Understanding and Preventing Violence, Volume 3: Social Influences (1994)

Candace Kruttschnitt

Are males really more violent than females? Although disagreement on this important question admittedly remains, a number of scholarly works have concluded that gender, perhaps more than any other variable, produces a dramatic and consistent difference in the extent and nature of interpersonal violence. For almost four decades now, men have dominated official reports of violent crime and, regardless of the data source, they appear to engage disproportionately in the most injurious acts of interpersonal violence. Race differences produce some variations among different data sets; gender differences, however, appear and reappear across time and different social contexts in crimes of violence and acts of aggression. From a public policy as well as a theoretical standpoint, this robust association is quite significant. Simply asked, what is it about being female that reduces the likelihood of aggressive or violent behavior? Unfortunately it appears that we know far more about the strength of the association between gender and interpersonal violence than about why it exists. Although numerous scholars have assessed the relationship between gender and crime (e.g., see Widom, 1978; Nagel and Hagan, 1982; Rutter

and Giller, 1983; Wilson and Herrnstein, 1985), virtually no integrated research reviews focus specifically on acts of violence. Instead the explanations for the relationship between gender and interpersonal violence have remained in disparate intellectual fields and in various theoretical camps, including biological, social psychological, structural, and methodological.

The purpose of this paper is to examine critically the research pertaining to gender and interpersonal violence with an eye toward providing a better understanding of the role of gender in producing different rates and types of interpersonal violence. Every attempt is made to present relevant data from all intellectual traditions.¹ However, as can be seen, the work on violence among males far exceeds the work on violence among females. As a result, readers will find that this review provides a synthesis of the omissions in our knowledge of the relationship between gender and interpersonal violence. We begin by defining the concepts that provide the framework for this analysis.

The Panel on the Understanding and Control of Violent Behavior limits its consideration of violent human behavior to interpersonal violence, which is defined as behavior that "threatens, attempts, or actually inflicts physical harm." This definition of violence is composed of the following three elements: (1) behavior, by one or more persons, that threatens, attempts, or inflicts physical harm (i.e., the harmful act need not be completed to be included in this study); (2) intentional infliction of physical harm (i.e., the definition excludes negligence and recklessness); and (3) one or more persons who are objects of the harmful behavior (i.e., the victims).

The term sex is used to refer to genetic sex or the chromosomal makeup of the individual. It is "sex" and not "gender" to which we refer when making the distinction between people who are biologically male or biologically female (Schur, 1984:10). By contrast, the term gender refers to the sociological, psychological, and cultural patterns that are used to evaluate and to shape male or female behavior. The evidence suggesting that gender is socially constructed is now well documented (e.g., see Macaulay, 1985; Bender, 1988:15, note 38; Epstein, 1988). However, because gender is imposed on sex by acculturation and socialization, it is not surprising to find that these two concepts are still used interchangeably (cf. Widom, 1984:5). The failure to distinguish sex

from gender can have serious implications for research examining behavioral differences between men and women. For example, criminologists frequently refer to sex as one of the most important demographic variables in their research because of the strong association observed between this variable and aggregate arrest rates. Yet a reference to sex differences in arrest statistics implies a biological basis for the disparate rates. Explanations for these differences, however, include socialization and opportunity, as well as biological factors. Similarly, models predicting sex differences in interpersonal violence can be misspecified when they fail to include biological variables; conversely, a gender-based explanation of violence should model exogenous social and cultural variables.

Further confusion arises with the related variables of gender role and gender identity. Gender roles are commonly perceived as a set of behavioral expectations based on an individual's sex in a particular social context, whereas gender identity usually refers to an individual's self-conception of being male or female. Each variable can have important and different etiological influences on crime. For example, female gender role socialization might constrain aggression or societal responses to it, whereas a masculine gender identity might encourage it (e.g., see Widom, 1984). However, in the relevant literature we find that (1) "gender/sex roles" often are employed as a generic concept for both social roles and personality traits (Norland and Shover, 1978), and (2) there are both considerable measurement variability (cf. Thornton and James, 1979; Norland et al., 1981; Horwitz and White, 1987) and questionable validity in the underlying constructs of masculinity and femininity (Spence and Sawin, 1985; Gill et al., 1987). Not surprisingly, then, our ability to draw from and build upon previous research pertaining to gender and interpersonal violence is hampered by this lack of conceptual and methodological clarity.

