Loinaz & Ma de Sousa, 2020). Knowledge of characteristics of aggressors has also led to propositions of various typologies, as has been done, for instance, in the case of intimate partner violence (i.e. Loinaz, Marzabal, & Andrés-Pueyo, 2018). In particular, Kuay et al. (2017) divided a group of CPV cases into widespread aggressors with features of emotional insensitivity (who have committed CPV as well as other violence outside the family) and specialized aggressors, whose assaults are limited to parents, with low insensitivity traits.
Despite the extensive amount of studies, few works specifically compare the psychosocial characteristics of boys and girls involved in CPV. Rosado et al. (2017), in a sample of secondary and high school students assessed with the SCL-90, found a significant positive relationship between interpersonal sensitivity in boys for assault on both parents; and in the case of girls, a positive relationship between the scale of obsessions and aggression toward parents, and a negative relationship between interpersonal sensitivity and aggression toward the mother. More impulsivity has also been reported in girls who have committed CPV compared to boys (Rico, Rosado, & Cantón-Cortés, 2017). Females incarcerated for CPV (Armstrong et al., 2018), come from more single-parent households, have greater histories of arrests for unarmed attacks, use more alcohol and other illegal drugs, and have suffered more emotional/physical/sexual victimization than their male counterparts. However, there were almost no differences in marijuana use, nor in indirect victimization at home. In addition, in terms of mental health, women had a higher tendency to show affective symptomatology, suicidal ideation and anxiety, while males had a higher prevalence rates of hallucinations.
The recent work of Beckmann et al. (2017), focusing on risk and protective factors, concludes that the relevance of factors does not substantially vary across child gender, noting that substance abuse predicts physical violence in both sexes while suicidal ideation is a risk factor only in boys. Narcissism, exposure to family violence, and greater rates of externalizing problems have also been shown to be predictors of CPV in boys (Calvete, Orue, et al., 2015). The bidirectionality of violence has been pointed out especially for male aggressors (Ibabe & Jaureguizar, 2011; Ibabe, Jaureguizar, & Díaz, 2009). Finally, anger is the most powerful predictor of CPV escalation for girls, while hostile attributions are the most important predictors for boys (Calvete, Gámez-Guadix, & García-Salvador, 2015). In general, the emotional component appears to be more important for girls than boys.
Current work
The aim of this study was to analyze the existence of differences in risk factors between boys and girls involved in CPV, evaluated with a new tool designed to assess the risk of this type of violence. Based on previous research, it was expected that boys and girls would show similar prevalence rates for risk factors. However, more impulsiveness, more anger, more drug use and more victimization were expected among girls. Furthermore, less severe violence and violence mainly directed towards mothers was expected for female offenders. Given these presuppositions, similar violence predictive capacity of the CPVR was anticipated in both sexes.
Method
Protocols
A total of 91 files of different youths were examined, 61 (67%) of which were assessed in clinical contexts (Euskarri and Amalgama 7) and 30 (33%) of which were assessed in judicial contexts (SMAT Barcelona and Fundación Pioneros), all of them in Spain. The average age was 17.07 years (SD = 2.4; range = 13-28), and most of them were Spanish (93.4%, n = 85). Boys (61.5% , n = 56) and girls (38.5%, n = 35) were of a similar age (16.20 girls; 17.30 boys; p = .156), and were also mostly Spanish (91.4% girls and 94.6% boys; p = .672), although the girls came from judicial contexts significantly more, χ2(1, N = 91) = 6.27, p = .012, than boys (48.6% vs. 23.2%).
Instrument
The Child-to-Parent Violence Risk assessment guide (CPVR) consists of 24 risk factors (organized into four blocks: type of violence, psychological characteristics of the aggressor, social adaptation of the aggressor and family factors), and 6 protective factors that are coded as present (Yes), partially present (?) or absent (No) for the present time and the past (one year prior and further in the past). Alongside these core variables, there are more than 20 possible risk factors in an initial section, including personal, family, history of violence and characteristics of the victim. The creation of the tool followed international standards for structured professional judgment guides, based on the available research, feedback from front line professionals for whom it was aimed to be in use by, and pilot applications (Loinaz et al., 2017). The tool has been shown to be useful in the differentiation of cases in judicial and clinical contexts, with good classification values for the type of offender (AUC = .830) and for the presence of injuries to the mother (AUC = .764) (Loinaz & Ma de Sousa, 2020).
