Calling the Police in Instances of Family Violence

Abstract

This study examines the impact of the victim–offender relationship on the willingness of victims to call the police in family violence incidents, with particular attention to the life stages of victims. Different stages of life have an impact on the decision to report criminal victimization. Family composition and the roles of family members change over life stages. When children are young, adults in the household have child-rearing responsibilities that shape the interpersonal dynamics in the household. When children approach adulthood and parents age, the parents may become more dependent on the children in a variety of ways. These changes in family composition and organization across life stages can affect the normative and cost–benefit considerations in deciding whether to call the police or not. Yet studies of responses to family violence have virtually ignored the influence of life stage on the decision to call the police. In addition, most studies of victims of family violence focus on marital or intimate relationships and fail to examine any other relationships in the family. It is not clear whether the findings from the general literature on domestic violence are applicable to intergenerational family violence. This article examines the impact of a broader range of victim–offender relationships across three age groups representing different life stages. The authors find that the factors explaining victims’ decisions to report victimization to the police vary across life stages.

Keywords

calling the police reporting victim–offender relationship life stage family violence

The current study addresses victims’ decisions to report to the police violent victimizations perpetrated by family members. Reporting of family violence has important implications. In most situations, reporting to law enforcement is required for any type of police or criminal justice response, and the deterrent effect of criminal sanctions more generally depends on the willingness of victims to report their victimization to the police. Even though studies of family violence are unclear about how well the different players in the criminal justice system respond to family violence, mobilizing the police is the first step to relief (Payne & Berg, 2003).

This study examines the influence of specific victim–offender relationships in conjunction with the age of the victim on the decision to report family violence to the police. There are many studies examining the impact of the victim–offender relationship on the victim’s decision to report a crime to police, but little research has been done specifically on reporting across life stages in the family. Most studies focus on age of the victim as a continuous variable and not as a discreet phenomenon that can have different meaning in the context of different life stages (Braithwaite & Biles, 1980; Felson, Messner, & Hoskin, 1999; Felson, Messner, Hoskin, & Deane, 2002; Gartner & Macmillan, 1995). Considering the current and impending changes in demographics, examining responses to family violence in certain life stages, and specifically those of older victims, is even more important. People are living longer because of advances in medicine and technology, and as the U.S. Senate Special Committee on Aging (1985–1986) reported, the population aged 55 or older will more than double between 1984 and 2050. The impact of moving toward an older society on the nature of, and responses to, family violence has seldom been considered. Perhaps the fact that the probability of victimization decreases as individuals age has led us to pay less attention to the victimization of older adults, in particular the elderly (Bachman, Dillaway, & Lachs, 1998; Bachman, Lachs, & Meloy, 2004; Bachman & Meloy, 2008).

More importantly, most studies of family violence or domestic violence focus exclusively on marital or intimate relationships and fail to examine other types of family relationships. This study expands the category of offenders to include children as well as intimate partners. The differential impact of these relationships on reporting abuse to the police is examined, and the effect of the relationship is assessed for victims in different age groups. In other words, this study tests whether our models for understanding the effect of victim–offender relations on reporting family violence to the police are generally applicable to intergenerational family violence and family violence in various life stages.

Literature Review

The prior relationship, or lack thereof, between victims and offenders is important in several ways for shaping the criminal justice response to victimization. It affects the victim’s decision to report an incident of abuse to the authorities, and prior relationship also has an impact on whether the response of the criminal justice system to the incident is appropriate or not. Family violence occurs disproportionately in private places, a fact that usually requires victims to file a police report for the authorities to be aware of the crime and respond to it. In addition, studies examining the impact of the victim–offender relationship on the reaction of the police and the courts have found a correlation between victim–offender relationships and ultimately the decisions made by police officers and prosecutors (Law Enforcement Assistance Administration, 1977a, 1977b).

The literature on why victims report crimes has emphasized both rational cost–benefit considerations and normative factors in the decision about whether or not to file a police report (Goudriaan, Lynch, & Nieuwbeerta, 2004). Cost–benefit explanations maintain that the greater the loss or injury and the greater the certainty of redress, the more likely an event will be reported. In other words, a crime is reported when the benefit of reporting (reduction in the chance of revictimization, recovery of property, etc.) outweighs the cost (e.g., time and efforts for reporting, aligned with the possibility for retaliation, etc.; Goudriaan et al., 2004; Schnebly, 2008). In contrast, normative theories hold that a crime is reported when this response is prescribed by society. Individuals’ decisions to report crimes are based on the normative contexts of their social environments. For example, those who believe that reporting a crime is their duty are more likely to do so even in the case of minor crimes. Norms can also influence the perception of what is appropriate to report to the police and what is not appropriate (Goudriaan et al., 2004; Schnebly, 2008). The normative prescription for reporting cybercrime, for example, is to contact an Internet provider or credit card company, not the police, whereas more traditional forms of theft are reported to the police and insurance companies.

One of the most significant factors affecting victims’ reporting decisions in the rational model is the seriousness of incidents. Victims are more likely to report incidents of more serious crimes involving both physical injuries and loss of property (Felson et al., 2002; Goudriaan et al., 2004; Skogan, 1976, 1984), and this finding is consistent across studies. In addition, the types of crime involved have an impact on victims’ decisions. Property crimes and violent crimes have different impacts on victims’ decisions. In the event of property crimes, victims are more likely to report incidents of thefts of insured property because of the higher probability of claiming compensation and insurance payment (Skogan, 1984).