Despite these limitations, this paper attempts to evaluate existing research and to suggest the most promising avenues for future work. The following section presents data on gender and interpersonal violence to address questions such as (1) Are men more violent than women regardless of offense, age, or residence? (2) Do victimization data present a comparable picture? (3) Do we have any evidence to indicate whether the gender gap in violent crime or violent victimizations is converging? (4) How do the violent criminal careers of men and women differ? Then we examine explanations for gender differences in aggression and violence,

with particular emphasis on gender role theory and the data relevant to this theory. Attention is also directed to correlates of female violence, and a preliminary set of explanatory hypotheses is offered for the most prominent patterns observed in gender and violent crime. Policy issues concerning the adjudication and sanctioning of violent offenders and sexual assault victims are discussed next. Finally, we conclude with a discussion of the research priorities that may lead to a better understanding of the relationship between gender and interpersonal violence.

Statistical data on gender and interpersonal violence are generally drawn from indicators of violent crime: arrest reports, victimization and self-report surveys, and public health agencies. In this section, we see that certain observations-most notably that females are underrepresented in the most serious/injurious types of violence-are remarkably consistent among all sources of data.

Perhaps the best-known data for assessing gender differences in interpersonal violence in the United States are the Uniform Crime Reports (UCR). These reports are collected annually from local law enforcement agencies throughout the country (although there is a higher rate of reporting among urban than rural agencies) by the Federal Bureau of Investigation. UCR data contain information both on offenses known to the police and on arrests. Only the arrest data, however, are broken down by the gender of the offender; these data have a number of well-known limitations: (1) they are limited to offenses that result in arrest; (2) the data vary considerably in the accuracy with which they reflect illegal behavior due to, for example, the misclassification of similar crime events or the nonrecording of a crime (Cressey, 1970; Erickson, 1975; Steffensmeier et al., 1979; Blumstein et al., 1986); (3) as summary statistics, they fail to distinguish multiple offenders from multiple events and include attempted offenses with completed offenses (Steffensmeier and Allan, 1988); and (4) broad offense categories, such as assault, may contain a set of heterogeneous criminal acts (more generally, see Reiss, 1981; Weis, 1986). Despite these problems, scholars argue that for certain purposes (e.g,, when serious

offenses are considered), UCR arrest data provide valid indications of the demographic distribution of criminal behavior (Hindelang et al., 1979; Gove et al., 1985).

The following analyses are based on five indices of violent crime for the year 1988: murder/nonnegligent manslaughter, rape, robbery, aggravated assaults, and other assaults. The overall measure of violent crime includes only the four indices used in the violent crime index by the UCR (i.e., it excludes other assaults). To calculate arrest rates, two types of computations were performed on UCR data. These computations were derived from Steffensmeier's extensive work on UCR data and gender differences in arrest rates (see Steffensmeier et al., 1979, 1989; Steffensmeier, 1980, 1982; Steffensmeier and Allan, 1988; Steffensmeier and Streifel, 1989).

First, 1988 arrest data from UCR were combined with census data to compute offense arrest rates that take into account sex distributions in the population.² Because few people under the age of 10 commit crimes, the rates are calculated for persons age 10 years and older (or for the population at risk). The formula used to compute the arrest rates is

where M = the arrest volume given in the appropriate UCR table, P = the estimated population volume figure from the same UCR table, N = the estimated number of persons who would be in the UCR table if coverage were complete (e.g., total U.S. population, total rural population; this figure is taken from the U.S. census); and T = the estimated number of persons in the target category for whom the arrest figures are given (e.g., females age 10 and over; this figure is from U.S. census data).

Second, to estimate the gender disparity in violent crimes, we calculate the female percentage of arrests (FP/A), controlling for the sex distribution in the target population. The FP/A is calculated as follows

where fn = female arrest rate/100,000 for offense (i) and year (j ) and mn = male arrest rate/100,000 for offense (i) and year (j). The FP/A also facilitates comparing these arrest data with Steffensmeier's earlier longitudinal analyses, ultimately suggesting whether the

relative gap in male and female violence has narrowed or widened.

