Child-to-Parent Violence during Confinement Due to COVID-19: Relationship with Other Forms of Family Violence and Psychosocial Stressors in Spanish Youth

: This study analyzes child-to-parent violence during strict confinement due to COVID-19 as well as its association with other forms of family violence (parent-to-child violence and exposure to violence between parents) and with different psychosocial stressors: Academics/work, family coexistence, finances, COVID-19 and physical and psychological health. The study included 2245 young people (52.8% females) aged between 18 and 25 years ( M = 21.52 years, SD = 2.07 years). The results show that more than half of the young people reported having performed at least one violent behavior toward their parents. Child-to-parent violence is significantly related to violence from parents to their young adult children, to exposure to violence between parents and to different psychosocial stressors: Academics/work, family coexistence and psychological health. The individual and additive predictive value of other forms of family violence and psychosocial stressors in child-to-parent violence is confirmed. Prevention and intervention programs for child-to-parent violence must take into account the role of other forms of family violence present as well as the influence of psychosocial stressors. of the on of


Introduction
The global pandemic caused by COVID-19 has abruptly and significantly changed the way of life of people around the world. It is an unusual situation, without recent precedents, with multiple sources of stress and with a strong impact on the daily activities of people, interpersonal relationships and, predictably, psychological well-being. Among the innumerable consequences, home confinement, as a measure to contain the pandemic, has forced many families into an intense and uninterrupted coexistence that, in many cases, has generated crises and tensions and a deterioration in family relationships and even triggered violent behaviors. For those who live in small houses or without open spaces, the stress generated has been greater.
In this situation, the home becomes a risk scenario for victims of family violence, as they are required to stay away from people who can validate their experiences and offer them help. Home confinement requires a forced and prolonged coexistence of aggressors and victims [1]. In this way, the decrease in available social support together with the difficult access to social resources offers the ideal situation in which aggressors in a family can exercise control over their victims (an essential characteristic of these forms of violence). Additionally, the pandemic has increased unemployment rates, generating serious economic problems in many households. Economic difficulties can increase stress and put more tension on relationships as well as reduce the opportunities for victims to leave relationships. Many victims of family violence have faced the "worst case scenario", that is, trapped in the home with a violent person without contact with the outside world [2].
In this sense, several studies carried out during the last year warn of family violence during the pandemic [2][3][4]. In Europe, as reported by Davies and Batha [5], charities for family abuse made a call to society to increase shelter resources for battered women in the face of a possible increase in gender-based violence during home confinement. In Spain, the number of calls to the gender violence hotline increased during the months of April, May and June of 2020 compared to the number of calls recorded in the previous two years [6]. Furthermore, the presence of adolescent and young adult children at home due to the closure of schools and universities during quarantine has also increased exposure to family violence, such as witnessing intimate partner violence [1]. This aspect is important if we take into account that the greater the exposure of young people to parental violence, the greater is the risk of developing physical, mental and social health problems in the future [7]. More specifically, exposure to parental violence can increase the risk that the young person will have subsequent violent experiences (e.g., [8,9]).
In the case of parent-to-child violence, while there has been a decrease in reports during the period of confinement, this may be due to a reduction in the opportunities of detection, due to, among other things, the closure of schools and access to other community services [2]. However, Lawson, Pie and Simon [10] report that the loss of parental work during the pandemic has increased the probability of child psychological abuse, especially when there was already a previous history of this type of family violence. Along these same lines, Lee et al. [11] point out that factors such as social isolation and loss of employment during this period have been associated with parental self-reports of physical and emotional neglect and verbal aggression toward their children.
Another type of family violence, less known than the previous ones, is the violence that adolescents and young adults exercise toward their parents, also called child-to-parent violence (CPV). For this type of violence, studies during the pandemic are practically nonexistent. In fact, we only found a recent study carried out by Condry et al. [12] in the United Kingdom. These authors evaluated mothers and fathers of adolescents who were aggressive in the home and analyzed parenting experiences during confinement (April-June 2020), finding that 70% of parents reported an increase in CPV episodes during this period.
Given the scarcity of studies that demonstrate the magnitude of family violence during confinement due to COVID-19, in this study, we investigate the presence of CPV, other forms of family violence and psychosocial stressors linked to this situation, as well as the associations between these factors. The exceptionality of the situation offers us a unique opportunity to determine the impact of external factors in this type of violence, an aspect that is little discussed, especially in the case of CPV.