Previous research has also indicated normative factors in the decision to report crimes to the police. To some extent, there is a public perception that crimes among nonstrangers are private matters and should not be reported to the police. Also, people believe that victims who engage in provocation are somehow complicit and should not call the police (Greenberg & Ruback, 1992). The place where a crime occurs also has an impact on the normative factors for victims’ decisions whether or not to report to the police. There is a general belief, for example, that crimes occurring at work or school should be reported to supervisors or teachers before being reported to the police (Lynch, 1987).

Crimes among nonstrangers are different from crimes among strangers in terms of both cost–benefit strategy and normative prescriptions for behavior. The interdependence of offenders and victims complicates victims’ reporting decisions. Sanctioning offenders could prove detrimental to victims or to third parties involved in the situation because of interdependency, as family members, between victims and offenders. In addition, offenders with immediate and constant access to victims are more likely to retaliate if a crime is reported. In terms of normative factors, the pervasive belief that family interactions are private and should not be reported can affect victims’ reporting decisions. Moreover, this norm may differ for parent–child relationships compared to intimate relationships. Parent–child relationships are thought to be the most enduring of all close relationships (Golish, 2000). In addition, the termination of a relationship is more difficult in situations of parent–child disputes. Older women, in particular, tend to have different perceptions about family issues. Many older women have been socialized with traditional values and attitudes that emphasize the role of gender and family values as well as the role of wives and mothers as a source of family cohesion (Zink, Regan, Jacobson, & Pabst, 2003). These values may lead older women who are being victimized by their children to refrain from reporting their victimization.

Although many studies have been conducted on victim–offender relationships, findings on the impact of victim–offender relationships on reporting decisions are inconsistent. Some studies show that violence by a nonstranger often goes unreported (Block, 1974; Braithwaite & Biles, 1980; Worrall & Pease, 1986). In a study by Block (1974), the cost–benefit calculation model of the victims’ decisions to notify was developed and tested. Self-reported victimization data compiled by the National Opinion Research Center in 1966 revealed that the closer the relationship, the less inclined the victim is to report victimization (Block, 1974). Most studies in this area, however, have found that the victim–offender relationship is not a significant predictor of crime reporting (Bureau of Justice Statistics, 1980, 1992; Lizotte, 1985; Skogan, 1984).

Also, there are inconsistencies in the literature about the impact of victim–offender relationships on reporting decisions for family violence and/or domestic violence. There is a common understanding that women who are being assaulted by their husbands or intimate partners are less likely to report their victimization to the authorities (Felson et al., 2002). However, studies have found mixed results as to whether victims of family violence are less likely to call police than are victims of violence by strangers. Studies using self-report data show a connection between the victim’s relationships to the offender and the decision to report (Block, 1974; Gartner & Macmillan, 1995). They found that when the offender is the victim’s partner or another family member, the victim is less likely to call the police. In contrast, Felson et al. (1999, 2002) analyzed National Crime Victimization Survey (NCVS) data and found that crime perpetrated by intimates and family members is more likely to be reported to the police.

Inconsistencies observed with regard to the effect of the victim–offender relationship on reporting to the police may occur because these effects are too complex to simplify. They may be highly dependent on, and conditioned by, other aspects of the context in which the crimes occur. Inconsistent findings about the victim–offender relationship, for example, could occur because different types of family relationships produce different outcomes in terms of reporting decisions. If all types of family relationships (e.g., intimate partner relationships, parent–child relationships, etc.) are considered as a single value on the variable (e.g., relatives), this may mask significant effects of specific relationships. Similarly, the effects of victim–offender relationships may differ depending on the social context, for example, families with elderly dependents. Although assaults on older persons perpetrated by the victim’s children may be routinely reported to the police, they may be less likely to be reported when the victim is financially or otherwise dependent on the offender. These effects of the victim–offender relationship on the decision to report will not be visible unless the relationship variable makes these fine distinctions or includes the social context of the families involved.

Older victims of family violence and particularly victims of intergenerational family violence may constitute a special group in the analysis of reporting decisions. Previous research has focused almost entirely on intimate partner relationships and failed to distinguish among the different types of familial bonds. Felson et al. (2002) classified the victim–offender relationship into “couple, other family, other non-stranger, and stranger.” Gartner and Macmillan (1995) grouped the victim–offender relationship into “spouse/ex-spouse, date or boyfriend, relative, work/friend/other, and stranger.” Considering children as a distinct offender group in family violence will provide specific information on the victim–offender relationship beyond that presented in previous studies.

The parent–child relationship may have different costs or benefits and normative implications than relationships among intimates. The norm of parents protecting their children from the criminal justice system may be much more significant than norms governing the protection of spouses. This could result in older parents not calling the police because they do not want to get their child in trouble. Even though the child is grown and has become an offender, the parents may still feel responsible for the child’s behavior. Conversely, parents may feel that they have a right to expect respect from their adult child, given that they devoted a great deal of time and energy to raising him or her. In this case, the thought of a child turning violent is abhorrent and generates extreme reactions, including reporting the child to the authorities. In terms of cost–benefit calculation, elderly parents may be even more dependent on their adult children than spouses are on each other. This may not be the case when parents are still healthy and employed, but as parents age they can become increasingly dependent on their children physically, emotionally, and financially. As discussed above, these distinctive characteristics of life stages may interact with the victim–offender relationship to affect the decision to report victimization to the police.

A recent study by the National Center on Elder Abuse at the American Public Human Services Association (1998) revealed that the major perpetrators of elder domestic abuse are the victims’ adult children. According to their survey, approximately 90% of elder abuse incidents are committed by family members, and in two thirds of the cases, the perpetrators are either the adult children or the spouses of the victims. Furthermore, this study also showed that 47.3% of perpetrators are the adult children of the victims, whereas 19.3% are spouses. When the definition of abuse is restricted to physical abuse, 48.6% of perpetrators are adult children and 23.4% are spouses.