Table 1 presents violent crime arrest rates for males and females for 1988. First, it appears that the male rate of arrest for index violent crimes is about seven times higher than the female rate. Second, although there is some variation in the rates at which men and women are arrested for various types of violent crime, in no case does the female rate exceed, or even approach, one-quarter of the male rate. Not surprisingly, the largest variation appears for the crime of rape and the smallest for other assaults.

Table 2 further disaggregates these data by age. When comparing the same ages and offense categories, the arrest rates of males are substantially higher than those of females. However, if we assume that these data indicate actual rates of offending among males and females, the peak ages of violent activity vary little by gender. For example, in the case of robbery or aggravated and other assaults, the arrest rates for both males and females are highest from the midteens to the late twenties. The only exception to this pattern is murder. Here we find that female involvement seems to continue at a relatively equal, albeit low, rate into the thirties, whereas the male rate drops off in the thirties. The greater tendency for women to engage in intrafamilial homicides

(Wolfgang, 1975; Mann, 1988) probably explains why murder rates have a slightly flatter age-related curve for women than men.

A relatively high proportion of the assaults (EP/As for aggravated and other assaults) also occur among the youngest group of females (10-14 years). Possible explanations for this finding encompass age, period, and cohort phenomena.³ First, because Steffensmeier and Allan's (1988:63) comparable analysis of UCR data for the years 1979-1981 reveals similarly high FP/As for assaults among the youngest age group (13-17 in their analysis), a cohort effect seems unlikely. Second, as to a period effect, one could argue that a gender convergence in crime is occurring due to changes in socialization: women who became parents during the 10-year period following the early stages of the women's movement (in the late 1960s and early 1970s) may have been especially sensitive to the issue of gender equality in the raising and socialization of their children. However, we have little confidence in a period explanation because other studies using various methodologies (Adler, 1975; Smith and Visher, 1980), and spanning a wide range of years, also find a more pronounced narrowing of the gender gap for adolescents than adults. An age effect, combined with social expectations, may be the most plausible explanation (more generally, see Farrington, 1986). These relatively high FP/As for assaults involving young adolescents in both 1979 (Steffensmeier and Allan, 1988) and 1988 may be due to (1) the natural tendency for females at this age to begin spending more time away from home and in the company of peers; (2) the visibility of this offense (Black, 1980:152); and (3) the greater willingness of parents and others to invoke legal authority when the crime involves a female (Hagan et al., 1985).

The final UCR data we present for 1988 involve an analysis of arrest rates for violent crimes by residence (urban and rural) and gender.⁴ Here we find that, regardless of gender, arrest rates are higher in urban than in rural areas and that, regardless of residence, the arrest rates of males again far exceed those of females (see Table 3). Notably, however, across most offenses the size of rural/urban differences in FP/A is small or negligible (see also Steffensmeier and Allan, 1988). The observation that residence adds little to our ability to predict proportional female involvement in violent crime may be explained by (1) the basing of arrest statistics on place of arrest rather than place of offender's residence, and (2) the apparent direction of a significant amount of female aggression and violence toward relatives and family members.

Since 1973, the Census Bureau has collected annual data for the National Crime Survey (NCS), using a sample of 60,000 households. The NCS collects data only from victims, age 12 and over, of six crimes, three of which are violent: rape, robbery, and assault (aggravated and simple). Initially, victimization surveys were designed to assess the extent of unreported crime; subsequently, a substantial difference was found in the number of crime victimizations and the number of offenses known to the police (President's Commission on Law Enforcement and Administration of Justice, 1967:21). Methodological differences between the UCR and the NCS in recording crimes may account for some of this discrepancy. For example, when a single robbery incident results in the victimization of more than one person, NCS records information on each victimization, regardless of the number of criminal incidents involved.⁵ The UCR, however, records information on offenses (and only the most serious offense if more than one occurs within a given crime event) and arrests, regardless of how many victims are involved. It is also important to remember that victimization surveys rely on the victim's judgment and memory about whether a crime has occurred, and analyses suggest that memory fade may vary with the relational distance between the victim and the offender (Law Enforcement Assistance Administration, 1972:Table 5). Nevertheless, NCS data provide a rich source of complementary crime data. They reveal not only the degree to which violent crime offending and victimization covary within the same gender and age groups, but in examining the perceived characteristics of offenders, they can also help to validate the patterns observed in UCR data.