CPV and Its Relationship with Other Forms of Family Violence
CPV has received increasing interest during the last decade due to the increase in reports filed by parents [13]. The current definition of the phenomenon includes the use of different forms of violence (physical, psychological and economic) to gain power and control [14] and to dominate parents [15]. Awareness of the perpetrator of violent behavior and its repeated nature is required, excluding isolated acts of violence [16].
While most studies have focused on the period between 12 and 18 years, CPV does not end at 18 years of age. Data from different countries reveal that at least half of young adults between 18 and 25 years old continue to live with their parents and warn that it may be a phenomenon that increases in importance considering the increasing age at which young people become emancipated in western countries [17]. However, studies on CPV in young adults are surprisingly still very scarce. A recent study by Simmons et al. [18] confirmed the high prevalence of CPV among young Australian adults (18-25 years). Specifically, one in seven was classified as abusive toward one parent. Ibabe et al. [19] found in a sample of Spanish young people between 18 and 25 years of age prevalence rates of physical CPV of 5% (technical abuse) and 14% (zero tolerance) and of psychological CPV of 67% (technical abuse) and 94% (zero tolerance) during the last year. Regarding the differences based on gender, Ibabe et al. [19] found that daughters exercise psychological CPV more frequently than do sons when the technical abuse criterion is applied, while Simmons et al. [18] report a higher frequency of abuse toward parents by sons than by daughters.
Numerous studies in the field of CPV have found that young people who abuse their parents are more likely exposed to violence in the family environment [8,[18][19][20][21][22][23][24][25]. Specifically, significant positive relationships are found between CPV and parent-to-child violence [8,20,21,[23][24][25]. In turn, abuse from parents predicts an increase in CPV. Positive relationships are also found between CPV and exposure to violence between parents. [8,19,21,23,25], being a significant predictor of CPV [8,19,21,[23][24][25]. Likewise, the results of various studies suggest that although both types of family violence are relevant in CPV, parent-to-child violence has a greater predictive capacity than does exposure to violence between parents [8,18,21,24]. A recent meta-analysis found that parent-to-child violence explains 16.8% of CPV, compared to 13.7% explained by exposure to violence between parents [26]. On the other hand, recent studies point to the need to also study the additive effect of different types of family violence in CPV, given its high co-occurrence. In this sense, Beckmann [8] finds that together, parent-to-child physical violence and parental intimate partner violence contribute to explaining CPV to a greater extent than each type of violence separately.

Family Violence and Its Relationship with Psychosocial Stressors
The measures adopted to contain the pandemic (e.g., closure of schools, universities, nonessential economic activities, etc.) have generated a variety of psychosocial stressors (economic problems, loss of employment, decrease in social support, etc.). Specifically, in the Spanish population, it has been found that during confinement, economic concerns stand out [27,28]. Recent studies conducted with young people in different countries also show the presence of academic and economic concerns [29], such as an overload of academic tasks, among others [30], and problems related to family coexistence during confinement [31]. They also highlight concerns related to the infection of a relative or acquaintance with COVID-19, both in studies carried out in Spain [27,28] as in other countries [29,32,33]. Likewise, other studies have confirmed the psychological impact of COVID-19. For example, studies carried out in Spain indicate that the younger population, compared to older age groups, had a greater psychological impact by COVID-19 in terms of stress, anxiety and depression [27,28,34,35], with women having higher levels [27,28,34]. Along those same lines, in the United Kingdom, it has been observed that stress levels in the adult population increased during the first months of the pandemic, especially in young people between 18 and 34 years of age and, particularly, in women [36]. In addition, problems related to sleep have been reported. For example, in a study conducted in Greece with an adult population, almost 40% of the participants had insomnia problems, particularly in the group of women [37]. In another study conducted in China, almost 20% of the adult population had sleep disturbances, especially the younger group (under 35 years) compared to the older age groups [38].
The consequences of the pandemic, in terms of psychosocial stressors, could have led to or favored domestic violence behaviors [39]. The classic literature on family violence highlights that the presence of various stressors can increase the risk of violence in family and intimate relationships when it is accompanied by other characteristics, such as poor coping skills and the justification of violence as an acceptable way to cope with stress [40]. In this sense, some authors have pointed out that there is a direct association between various stressful life events and gender-based violence (see review by [41]). More recently, it has also been found that the presence of economic stressors (e.g., economic difficulties paying for rent, food, etc.) is associated with a greater probability of physical intimate partner violence [42]. In the context of intimate partner violence, differences have also been found in the nature of psychosocial stressors based on the gender of the aggressor, observing that work-related stressors (e.g., unemployment, layoffs, job changes, etc.) are associated with violence by men, while work-related stressors and interpersonal stressors (e.g., death of a loved one and loss of something of value) are associated with violence by women [43]. In addition, the experience of stressful life events (e.g., death of a close family member or friend, own illness or of a loved one, etc.) has also been positively associated with an increased risk of dating violence during adolescence [44].
Regarding child abuse, while it is understood that it is a multicausal phenomenon, the classic theoretical models on this topic have already highlighted the role played by various stressors, such as housing conditions, including the lack of space [45], and parental unemployment or other economic problems [45,46]. More recent studies also highlight the role of these economic stressors in the prediction of child abuse. Specifically, it has been found that low income is a risk factor [47] and that unemployment is linked to behaviors of neglect and physical abuse [48].
In the case of CPV, there are no studies to date that have analyzed the influence of psychosocial stressors on this type of violence [17]. However, the associations between different types of psychosocial stressors and different forms of violence during childhood and adolescence have been examined. Some stressors related to family coexistence, economic difficulties, etc. are related to behavioral problems such as aggression during adolescence [49]. In this regard, with respect to the conditions of family coexistence, Makinde et al. [50] found that overcrowding in the home is a predictor of antisocial behaviors in adolescents, violence by siblings and parents and observing violence between family members. A positive relationship has also been observed between sleep problems and aggression in incarcerated adolescent and young offenders [51] as in adolescents from the community population (see review [52,53]).
Previous research on family violence has indicated that analyzing the accumulation of risk factors is more effective in predicting child abuse than analyzing the individual contribution of each factor [54,55]. In this regard, it has been found that the parental perception of feeling overwhelmed by the accumulation of different stressors increases the risk of violence toward their children [56]. Specifically, the cumulative risk model establishes that as the number of risk factors increases, regardless of their nature, the greater is the potential for negative results [54]. Studies from other areas also provide results that are along these lines; it has been observed that the accumulation of risk factors predicts an increase in behavioral problems during adolescence [57].
In short, it is necessary to provide data on the impact of confinement due to COVID-19 and the associated psychosocial stressors in the presence of violent behaviors in the family, especially CPV, because there are practically no studies on this topic. The global pandemic offers a unique opportunity to better understand how external factors increase the risk of family violence.