Previous studies on family violence or domestic violence victimization, especially against women, consistently show the impact of certain risk factors on the chance of victimization. Age, gender, race, and socioeconomic status are correlated with the chance of victimization in general as well as with specific forms of violence within different types of victim–offender relationships (for a review of previous studies, see Lauritsen & Schaum, 2004). However, findings on the impact of victim characteristics and victim–offender relationships on the victims’ decisions to report their victimization are not as consistent. More importantly, the impact of victims’ life stage has not been thoroughly examined in previous studies. The current study tests whether the victim’s decision to report family violence to the police is influenced by the interaction of life stages and the victim–offender relationships.

Test of Empirical Model and Discussion

For the purposes of this study, the term family violence is defined as violence among both present and former family members. The U.S. government defines the term family as “a group of two or more persons related by birth, marriage, or adoption and residing together in a household” (U.S. Bureau of the Census, 1992). Intimate partners, both former and current, who are not legal spouses are also considered as family members in this study. Violence is defined as an intentional attack on or threat of attack on another.

To test the effects of different normative expectations among family members across life stages, separate models were estimated for three different age groupings of victims: young victims, adult victims, and older adult victims. The young group includes victims of family violence who were between 18 and 46 years old. This age range was chosen because these individuals are more likely to have dependent children, and this may shape normative expectations and relations among family members. In the United States between 1991 and 2000, women were, on average, approximately 27 years old when their first child was born, and men were, on average, approximately 28.7 years old (National Vital Statistics System, U.S. Department of Health and Human Services, 2002). It follows that the average age of parents with 18-year-old children is 47, and in these households the roles of family members will be largely defined by child-rearing obligations. Individuals who have children younger than 18 are obliged to support and nurture them. Moreover, adults younger than 47 are more likely to be financially and emotionally independent of other family members because they are able to work and are less likely to be ill than older adults. It is also the case that the children or stepchildren of victims in this age group are not generally able to abuse their parents physically, although there may be some exceptions.

The age range of 47 to 54 was used to define adult victims because the roles within the family change as the children reach adulthood, but the adults in this age range are still independent. Presumably, the role of nurturing children diminishes during this life stage, and normative prescriptions about protecting children may also be less influential, and adult parents may not be reluctant to call the police when older children engage in assaultive behavior. Norms regarding spousal interactions could also change so that attachments to spouses increase as the children remove themselves from the household. Older children leaving the household may have implications for dependency as the need to “stay together” to support children financially decreases and spouses become less economically dependent on one another. Presumably, this would result in greater willingness to report incidents of victimization perpetrated by spouses.

The older adult group includes victims aged 55 or older. Victims in this group are even further removed from the child-rearing years than are adult victims. They are more likely to be retired, which reduces income and also reduces contacts with friends and coworkers. Instead, they are more likely to stay with family members including spouses and children, and their activities are changed as well. In this life stage, norms and expectations regarding family members are different from those of young individuals and adults. Older adults may become more dependent on their adult children socially, emotionally, and financially. This might have an impact on victims’ reporting decisions in terms of rationality and even, perhaps, normative prescriptions. As the interdependence among family member increases, the relative benefits of calling the police may decrease.

Research Questions and Hypotheses

The three main research issues to be investigated are (a) the impact of victim–offender relationships on the reporting of family violence, (b) the impact of life stages on the probability of reporting, and (c) the interaction of life stage and other factors (especially victim–offender relationship) on reporting. We are testing five specific research hypotheses.

Hypothesis 1: Victims of family violence will be less likely to report when the offenders are their children than when the offenders are their intimate partners or some other relative.

This assumes that there is a strong norm of protecting one’s children from legal intervention across all life stages.

Hypothesis 2: Older adult victims of family violence will have a lower probability of reporting to the police than either young victims or adult victims.

This hypothesis is based on the assumption that the increased dependence of older persons on their family will tip the cost–benefit calculation in favor of not reporting family violence to the police. Police intervention will put at risk the support of their family that they so desperately need.

Hypothesis 3: The effect of victim–offender relationship on reporting family violence to the police will differ across life stages.

Hypothesis 3a: Younger victims will be less likely to call the police when the offender is a spouse or ex-spouse than victims in other life stages, but more likely when the offender is their child.

This hypothesis is based on the assumption that during this life stage more victims will feel the obligation to keep the family unit together for child-rearing purposes and will protect their spouse from legal intervention because of his or her contributions to that end. On the other hand, children who attack parents have crossed the line and pose a risk to the family unit, possibly including other children. For this reason, when the child is the perpetrator, the victim will be more likely to call the police.

Hypothesis 3b: Adults and older adult victims will be less likely to call the police than younger victims when the offender is a child than when the offender is a spouse.

In the adult and older adult life stages, the importance of the family as a child-rearing entity declines and the immunity that assaultive spouses enjoyed during earlier life stages is reduced and the probability of calling the police in response to attacks by spouses increases.

Hypothesis 3c: Older adult victims will be less likely to call the police in response to family violence committed by their children than adult victims or younger victims.

This is a strictly an instance of cost–benefit calculation. The greater the dependence, the less willing these older victims are to jeopardize the support coming from their adult children.