We begin by presenting victimization rates, for 1987, for the three crimes of violence by gender.⁶Table 4 reveals that, with the exception of rape, males are more likely to be the victims of criminal violence than females. In the case of completed simple assaults with injury, however, the male and female rates are very comparable. Although the proportion of these victimizations that involve domestic violence remains unknown, it seems likely that spouse abuse is an explanatory factor. More generally, gender appears to have little effect on the differences between rates of attempted and completed victimizations: for both men and women, assaults are less likely to be completed than robberies. Finally, it is interesting to note that the rates of attempted and completed rape are virtually the same. It is possible, however, that if knowledge of the victim-offender relationship were available, the data

would indicate a higher completion rate for rapes involving known offenders (see Russell, 1984:59).

Table 5, which examines gender-age patterns of victimization, reveals that the peak age at which men and women are violent crime victims varies only slightly. Attempted and completed violent crime victimizations peak in the mid-to late teens for men (ages 16-19) and in the early twenties for women. Moreover, rates of completed victimizations are virtually identical for males and females age 25 to 49, and age 65 and over. Thus, regardless of gender, the peak ages for both offending and victimization appear to be from the midteens to the midtwenties (see also Russell, 1984, for comparable self-report data on rape victims and offenders).

Finally, as with the UCR data, we examine violent crime victimization rates by gender and location (metro/central city versus nonmetro). Locality of residence in the NCS data pertains to the place where the person lived at the time of the interview, not to the place where victimization occurred. Table 6 reveals that, regardless of gender and type of violence, victimization rates are higher in metro than nonmetro areas. The victimization rates of women also vary somewhat less by location than those of men, which suggests again that locality/residence has less impact on the frequency with which women are involved in violence (as either victims or offenders) by comparison to men. Further, as can be seen in Table 6, women's violent victimization rates are not always lower than the violent victimization rates for men. In fact, the completed violent crime victimization rate of metro women (13.76) exceeds the comparable rate for nonmetro men (9.66). Although these aggregate violent crime victimization rates for metro women and nonmetro men are significantly influenced by the skewed gender distribution of rape cases, comparable data appear for robbery rates.

A Closer Look at the Data Thus far, the NCS and UCR data on violent crime by gender suggest a considerable degree of parity between victims and offenders. Generally, women are much less likely to be involved in crimes of violence as either offenders or victims than men. The only exceptions to these patterns involve rape victimizations and the interaction of gender and crime location in robbery victimizations. The NCS data, however, can also be utilized to shed light on the validity of UCR data. Because this survey elicits information from crime victims about characteristics of the offender and whether the crime was reported to the police, we can begin to examine (1) the degree to which male dominance in crimes of violence is related to discretionary behavior on the part of police toward female offenders, and (2) whether females underreport victimization experiences as a result of the often-noted relational quality of their aggression.⁷

Table 7 presents data on the perceived sex of the offender from crime victims' reports. These data, again, suggest an overwhelming concentration of male offenders in crimes of violence (85.6% male and 13.7% female). The only offense category in which women even approximate roughly one-fifth of the offender population is simple assaults (males 82.8%, females 16.5%); the FP/A for other assaults in the UCR data was 14.5 percent. Thus, consistent with Hindelang's (1979:147) findings, it appears that

the net gender-linked selection bias through the arrest stage is very small.

Table 8 illustrates the percentage of violent crime victimizations reported to the police by the victim's gender and the victim-offender relationship. First, these data indicate that only about one-half of all violent crimes are reported to the police and that females are slightly more likely than males to report being victimized. Gender-based discrepancies in reporting are particularly notable in the case of robbery (47.5 and 69.6% of the victimizations are reported by men and women, respectively) and simple assault (33.9% of the men report and 47.3% of the women). The greater reporting on the part of female victims of simple assault may be a function of increased recognition and reporting of domestic violence incidents. In the case of robbery, this may be due to differences (perceived or actual) in the seriousness of the event because such factors as extent of physical injury, financial loss, or weapon use are important determinants of crime reporting (Hindelang and Gottfredson, 1976; Gottfredson and Hindelang, 1979). This would also be consistent with our finding that proportionately

more robberies are completed with female victims than with male victims (Table 4).