Objectives and Hypotheses of this Study
In this study, we propose, first, to examine in Spanish youth the frequency of CPV behaviors toward the mother and the father and other forms of violence in the family context (parent-to-child violence and exposure to violence between parents) as well as the frequency of various psychosocial stressors associated with confinement, studying the differences based on gender. We hypothesize that psychological CPV is more frequent in females than in males [19] and that the exposure to violence by the father toward the mother is equal to or slightly higher than that of the mother toward the father [19,58]. Additionally, a high presence of stressors related to academics/work [29,30], family coexistence [31], finances [27][28][29] and the infection of a family member or close friend with COVID-19 [27,28] is expected, as is a high frequency of stressors related to mental health [27,28,34] and sleep disturbances [37,38], with higher rates among females than males [27,28,34,37].
Second, the relationship between CPV and other types of family violence is examined as well as its relationship with different psychosocial stressors. We hope to find a positive and significant relationship between CPV and parent-to-child violence [8,20,24] and with exposure to violence between parents [8,19]. A positive and significant relationship between CPV and the presence of psychosocial stressors is expected, both when analyzed individually and cumulatively [49,50,53].
Finally, we aim to determine the individual and additive contributions of the different types of family violence and of the different types of psychosocial stressors in the prediction of CPV. It is expected that, independently, parent-to-child violence [8,18,20,24], exposure to violence between parents [8,18,19,24] and psychosocial stressors are significant predictors of CPV [49,50,53] and that, together, they contribute to significantly improving the prediction of CPV [8,54,57], with parent-to-child violence being the predictor with the greatest explanatory capacity [8]. Finally, we also analyze the effect of the accumulation of stressors on the prediction of CPV.

Participants
Applying intentional nonprobability sampling, a total of 2,245 young people (52.8% females) aged between 18 and 25 years (M = 21.52 years, SD = 2.07 years) of Spanish nationality (98.2%) or residents in Spain (1.8%) participated in this study. The majority of the participants were from Andalusia, a region of southern Spain (82.5%), with the others living in other regions of the country (17.5%). The minimum sample size required with a confidence level of 95% and a maximum marginal error of 2.5% was 1,537 participants.
The academic level of the sample was distributed as follows: Graduate university studies (5.3%), undergraduate university studies (50.7%), baccalaureate studies (11th and 12th grades) (35.6%), secondary studies (7th to 10th grades) (6.7%), primary studies (1.2%) and non-education (0.5%). Compared with the general population of young people in the region of Spain where the study was mainly conducted (Andalusia), the present sample presents a higher percentage of university students (56% compared to 34.9%), similar percentage of young people with baccalaureate studies (35.6% compared to 49%) and a lower percentage of young people with primary and secondary studies (8.4% compared to 37.4%) [59].
The majority of participants (83.1%) reported that their parents were married, while 12% reported that their parents were divorced or separated. Nine percent of the sample was an only child, 59% had one sibling, 24.2% had two siblings and the remaining 7.8% had three or more siblings. Finally, 18% of mothers and 16.4% of fathers had completed university studies, 38.8% of mothers and 36.8% of fathers had a secondary education, 36.5% of mothers and 38.8% of fathers had a primary education and 6.7% of mothers and 8% of fathers had no education.