Data

This study uses data from NCVS, a national-level probability survey of the residential population of the United States. It is a stratified, multistage, cluster survey with a rotating panel design. NCVS is one of the most important sources of information regarding crime victimization, and data are collected from a sample of 77,200 households comprising nearly 134,000 persons each year. The survey is used to estimate victimization rates for different types of crime (e.g., rape, sexual assault, robbery, assault, theft, burglary, motor vehicle theft) for the population as a whole as well as for subgroups of the population. The NCVS provides information on the impact of crime victimization, characteristics of violent offenders, and victims’ responses to crime (see the U.S. Department of Justice, Bureau of Statistics Web page for more information <http://bjs.ojp.usdoj.gov/>). The abundant information available in the survey as well as the large sample make it possible to examine victims’ reactions to family violence incidents across different life stages. From public NCVS data files from 1992 to 2003, an analysis file that included all family violence incidents reported was created. The analysis is based on the 3,662 incidents of family physical abuse that involved victims 18 years or older.

Measures

The dependent variable in this analysis is whether or not victims called the police to report incidents of physical abuse by a family member. Calls from a third party to the police are excluded because this study focuses on the victim’s reporting decision.¹ The primary independent variable is the victim–offender relationship. Consistent with our theoretical frame, different categories of the victim–offender relationship—(a) spouse or ex-spouse, (b) (ex-)boyfriend or girlfriend, (c) child, and (d) other relative—are used. The “child” designation in the NCVS includes both children and stepchildren. As we noted above, life stage is measured by the age of the victim, with those 18 to 47 years of age classified as younger victims, those 48 to 54 as adult victims, and those 55 and older as older adults.

The analysis also includes several other independent variables that may affect victims’ decisions to report. The seriousness of incidents is included in the model as a dummy variable—assaults versus threats. The category of assault includes simple assault, aggravated assault, sexual assault, robbery with injury, attempted robbery with injury, and so on.² Threats include verbal threats of physical assault, sexual assault, and rape. A dichotomous variable indicating a history of prior attack by the same offender, or any other family members, is also included in the model. The history of prior attacks variable is included because in cases of repeat victimization victims may “make a routine” of the experience and cope in ways that do not involve calling the police. It is also possible that repeat victimization by the same offender makes the victim even more disturbed, and the crime is more likely to be reported. Either way, a history of prior attacks can be an important consideration in the decision to call the police. The victim’s socioeconomic status is measured by the household income, education level, and most recent work experience of the victim. Individuals’ income level and the sources of income would be the most desirable indicators of their economic status, but the NCVS asks about only household income and not personal income. Whether the victim was in the labor force and whether the victim lived in a high-income household are used as measures of socioeconomic status under the assumption that someone who is working or has income will be less economically dependent. In addition, individuals with a higher education level will be less economically dependent. The ordinal variable of offender age is included in the analysis as an independent variable. The specific variables used in the analysis and the distributions of the values of these variables are presented in Table 1.

Table 1.

Descriptive Statistics of Sample—Pooled Model.

Variable	Attributes	Coded	Valid Percentage
Calling the police	Yes	1	46.3
	No	0	53.7
Seriousness of crime	Assault	1	78.8
	Threat	0	21.2
Victim–offender relationship	Spouse or ex-spouse	1	41.8
	(Ex-)boyfriend or girlfriend	2	39.5
	Children	3	8.0
	Other relatives	Reference	10.7
History of prior attack	Yes	1	29.0
	No	0	71.0
Victim’s job status (had work last week)	Yes	1	63.9
	No	0	36.1
Gender of offender	Male	1	84.7
	Female	2	15.1
Gender of victim	Male	1	15.5
	Female	2	84.5
Offender’s age	Younger than 12	1	0.5
	12–14	2	1.6
	15–17	3	3.7
	18–20	4	6.4
	21–29	5	31.8
	30 or older	6	55.3
Victim’s age	M	33.65
	SD	11.73
Total number of cases		3,662^a

Victims of family violence, physical abuse only, aged 18 or older.

Method

To investigate the differences in reporting among different stages of life across age groups, the analysis is conducted in two stages. First, a pooled model is used to examine the impact of the independent variables, including victim’s age as a continuous variable on the victim’s decision to report. Then, three different models specific to different life stages are examined to compare the effects of predictor variables on the reporting decisions of the young, adult, and the older adults. The comparison of the different life stage groups of victims will indicate the effect of life stages on the decision to report, and we expect that the reporting decision of adult victims and the older adults will be different from that of younger victims. It is expected that the older adult victims will be influenced by normative prescriptions against reporting their own children to authorities and by their reliance on the interrelationship among family members, whereas young and adult victims will be deterred from reporting only by the normative prescription of protecting their spouses.

The logistic regression analysis is based on unweighted data. Scholars use both weighted and unweighted data from the NCVS, and both have their own strengths and weaknesses. Unweighted data have been used to model victimization risk, whereas weighted data are more often used for national estimates of levels and changes in crime. We use the unweighted data to model the victims’ reporting decision because doing so was logistically simpler and will not substantially affect the analysis.³ To test the effects of using weighted data in the analysis, we estimated the model with the pooled data applying the incident weights and not applying the weights. (See the NCVS codebook [Bureau of Justice Statistics, 2004] for a detailed discussion of weighting in the survey.) There is no significant difference between the weighted and unweighted analysis results except for the fact that the victim’s education variable becomes nonsignificant in the weighted pooled model analysis. Overall, the weighted and unweighted analyses are similar in model significance as well as the coefficients of predictors.

