Introduction

The relationships between parents and their children can occasionally escalate to high levels of family conflict (^{Omer, 2004}). Several parent-child relationships transcend the limits of “conflictive relationships” to become “abusive relationships” of children to parents. These relationships have been defined as child-to-parent aggression (CPA) (^{Cottrell & Monk, 2004}). Although CPA is not a new problem, as indicated in research from the 1980s (^{Harbin & Madden, 1979}; ^{Peek, Fisher, & Kidwell, 1985}), there is still scarce empirical evidence available (^{Aroca-Montolío, Lorenzo-Modelo, & Miró-Pérez, 2014}; ^{Hong, Kral, Espelage, & Allen-Meares, 2012}; ^{Romero, Melero, Cánovas, & Antolín, 2007}; ^{Walsh & Krienert, 2007}). However, in recent years, the number of international studies focusing on CPA is increasing (^{Calvete, Orue, Gámez-Guadix, & Bushman, 2015}; Hong et al., 2012), as a result of the considerable interest that CPA is receiving because of the increasing frequency (^{Coogan, 2011}; International News Agency (EFE), 2009) as well as the serious consequences for its victims (^{Paterson, Luntz, Perlesz, & Cotton, 2002}).

The severity of the problem and the increasing number of cases emphasize the need to understand the factors involved in the development of CPA. CPA, like other types of family violence, has remained hidden (^{Robinson, Davidson, & Drebot, 2004}). Familiar factors are important aspects for understanding CPA (^{Calvete et al., 2015}; ^{Gámez-Guadix, Jaureguizar, Almendros, & Carrobles, 2012}; ^{Hong et al., 2012}). In the last three decades, several authors have emphasized the importance of evaluating the disciplinary practices of parents regarding the occurrence of CPA (Calvete, Orue, Gámez-Guadix, Del Hoyo-Bilbao, & López de Arroyabe, 2015; ^{Cottrel, 2005}; ^{Harbin & Madden, 1979}; ^{Ibabe, 2015}). Although recent reviews have indicated mixed results (Hong et al., 2012), several studies have shown that parenting practices are related with CPA (Calvete, Orue, & Sampedro, 2011; Calvete et al., 2015; Gámez-Guadix & Calvete, 2012; Ibabe, 2015; ^{Pagani, Tremblay, Nagin, Zoccolillo, Vitaro, & McDuff, 2009}).

Regarding parental practices, the strategy that has received most empirical attention is corporal punishment (CP). CP is likely to be the most controversial issue related to parental discipline (^{Berlin et al., 2009}; ^{Gámez-Guadix, Straus, Carrobles, Muñoz-Rivas, & Almendros, 2010}). CP is defined as the use of physical force with the intention of causing pain or discomfort in a child to correct or control the child’s behavior (Straus, 2001). In Spain, despite the scarcity of the literature that has focused on the consequences of CP, the European DAPHNE project indicated that between 50 and 60% of European families justify corporal punishment (Red de Universidades Valencianas para el fomento de la Investigación, el Desarrollo y la Innovación, 2013), and more than 60% reported having received CP in childhood (Gámez-Guadix, Straus et al., 2010). The prevalence of the acceptance and use of CP in Spain represent a problem because empirical evidence has shown the link between CP and negative consequences in the short and long term. For instance, CP has been associated with a greater likelihood of aggressive and antisocial behavior (Berlin et al., 2009; Gámez-Guadix, Straus, & Hersberguer, 2011; Straus, 2001, 2008; Gámez-Guadix, Straus et al., 2010; ^{Simons, Simons, & Wallace, 2004}). ^{Paolucci and Violato (2004}), in a meta-analysis of 70 studies, found that children who had received CP demonstrated significantly more behavioral problems. Additionally, in a meta-analysis of 88 empirical studies, ^{Gershoff (2002}) reported a relationship between CP and numerous negative consequences, including externalizing and internalizing problems in children. Furthermore, in a recent meta-analysis in which 45 longitudinal studies were analyzed, ^{Ferguson (2013}) found that CP was associated with internalizing and externalizing negative consequences, as well as with worse cognitive efficiency, although the effect size of the relationship was small.

As previously indicated, the available results are unclear about the relationship between CPA and disciplinary practices. This also occurs in the available evidence between CPA and CP. For example, in two studies, it was found that CP was positively associated with CPA (^{Gámez-Guadix et al., 2012}; ^{Pagani et al., 2009}). Similarly, in a recent qualitative study of families seeking assistance from a specialized service aimed at treating CPA problems, a high percentage of participants (67%) reported the use of CP in their family context (^{Calvete et al., 2015}). However, a correlational study with a community sample of 1427 adolescents found that the use of punishment by the parents, in general, was associated with less aggression in children (Calvete et al., 2011). Nevertheless, the authors of the study concluded that it is not the application of CP what is related to CPA. Adolescents who exercise CPA informed to have been exposed to fewer acts of discipline, including both functional and dysfunctional practices (Calvete et al, 2011).

