This study examines the nature of all domestic violence incidents involving parents and their minor children to which police in Philadelphia responded during the 2013 calendar year. We use a retrospective design to explore the nature and outcome of parent-child incidents to which police are summoned. Incidents that officers determined met the state statute definition of child abuse are not included. Of 54,456 domestic violence incidents in the city of Philadelphia in 2013, 2,361 involved a verbal incident or physical altercation between a minor child and at least one parent. Most reports (83.3%) identified the child as the offender and were for verbal incidents (89.6%), suggesting police were called to resolve conflict in the home. When a child was the offender, boys were the most common offenders and mothers the most common victims. When a parent was the offender, mothers were the most common offenders and daughters the most common victims. Parent-offender incidents were far fewer (16.7%) but more likely than child-offender incidents to involve physical violence (AOR = 6.19) and to result in arrest (AOR = 3.67). Parent-child incidents that are not child abuse constitute about 20% of all domestic violence incidents to which police are summoned. Parent-child incidents are an under-researched and perhaps under-served issue. We know of few resources beyond law enforcement for on-the-scene crisis intervention and, as such, officers appear to serve as mediators in these mostly verbal disagreements. The appropriateness and cost of such intervention merits investigation and discussion.
Conflict is an expected part of the parent-child relationship, especially during adolescence, as children wrestle with issues of identity development and strive for independence. Whereas the majority of conflicts between parents and children can be resolved in the home between the involved parties, some portion of the incidents spill over and police are summoned.
This study examines all domestic violence incidents involving parents and their minor children to which police in Philadelphia responded during a single calendar year—2013. Although the term “domestic violence” commonly is used to describe violence between intimate partners, law enforcement agencies generally use a wider definition that includes any incident that is “characterized by intimacy, familial ties, or a shared household” (Dempsey 2006).
The present study investigates service use, that is, the use of law enforcement services in parent-child conflict. It addresses both parent-to-child and child-to-parent incidents. It is not nor does it purport to be an investigation of child abuse. In fact, the data made available to the researchers do not include cases that the responding officers, using the definition determined by state statute, deemed to merit investigation as potential child abuse.Footnote 1 The present study also is not an investigation into the reasons that community members choose to call or not call law enforcement when faced with a troubling circumstance involving a parent and child. Our study addresses three key questions:
What portion of domestic violence calls (that are not considered reportable child maltreatment) involving parents and their children involve minors as victims or offenders?
What were the characteristics of such incidents?
What was the law enforcement response?
The three questions contribute to the knowledge base on both family violence and how police respond to conflict and violence between parents and children. Given that law enforcement resources often are stretched thin and that officers consider certain crimes to bring more prestige (assaults, robberies, etc.), family-related calls can be viewed as a lower priority (Stanko 1989: 51). Furthermore, officers do not have the resources necessary to respond effectively to cases of family conflict, especially in those incidents that do not meet the high bar of what constitutes reportable child abuse in Pennsylvania. This further complicates the handling of these incidents.
The information presented in this paper can help policy makers and legislators better understand what is being asked of police officers who respond to domestic violence incidents. In particular, this research considers incidents of parent-child conflict and violence that come to the attention of the police, what this might tell us about how police are utilized by the community, and offers possible solutions to the current lack of resources available for intervention.
Prior Research on Incidents between Parents and Children
Systematic research on violence between parents and children began nearly 40 years ago and has largely focused on the areas of corporal punishment and child abuse (i.e., parents’ behavior towards a child). A large portion of this literature has examined the role corporal punishment and child abuse play in determining short and long term health and psychological outcomes of children (Gershoff 2002). However, a few studies have examined the police response to incidents of child abuse and corporal punishment (Shireman et al. 1981; Willis and Wells 1988). Shireman et al. (1981) suggested in their study that the criteria police use to identify child abuse is highly variable, both by state context and between individual police officers. Willis and Wells (1988) found in their study of police reporting of child abuse in North Carolina that as the perceived severity of incidents between children and their parents varies significantly, police officers often make a choice as to whether or not an incident is serious enough to be escalated further to child welfare agencies. Willis and Wells (1988) found that as many of the incidents police respond to between parents and children take on a more “social work” as opposed to “crime fighting” nature, the greatest likelihood of reporting child abuse occurred only when the police could determine “criminal intent” or serious neglect, with less “serious” incidents even being deemed appropriate parental disciplinary action as opposed to abuse.
There has also been a fair amount of research on the phenomena of child-to-parent violence, adolescent-to-parent violence, and “parent abuse” (Cornell and Gelles 1982; Evans and Warren-Sohlberg 1988; Erez and Tontodonato 1989; Hutchison et al. 1994), that is, violent incidents in which 10- to 17-year old children are perpetrators. The first study to assess the national prevalence of the phenomenon surveyed 2,143 families across the United States and estimated that 2.5 million adolescents had struck a parent at least once in the previous year, and that about 900,000 had engaged in acts classified as “severe violence” (e.g., kicking, punching, or using a weapon) against at least one parent (Cornell and Gelles 1982). Children who had been victims of parent-perpetrated violence themselves or who had witnessed intimate partner violence between their parents were the most common perpetrators. For nearly every age group, males were more likely than females to be violent towards their parents, and mothers were more likely than fathers to be victims of the violence (Straus et al. 1980).