Procedure
As part of the process of creating the tool, two clinical centers (Euskarri -Bilbao- and Amalgama 7 -Barcelona-) and two sources of judicial cases (Fundación Pioneros -La Rioja- and SMAT -Barcelona-) collaborated in its application. Cases from Euskarri and Pioneros were assessed by the professionals working on each case, with the assistance of the first author of the tool. The cases of Amalgama 7 and SMAT were coded by a research team with the information collected by the professionals of the centers during their daily activities. Each main factor was coded as present, partially present or absent following the guidelines of the tool and justifying the information on which the response was based (see also Loinaz & Ma de Sousa, 2020).
Coding
The interrater reliability of CPVR codification was estimated with the true kappa, calculated like Cohen’s kappa but confirming the correspondence between the sources used for coding each item (Vilariño, Arce, & Fariña, 2013). Whiting-rater concordance (test-retest reliability with the same case and information), and between-context concordance (estimating whether other encoders, equally trained, would get similar results) were also analyzed. In this research line (Loinaz & Ma de Sousa, 2020), two researchers rated 30 CPVR each, repeated their assessment of three cases, and rated three cases of the other researcher. Inter-rater and the intra-rater (test-retest) reliabilities were high (> .9) in both cases. In addition, one of the researchers also showed reliability coding done in a previous study (Loinaz, et al., 2017).
Data analysis
The data was analyzed using the IBM SPSS Statistics 25 statistical package. Contingency tables (chi-squared test) were used for the comparison of the prevalence of risk factors, with “Yes” and “?” coding added to facilitate 2 X 2 comparisons. Quantitative variables were compared with t -tests, and ROC curve analysis was used for violence prediction using the total score of the tool. Effect size was computed in d (Heege’s formula), phi, Odds Ratio and Area Under the Curve (AUC). As for the comparisons of the effect sizes, tables from Salgado (2018) were employed. The interpretation of the magnitude of the effects of the results in terms of the AUC was made in terms of probability superiority (PSTE) over all possibilities (Vilariño, Amado, Vázquez, & Arce, 2018). To obtain a numerical risk level, risk factor codifications were translated to coding of 2 (Present), 1(?), and 0 (No).
Results
Main factors of the CPVR
The comparison of the 30 central factors of the CPVR in the last year (present) (see Table 1) showed few differences between boys and girls. There were some significant differences in risk factors, with, greater rates of escalation, among boys, as well as self-esteem problems, violence between parents, cohabitation problems other than CPV and problems in parents, which were more prevalent among girls. Protective factors on the other hand differed only in family involvement in therapy, which was significantly more common in boys. An extra factor, included in this section, is the possibility of being a victim of intimate partner violence (IPV), initially intended for women. Although there was a high prevalence of omitted responses in both samples, one boy (4.2%) and 8 girls (27.6%) had been victims of IPV during the last year. This difference was significant χ2 (1, N= 53) = 5.11, p = .024.
Table 1. Differences in risk factor prevalence between boys and girls in the present (up to a year before the assessment).
As some of the risk factors may reflect problematic histories or possible precursors to CPV, the same comparison of coded factors was made for the past, beyond the one year before the assessment (see Table 2). When analyzing the presence of variables in the past, some statistically significant new differences emerged between sexes. The girls had suffered significantly more domestic violence (bidirectionality), had significantly more CPV complaints, increased presence of self-esteem issues, more prevalence of violence between parents, problems of cohabitation other than CPV and had parents with more problems themselves. For boys, problems of substance abuse in the past were significantly more prevalent than in girls. Unlike the assessment for the present moment, boys and girls presented the same history of escalation in the CPV and there were no differences in the protective factors. Regarding past intimate partner violence victimization, the percentage of girls increased to 34.5% (n = 10).
Other variables of interest
The comparison of other variables from the initial section of the CPVR (see Table 3) showed no statistically significant difference. Boys and girls were the same in aspects such as their academic/work situation, family variables such as single parenthood, adoption, migration and parental criminal history. This final variable almost tripled in prevalence for girls (10.6% vs. 28.0%) but the difference also did not reach statistical significance. Regarding parents, girls (M = 45.00, SD = 6.25) had significantly younger mothers, t(49) = 2.43, p = .019, d = 0.70, than boys (M = 49.09, SD = 5.53). The age of the fathers, on the other hand, was the same for both groups, t(37) = 1.10, p = .278, d = 0.36, (M = 52.03, SD = 5.68 in boys, M = 49.46, SD = 8.88 in girls).