Second, although both men and women show greater willingness to invoke the police when victimized by strangers, this reporting pattern is relatively weak and sometimes variable (cf. Black, 1976:40-48). Perhaps what is most surprising is the lack of variation in women reporting rape to the police based on the victim-offender relationship (cf. Russell, 1984:96). Recent analyses of the characteristics that predict rape reporting suggest that factors that elevate the offense to a higher level of seriousness, or make prosecution easier, may tip the scales in favor of reporting the incident to the police (Lizotte, 1985). In effect, the victim analyzes the strength of her case before deciding whether to report the offense. For example, in addition to the familiarity of the offender (i.e., the more familiar, the less likely the report), the more property a rapist steals and the more serious the physical injury, the more likely it is that the victim will report. Hence with the considerable amount of attention given to the handling of rape cases by criminal justice personnel over the past decade

(e.g., see Polk, 1985), it may be that factors other than victim-offender familiarity are influencing a victim's decision to notify the police.

Finally, although NCS data provide information on the gender and the race of violent crime victims, the UCR fails to provide comparable information in arrest data. Because UCR data indicate that a disproportionate amount of violent crime occurs among nonwhites, it is certainly possible that the interaction of gender and race may alter the picture we have of male dominance in crimes of violence. Analysis of other official sources (e.g., health statistics, police files in selected cities) provides an unsystematic but informative examination of the interrelationships among gender, race, and violent offending and victimization.

Race and Gender Virtually all of the research pertaining to the interaction of gender, race, and violent crime focuses on murder. Studies spanning the last decade uniformly suggest that race may be a better predictor of homicide than gender. For example, using homicide data from Detroit police records, Letcher (1979) found that, of the women arrested for murder, black women are more frequently involved in acts of lethal violence, both as victims and as offenders, than are white women. Wilbanks's (1982) analyses of the National Center for Health Statistics data for 1975 confirmed Letcher's finding and further suggested the following rank ordering of homicide victimization rates for the four race-sex groups: nonwhite males, nonwhite females, white males and white females. Mann's (1987) ongoing field research, which involves collecting data from centralized homicide files in the police departments of Chicago, Houston, and Atlanta, also reveals that, of the women arrested for murder, black women are found predominantly in cleared murder cases. Their proportions among total murder arrestees in these cities range from a low of 8.8 percent (Houston, 1979) to a high of 14.2 percent (Atlanta, 1983). The comparable range for white female murders is a low of 1 percent (Chicago, 1983) to a high of 5.2 percent (Atlanta, 1979) (Mann, 1987:177). Finally, in a review of both national and local studies that examine race, gender, and homicide, Riedel (1988a:9-10) finds that (1) the interaction of gender and race, in the case of both homicide offenders and homicide victims, produces the same rank ordering for the four race-sex groups as described by Wilbanks; and (2) the current black female homicide victimization rates are most similar to the rates of white males.

The limited data available on race-gender subgroups pertaining

to other crimes of violence generally underscore the patterns observed in studies of homicide. For example, Steffensmeier and Allan (1988) used 1979-1981 statewide statistics, gathered by the Pennsylvania UCR system, to make black-white comparisons in the arrest rates and the FP/As for a full range of offenses. The violent crime (i.e., murder, aggravated and simple assault) arrest rates reveal the same rank ordering by race and gender as previously noted. Offense rates for black females, although closest to those for white males, actually exceed offense rates for both white males and white females. Comparable figures also are reported in NCS data on violent juvenile offending from 1973 through 1981 (Laub and McDermott, 1985). To illustrate more fully these race-gender interactions, Laub and McDermott also computed offending ratios by race and gender subgroups. Regardless of the offense, the lowest ratios appear for white males/black females and the highest for black males/white females. Within racial groups the gender ratio for assaults is reported to be lower for blacks than for whites. Although Laub and McDermott draw attention to this latter finding as a crime-specific pattern, the calculation of gender participation ratios for blacks and whites reveals an identical pattern across a broader range of offenses (e.g., robbery, auto theft, burglary), reflecting greater racial differences in female than male participation (Visher and Roth, 1986:251).