Subsequent studies examined the role of police utilization and intervention in child-to-parent violence (e.g., Evans and Warren-Sohlberg 1988; Erez and Tontodonato 1989). In a study of one region of Washington State, researchers examined 65 police reports of child-perpetrated violence against a parent (Evans and Warren-Sohlberg 1988). In a substantial majority of the cases, the child was physically violent toward the parent, and incidents were related to household responsibilities (chores, room cleaning, etc.), money, and alcohol consumption. The authors proposed that the high proportion of son-to-mother abuse was due to mothers (vs. fathers) serving as the primary disciplinarian in the child’s life and that, overall, mothers have more interaction with their children than do fathers, increasing their time at risk. Half of the incidents involved some type of police action, such as arrest, placing the child in temporary custody of child protection services, or temporarily moving the child to another residence (e.g., the home of another family member). One-third of the incidents involved an arrest, which is much higher than expected given the low level of felonious assault reported. The authors concluded that there is lack of standardization in police response to incidents between children and parents, and that the decision to make an arrest relies heavily on the discretion of individual officers.
Researchers made similar observations in another study: among the incidents in which the child was the offender, most involved a male child offender and a mother victim (Erez and Tontodonato 1989). When physical assault cases alone were analyzed, however, the most common pattern was for fathers to victimize their daughters. The majority of child-offender-parent-victim cases reported to police were verbal in nature, whereas most parent-offender-child-victim cases were physical in nature. Police intervention (defined as making an arrest) occurred infrequently. An arrest was made in only 3% of child-perpetrated verbal incidents, and arrest rates remained low in violent incidents; only 17% of violent incidents involving bodily weapons (hands, fists, and feet) and 27% of violent incidents involving a criminal weapon involved an arrest. The decision to make an arrest was associated with factors including the severity of the incident (verbal vs. physical, victim injury resulting in hospitalization) as well as the race of the offender (Black offenders were more commonly arrested than White offenders, regardless of the severity of the offense). Additionally, police typically defined the incidents as “domestic disturbances” rather than crimes, even when the cases involved physical force or weapons.
More recently, Walsh and Krienert (2007) examined 11 years of fatal and non-fatal cases of child-to-parent violence recorded in the National Incident-Based Reporting System.Footnote 2 In nearly 18,000 cases involving offenders ranging in age from 7 to 21, most child offenders were males aged 14–17 years old. White families were more likely to report incidents of child-to-parent violence than Black families. Whereas it was somewhat unclear whether Whites were overrepresented in the data because they were more likely to report incidents to police or if Whites simply have a higher prevalence of child-to-parent violence, researchers believed overrepresentation was a more likely explanation. The authors noted that Blacks and other minority group members in poor neighborhoods were less likely to report crime to police due to mistrust of the criminal justice system and fear that reporting family violence may result in the removal of children by agencies such as Child Protective Services. Most incidents perpetrated by children involved weapons, but few resulted in serious injury or any injury at all. The study did not report what actions police took.
Other factors involved in parent-child violence include divorce, child living arrangements, and custody arrangements (Ellwood and Jencks 2004; Nock and Kazdin 2002; Ibabe and Bentler 2016; Margolin and Baucom 2014). Disruptions to home life, such as divorce, spousal domestic violence, and early parent-to-child aggression, each are associated with an increased likelihood of parent-directed child aggression.
Some research on incidents between parents and children has focused on a child’s exposure to aggression in the family of origin (e.g., spousal violence between parents) and the likelihood of perpetrating parent-directed aggression during adolescence. In a sample of 93 non-clinical and non-court referred adolescent volunteers and their parents, 75% of the incidents analyzed were cases of verbal aggression. Verbal and physical child-to-parent aggression was positively associated with previous parent-to-child aggression and, thus, posited to be a learned, reactive behavior (Margolin and Baucom 2014). Parent-to-child aggression in the child’s family of origin was associated with a higher likelihood of the adolescent child perpetrating both verbal and physical aggression towards the parent.
The phenomenon is not limited to the United States. In a sample of 485 children aged 12–18 in Basque country, Spain, low familial cohesion and families in which “power-assertive” parenting strategies, such as penalty and intense supervision, were associated with the occurrence of adolescent-to-parent aggression (Ibabe et al. 2013). Children who had experienced parent-to-child violence and inter-parental violence were found to be at higher risk of perpetrating child-to-parent violence (confirming the authors’ hypothesis of bi-directionality in family violence). Whereas there was no difference in the likelihood of children physically victimizing their mothers vs. their fathers, mothers were more commonly victims of emotional and verbal abuse. Boys were more likely than girls to be physically abusive, and were also more likely to physically abuse their mothers if they had witnessed their mothers being physically victimized by their fathers.