Tools
Sociodemographic data: An ad hoc sociodemographic questionnaire was designed that included questions regarding the characteristics of the participants (gender, age, origin and academic level) and the characteristics of the family structure (marital status and academic level of the parents and number of siblings).
The following tools were applied to obtain data only for the period of strict home confinement in Spain (from March 14 to May 10, 2020).
Child-to-parent violence: The Child-to-parent Violence Questionnaire, youth version (CPV-Q-J [60]) is composed of 19 parallel items (mother and father separately) distributed in four subscales that measure the frequency of behaviors of physical (5 items), psychological (6 items), economic (4 items) violence and behaviors aimed at exercising control and dominance over parents (4 items). The Likert response scale includes the following values: 0 = never (has never occurred); 1 = rarely (occurred once a month); 2 = a few times (occurred 2-3 times a month); 3 = quite a few times (occurred 4-5 times a month) and 4 = very often (occurred six or more times a month). The internal consistency in this study was α ordinal = 0.930 for the maternal figure and α ordinal = 0.923 for the father figure.
Family Violence: The Violence Exposure Scale (VES) [61] evaluates exposure to violence in different contexts. In this study, the subscale of exposure to domestic violence, which is composed of two subscales, was used. The first measures the frequency of physical (1 item), psychological (1 item) and verbal (1 item) violence by parents. For this study, violence by the mother was differentiated (α ordinal = 0.922) from that by the father (α ordinal = 0.972). The second subscale evaluates the frequency of exposure to physical (1 item), psychological (1 item) and verbal (1 item) violence between parents. Additionally, in this case, the exposure to violence by the mother toward the father (α ordinal = 0.910) and from father toward the mother (α ordinal = 0.874) were differentiated. Both subscales have a Likert response format that includes the following values: 0 = never; 1 = once; 2 = sometimes; 3 = many times; 4 = every day.
Psychosocial stressors: An ad hoc questionnaire for psychosocial stressors related to confinement due to COVID-19 was designed, composed of 18 items that evaluate the presence (1) or absence (0) of stressors classified into six types: Academics/work (5 items), family coexistence (2 items), finances (2 items), COVID-19 (4 items), physical health (2 items) and mental health (3 items). In addition, the total score of accumulated psychosocial stressors was obtained by summing the scores for all items. The internal consistency of the scale was α ordinal = 0.637.

Procedure
The assessment protocol was conducted in accordance with the Declaration of Helsinki and was approved by the Ethics Committee of the University of Jaén (Reference: OCT.19/1.PRY). Data collection was carried out during the months in which the government of Spain decreed a state of alarm in the country and strictly mandated home confinement throughout the national territory due to the COVID-19 pandemic. Specifically, the data for this study were collected between 17 April and 10 May 2020, both included. Given the restrictions of social contact, the data were collected online through a questionnaire designed on the Google Forms platform. The participants were contacted by email and by various social networks (Facebook, WhatsApp and Twitter), applying the snowball technique. Before responding to the questionnaire, participants received information about the study and informed consent was obtained. Participation was voluntary, anonymous and confidential. The inclusion criteria were age between 18 and 25 years old, Spanish nationality or residence and living with at least one parent during the period of confinement. The design of the research is framed within descriptive studies through crosssectional surveys [62].

Data Analysis
The internal consistency of the questionnaires was carried out using the ordinal alpha coefficient, considered the best statistic of reliability of instruments when the data are ordinal and nominal in nature and present an asymmetric distribution [63]. This is so given that this coefficient is based on polychoric and tetrachoric correlations [64].
Descriptive, correlational and predictive analyses were performed to contrast the proposed hypotheses. A significance level of p < 0.05 was considered in all analyses.
First, contingency analyses were carried out to examine the frequencies of CPV, other types of family violence and psychosocial stressors related to confinement. To determine the frequency of CPV and other types of family violence, a zero tolerance criterion was used, referring to the presence of one or more violent behaviors per month (any response other than 0 on the response scale). The presence of one or more stressors was taken into account to determine the frequency of each type of stressor, and the presence of two or more stressors was used to determine the frequency of accumulated stressors. Chisquared analyses were performed to examine whether there were significant differences between females and males in both types of violence and psychosocial stressors, and effect sizes were calculated.
Second, the possible relationships between CPV toward mothers and fathers and the other types of family violence and psychosocial stressors were examined using bivariate Spearman correlations, given that the corrected Lilliefors significance of the Kolmogorov-Smirnov test established that the distribution of the sample in the study variables did not follow a normal distribution, something common in the study of violent behaviors in community populations.
Finally, two sets of multiple linear regressions were carried out to examine the individual and additive effects of the types of family violence and stressors (independent variables) on the prediction of CPV toward the mother and CPV toward the father (dependent variables). The assumptions of the regressions were tested and both the effect sizes [65] and the statistical power [66] of the tested models were calculated. In the first set of regressions, the predictor variables were introduced in the model hierarchically following a theoretical order of relevance [67]. In the first block, the predictors referring to motherto-child violence and father-to-child violence were introduced. In the second block, the predictors referring to exposure to mother-to-father violence and exposure to father-to mother violence were added and in the third block, the six predictors referring to the six types of psychosocial stressors were added. The significance of the change in R 2 between blocks was included. This methodology allows determining the additive effect of each block of predictors on the final model. To provide additional information, a second set of regressions was performed in which the same procedure described above was applied with some modifications. Regarding the predictors referring to other types of family violence, only those that were significant in the final model of the previous regressions were introduced. Regarding the predictors referring to psychosocial stressors related to confinement, the single predictor referring to the accumulation of stressors was added, which allowed determining the cumulative effect of several psychosocial stressors on CPV toward mothers and fathers.