Self-Report Studies Numerous cross-sectional self-report studies have been conducted over the past 30 years, again in an attempt to assess the amount of crime that goes undetected. These data pertain almost uniformly to youthful offenders. Smith and Visher's (1980) meta-analysis of 44 self-report (and some official) studies encompassing the years 1940-1975 treated the magnitude of the association between gender and deviance/crime (gamma) derived from each study as a dependent variable to be explained by the characteristics of the study. By so doing, the gender-crime association appeared to be less for (1) self-report versus official data (a point to which we will return), (2) personal and youth offenses versus violent and property offenses, and (3) nonwhite versus white populations. The relationship was greater for adults and urban samples as opposed to youth and rural samples. Comparable patterns were also uncovered during 1976-1980 by the National Youth Survey. The most pronounced gender differences for both prevalence and incidence rates were for violent offenses (Elliott et al.,

1983); a significantly higher proportion of black than white females reported involvement in violent crimes (Ageton, 1983). Irrespective of age, then, the participation of white females in crime appears to be concentrated in minor acts of deviance rather than serious criminal behavior.

Relational Violence That the magnitude of the gender-violence association may depend on the severity of the behaviors measured is also evident in studies of domestic violence. For example, studies of courtship and marital violence produce rates (per 100 people) that vary from a high of 37 to a low of 1.8 for the husbands/male partners and from a high of 24 to a low of 0.02 for the wives/female partners (Frieze and Browne, 1989:178). Variations in the questions asked of respondents (e.g., whether injuries were sustained, whether the violence was mutual or in self-defense) and difficulties in obtaining accurate recall of these events (especially when they occur frequently and information is elicited from only one party) probably contribute to the wide range of estimates (see also Kurz, 1989). However, what does appear consistently from these studies is that men have higher rates of using the most dangerous and injurious forms of violence (Stark et al., 1979; Berk and Loseke, 1981; Berk et al., 1983; Makepeace, 1983, 1986; Stets and Pirog-Good, 1987; Kratcoski, 1987; Aizenman and Kelley, 1988; Frieze and Browne, 1989).

In a related vein, research on the physical abuse of children frequently reports that mothers are at least as abusive toward their children, if not more so, than fathers (Bennie and Sclare, 1969; Steele and Pollock, 1974; Gil, 1970; Parke and Collmer, 1975; Straus et al., 1980; U.S. Department of Health and Human Services, 1981; Gelles, 1982; Widom, 1987; Gelles and Straus, 1988). Others, however, suggest that sample selection bias (i.e., researchers' and clinicians' focusing only on mothers) and a failure to consider the ''time at risk" factor may explain the unusually high rates of female violence in this setting (Pagelow, 1984:187-190). Still others (Gelles, 1979) point out, again, that the most dangerous and potentially injurious acts are performed more by men than by women (for a more complete discussion of the methodological issues involved in analyzing domestic violence data, see Fagan and Browne, in this volume).

Taken together, both official incidence and self-report data paint a very consistent picture of the relationship between gender and interpersonal violence. Women are substantially underrepresented in crimes of violence, most notably when attention is directed to

the most serious/injurious acts of violence. Further, although the gender-race interaction produces considerable variation in the rates of serious homicide, less is known about how this interaction affects other types of interpersonal violence. Whether these relationships between gender and violent offending are part of a larger pattern of stability or change remains to be seen.

Political and social changes in the lives of women, especially since the rebirth of the women's movement, prompted some scholars to hypothesize that significant increases are occurring, and will continue to occur, in female crime (Adler, 1975; Simon, 1975). In this section we consider whether, since 1960, (1) there has been an increase in women's levels of violent crime in this country, and (2) there has been an increase or decrease in the proportional involvement of females in crimes of violence.

To address these issues, Table 9 presents data from three decades (1960, 1977, 1988) on male and female arrest rates (per 100,000) and the female percentage of total arrests (FP/A) for violent crime.⁸ The 1960 and 1977 data are extracted from previous research (Steffensmeier et al., 1979), and the 1988 data appear in Table 1. With regard to the first question, we find a moderate increase from 1960 to 1977 in the rates of female violence, regardless of the specific offense. However, comparable data for males indicate much larger increases across all offenses, suggesting that the relative gender gap has remained stable. Over the next decade (1977-1988), the female murder and robbery rates actually declined, whereas assault rates showed moderate increases. Identical patterns, by offense type, appear for males over the years 1977-1988.