The current study expands upon a small body of largely outdated literature to examine conflict and violence between parents and their children to which police were summoned. In addition to examining the nature of such conflict, the study also examines law enforcement response to such incidents, using the population of police-reported cases of parent-child incidents during one year.
The Present Study
In comparison to the extensive literature that exists on forms of domestic violence such as intimate partner violence and child maltreatment, there is a paucity of scholarship that examines parent-child violence in the home that is not considered child abuse. There is even less research on any form of violence and conflict in which the parent is the victim. Furthermore, what little research there is on child-perpetrated violence against parents largely ignores: 1) verbal incidents to which police are called to respond where the offender is a minor child, and 2) incidents in which minor children call the police on their parents for reasons other than to report child maltreatment (physical abuse, neglect, or sexual abuse).
We employed a retrospective study design using law enforcement data drawn from the city of Philadelphia Police Department forms during calendar year 2013; Philadelphia was the fifth largest city in the U.S. at the time. The data used herein were obtained by the second author as part of a larger study following the execution of a confidentiality and data-use agreement between the University of Pennsylvania and the Philadelphia Police Department. Institutional IRB approval was obtained prior to conducting any data cleaning, coding, or analysis.
Data were drawn from the detailed form that the Philadelphia Police Department mandates officers to complete when they respond to a domestic violence call for assistance. The forms include an open-ended field for the officer’s description of the incident, checkboxes regarding basic demographic characteristics of the offender and victim, outlines of the human body where officers noted injuries, and checkboxes about specific actions by victim, offender, and officer. The forms are potentially useful for investigating officers (i.e., detectives who follow through on cases when indicated), the District Attorney’s office (which determines what charges if any are filed), the officer himself or herself (who might be called to testify), and for researchers interested in criminal justice intervention in the community. We used the open-ended narrative and checkbox sections of the forms.
As noted previously, law enforcement defines “domestic violence” broadly to include any incident that occurs within the home, and there were 54,456 such cases in 2013, of which 10,953 involved parents and their children. Cases that were not relevant to the present investigation – that is, incidents that did not involve parents and their minor children – were excluded from analysis. Incidents included in the analysis were those involving minor children and their parents. We defined “parents” as foster, step-, adoptive, and birth parents. We did not code for what type of parent because the information was not consistently reported in the records. To avoid making unwarranted assumptions regarding family structure and roles, grandparents and other family members were excluded from analysis unless the police report explicitly stated that they were the child’s legal guardian.
We restricted analysis to children under the age of 18 because of the legal ties between parents and their children, including the fact that children of minor age typically, albeit not always, reside in the same household as their parent or other guardian. Moreover, such adults have a legal requirement to tend to their child.
As noted previously, the cases provided for analysis were defined by responding officers as incidents of domestic violence rather than as child abuse. Child abuse cases are not recorded on the Department’s domestic violence forms; thus, the Philadelphia Police Department did not provide information on any incident that they considered to constitute child abuse under Pennsylvania statute.
The police form refers to the involved parties as “offenders” and “complainants” rather than offenders and victims or suspects and complainants. For ease of exposition, we will use the terms “offenders” and “victims.”
Forms were converted from hard copy to an electronic format via an emerging digitization process. The digitization agency reported an accuracy rate of 99%.
The officer’s description of the incident (the narrative) was reviewed and coded by the lead author and two research assistants under the supervision of the second author. To ensure the quality of the coding, coders were trained on a subset of the data until their work was deemed to be of sufficient quality (95% or greater agreement with the coding of the senior investigator), after which they were allowed to work on the full data set. The most experienced coder checked a random 10% sample of each assistant’s coding throughout the entire coding period, and provided feedback to the other coders who made the appropriate changes to their work to ensure quality consistency. New variables based on information abstracted from the narrative section of the form were created for analysis.
We used simple descriptive statistics (e.g., frequencies, percentages) as well as cross-tabulations and multivariate logistic regressions to assess: a) the proportion of domestic violence incidents to which police are called to intervene that involve a minor child and a parent; b) differences in the nature of the incident when a minor child is the offender versus when a parent is the offender; and c) police actions at the scene.
Demographic and situational characteristics associated with parent-child incidents considered in the multivariate regressions included race, Hispanic ethnicity, gender, and age of the victim and the offender and whether the offender was on probation, had a history of substance abuse, had a history of domestic violence, or had ever had a domestic violence report filed against him or her. As is common in administrative data, information was occasionally missing (e.g., the gender of the minor child offender was not recorded in 2.3% of the incidents). Therefore, to avoid contaminating comparisons (i.e., when missing or otherwise not ascertained data were not grouped with “No” or “Yes”) and to avoid potential bias by including only cases with complete data, a “not ascertained” category was created for variables and included in the regressions.