Frequencies of CPV and other Types of Family Violence
A total of 65.2% of the young people exercised some violent behavior toward their mother and 59.4% toward their father, with statistically significant differences based on gender (see Table 1). Specifically, a significantly higher proportion of females than males reported having exercised CPV toward mothers χ 2 (1, 2221) = 22.5, p < 0.001 and toward parents χ 2 (1, 2181) = 10.4, p = 0.001. By type of CPV, the most frequent was psychological violence, with prevalence rates that ranged from 40.1% to 61.3%, followed by control and domination behaviors (36.5-43%), economic violence (12-16.6%) and physical violence (1.7-3.3%). Likewise, significant differences were found between the proportions of females and males regarding CPV types. Specifically, more daughters than sons exercised psychological CPV toward mothers χ 2 (1, 2221) = 33.1, p < 0.001 and toward fathers χ 2 (1, 2181) = 28.9, p < 0.001 as well as control/dominance behaviors toward mothers χ 2 (1, 2221) = 4.0, p = 0.040. On the other hand, a significantly higher proportion of sons than daughters exercised physical violence χ 2 (1, 2181) = 6.0, p = 0.014 and economic violence CPV χ 2 (1, 2181) = 8.2, p = 0.004 toward fathers.  The analysis of the other types of family violence (see Table 1) indicated that 16% and 14.9% of the participants suffered at least one violent behavior by their mother and father, respectively. The most frequent type was psychological violence (12.4-17.1%), followed by verbal violence (4.6-5.7%) and physical violence (2.8-3.4%). Significant differences were only found between females and males in psychological violence by the mother. Specifically, more daughters than sons reported being victims of psychological violence from their mothers χ 2 (1, 2221) = 9.7, p = 0.002. Likewise, 11.6% and 11.9% of young people observed violence from the mother toward the father and from the father toward the mother, respectively. Again, exposure to psychological violence between parents was the most frequent (9.9-13.4%), followed by exposure to verbal violence (0.4-1.2%) and exposure to physical violence (0.9-1%). Additionally, in this case, significant differences were found between females and males. Specifically, a greater proportion of daughters than sons witnessed psychological violence from the father toward the mother χ 2 (1, 2159) = 6.5, p = 0.010.

Frequencies of Psychosocial Stressors Related to Confinement
By type of psychosocial stressor, approximately half of the sample (44.6%) experienced some stressor related to academics/work, with differences between females and males being more frequent among females χ 2 (1, 2245) = 19.7, p < 0.001. Fear regarding the results of the academic year (95.9%) and fear regarding the continuity of employment (88.9%) stand out (see Table 2). A total of 29.3% of young people experienced some of the stressors related to family coexistence, being more frequent among females than among males χ 2 (1, 2245) = 10.5, p = 0.001. Specifically, 21.8% experienced overcrowding or lack of personal space during confinement and 12.6% had difficulty spending time in confinement with the family. On the other hand, 12.8%, females and males in equal proportion, experienced some stressors related to economic reasons. For example, almost 12% of young people had serious economic problems during this period. Likewise, 45.1% of the participants experienced some stressor related to COVID-19, the most frequent being the infection (43.4%) and the death (9%) of a family member or close friend, with differences according to the gender. Specifically, a significantly higher proportion of females than males experienced both stressors (disease χ 2 (1, 2245) = 32.5, p < 0.001, and death χ 2 (1, 2245) = 12.4, p < 0.001). A total of 10.7%, females and males in equal proportion, experienced some stressor related to physical health. Specifically, 8% and 4.4% reported the illness and the death of a family member or close friend, respectively, not due to COVID-19. Finally, psychosocial stressors related to mental health were the most frequent among the young people in the study. A total of 96.6% experienced some of them, with differences according to the gender. Specifically, more females than males experienced all the stressors included in this typology χ 2 (1, 2245) = 31.8, p < 0.001. The psychological impact on a family member (88.4%) was the most frequent stressor, followed by their own psychological impact (87.1%) and sleep disturbances (77%). Almost all young people (94.7%) experienced an accumulation of psychosocial stressors during this period (two or more), with significant differences between females and males (see Table 2). Specifically, more females than males experienced cumulative stressors χ 2 (1) = 18.6, p < 0.001.