To answer the second question, we examine the female percentage of total arrests (FP/A) for violent crimes from 1960 to 1988. These data replicate some of the same patterns initially identified by Steffensmeier and his colleagues (1979:222). The relative gap between males and females in rates of violent crimes narrowed slightly for robbery and other assaults but widened for homicide; aggravated assaults remain relatively stable over this 28-year period (see also Steffensmeier and Streifel, 1989). Similar analyses of rural and urban violent crime arrest data also produced no significant shifts in the relative gender gap (see Steffensmeier and Jordan, 1978; and Table 3).

Finally, we examine whether the involvement of adolescent females in violent crime changed materially over the period in question. Again, by relying on previously published research for our base of comparison, we can determine whether the gender gap in crimes of violence among youth is closing at a faster pace than it is among adults. Specifically, Steffensmeier and Steffensmeier (1980) calculated gender-specific arrest rates on all UCR offense categories (except rape) for the aggregate age group 10 through 17; for comparable ages, our data are disaggregated into two age groupings (10-14 and 15-19); therefore comparisons between our 1988 data and the Steffensmeiers' data should be interpreted cautiously. Only the FP/As for ages 15 to 19 years are used because crime rates tend to be higher among this age group than among 10 to 14 year olds; any decreases in the proportion of adolescent female involvement in violence, then, cannot be attributed to focusing on the least crime-prone group. Examination of the FP/As for 1965, 1977, and 1988 (Table 10) for youth involved in crimes of violence reveals a small narrowing of the gender gap in the case of robbery and other assaults but relatively stable patterns in the case of murder and aggravated assaults. The narrowing of the gender gap for robbery and other assaults, however, largely occurred from 1965 to 1977; from 1977 to 1988 the gender gap widened slightly (see also Ageton, 1983).

Thus, irrespective of age, stability rather than change is reflected in official data on female violent offending. Although rates of female violence show some modest increases, they are far outweighed by the increases in male violence. Not surprisingly,

then, the absolute gender gap in violent crime (male rate minus female rate) continues to widen (see Steffensmeier et al., 1979; and Table 1). Whether this gender-based pattern of violent crime exists in other societies is difficult to determine. Few scholars have simultaneously addressed cross-cultural and longitudinal trends in female violent crime, in part because of the difficulty of obtaining reliable definitions of types of crimes (see Archer and Gartner, 1984, for an excellent critique of INTERPOL data). Nevertheless, what limited evidence exists suggests that women in both industrialized and nonindustrialized nations continue to play relatively minor roles in violent criminal activities (Adler, 1981; Simon and Baxter, 1989).

Far less attention has been devoted to gender-specific changes in victimization rates. One recent study, however, utilizing NCS data (on robbery, larceny, and assault) and UCR data (on homicide) found a definite increase in the proportion of female robbery victimizations (rising from 28% in 1973 to 37% in 1982) (Smith, 1987b; see also Bowker, 1981). Changes in the routine lifestyles of women may explain this finding. "Women now may be seen as more accessible and profitable targets than in the past, targets posing less threat for the offender than might confrontations with males" (Smith, 1987b:298). Homicides and assaults, by contrast, are less likely to involve changes in lifestyle because they frequently occur between individuals who are known to one another. Trends in black and white female homicide victimization rates further support this supposition because they have remained relatively constant since 1940 (see Riedel, 1988a:10).

An important qualification to this "lifestyle" explanation of gender differences and changes in violent crime victimization rates can be found in the work of Gartner and her colleagues (1987). Their study is also one of the few longitudinal analyses of victimization rates that employs cross-cultural data. Specifically, using pooled time-series data from 18 industrialized nations for the period 1950-1980 to examine and explain changes in females' risk of homicide victimization, they found that (1) the victimization rates for both males and females increased over the 30-year period in almost every nation; (2) there is a wide range of variation in homicide rates across nations, with the United States having the highest rate, regardless of gender, across the 30 years; and (3) although women's nontraditional activities do increase their risk

of victimization, changes in lifestyle are not a sufficient explanation for such an increase. In the context of modern, industrialized societies, it appears that women's risk of victimization increases only when nontraditional activities and lifestyles are matched with traditional patriarchal expectations.