One likely explanation for these inconsistent results regarding the effects of CP is the argument that using CP in a positive parenting context has no negative effects on the child (^{Benjet & Kazdin, 2003}; Larzerele, 2000). However, in addition to being inconsistent in their conclusions, there are few studies that analyze the moderating role of positive parenting with respect to the effects of CP on children (e.g., ^{Harper, Brown, Arias, & Brody, 2006}). Furthermore, a recent study with a community population that analyzed the relationship between CP and behavioral problems in the long term demonstrated that the negative effects of CP were not moderated by positive parenting (^{Gámez-Guadix, Straus et al., 2010}). In other words, CP was related to an antisocial orientation in children regardless of whether CP was administered in a context of positive parenting.

Therefore, given the previous inconsistent results regarding the consequence of CP and its habitual use and justification among parents, it is important to analyze its relationship with CPA. Additionally, the role of CP may be moderated not only by positive parenting but also by the adolescent’s age (Larzerele, 2000). Larzerele (2000) carried out a literature review in which he found that, when CP is administered at younger ages, it does not have negative consequences. Another important aspect for understanding the possible consequences of administrating CP is the sex of the child. It seems that the child’s sex does not moderate the relationship between CP and externalizing problems, as was concluded in a transversal study conducted in the Spanish population (^{Gámez-Guadix, Straus et al., 2010}). Furthermore, there is no agreement about the role of the child’s sex and CPA. Although some studies have indicated that there are no sex differences (Gámez-Guadix & Calvete, 2012; ^{Pagani et al., 2004}, 2009), other studies have suggested that girls are more likely to have higher scores on psychological CPV (e.g., Calvete et al., 2015; Calvete et al., 2013; Jaureguizar, ^{Ibabe, & Straus, 2013}) and boys on physical CPV (e.g., ^{Brezina, 1999}; Calvete et al., 2011; Calvete et al., 2015; Ibabe, Jaureguizar, & Bentler, 2013).

Moreover, the majority of the studies that focused on the analysis of family strategies and CPA are cross-sectional (for exceptions, see ^{Calvete et al., 2015}; ^{Brezina, 1999}), which is a major limitation of the research to date because it does not allow for the establishing of longitudinal relationships between variables (Calvete, Gámez-Guadix, & Orue 2014; Gámez-Guadix et al., 2012).

From the results of the available literature, the present study has the following aims. The first aim of this study was to analyze the longitudinal relationship between CP and CPA (both physical and psychological) in an adolescent Spanish sample. Because previous studies have reported a cross-sectional relationship between severe discipline and CP with CPA (^{Gámez-Guadix et al., 2012}; ^{Pagani et al., 2009}), we hypothesized that receiving CP within the last year will increase the likelihood of child-to-parent aggression perpetrated 6 months later. The second aim of this study was to analyze the moderating role of positive parenting in the relationship between CP and CPA. Despite the controversy regarding the consequences of CP (e.g., ^{Benjet & Kazdin, 2003}), drawing on the findings of studies carried out in the Spanish context on the moderating role of positive parenting in the relationship of CP and other externalizing behaviors (Gámez-Guadix, Straus et al., 2010), we hypothesized that positive parenting would not moderate the relationship between CP and future CPA. The third aim of the study was to analyze whether sex moderates the association between CP and CPA. As in the previous point, there are inconsistencies in previous studies regarding the role of the child’s sex in CPA (e.g., Gámez-Guadix & Calvete, 2012; Jaureguizar, ^{Ibabe, & Straus, 2013}). Nevertheless, attending to one of the few studies conducted in the Spanish population in which sex did not moderate the association between CP and behavior problems in adolescents, we hypothesized that the child’s sex would not moderate the predictive relationship between CP and future CPA (Gámez-Guadix et al., 2010). Finally, the fourth aim of the study was to analyze whether the adolescent’s age moderates the predictive association between CP and CPA. Given the absence of previous studies that analyzed the moderating role of adolescents’ age on this relationship, we analyzed this issue exploratorily.