We calculated diagnostic statistics (correlations and Variance Inflation Factors [VIFs]) prior to conducting multivariate analyses. All were in an acceptable range. The primary variable of interest in the analysis was whether the parent or the minor child was the offender. To adjust for multiple tests in the multivariate logistic regressions, we applied a Bonferroni correction resulting in a p-level of 0.0005.
Of the 10,953 domestic violence incidents involving parents and their children (of all ages), 2,979 (27.2%) involved parents and their minor children. A substantial majority (81.9%) of such incidents involved a child offender and parent victim. These cases will be referred to as COPV incidents. The remaining incidents involved child victims and parent offenders. These cases will be referred to as POCV incidents. COPV cases, the most common type of incident, are described first followed by POCV cases. Finally, multivariate results comparing the two are reported.
Figures 1 and 2 show the similarities in the racial and ethnic distribution of parent and child offenders and victims in COPV and POCV incidents. In COPV incidents, the racial and ethnic composition of the child offenders was as follows: 70.6% were non-Hispanic Black, 13.7% were Hispanic of any race, 12.0% were non-Hispanic White, and 3.7% were identified as belonging to an “other” category of race/ethnicity. The racial composition of parent victims closely mirrored that of the child offenders, as was expected.
In POCV incidents, the racial and ethnic makeup of the parent offenders was similar to that observed in COPV incidents: 68.8% of the parent offenders were identified as non-Hispanic Black, 14.0% were non-Hispanic White, 11.7% were Hispanic of any race, and 5.6% were identified as belonging to an “other” race/ethnicity. As in the COPV incidents, the racial/ethnic makeup of the child victims closely mirrored that of the parent offenders.
In 82.0% of COPV incidents, the mother was the victim, whereas only 16.0% of COPV cases involved the father as the victim. COPV incidents had more male than female child offenders (56.2% versus 43.8%). The gender of the child offender and parent victim was not ascertained in about 2% of incidents.
Table 1 reports the distribution of the nature of all COPV incidents and the context of verbal incidents in COPV incidents. Most (89.6%) of the COPV incidents attended to by police were verbal in nature. COPV verbal incidents largely centered on behavior and discipline (38.2%) and home life concerns such as living arrangements (31.0%). A total of 10.4% of the COPV incidents involved physical violence (6.6% of total incidents involved a bodily weapon [hands, fists, or feet] and 3.8% involved an external weapon).
The time of day and day of the week of such incidents is relevant to the allocation of law enforcement resources. As shown in Fig. 2a, police most often responded to calls of COPV violence on Mondays and Tuesdays between the hours of 8:00 a.m. and 3:00 p.m. Police responded to the third largest number of calls on Sundays between the hours of 8:00 p.m. and 01:00 a.m. Police responded to the fewest number of calls on Fridays and Saturdays, with the fewest overall number of calls occurring between the hours of 1:00 a.m. and 8:00 a.m. and 3:00 p.m. and 5:00 p.m. on Saturdays.
As shown in the second row of data in Table 2, slightly fewer than half of the child offenders were female, and more than two thirds were non-Hispanic Black. A total of 13.8% of parent victims indicated that there had been prior domestic violence perpetrated by the offender and about one in ten said that prior domestic violence reports had been made to police (9.9%). Substance use was reportedly involved in 3% of COPV cases. Few child offenders were on probation or under a protection from abuse order (1.5% and 0.3%, respectively). As shown in the second row of data in Table 3, the majority of parent complainants in COPV incidents were female (81.7%). About one-third of the parent victims reported feeling distraught, and about 15% were tearful. Far fewer parent victims felt frightened (4.6%) or were observed with disarrayed clothing (1.6%). Injuries were uncommon: about 5% of parent victims were injured – 3.6% had visible injuries and 1.3%, had injuries that were not visible (e.g., under clothing).
The second row of data in Table 4 reports scene and incident characteristics for COPV incidents. More than one in ten incidents involved physical violence. Blood was present at the scene in fewer than 2% of the incidents, and officers observed disarrayed furniture in 2.6% of the incidents. Nearly one in twenty incidents involved the involuntary commitment of a minor child offender, and mental illness and substance use were involved in fewer than one in thirty incidents (3.0% and 2.9%, respectively). Less than 1% of the cases involved the violation of a protection- from-abuse-order or an incident about child custody (0.3% and 0.5%, respectively).
Aggressive behaviors were uncommon in COPV incidents. As shown in Table 5, in fewer than one in ten incidents, the child offender pushed or shoved the victim. The child offender grabbed, punched, or threatened the victim in about 5% of incidents (5.6%, 4.6%, and 4.5%, respectively). The offender pulled the victim’s hair, kicked, or slapped the victim in approximately 2% of the incidents (2.1%, 2.4%, and 1.8% respectively). Fewer than 1% of child offenders strangled their parent victims. A total of 3.8% of child offenders threw objects, and 3.4% damaged property.