The theoretical model being tested is presented in Figure 1. As the literature review indicates, there are several different factors that affect victims’ decisions to report instances of family violence. Factors that have been shown to significantly affect reporting decisions are considered as major independent variables in this study. These factors are the seriousness of the incident, history of prior attacks by the same offender, and the socioeconomic status of the victim. Age plays an important role in differentiating the impact of these predictors. Different victim–offender relationships are considered in this study, which involves victims’ expectations about normative prescriptions because of their different stages of life. The socioeconomic status of the victim is also considered with regard to financial dependence on family members. Other forms of dependence, such as psychological and emotional dependence, are difficult to operationalize even though they are crucial to victims’ decisions to report family assault. In addition, there are no variables to measure types of dependency other than financial dependency in the data.

Figure 1.

Description of model.

Results: The Changing Nature of Family Violence Across the Life Stages

Model I: All-Ages Pooled Model

In the first model, the full sample of 3,662 victims of family violence is used, and descriptive statistics for the variables in the first model are presented in Table 1. The age of the victims ranges from 18 to 90 years old, with a mean of 33.65 (SD = 11.73). Fewer than 50% of the victims (46.3%) called the police. In terms of the seriousness of the incidents, assaults composed approximately 80% of the abuse incidents. The most popular offender category was spouse or ex-spouse (41.8%), followed by current or former boyfriend or girlfriend (39.5%). Only 8.0% of offenders were the children or stepchildren of their victims. Approximately 30% of victims experienced prior attacks by the same offender, and two thirds of victims were employed the week before the survey. More than 90% of offenders were aged 18 or older, and this might be related to the fact that the majority of offenders were (ex-)spouses or (ex-)boyfriends or girlfriends. As expected, family violence is mainly male violence toward female victims. White is the most frequent race category for both victims and offenders because of the demographics of the general population. More than 80% of victims are White, and victim’s race is not considered in the model because of the very small number of non-Whites.

The model predicting reporting to the police with 10 independent variables is statistically significant, χ²(12) = 76.198, p < .001, but does not fit the data very well (pseudo-R² = .028). Table 2 presents the pooled model of the victims’ decisions to notify police—all victims aged 18 or older are included in this model. In the offender categories that represent the victim–offender relationship, “other relatives,” which generally refers to the most distant relationship in the family (e.g., uncles), is a reference category. The correlation matrix including the dependent variable and predictors is presented in Appendix B. The largest bivariate correlation among these variables is .187 for the association between victim’s age and offender’s age. The variance inflation factor score ranges from 1.010 to 1.47, and no tolerance score is lower than 0.69. Multicollinearity is not a concern for these variables.

Table 2.

Logistic Regression Coefficients—Pooled Model.

Independent Variable	B	SE	Wald	df	Sig.	Exp(B)
Assault	0.258	0.085	9.249	1	.002	1.294
Offender category			19.444	3	.000
Spouse or ex-spouse	0.319	0.125	6.511	1	.011	1.375
Boyfriend or girlfriend	0.163	0.125	1.685	1	.194	1.176
Children	0.733	0.189	15.066	1	.000	2.082
Prior attack	0.321	0.076	17.618	1	.000	1.378
Job status	−0.123	0.072	2.938	1	.087	0.884
Gender of offender (being female)	0.066	0.117	0.323	1	.570	1.069
Gender of victim (being female)	0.442	0.117	14.372	1	.000	1.556
Household income	0.001	0.001	0.588	1	.455	1.001
Victim’s education level	−0.008	0.003	7.425	1	.006	0.992
Offender’s age	0.091	0.048	3.595	1	.058	1.096
Victim’s age	0.001	0.004	0.037	1	.848	1.001
Constant	−1.899	0.423	20.167	1	.000	0.150
χ²(12)	76.198
p	< .001
−2 log likelihood	4870.68
Nagelkerke R²	.028

There are five positive and significant determinants for crime reporting, and there is one negative significant predictor. For the categorical variable of victim–offender relationship, both categories—spouse or ex-spouse and children—are significant predictors. When the offender is a spouse or an ex-spouse, the victim is 1.4 times more likely to report the incident to the police than when the offender is another relative (Exp(B) = 1.375). Victimization at the hands of a child increases the odds of reporting by a factor of 2 (Exp(B) = 2.082). A victim of assault, compared to a victim of a threat, is 1.3 times more likely to call the police (Exp(B) = 1.294), so the seriousness of the crime matters. The experience of prior attacks has a positive impact on the victim’s decision to call the police (Exp(B) = 1.378). Female victims are more likely to report an incident, and a higher education level of the victim decreases the odds of reporting.

Model II: Models Differentiated by Age Group

The descriptive statistics for the three victim groups—young, adult, and older adult victims—are compared in Table 3. In all of the groups, fewer than 50% of victims called the police. This reaffirms the common understanding that a substantial proportion of intrafamilial violence is not reported to the police. The probabilities of reporting across age groups are not significantly different even though the adults are a little less likely to report compared to young and older adult victims.

Table 3.

Descriptive Statistics of Samples—Age-Divided Models.

		Young Adults (18–47)	Adults (48–54)	Older Adults (55 or Older)
Variable	Attribute	Percentage	Percentage	Percentage
Reported to police	Yes	46.3	44.6	49.2
	No	53.7	55.4	50.8
Offense seriousness*	Assault	79.9	70.9	71.6
	Threat	20.1	29.1	28.4
Victim–offender relationship*	(Ex-)spouse	42.5	43.9	25.7
	(Ex-)boyfriend or girlfriend	42.6	18.2	18.0
	Children	5.4	24.2	28.4
	Other relatives	9.5	13.7	27.9
Prior attack	Yes	29.3	25.6	30.0
	No	70.7	74.4	70.0
Have a job*	Yes	65.3	64.3	39.0
	No	34.7	35.7	61.0
Gender of offender*	Male	85.6	80.7	76.0
	Female	14.3	18.9	24.0
Gender of victim*	Male	14.5	21.4	23.6
	Female	85.5	78.6	76.4
Offender’s age*	Younger than 12	0.4	1.1	0.5
	12–14	1.4	3.5	1.6
	15–17	3.4	6.3	4.9
	18–20	6.6	4.9	2.2
	21–29	34.7	14.4	9.3
	30 or older	52.8	68.8	78.7
Victim’s age*
	M	30.45	49.83	65.23
	SD	7.84	2.21	10.98
Total number of cases		3,194	285	183

Age group differences are significant at the .05 level.