Method

Results

Descriptive Analyses

First, the prevalence of CPA within the last year was calculated. The total prevalence for psychological CPA was 91.2% at T1 and 90.9% at T2. For physical CPA, the total prevalence was 9% at T1 and 8.2% at T2. To estimate the total psychological CPA and total physical CPA, we included all the adolescents who admitted to at least one act of these categories. As these percentages also included cases of a single incident occurring during the past year, this may lead to overestimation of the prevalence of CPA. Therefore, we estimated severe forms of CPA, following the procedure indicated by the authors of the questionnaire (^{Calvete et al., 2013}). Incidents of severe psychological CPA included adolescents who reported performing threatening behavior, insulting, blackmailing, taking money without permission, doing something to annoy the parent, and disobeying a parent’s important request more than 6 times within the past year (Calvete et al., 2013). Thus, the prevalence of severe psychological CPA was 14% at T1 and 13.1% at T2. To estimate severe physical CPA, the authors of questionnaire proposed to calculate the percentage of adolescents who reported having committed an act of physical aggression against their parents at least 3-5 times in the past year. In this study, the prevalence of severe physical CPA was 1.8% at T1 and 2.8% at T2.

Regarding CP, to consider the existence of CP, we used the criteria indicated by the questionnaire’s authors (^{Straus & Fauchier, 2007}). They consider the existence of CP when the adolescents reported their parents’ shaking or grabbing them to get their attention; spanking, slapping, smacking, or swatting them; using a paddle, hairbrush, belt, or other object to punish them at least once within the past year (1, 2, 3, 4, and 5 on the response scale). Thus, 43% of the adolescents admitted that their parents used CP (e.g., spanking, slapping, smacking, or swatting).

Table 1 shows the correlations and descriptive statistics (mean and standard deviation values) for the variables in the study. As shown, the highest correlations are established between the same variable at T1 and T2, indicating a high stability between CPA from T1 to T2. Additionally, the correlation between psychological CPA and physical CPA at the same time point (e.g., at T1) indicate coexistence between the two types of violence. CP at T1 showed a positive and statistically significant relationship relevant to all types of CPA at both T1 and T2 (Table 1).

Table 1: Pearson correlations and descriptive statistics (means and standard deviations) for the variables in this study. 

Analyses of the relationship between corporal punishment, child-to-parent violence and the interaction of age, sex and positive parenting

To analyze the relationships between the variables in the model, we used the statistical software EQS 6.1 (^{Bentler, 2005}). We used the robust maximum likelihood (ML) estimation method with the Satorra-Bentler scale chi-squared (S-B χ²) because the data did not meet the assumption of normality (the normalized Mardia’s coefficient = 273.58). To study the adequacy of the estimated models, we used the comparative fit index (CFI), the standardized root mean square residual (SRMR), and the root mean square error of approximation (RMSEA). For the NNFI and the CFI, values > .90 indicate an acceptable fit. Values on the SRMR and the RMSEA < .08 indicate an acceptable fit (^{Byrne, 2006}; ^{Hu & Bentler, 1999}).

First, we estimated a model that included the analysis of the relationship between age and sex of the child, CP, and positive parenting at T1, and psychological and physical CPA at T2. The child’s sex was treated as effect codification (-1 and 1) as recommended by ^{Frazier, Tix, and Barron (2004}), because it is a dichotomous variable with mutually exclusive categories. Furthermore, we used three interaction terms, CP × positive parenting, CP × age, and CP × sex. To create the interaction terms we followed the steps suggested by Homlbeck (1997). First, continuous variables (e.g., positive parenting) were standardized. Then, the interaction term was created by multiplying CP with positive parenting and age and sex of the child. The model also included autoregressive paths between the same variable in T1 and T2. This strategy allowed us to analyze whether the change in CPA at T2 could be explained by the remaining predictors once the baseline of this same variable was controlled at T1.

The estimated initial model showed that several paths were not statistically significant. For example, the relationship between CP at T1 and physical CPA at T2 was not significant. Additionally, the interaction between CP, age and sex of the child and positive parenting at T1 was not related to either physical or psychological CPA at T2. These paths were removed from the model, which was subsequently re-estimated with only the significant paths. The fit indices for the estimated final model were adequate: χ² (3, N = 896) = 21.86; NFI = .97; CFI = .98; RMSEA = .058; and SRMR = .016.

Figure 1 shows the hypothesized theoretical model and Figure 2 shows the standardized parameters of the final model. At the longitudinal level, all of the autoregressive paths were high and significant: .55 and .70 (all, p < .001). Sex, specifically being female, was associated with increased psychological CPA at T2. Furthermore, positive parenting at T1 decreased both types of CPA at T2. Additionally, the results showed that CP at T1 significantly predicted increased psychological CPA at T2. This relationship was significant regardless of the child’s age or sex and the context of positive parenting in which CP was used. Finally, psychological CPA at T1 significantly increased physical CPA at T2.

Figure 1: Hypothesized theoretical model. Note. CPA (Child-to-parent aggression). 