Table 6 reports the variety of actions taken by responding officers at the scene of COPV incidents (row two). Officers infrequently made arrests (3.5% of COPV incidents). It was also uncommon for officers to gather evidence or take statements, either from an adult or child witness (1.9%, 1.2%, and 0.9%, respectively). Police officers most often took action in the form of checking the protection-from-abuse (PFA) registry in 7.0% of cases. Although it is illegal for minor children to carry firearms, police reported checking the child offender’s carry permit in 7.1% of cases. The parent victim’s carry permit was checked in 2.0% of cases.
Table 7 reports types of assistance that police provided at the scene of the incident (line two). In almost one-third of COPV incidents, police provided information to the victim on obtaining a protection-from-abuse (PFA) order. In nearly 1 in 5 cases, officers provided victims with a community resource card. In 13.7% of cases, officers provided victims with the victim-assistance emergency phone number, and in 7.4% of cases, officers provided victims with contact information for a domestic violence advocate. As COPV cases were largely non-violent, only 2% of COPV cases required police transport to a hospital.
In POCV incidents, the majority (67.8%) of child victims were daughters. In contrast to COPV incidents, a majority of the cases (52.3%) were perpetrated by mothers. Given the high proportion of Black families reported to reside in single-parent households in Philadelphia in 2013 (nearly 30%), we suspect the predominance of mother perpetrators of POCV incidents may be linked to the high prevalence of female-led households (U.S. Census Bureau 2013a).
Table 1 reports the distribution of the nature of all POCV incidents and the context of verbal incidents in POCV incidents. In contrast to COPV incidents, of the POCV incidents, 44.2% involved physical violence (29.7% of total incidents involved bodily weapon, 14.5% involved an external weapon). A total of 55.8% of all POCV incidents were verbal incidents, of which most (47.1%) were about issues relating to “home life.”
Figure 2b shows the distribution of the day and time of police response to POCV incidents. In contrast to COPV incidents, police most frequently responded to POCV incidents on Wednesdays between the hours of 8:00 a.m. and 3:00 p.m. and 5:00 p.m. and 8:00 p.m. The day of the week with the second highest frequency of police responses to POCV incidents was Friday between 5:00 p.m. and 8:00 p.m., followed by Monday between 8:00 p.m. and midnight. Police least frequently responded to calls on Saturdays between the hours of 1:00 a.m. and 8:00 a.m.
As shown in the first row of data in Table 2, over half of the parent offenders were female and more than two thirds were non-Hispanic Black.Footnote 3 Nearly one in five indicated that there had been prior domestic violence perpetrated by the offender and about one in ten said that prior domestic violence reports had been made to police. Substance use was reported for about one tenth of the parent offenders. Far fewer were on probation or under a protection-from- abuse-order (1.5% and 0.5%, respectively). As shown in the first row of data in Table 3, the majority of complainants in COPV incidents were female (81.7%). About one-fifth of child victims were distraught, and more than one-third were tearful/crying at the scene. Sixteen and one-half percent (16.5%) of child victims reported feeling frightened, and about 4% were reported to have disarrayed clothing. Injuries were relatively uncommon: 15.7% of child victims had visible injuries, while 6.9% had injuries that were not visible to the officer. Row one of Table 4 shows scene and incident characteristics for POCV incidents. In a small percentage of cases, blood was present at the scene and furniture was observed to be in disarray (4.8% and 3.1%, respectively). Physical violence was involved in nearly half of all POCV incidents (48.2%). Almost one in ten incidents involved a drugs or alcohol (7.1%). Mental illness was infrequently involved in POCV incidents (1.0%), and involuntary commitments of parent offenders were rare (0.5%). Few cases involved an incident over child custody (1.8%), and a protection-from-abuse-order was violated in only 0.5% of incidents.
Table 5 reports offender behavior during the incident. In more than one in five cases, parent offenders grabbed, punched, or pushed/shoved their child (23.6%, 22.1%, 21.6%, respectively). A total of 15.2% of parent offenders slapped their child. Less commonly, parents threatened, pulled hair, strangled, or kicked the victim (8.1%, 6.6%, 6.1%, and 3.3%, respectively). Few parent offenders engaged in aggressive behaviors such as throwing objects (4.1%) and damaging property (3.8%).