Relationships between CPV and other Types of Family Violence and Psychosocial Stressors Related to Confinement
CPV toward mothers and fathers was positively and significantly related to the types of family violence included in this study (see Table 3). The highest coefficients were found for the relationship between CPV toward mothers and mother-to-child violence (2159) = 0.37, p < 0.001 and in the relationship between CPV toward fathers and father-to-child violence (2159) = 0.37, p < 0.001; also notable are the coefficients of the relationship between mother-to-child violence and exposure to violence by the mother toward the father (2159) = 0.43, p < 0.001 and between father-to-child violence and exposure to violence by the father toward the mother (2159) = 0.48, p < 0.001. Table 3. Spearman correlations between CPV and other types of family violence and psychosocial stressors related to confinement. Likewise, CPV was positively and significantly correlated both with the type of psychosocial stressor (academics/work, family coexistence and mental health) and with the accumulation of psychosocial stressors (mother CPV (2245) = 0.16, p < 0.001 and father CPV (2245) = 0.15, p < 0.001).

Multiple Linear Regressions of CPV toward Mothers and Fathers
Both sets of regressions had good indicators of goodness of fit [68]. Likewise, medium effect sizes stood out [65] in all tested models, with perfect statistical power [66]. Values for the variance inflation factors lower than five determined the absence of multicollinarity between the predictors, and the Cook distances indicated the absence of cases that contributed disproportionately in the regressions. The values for the Durbin-Watson statistic allowed the determination of the independence of the errors of the residuals. However, the graphs, histograms and the Kolmogorov-Smirnov test for the standardized and nonstandardized residuals led to the acceptance of their nonnormality and heteroscedasticity. In this sense, it has been suggested that failure to comply with these two assumptions should not be an impediment to the selection of parametric statistical techniques if the sample size is greater than 30 and the nature of the study is cross-sectional [69,70].
The first set of regressions (see Table 4) indicates that 15.8% of CPV toward the mother is explained by mother-to-child violence (β = 0.392, p < 0.001, R 2 = 0.158). However, the analysis also indicates that the variance in CPV toward the mother significantly increases by exposure to violence by the mother toward the father (β = 0.120, p < 0.001, R 2 = 0.168) and by psychosocial stressors related to academics/work (β = 0.065, p = 0.001), family coexistence (β = 0.093, p < 0.001) and mental health (β = 0.057, p = 0.005). The final retained model F (7, 2158) = 70.91, p < 0.001, R 2 = 0.187 shows that the additive effect of different types of family violence and certain stressors experienced during confinement predicts 18.7% of CPV toward mothers.  Regarding fathers, 17.5% of CPV toward the father was explained by father-to-child violence (β = 0.383, p < 0.001, R 2 = 0.175). Similar to the previous results, the variance in CPV toward the father significantly increases by the exposure to violence by the mother toward the father (β = 0.139, p < 0.001, R 2 = 0.188) and by psychosocial stressors related to academics/work (B = 0.053, p = 0.007), family coexistence (β = 0.077, p < 0.001) and mental health (β = 0.086, p = 0.025). The final retained model F (7, 2158) = 77.44, p < 0.001, R 2 = 0.201 indicates, also in this case, that the additive effect of different types of family violence and certain psychosocial stressors experienced during confinement predicts 20.1% of CPV toward the father (see Table 4).

Multiple Linear Regressions of CPV toward Mothers and Fathers that Include the Effect of the Accumulation of Psychosocial Stressors
The last analysis addresses the effect of the accumulation of psychosocial stressors on the prediction of CPV toward mothers and fathers during confinement (see Table 5). While the new regressions did not improve the proportion of variance explained with respect to the previous regressions, the accumulation of various psychosocial stressors had a more important effect on the prediction of CPV toward both mothers (β = 0.107, p < 0.001) and fathers (β = 0.095, p < 0.001) with respect to the effect provided by each stressor separately.