Figure 2: Estimated final model. Note. CPA (Child-to-parent aggression) * p < .05; **p < .01; ***p < .001 

Discussion

This study is the first to analyze the longitudinal relationship between receiving CP and perpetration of CPA in adolescents. At the cross-sectional level, CP was associated with both physical and psychological CPA, and at the longitudinal level, CP increased the likelihood of psychological CPA. Positive parenting, sex and age of the child did not moderate the relationship between CP and CPA. We subsequently discussed the primary theoretical and practical implications of these results.

Consistent with previous studies (^{Gámez-Guadix, Straus et al., 2010}; Straus, 2001), we found high rates of CP by parents. Specifically, more than 40% of the adolescents were physically hit during the previous year. The prevalence rates observed in the present study are lower than those found in previous studies with Spanish samples, which showed prevalence rates of 60% more than a decade ago (for a revision, see Gámez-Guadix, Straus et al., 2010). This finding could indicate that the use of CP is decreasing in recent years in Spain. Even so, the rate of CP remains considerably high and should be the focus of educational interventions.

In this study, CPA rates reported by the adolescents were also relevant. In particular, between 8.2% and 9% of the adolescents admitted to having perpetrated at least one act of physical aggression against their parents in the assessment period. Repeated psychological violence occurred more frequently as between 13.1% and 14% of the participants reported performing threatening behavior, insulting, blackmailing, taking money without permission, doing something to annoy the parent and disobeying a parent’s important request more than 6 times in the last year. These results are similar to those obtained in several epidemiological studies (^{Calvete et al., 2013}; Calvete et al., 2011; ^{Pagani et al., 2004}, 2009; ^{Ulman & Straus, 2003}), and they highlight the magnitude of this modality of family violence in society and the need to develop preventative actions.

CP was found to be associated with an increased likelihood of psychological CPA over time regardless of whether there was positive parenting. This finding is an important result because it provides empirical evidence contradicting the claim that CP has no negative consequences when CP is used by loving and supportive parents (^{Larzelere, 2000}; ^{Ripoll-Núñez & Rohner, 2006}). Overall, the results are congruent with the idea that although CP might produce conformity in the immediate situation, in the long term CP may increase the likelihood of deviance (^{Straus, 2001}; Straus, Sugarman, & Giles-Sims, 1997), including aggressive behavior and CPA. Several mechanisms, such as less child internalization of the parents’ values and low self-control, may explain the relationship between CP and future aggressive behavior in children (^{Gottfredson & Hirschi, 1994}). Similarly, a child’s sex and age do not moderate the relationship between CP and psychological CPA. Thus, our findings have suggested that CP increases the likelihood of psychological CPA regardless of the sex and age of adolescents.

This study supported the hypothesized relationship between CP and physical CPA at the cross-sectional level but not at the longitudinal level. The type of sample could likely have influenced this result because the levels of physical violence against parents are low in adolescents of community samples, and it may have attenuated the relationship between these variables. The relationship between physical punishment and CPA could emerge with greater time intervals (e.g., one year). Future studies should specify the conditions in which CP and physical CPA are associated.

Importantly, psychological CPA at T1 increased the likelihood of physical CPA at T2, suggesting that CPA is stable over time. Thus, psychological aggressions against parents should not be ignored. Instead, these acts could be indicative that something in the family relationships is wrong. Early educational interventions may be sufficient to reorient the patterns of parent-child relationship.

Conclusions, limitations, implications for practice and future prospects of the study

The results of this study should be interpreted in the context of third important limitations. First, the data were based on information provided by adolescents using self-report measures. It is recommended that future studies include information obtained from parents and use other assessment strategies (e.g., interviews). Second, although this is the first study to analyze the longitudinal relationship between CP and CPA, the relationship between these variables could emerge over periods exceeding six months. Future studies should examine these relationships using other larger time intervals and more time points. It would also be important to include measures of earlier CP because the incidence of CP is likely to decrease as children grow older and can defend themselves against their parents. Third, the variables CP and positive parenting are not included at Time 2. Thus, future studies should include these variables at Time 2 with the aim of studying the possible reciprocal relationships between these variables and CPA.

The results demonstrate that CP plays a negative role in the behavior of children towards their parents even when this occurs in a positive parenting context. Preventive interventions should inform and sensitize families concerning the negative consequences of CP. Additionally, families should be provided with adaptive strategies of conflict resolution and discipline. Furthermore, the findings highlight the importance of positive educational practices that do not involve the use of physical punishment. Therefore, strategies based on monitoring and control, such as positive reinforcement of appropriate behavior and the use of non-aversive punishment (e.g., penalty task, restorative behavior and deprivation of privileges) (^{Straus & Fauchier, 2007}), and the adequate management of parental attention can contribute to the prevention of CPA.