Mothers were the most victimized in all three forms of violence (physical, psychological or economic), with them being aggressed by 96.4% of the boys and 94.3% of the girls. Fathers were victims in a much smaller proportion, 51.8% of the time for boys and 45.7% of the time for girls. There was no difference between the sexes in terms of victimization of one parent or the other. Although there were no statistically significant differences, physical violence was exercised to a greater extent by boys (32.1% vs. 17.1%), but physical violence against the mother was more prevalent in girls (68.6% vs. 48.2%). Only 3.6% of boys injured their father while no females caused injuries to their fathers. For mothers, injuries were caused more often by daughters (25.7%) than by sons (16.1%), although this difference was not significant. The age of onset of the violence was also the same, t(71) = 0.16, p = .875, d = 0.04, for boys (M = 12.32, SD = 3.59) and girls (M = 12.46, SD = 3.03). The duration of violence until the beginning of the current intervention was also the same, t(67) = 1.21, p = .229, d = 0.18, for boys (M = 5.38, SD = 3.86) and girls (M = 4.26, SD = 3.38).
Differences in scores
The total level of risk (score 0-48) at the time of the assessment was similar, t(89) = .463, p = .645, d = 0.10, for boys (M =17.76, SD = 10.1) and girls (M =18.74, SD = 9.3). Taking this score into account, the prediction of injuries toward the mother was significant for male offenders (AUC = .842, 95% CI (.704 - .930), p = .001), but was not significant for female offenders (AUC = .660, 95% CI (.451 - .870), p = .157). The magnitude of the effect for the prediction of injuries towards their mothers for male offenders was higher than the 68.26% (PSTE = .6826) of all the potential effect sizes relating male offenders and mother victimization. No comparison was made between female and males for injuries to the father, given that daughters in this sample were only physically violent towards their mothers. The prediction of physical or psychological violence towards the father or mother did not prove significant in boys or girls.
Discussion
This paper aimed to describe the differences and similarities in risk factors for boys and girls involved in CPV, as well as to analyze the usefulness of a specific violence risk assessment tool for this type of violence, the CPVR (Loinaz et al., 2017). CPV has attracted great academic and professional interest in recent decades. However, despite various work and some consensus on the potential risk factors involved, few studies have addressed the comparison of these factors between boys and girls (Beckmann et al., 2017; Rico et al., 2017; Rosado et al., 2017). Those who engage in more violence or more serious violence, on the other hand, has been an issue more frequently addressed in the literature.
Type of violence and parent assaulted. The overall results regarding the use of violence have not been confirmed. Though there were no statistically significant differences, boys committed slightly more psychological abuse towards both mothers and fathers. Physical violence, on the other hand, was slightly more prevalent from daughters to mothers and from sons to fathers. Overall, violence was more common towards mothers (Aroca, Lorenzo, & Miró, 2014; Calvete & Orue, 2016; Condry & Miles, 2014; Ibabe & Jaureguizar, 2011; Lyons et al., 2015), and girls caused more injuries to mothers (10% more than boys) though this difference was not significant (Calvete & Orue, 2016; Del Hoyo-Bilbao et al., 2018; Walsh & Krienert, 2007). When violence is physical, our data show, in accordance with previous work, that it is more common that males assault their fathers (Simmons et al., 2019). However, we did not replicate the finding that the most serious assaults are directed towards the father (Simmons et al., 2019), since in this same, injuries were almost entirely suffered by the mothers.
In this sample, women had been reported to law enforcement significantly more, unlike the international literature that show a greater tendency to denounce male aggressors for their harm potential and greater fear by parents of serious consequences of their attacks (Armstrong et al., 2018; Miles & Condry, 2016; Simmons et al., 2018; Strom et al., 2014; Walsh & Krienert, 2007). As we will mention in the limitations, this result may be biased by the fact that more women in the sample came from judicial contexts which directly affects the generalizability of the results.