Discussion
This study analyzes the impact of strict confinement due to COVID-19 and the associated psychosocial stressors on the presence of violence in the family environment, especially CPV. To date, while several studies have warned of the risk of this type of violence during confinement, there are few studies that provide data on family violence and even less on CPV.
The results regarding CPV frequency indicate that more than half of the young people reported having exercised at least one violent behavior toward their parents within a month (65.2% toward the mother and 59.4% toward the father). By type of violence, physical violence ranged between 1.7% and 3.3%. This is a very high percentage when taking into account the short time frame used in this study to evaluate the frequency of violent behaviors (number of violent behaviors in a month) and when comparing the obtained values with the data usually reported with samples of adolescents (5-22%) with a wider time margin, e.g., referring to the last 12 months [17]. Ibabe et al. [19], using the criterion of zero tolerance, found that 24% of young people exhibited physical violence. In this study, psychological violence among the study population ranged from 40.1% to 61.3%. These percentages are lower than those found by Ibabe et al. [19], i.e., 94%. Along with the differences between the studies regarding the time frame, it is necessary to add that it is common to find considerable disparity in the prevalence of this type of violence in CPV studies due to the different behaviors and criteria included in each study. Economic violence ranged from 12% to 16.6%, and behaviors of control and dominance over parents oscillated between 36.5% and 43%. By gender, more females than males exercised psychological violence toward fathers and mothers and controlling/dominating behaviors toward mothers. In contrast, more males than females exercised physical and economic violence against their parents. In samples of Spanish adolescents and young adults it is common to find more psychological violence by females than by males, with no gender differences in physical violence [19,71,72]. However, other studies have found higher rates of violence toward parents in young males [18], leading to the hypothesis that patterns of CPV change with the age of the young person. Likewise, more research is needed to delve into the role of gender in this type of violence.
This study also provides the frequency of other types of family violence. With respect to the frequency of violent behaviors of parents toward their young adult children, we observed that 16% suffered some type of violence from the mother and 14.9% from the father. While physical violence ranged from 2.8% to 3.4%, psychological violence ranged from 12.4% to 17.1%. When comparing these data with those obtained from samples of Spanish adolescents, we observed similar rates of psychological victimization, 15.9% (both parents), and lower rates of physical victimization of the mother, 27.1%, and of the father 27.5% [24]. In this regard, and in the case of a sample of young people, it is expected that victimization rates will decrease as the age of the young people increases. Regarding gender differences, the only significant finding is that females report psychological violence by mothers more frequently than do males. In contrast, Izaguirre and Calvete [24] found higher rates of physical victimization of parents (fathers and mothers) by adolescent males than by adolescent females.
Regarding the frequency of exposure to violent behaviors among parents, physical violence ranged from 0.9% to 1%, psychological violence ranged from 9.9% to 13.4% and verbal violence ranged from 0.4% to 1.2%. Rates of exposure to physical violence were lower than those found in other studies [19,24,58]; among those studies, in turn, there were significant differences, with rates of 14%, 9.7% and 2.1%, respectively. The levels of exposure to violence by the mother toward the father (11.6%) and exposure to violence by the father toward the mother (11.9%) were similar. These data are consistent with other studies that report similar frequencies of exposure to both types of violence (e.g., [19,58]). Of interest, we found that daughters report more frequent exposure to psychological violence from the father toward the mother. Other studies also indicate that more females than males report having witnessed violence toward their mothers [24].
As expected, almost half of the young people revealed a high concern related to academics/work, specifically regarding the results of the academic year and the continuity of employment, finding that coincide with the results of other research [29,30]. Almost 22% experienced difficulties related to family coexistence, especially the lack of space and spending time with the family; these data are consistent with previous literature [31]. Economic concerns were also present among young people, as indicated by other studies on the subject [27][28][29]. Almost half of the sample showed a high concern for the virus, especially for the infection of a family member, in line with recent studies [27][28][29]32,33]. The results also show a high percentage of stressors related to mental health (more than 90%). Specifically, the majority of the young people recognized that confinement had psychologically affected them and family members, and many also presented sleep disturbances. Previous studies have also found that psychological impacts [27,28,34,35] and sleep disturbances [37,38] are highly prevalent among adult populations during the pandemic. In addition, in line with previous studies, it was observed that, in general, females presented a higher proportion of stressors related to mental health than did males [27,28,34,36,37]. Finally, almost all of the young people presented an accumulation of various stressors (more than 90%), with females having higher percentages.
Furthermore, as expected, positive and significant relationships were found between CPV and parent-to-child violence [8,20,24] and between CPV and exposure to violence between parents [8,19]. The most intense correlations were between child-to-mother violence and mother-to-child violence (0.37) and between child-to-father violence and fatherto-child violence (0.37), suggesting a certain reciprocity in these types of violence. Similar data are reported by Izaguirre and Calvete [24], with correlations ranging from 0.41 to 0.49 for victimization by the mother and child-to-mother aggression and a correlation of 0.41 for victimization by the father and child-to-father aggression. In particular, the relationship between mother-to-child violence and exposure to mother-to-father violence (0.43) and, on the other hand, the relationship between father-to-child violence and exposure to father-to-mother violence (0.48) stand out. These data indicate that violent behavior, if it occurs, extends to both the children and the partner. Likewise, CPV toward both the father and the mother was significantly and positively related to psychosocial stressors related to academics/work, family coexistence and mental health, as well as to the accumulation of various psychosocial stressors. In fact, the intensity of the correlations was greater between CPV toward the mother and father and accumulated psychosocial stressors than between CPV and the stressors considered independently.