In row one of Table 6, officer actions for POCV incidents are reported. About 13% of POCV involved an arrest. Officers checked the state protection-from-abuse-registry and the offender’s carry permit in about one of every ten cases. Although child victims in POCV cases are minor children, officers reported checking the victim’s carry permit in 2% of cases. Officers reported taking statements infrequently; statements taken from anyone were reported in 8.1% of cases, and statements taken from children were reported in 4.3% of cases. Officers reported collecting evidence in only 2.3% of cases. Table 7 reports the types of assistance officers provided at the scene. In nearly one-third of all POCV cases, child victims were given information on obtaining a PFA order. In less than one in twenty cases, victims were provided a community resource card. Less frequently, victims were given victim assistance contact information or provided information on obtaining a domestic violence advocate (14.0% and 7.8%, respectively). In nearly one in ten POCV cases, victims required police transportation to a hospital (7.6%).
Many of the differences between COPV and POCV incidents observed in bivariate analysis remained statistically significant in multivariate analysis. Confidence intervals are reported in the tables; in the text, we provide the adjusted odds ratios that were statistically significant at p < 0.0005.
Table 2 shows that parent offenders were more likely than child offenders to be female (AOR = 1.64,) and to have a history of substance use and domestic violence (AORs = 3.46 and 2.33, respectively). Table 3 shows that, although female victims predominated in COPV and POCV incidents, the victims had a higher odds of being male than female in POCV incidents (AOR = 2.36). In the POCV incidents, the child victim was more likely to be tearful/crying or frightened (Table 3; AORs = 3.90 and 3.98, respectively). In addition, child victims were more likely to have sustained visible injuries as well as injuries that were not visible (e.g., hidden by clothing) (AORs = 4.71 and 5.47, respectively). Table 4 shows that physical violence was substantially more likely to have occurred in POCV incidents than in COPV incidents (AOR = 6.19). As shown in Table 5, the offender generally was more aggressive in POCV than in COPV incidents. Specifically, in POCV incidents, the offender was more likely to have pushed and shoved, grabbed, pulled hair, slapped, punched, or strangled (AORs = 2.97, 4.94, 3.26, 9.20, 5.24, and 8.92, respectively). For certain behaviors, such as threatening, kicking, throwing objects, and damaging property, there was no difference in the likelihood of occurrence when the parent was the offender versus when the child was the offender.
The responding officer was more likely to make an arrest and take statements (both in general and from children) in POCV incidents, as shown in Table 6 (AORs = 3.67, 7.07, 4.11, respectively). In incidents with a child (vs. parent) victim, police were more likely to have provided hospital transportation (AOR = 4.17). There was no difference in the likelihood that police would provide information about domestic violence advocacy services or protection-from- abuse-orders or to provide an information card about community resources or the phone number of the victim-assistance program.
Our examination of parent-child incidents to which police were summoned and responded sheds light on the little-studied topic of minor-child-to-parent conflict and violence and adds to our understanding of parent-to-child violence that is not considered child maltreatment by the responding officer. A substantive finding of this research is that one in twenty domestic violence incidents to which officers are summoned involve minor children and their parents. We infer from this finding that some members of the general public view law enforcement as a resource when a situation involving parents and their children becomes unmanageable.
Our study shows very little child-to-parent violence to which police officers respond results in further criminal justice system intervention. When police are summoned and respond to minor-child-to-parent violence, most often the incidents involve verbal conflict and no physical altercation. Most of the victims of child-to-parent domestic conflict and violence are mothers and the offenders are their sons. Whereas one might believe that parents and children would be more likely to engage in incidents when they are more exposed to one another, we found that police are called to respond most frequently during what would be considered school hours—weekdays between 8:00 am and 3:00 pm. Perhaps the child not attending school is, in and of itself, a source of conflict that escalates to the point that police are summoned. Overall, the picture that emerges is of the police responding to verbal and some physical conflict that they do not consider to constitute criminal activity. The police appear to be viewed by those making the call – parents, children, other family members, neighbors, or others – more as first-responder social workers than as agents of the criminal justice system. Arrest occurs rarely in minor-child-to-parent incidents, and we would venture that arrest likely was not a desired outcome by most of those calling for assistance.
Parent-to-child incidents are mostly physical violence perpetrated by mothers against their daughters. The findings from our police data are inconsistent with epidemiological studies that identify fathers as the most common offenders. The demographics of Philadelphia may partially explain the difference; Philadelphia has a large Black population with a substantial proportion of households led by mothers; father-perpetrated violence may be more likely to lead to child maltreatment reports.
Even though the parent-to-child conflict is frequently violent, arrest is a low probability intervention. Again, the police serve the function as first responder conflict resolution agents as opposed to agents of social control.
Overall, police responses to incidents of child and parent conflict and violence seem to occur because offenders, victims, and/or other observers/witnesses have no one else to call to intervene and resolve verbal and physical conflict. Both victims and offenders may be reluctant to call Child Protective Services because of assumptions that CPS will take the child into custody and place the child in out-of-home care, irrespective of who might be the offender.