Characteristics of offenders and victims. The results have confirmed that boys and girls involved in CPV show little differences in their risk factors (Armstrong et al., 2018; Beckmann et al., 2017). The results of this work do not replicate previous findings of the larger proportion of single-parent families in boys (Armstrong et al., 2018), or the more impulsivity (Rico et al., 2017), and more prevalent substance use in girls (Armstrong et al., 2018). In fact, in our sample, substance abuse was found to be significantly higher for boys during the past (one year before and beyond the time of assessment).
One of the most noted and impactful variables in CPV, direct or indirect victimization at home (Gallego et al., 2019), was shown to be significantly more prevalent in girls, both in its direct form (bidirectionality of violence), as well as in indirect form: violence between parents or even problems of cohabitation other than VFP. Previous studies have pointed to the relevance of these variables for male aggressors (Calvete, Orue, et al., 2015; Ibabe & Jaureguizar, 2011; Ibabe et al., 2009). However, victimization at home has also been noted for its significant relationship to later IPV (perpetration and victimization) by girls, as well as to CPV (Izaguirre & Calvete, 2017). In the sample analyzed, it can be confirmed that female aggressors are more often victims of other forms of violence (Armstrong et al., 2018), in intimate relationships, in direct or indirect forms in the family, or at school (bullying). Girls were also more victims of IPV, though it is important to keep in mind that the search for information may already have been skewed by the sex of the person assessed as far as it is a variable designed, initially, for girls. Violence exposure at home affects social cognitive processing and should be included in CPV prevention and treatment plans (Contreras, León, & Cano-Lozano, 2020).
Both sexes showed a similar risk profile, with very few significant differences. In summary, girls come from more problematic contexts (with bidirectionality of violence, violence between parents, cohabitation problems and have parents with more problems) and have significantly lower self-esteem. Although the difference is not significant, they are also more often than boys (by more than 10%) victims of bullying, victims of IPV, have more anger problems, and have suffered more problematic educational styles as well as nonviolent parental conflicts. Boys have significantly more histories of substance abuse problems and further escalations of violence over the year before the assessment. They are more antisocial, though not significantly so.
It is necessary to remember some limitations when considering these results. The sample may be considered small and should be extended to confirm the conclusions of this work. In addition, although the differing origins of cases (clinical and judicial) is positive point in the aim of generalizability, the girls in the sample came significantly more from judicial contexts, which likely influenced some aspects of the results. Recent work confirmed the existence of differences among cases coming from clinical or judicial settings, with, for example, more complex and conflictive families in perpetrators from the judicial contexts (Loinaz & Ma de Sousa, 2020). On the other hand, although the tool was coded based on professional reports or even by the professionals working with the perpetrators themselves, there were not psychometric results that would allow a more accurate comparison of some of the risk factors of the tool as well as possible gender differences.
One possible future research line is the analysis of the capacity of the tool to measure therapeutic change, or how the ability of a given treatment to modify the risk factors presented in boys and girls. It is essential to expand the samples to confirm the risk variables and gender differences. It would also be of interest to complement the use of the CPVR (Loinaz et al., 2017; Loinaz & Ma de Sousa, 2020) with other tools designed to assess CPV, such as the Child to-Mother Violence Scale (Edenborough, Wilkes, Jackson, & Mannix, 2011), the Child-to-Parent Violence Questionnaire (CPV-Q) (Contreras, Bustos-Navarrete, & Cano, 2019), the Child-to-Parent Aggression Questionnaire (CPAQ) (Calvete et al., 2013; Del Hoyo-Bilbao et al., 2018) or the Abusive Behaviour by Children-Indices (ABC-I) (Simmons et al., 2019).
Results of studies examining CPV are highly heterogeneous, as evidenced by various available reviews (Holt & Shon, 2018; Moulds & Day, 2017; O’Hara et al., 2017; Simmons et al., 2018) and there is insufficient information regarding gender differences. This work makes it possible to state that, broadly speaking, while boys and girls who commit CPV are very similar in terms of their characteristics, and their violence is very similar in all forms (although only boys inflicted physical injures on their parents), the family backgrounds differ between both genders - families of female perpetrators appear to be more problematic than those of males. These results, as well as previous research, justify the use of common treatments and tools for male and female CPV perpetrators, while keeping in mind the differences that do exist between sexes that this work has highlighted.