The data from this study confirm that independently, parent-to-child violence and exposure to violence between parents predict CPV toward both the father and the mother. These results coincide with those obtained in various studies [8,[18][19][20]24]. More specifically, CPV toward the mother is predicted mainly by mother-to-child violence (15.8%) and, although to a lesser extent, by exposure to violence by the mother toward the father, jointly explaining 17% of CPV toward the maternal figure. In the case of CPV toward the father, this is predicted mainly by father-to-child violence (17.5%) and, although to a lesser extent, by exposure to violence by the mother to the father, jointly explaining almost 19% of CPV toward the paternal figure. These data are consistent with the perspective of the intergenerational transmission of violence (e.g., [73,74]). Through observational learning and imitation of adult models [75], children who come from violent homes are more likely to become violent adults. Regarding the role of direct victimization, previous longitudinal studies also indicate that direct victimization by the mother predicts an increase in childto-mother violence and that direct victimization by the father predicts an increase in childto-father violence [24]. These data suggest reciprocity in CPV and that the aggressive behaviors of young people toward their parents could represent responses to previous aggressions by the parents. In addition, it has been confirmed that the predictor with the greatest predictive capacity is parent-to-child violence, an aspect noted in various studies [8,18,24]. Regarding exposure to violence between parents, it is noteworthy that the variable that has predictive value for CPV both toward the mother and toward the father is violence by the mother toward the father. These data suggest that the violent behavior of the mother has greater weight in explaining CPV.
Regarding the predictive value of the psychosocial stressors analyzed, the results showed that stressors related to academics/work, family coexistence and mental health significantly predicted CPV toward both the mother and toward the father. In the case of CPV toward the mother, the stressors related to family coexistence are those that have a greater predictive capacity, while in the case of CPV toward the father, the stressors related to mental health are those that have a greater predictive value. Likewise, the significant predictive effect of the accumulation of various psychosocial stressors in CPV is greater than the predictive effect of each stressor independently. These results are consistent with those found in other studies on youth violence [49,50,53].
Finally, it is confirmed that both family violence type and the psychosocial stressors analyzed are predictors that, together, contribute to explaining CPV to a greater extent than each stressor separately, consistent with what has been reported by various studies of violent behavior in adolescence [8,54,57].
In this study, four important limitations can be identified that should be taken into account when interpreting the results. These refer to the design of the study, the use of a single informant in the study, the type of sample used and the way to evaluate psychosocial stressors. The cross-sectional design of the study does not allow identification of causal relationships between the factors analyzed and the violent behavior of young people. To address this type of objective and to determine the temporal sequence of the relationships, it would be necessary to have longitudinal data. In this study, the informants were the young people. It would also be necessary to incorporate reports from parents to obtain a more complete view of the problem. The results are based on a sample of Spanish youth, thus limiting the generalization to other age ranges or countries. In addition, the sample shows a slight over-representation of university students and under-representation of young people with primary and secondary studies. While little is known about cultural differences in CPV patterns, cultural factors regarding whether authoritarian/permissive parenting styles are favored, the role and rights of women and the age at which children become independent from home possibly influence the frequency and pattern of CPV [17]. Finally, another limitation of the study arises from the fact that the psychosocial stressors questionnaire included few items to evaluate each type of stressor. While this approach was taken to avoid excessively lengthening the time spent completing questionnaires and to avoid fatigue, it would be advisable to examine these aspects in future studies with previously validated evaluation tools.
Despite the limitations mentioned, these findings, in addition to providing data on the frequency of different types of family violence and psychosocial stressors related to strict confinement due to COVID-19, corroborate the existing research that demonstrates the solid relationships between different types of family violence and CPV. Furthermore, novel data are provided regarding the relationship between CPV and the existence of external factors such as the influence of psychosocial stressors, both the specific role of each and the cumulative effects of various stressors. This aspect, while it has been addressed in other forms of violence, has not been specifically addressed in the previous literature on CPV. In the same way, not only is the independent contribution of each of these factors (parent-to-child violence, exposure to violence between parents and psychosocial stressors) to CPV analyzed, but the joint contribution of all these factors is assessed in the context of a unique and exceptional situation, i.e., strict confinement due to COVID-19.
Likewise, the findings suggest various practical implications. Taking into account the high rates of family violence detected during the time of strict confinement due to COVID-19, it is necessary for institutions to provide additional resources to families in this type of situation or similar situations in which the risk of this type of violence may increase. Professionals who serve families should refer families to community resources to expand their support networks. Health prevention services should be especially alert in these situations to detect families at risk of this type of violence. Furthermore, care services that serve families can benefit from knowledge about the psychosocial stressors that young people most frequently experience in these types of situations and how these stressors contribute to CPV. Finally, prevention and intervention programs for CPV must take into account the important role that other forms of family violence may have as well as the influence of psychosocial stressors.
This study provides preliminary information on the factors that affect CPV during strict confinement due to COVID-19 in the short-term. Future research is needed to verify the long-term effects of the pandemic on the different forms of family violence.  Informed Consent Statement: Informed consent was obtained from all subjects involved in the study.

Data Availability Statement:
The data of this research are available upon request to the corresponding author.

Conflicts of Interest:
The authors declare no conflicts of interest.