These data are roughly consistent with our argument that the face of family violence changes across life stages. The characteristics of family violence among younger victims suggest that it is more centered on sexual intimacy and child rearing, which is not the case for the adults and older adults. First, as we can see from Table 3, the proportion of assaults perpetrated by intimate partners [(ex-)spouses and (ex-)boyfriends or girlfriends] is higher among younger victims than other groups, and the proportion of assaults perpetrated by children and other relatives increases as the age of the victim increases. The percentage of offenders in the “other relatives” category also increases as the age of the victim increases. This implies that the face of family violence is centered on intimate partner violence for young victims, whereas it is more on family violence for the older adults. Second, there is a substantial increase in financial and emotional dependency as the victim’s age increases, as indicated by the decrease in the proportion of victims who have jobs. Of the young adult victims, 65% had jobs when they answered the survey, compared to only 39% of the older adults. Job status can also have an impact on one’s social life. Third, there are differences in the offender’s and the victim’s gender depending on the age variable. The majority of offenders are male in all life stage groups, but there is an increase in female offenders for adult victims and the older adults. There is also an increase in male victims as victim’s age increases, although the majority of victims are still female. Last, as victim age increases, offender age increases. For the adult and the older adult victims, there is a substantial increase in the percentage of offenders who are older than 30. These results are an indication of the change in the social context of family violence across life stages.

These descriptive statistics suggest that the social context of family violence changes across the life stages, and this is consistent with the expectation that the decision to call the police in response to family violence may also change across life stages. The changing functions and composition of families over the life stages are consistent with increases in non–intimate partner violence over the life stages. The same is true for the increases in family members not in the labor force.

To examine the different determinants of the victim’s decision to notify the police among the three age groups, logistic regression models are estimated with the same 10 independent variables that were included in the pooled model. All three models are statistically significant in distinguishing between reporting and nonreporting. In the young victims model (victims aged 18 or older and younger than 47), there are some determinants in common with the pooled model (see Table 4). Assaults, compared to threats, are 1.3 times more likely to be reported to the police (Exp(B) = 1.302). All three of the victim–offender relationship variables are significant predictors. If the offender is a spouse or intimate partner, as opposed to another relative, the chance of reporting increases. In addition, children, compared to other relatives, are 1.7 times more likely to be reported to the police (Exp(B) = 1.737). This is the most powerful predictor in the model. The experience of prior attacks by the same offender also increases the likelihood of reporting (Exp(B) = 1.326), and female victims are more likely to call the police (Exp(B) = 1.648).

Table 4.

Logistic Regression Coefficients—Age-Divided Models

	Young Victims			Adult Victims			Older Adult Victims
Independent Variable	B	SE	Exp(B)	B	SE	Exp(B)	B	SE	Exp(B)
Assault	0.264	0.092	1.302**	0.454	0.295	1.575	−0.143	0.391	0.867
Offender category
Spouse or ex-spouse	0.528	0.141	1.679**	−0.453	0.461	0.636	−1.077	0.512	0.341**
Boy(girl) friend	0.317	0.139	1.372**	−0.315	0.498	0.730	−0.674	0.595	0.510
Children	0.552	0.276	1.737**	0.737	0.467	2.089	0.494	0.467	1.639
Prior attack	0.282	0.082	1.326**	0.967	0.302	2.629**	0.261	0.375	1.298
Job status	−0.084	0.078	0.920	−0.230	0.277	0.794	−0.534	0.359	0.587
Gender of offender	0.037	0.137	1.037	0.699	0.357	2.012	−0.188	0.428	0.829
Gender of victim	0.499	0.139	1.648**	0.187	0.337	1.206	0.565	0.421	1.759
Household income	0.001	0.001	1.001	−0.004	0.005	0.996	0.006	0.006	1.006
Victim’s education level	−0.007	0.003	0.993**	−0.013	0.010	0.987	−0.042	0.019	0.959**
Offender’s age	0.035	0.063	1.036	0.228	0.157	1.256	0.080	0.179	1.083
Victim’s age	0.005	0.006	1.005	0.036	0.059	1.036	−0.009	0.018	0.991
Constant	−1.978	0.449	0.138	−4.403	3.172	0.012	0.424	1.844	1.527
χ²(12)	67.588	30.481	23.770
p	< .001	.002	.022
−2 log likelihood	4262.990	352.843	207.735
Nagelkerke R²	.028	.139	.177

Significant at the .05 level.

In the adult model that contained a sample of victims aged 47 to 54, there is only one significant determinant after controlling for all other variables. None of the variables in the offender category is a significant predictor of the victim’s decision to report. Experience of a prior attack by the same offender is a significant predictor. Victims who had experienced prior attacks by the same offender, compared to those who had not, were 2.6 times more likely to call the police (Exp(B) = 2.629), and this is consistent with the young victims model. The overall fit of the adult model is better than that of the model predicting reporting among the young.