Our multivariate analyses show that there is no difference in the likelihood of police providing some type of information (a DV community resource card, phone number for the DV advocate, etc.) to families when there is a child offender versus when there is a parent offender. This may be because officers are unaware of organizations that can handle parent-child incidents, the lack of such organizations providing information cards to police for on-the-scene distribution (agencies serving intimate partner violence victims prepare such information for the department), and, perhaps most likely, the lack of organizations capable of providing emergency intervention in such circumstances.
The Pennsylvania Department of Public Welfare Annual Report noted that in 2013, the year of our study, 4,546 cases of child maltreatment were reported for Philadelphia. In the entire state of Pennsylvania, 39% of all child maltreatment reports were referred by law enforcement officials (Pennsylvania Department of Public Welfare 2014). It is clear from these statistics that police have an important role in reporting cases of child abuse. However, in Pennsylvania, the bar for what is considered child abuse is set relatively high; although a substantial number of incidents to which police are called exhibit problematic behavior between parents and minor child, such incidents do not legally constitute reportable child abuse.
The police response to parent-child incidents today is similar to their response to intimate partner violence in the early 1980s; there is virtually no knowledge base regarding what constitutes an effective intervention. Researchers must continue work in this area of family violence in order to inform police, who act as the first responders to parent-child conflict, as to what could be the most effective means of intervening and mediating in cases of parent-child conflict.
Strengths and Limitations
A key strength of this study is that it was based on analyses of data of an entire population of domestic violence calls responded to by police in a large city in the United States in a single calendar year. This large number of cases increases the confidence in our findings.
There are a few limitations to this study. As with all research based on administrative data, there are inherent quality concerns (e.g. missing data and lack of independent verification). Nonetheless, administrative data provide a convenient and economical way to examine parent-child incidents that come to the attention of authorities. Another caveat is that data from domestic incidents are from only one year of police reports in one city. Therefore, it is possible that domestic incidents occurring in that single year are not representative of the trends in parent-child incidents that occur in other years. Also, as these incidents occurred in Philadelphia, it is uncertain the degree to which our findings in this study may generalize the incidents in other locations, especially in smaller cities, towns, rural areas, and in different regions of the United States. Given that we did not study incidents to which police were not summoned, findings cannot be extended to incidents for which service was not sought.
Our research shows that a substantial, but not overwhelming, number of incidents responded to by police were between parents and their minor children. As such, further research could be done to examine the nature of such domestic incidents; especially those that are non-violent, yet still involve police mediation. More qualitative analysis of these incidents may be carried out in the future in the form of surveys or interviews of parents and children involved in these incidents so that we may better understand who (parents, children, neighbors, etc.) makes reports to police and what leads to this reporting.Footnote 4 Similarly, police actions at the scene of these incidents, as well as their decision-making process in regard to providing assistance to families, should also be further examined in order to understand how police choose to respond to parent-child incidents, and ways in which their response could be altered and improved.
Lastly, this study provides a cross-sectional view of incidents occurring between parents and their minor children in one year in Philadelphia. In order to understand how such incidents factor into the overarching timeline of incidents that occur between parents and their minor children, more research must be done. For example, it is important to understand if verbal incidents escalate into violent incidents to which police officers are repeatedly called to intervene, and if so, to determine at what point verbal incidents occur in the history of incidents between offender and victim.
Parent-child incidents that involve police intervention constitute a significant portion of all domestic violence incidents. Previous estimates demonstrate that in a national sample, around 15% of domestic incidents were between a parent and child (Hutchison et al. 1994). The present study shows that in Philadelphia in 2013, around 20% of all domestic violence calls involved parents and their children. Parent-child incidents are an under-researched area of study, but the extent of such incidents and the cost of police intervention in such incidents merit more attention. For example, qualitative research designs could be used to provide insight into the reasons why police are contacted when there are non-violent incidents between parents and their children. Other research could examine police decision-making to shed light on how and why officers determine that some cases but not others meet the statutorily-defined criteria for investigation for possible child abuse. And, yet another line of research might examine the nature and frequency of incidents between children and their parents in terms of whether such incidents escalate, what external resources (including police) are called upon and are effective, as well as what actually constitutes “effective.”
In terms of whether the police are the most appropriate resource for responding to parent-child conflict and violence, domestic violence organizations, Child Protective Services, and other social services can perhaps provide more targeted, long-lasting assistance to resolve verbal incidents between parents and their children, and although officers play an important role in serving as the first responders to COPV incidents, they cannot be charged with resolving the more pervasive, underlying issues causing such incidents. Therefore, it is key that officers move these incidents along to other organizations which can better address and resolve these incidents, which will in turn lessen the likelihood that police will need to respond multiple times to the same household in the future.
In Pennsylvania, child abuse is defined as: 1) any act or failure to act by a perpetrator which causes or puts the child at risk for non-accidental serious physical or mental injury to a child, or causes or puts the child at serious risk of sexual abuse or exploitation, and 2) serious physical neglect to provide medical care and other “essentials of life” which endangers or impairs a child’s functioning (The Child Protective Services Law (n.d.)).