In the older adult model that includes the sample of victims aged 55 or older, there are two significant determinants after controlling for all other variables. Only one variable in the offender categories is significant. Contrary to the previous models, being victimized by a spouse or ex-spouse decreases the odds of reporting by approximately 66% (Exp(B) = 0.341). Unexpectedly, however, having an offender who is the victim’s child does not affect the victim’s decision of reporting. A higher level of education decreases the chances of reporting (Exp(B) = 0.959). The overall fit of the model for the older adult victims is better than the fit of the models for the young adult and adult victims.

Discussion and Conclusion

It is clear that the victim–offender relationship does affect reporting decisions. In a pooled sample of family violence incidents, we found that when the offender is a spouse or ex-spouse or when the offender is a child (including stepchild), the likelihood of reporting is higher than when the offender is in the reference category of “other relative.” These results are not consistent with the research hypothesis. Rather than being protective of children or spouses and not reporting them to the police, it seems that victims may be outraged or fearful that close relatives are assaulting them and report at higher rates. One possible explanation for this can be that assaults by their children make victims angry about the breach of the normative prescription of the relationship between parents and children, which results in a higher probability of reporting.

The second research question asked whether the life stages of individuals differentiate reporting behavior. We hypothesized that the different stages of life would affect victims’ rational calculations of the cost–benefit analysis of reporting and normative prescriptions of reporting as well. Our findings support this research hypothesis in some ways. The overall reporting rates differ across the life stage groups, but these differences are not statistically significant. Adults (44.6% reporting rate) are less likely to report their victimization compared to young victims (46.3%) and older adults (49.2%).

The factors that encourage and discourage reporting do differ across life stages. A history of prior attacks encourages reporting behavior for young victims and adults but not older adults. In addition, being female encourages the chance of reporting only for young victims, and this gender effect diminishes as victims age. It is also clear that the impact of victim–offender relationships on reporting decisions differs across the life stages. When offenders are spouses or intimate partners, the odds of reporting for young victims increase significantly, but this effect on reporting decreases through the life stages to the point where older adult victims are less likely to report assaults by spouses. The progression from a positive effect to a negative effect is apparent in the adult model, where the sign of the coefficient is negative but the coefficient is not significant, and even more apparent in the older adult model, where the sign remains negative and the coefficient becomes significant. This finding is not consistent with our expectation, but it is consistent with the idea that as child rearing becomes less of a central concern, spouses may become more important to each other and less willing to involve the police in assaults. Boyfriends and girlfriends show a similar pattern with reporting over life stages, but it is not as strong.

Reporting decisions in assaults perpetrated by children are more consistent across life stages. When children assault parents, the parents are more likely to report the attack to the police. The coefficient is positive, large, and significant for young parents, and it remains large and positive but is insignificant for the group of older adult victims. This may be in part because of the smaller number of cases in the older adult victims group.

We hypothesized that the reliance of older adult victims on their abusive children and their normative prescription about the relationship between them and their children would make it unlikely for the older victims to report instances of assault, but our findings do not support this hypothesis. The dramatic change in the effect of spousal relationships across the age groups suggests that older adults may be more dependent on their spouses than on their children, and for this reason they are less likely to report incidents of spousal abuse.⁴

None of the indicators of independence seem to have had the expected effect. The employment status of victims has no effect on reporting in any of the models, and the same is the case for household income. If being employed or well-off reduces financial dependency on family members, this does not translate into a greater likelihood of reporting familial violence to the police. The educational attainment of victims, which should give them added independence, is negatively related to reporting in all life stages and significantly so among younger adult victims and older adults. All of this suggests that if dependence is influencing the reporting of family violence to the police across life stages, then it may be logistical and emotional dependence rather than simple financial dependence.

Finally, experiences of prior victimization by the same offender encourage victims’ reporting for both young victims and adult victims. Instead of making their victimization routine and being less likely to report, victims are more likely to report.

One encouraging finding of the current study is that although increased dependence on spouses may make victims less likely to call the police in cases of spousal assault, the proportion of family violence committed by spouses decreases sharply among older adult victims (see Table 2).

These findings regarding the reporting of family violence across life stages are limited by the fact that we do not have very detailed and precise measures for the characteristics of life stage and dependency. We have inferred stages of life from age and dependency from crude socioeconomic indicators. The sample size also limits what we can say about older adult victims and their reporting behavior. In spite of the relatively massive size of the NCVS sample and the accumulation of cases over the years, family violence is a relatively rare event among older adults, and large samples are necessary for reliable analyses. The sample size in this analysis required that we define older adult victims as being 55 years of age or older. This definition may include many people who are far from elderly by any definition based on level of functioning and therefore dependency. If we had been able to define the older adult life stage as 75 and older and had had enough victims for analysis, our results may have been very different. With that said, there was a marked difference in lifestyle across the age groups that we did examine (see Table 3), and the age variable within the older adult victim model did not have a significant effect on the decision to report to the police. This suggests that although our definition of life stage was not ideal, it was picking up variation in the lifestyles and can provide a conservative test of the effects of life stage and victim–offender relationship on the reporting of family violence to the police.⁵

Footnotes

Appendix

Appendix B

Correlation Matrix Among Variables

	Report	Assault	Prior Attack	Job	Offender Sex	Victim Sex	Victim Age	Income	Education
Assault	.050**
Prior attack	.058**	.022
Job	−.039*	−.042*	.029
Offender sex	−.048**	−.004	.054**	.058**
Victim sex	.074**	.018	−.078**	−.110**	−.463**
Victim age	.014	−.060**	−.036*	−.039*	.061**	−.089**
Income	.010	−.017	.002	.045**	.010	−.032	.069**
Education	−.049**	.001	−.011	.103**	.023	.004	.101**	.071**
Offender age	.006	−.054**	−.055**	.012	−.020	.061**	.187**	.046**	.075**

Note: Bivariate correlation is not applicable for categorical variables. For the categorical variable in the analysis model, the victim–offender relationship category, cross-tabulation with dependent variable was conducted. All of the subdivided categories for the association between offender category and reporting decision have a substantial number of cases.

Correlation is significant at the .05 level, two-tailed. **Correlation is significant at the .01 level, two-tailed.

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the authorship and/or publication of this article.

Financial Disclosure/Funding

The author(s) received no financial support for the research and/or authorship of this article.

Notes

Author Biographies

Ji Hyon Kang received her PhD at the Graduate Center/John Jay College and is an associate research fellow at the Korean Institute of Criminology, Seoul, Korea.

James P. Lynch received his PhD in sociology at the University of Chicago and is a distinguished professor at the University of Maryland, College Park, USA.

References

Bachman

Dillaway

Lachs

M. S.

(1998). Violence against the elderly: A comparative analysis of robbery and assault across age and gender groups. Research on Aging, 20(2), 183-198.

Bachman

Lachs

M. S.

Meloy

(2004). Reducing injury through self-protection by elderly victims of violence: The interaction effects of gender of victim and the victim/offender relationship. Journal of Elder Abuse & Neglect, 16(4), 1-24.

Bachman

Meloy

M. L.

(2008). The epidemiology of violence against the elderly: Implication for primary and secondary prevention. Journal of Contemporary Criminal Justice, 24(2), 186-197.

Block

(1974). Why notify the police? The victim’s decision to notify the police of an assault. Criminology, 11, 555-569.

Braithwaite

Biles

(1980). Empirical verification and Black’s The behavior of law. American Sociological Review, 45, 334-338.

Bureau of Justice Statistics. (1980). Intimate victims: A study of violence among friends and relatives. Washington, DC: U.S. Department of Justice.

Bureau of Justice Statistics. (1992). Criminal victimization in the United States, 1991. Washington, DC: U.S. Department of Justice.

Bureau of Justice Statistics. (2004). National crime victimization survey, 1992-2003 [Data file and code book] (Conducted by U.S. Department of Commerce, Bureau of the Census, ICPSR ed.). Ann Arbor, MI: Inter-University Consortium for Political and Social Research.

Felson

R. B.

Messner

S. F.

Hoskin

(1999). Victim–offender relationship and calling police in assaults. Criminology, 37, 931-947.

10.

Felson

R. B.

Messner

S. F.

Hoskin

A. W.

Deane

(2002). Reasons for reporting and not reporting domestic violence to the police. Criminology, 40(3), 617-647.

11.

Gartner

Macmillan

(1995). The effect of victim–offender relationship on reporting crimes of violence against women. Canadian Journal of Criminology, 37, 393-429.

12.

Golish

T. D.

(2000). Changes in closeness between adult children and their parents: A turning point analysis. Communication Reports, 13(2), 79-97.

13.

Goudriaan

Lynch

J. P.

Nieuwbeerta

(2004). Reporting to the police in Western nations: A theoretical analysis of the effects of social context. Justice Quarterly, 21(4), 933-969.

14.

Greenberg

M. S.

Ruback

R. B.

(1992). After the crime: Victim decision making. New York: Plenum.

15.

Lauritsen

J. L.

Schaum

R. J.

(2004). The social ecology of violence against women. Criminology, 42, 323-357.

16.

Law Enforcement Assistance Administration. (1977a). Forcible rape. A national survey of the response by police (Vol. 1). Washington, DC: Government Printing Office.

17.

Law Enforcement Assistance Administration. (1977b). Forcible rape. A national survey of the response by prosecutors (Vol. 1). Washington, DC: Government Printing Office.

18.

Lizotte

(1985). The uniqueness of rape: Reporting assaultive violence to the police. Crime & Delinquency, 31, 169-190.

19.

Lynch

J. P.

(1987). Routine activity and victimization at work. Journal of Quantitative Criminology, 3, 283-300.

20.

National Center on Elder Abuse at the American Public Human Services Association. (1998). The National Elder Abuse Incidence Study. Washington, DC: Author.

21.

National Vital Statistics System, U.S. Department of Health and Human Services. (2002). National vital statistics report. Washington, DC: Author.

22.

Payne

B. K.

Berg

B. L.

(2003). Perceptions about the criminalization of elder abuse among police chiefs and ombudsmen. Crime & Delinquency, 49, 439-459.

23.

Schnebly

S. M.

(2008). The influence of community-oriented policing on crime-reporting behavior. Justice Quarterly, 25, 223-251.

24.

Skogan

W. G.

(1976). Citizen reporting of crime: Some national panel data. Criminology, 13, 535-549.

25.

Skogan

(1984). Reporting crimes to the police: The status of world research. Journal of Research in Crime and Delinquency, 21, 113-137.

26.

U.S. Bureau of the Census. (1992). Statistical abstract of the United States. Washington, DC: Government Printing Office.

27.

U.S. Senate Special Committee on Aging. (1985–1986). Aging America: Trends and projections. Washington, DC: U.S. Department of Health and Human Services.

28.

Worrall

Pease

(1986). Personal crimes against women: Evidence from the 1982 British Crime Survey. Howard Journal, 25, 118-124.

29.

Zink

Regan

Jacobson

C. J.

Pabst

(2003). Cohort, period, and aging effects: A qualitative study of older women’s reasons for remaining in abusive relationships. Violence Against Women, 9, 1429-1441.