The National Incident-Based Reporting System (NIBRS) is a reporting system used by law enforcement that collects information on crimes reported to police at the incident level.
Estimates from the 2013 American Community Survey show that for Philadelphia, 42% of respondents reported their race as “White alone,” 43% reported themselves as “Black or African American alone,” around 7% reported themselves as “Asian alone,” with the remaining 8% reporting either as “American Indian or Alaska native alone,” or “Some other race alone,” or “Two or more races” (U.S. Census Bureau 2013b). The over-representation of Blacks in these data is consistent with the over-representation of Blacks in criminal justice data in general.
Whereas officers reported who was the offender and the victim (parent or child), our data did not provide information on who was responsible for reporting the incident to police.
Cornell, C. P., & Gelles, R. J. (1982). Adolescent to parent violence. Urban and social change review, 15(1), 8–14.
Dempsey, M. M. (2006). What counts as domestic violence? A conceptual analysis. William and Mary Journal of Women and the Law, 12, 301.
Ellwood, D. T., & Jencks, C. (2004). The spread of single-parent families in the United States since 1960.
Erez, E., & Tontodonato, P. (1989). Patterns of reported parent-child abuse and police response. Journal of Family Violence, 4(2), 143–159.
Evans, E. D., & Warren-Sohlberg, L. (1988). A pattern analysis of adolescent abusive behavior toward parents. Journal of Adolescent Research, 3(2), 201–216.
Gershoff, E. T. (2002). Corporal punishment by parents and associated child behaviors and experiences: A meta-analytic and theoretical review. Psychological Bulletin, 128(4), 539–579.
Hutchison, I. W., Hirschel, J. D., & Pesackis, C. E. (1994). Family violence and police utilization. Violence and Victims, 9(4), 299–313.
Ibabe, I., & Bentler, P. M. (2016). The contribution of family relationships to child-to-parent violence. Journal of Family Violence, 31(2), 259–269. https://doi.org/10.1007/s10896-015-9764-0.
Ibabe, I., Jaureguizar, J., & Bentler, P. M. (2013). Risk factors for child-to-parent violence. Journal of Family Violence, 28, 523–534. https://doi.org/10.1007/s10896-013-9512.
Margolin, G., & Baucom, B. R. (2014). Adolescents' aggression to parents: Longitudinal links with Parents' physical aggression. Journal of Adolescent Health, 55(5), 645–651. https://doi.org/10.1016/j.jadohealth.2014.05.008.
Nock, M. K., & Kazdin, A. E. (2002). Parent-directed physical aggression by clinic-referred youths. Journal of Clinical Child and Adolescent Psychology, 31(2), 193–205.
Pennsylvania Department of Public Welfare (2014). 2013 Annual Child Abuse Report. Retrieved from http://www.dhs.pa.gov/cs/groups/webcontent/documents/report/c_086251.pdf.
Shireman, J., Miller, B., & Brown, H. F. (1981). Child Welfare Workers, Police, and Child Placement. Child Welfare, 60(6).
Stanko, E. A. (1989). Missing the mark? Policing battering. Women, policing, and male violence (pp. 46–69).
Straus, M. M. A., Gelles, R. J., & Steinmetz, S. K. (Eds.). (1980). Behind closed doors: Violence in the American family. Transaction Publishers.
The Child Protective Services Law, 23 Pa.C.S. § § 3490.4 (n.d.) https://www.pacode.com/secure/data/055/chapter3490/s3490.4.html.
U.S. Census Bureau (2013a). Selected population profile in the United States 2013 American community survey 1-year estimates. Retrieved from https://factfinder.census.gov/faces/tableservices/jsf/pages/productview.xhtml?pid=ACS_13_1YR_S0201&prodType=table.
U.S. Census Bureau (2013b). Selected population profile in the United States 2013 American community survey 1-year estimates. Retrieved from https://factfinder.census.gov/faces/tableservices/jsf/pages/productview.xhtml?pid=ACS_13_1YR_B02001&prodType=table
Walsh, J. A., & Krienert, J. L. (2007). Child–parent violence: An empirical analysis of offender, victim, and event characteristics in a national sample of reported incidents. Journal of Family Violence, 22(7), 563–574.
Willis, C. L., & Wells, R. H. (1988). The police and child abuse: An analysis of police decisions to report illegal behavior. Criminology, 26(4), 695–716.
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
This work was supported, in part, by funding from the Ortner Center on Violence & Abuse in Relationships. We thank the Philadelphia Police Department for making the data available for analysis.
About this article
Cite this article
Schut, R.A., Sorenson, S.B. & Gelles, R.J. Police Response to Violence and Conflict between Parents and their Minor Children. J Fam Viol 35, 117–129 (2020). https://doi.org/10.1007/s10896-019-00088-6
- Adolescent aggression
- Child-to-parent aggression
- Domestic violence
- Family violence
- Police intervention