Criminal Careers Another way of looking at trends and patterns in violent crime is to focus on the criminal career of the individual violent offender. "A criminal career is the characterization of the longitudinal sequence of crimes committed by an individual offender" (Blumstein et al., 1986:12). Over the last decade, scholars have increasingly moved away from relying on aggregate crime statistics and have started to examine two components of these statistics as key elements in the analysis of a criminal career: participation, the distinction between those who commit crime and those who do not; and frequency, the rate of activity of active offenders (Blumstein et al., 1986:12; see also Blumstein et al., 1988a,b; Weiner, 1989). Analyses of participation and frequency allow one to determine, for example, whether changes in crime statistics are due to a greater proportion of the population initiating criminal activity or merely to changes in the frequency of offending by active criminals. Examination of criminal career data can also answer such important policy issues as age at initiation, escalation and desistance of criminal activity, and causal factors associated with each of these stages of a criminal career.

Unfortunately, the bulk of the data that would allow one to assess such issues as the initiation, specialization, escalation, and termination of violent criminal careers pertains only to males (e.g., see Farrington, 1982; Weiner, 1989). The general tendency to exclude female offenders from longitudinal research on delinquency and antisocial behavior may be due, at least in part, to the greater frequency of violent offending among males. However, even among those studies that contain both males and females in the sample populations, the respondent's gender is often ignored; the researchers simply fail to disaggregate the data by gender (e.g., see Hamparian et al., 1978).

Weiner's (1989) recent analysis of violent career criminals represents an important exception to this pattern. Utilizing research that involved individual data on violent criminal careers from multiple sources (e.g., analyses of arrest histories of an urban birth cohort and self-reports of offending by high school students), Weiner

found substantial gender variation in participation, age at initiation, recidivism, and desistance of violent career criminals:

1. The cumulative violent juvenile participation of females is well below that of males, irrespective of the type of violent crime and irrespective of whether the data are self-reports or official records (Weiner, 1989:49, 56).

2. Age-specific participation rates for males exhibit a pattern of initial increase over the juvenile years (12-18) followed by a decline after either the more advanced juvenile or the young adult years (19-21).

3. By contrast, serious violence among females (based on Elliott et al., 1986), which is uniformly below the male rate, declines from the early juvenile years (12-14) through young adulthood (age 21) (Weiner, 1989:63-64).

4. Age-specific violent career hazard rates (based on Elliott et al., 1986) also indicate that violent female careers appear to both begin and peak earlier than do those of males (Weiner, 1989:102).

5. Females are at a much lower risk than males of recidivating violently, and females desist from violence at a much greater rate than do males (Weiner, 1989:108-109).

6. Generally, violent offenses constitute a modest proportion of the total offense accumulation of all offenders in a population; females accumulate an even lower proportion of violent offenses than males, and these offenses tend to be concentrated in the younger age groups (Weiner, 1989:121).

Unfortunately, no systematic attention has been given to gender differences in the various studies of violence specialization or escalation (see also Farrington, 1986).

The National Research Council's Panel on Research on Criminal Careers, which focused primarily on criminal careers that involve robbery, burglary, and aggravated assault, arrived at comparable findings with regard to gender and participation. The most consistent pattern with respect to gender was the extent to which male criminal participation in serious crimes exceeds that of females, regardless of data source, crime type, level of involvement or measure of participation (Blumstein et al., 1986:40). However, that panel's analysis of gender variation in the individual frequencies for active offenders led it to conclude that active female offenders commit crime at rates similar to those of active males (Blumstein et al., 1986:67). By contrast, Weiner (1989:82) finds that, among those who are actively participating in violent crime,

the male rate outstrips the female rate. This discrepancy in gender ratios may be, in part, a function of relying on different data sets. The panel's conclusion was based on two studies: Inciardi's (1979) analyses of self-report data from active heroin users and the National Youth Survey (NYS; Elliott et al., 1983). Weiner, focusing only on violent offenders, largely relies on a latter analyses of the NYS data (Elliott et al., 1986), which constrained individual offense rates for "serious violent" youngsters to three violent crimes. Nevertheless, these findings warrant further attention. As the panel suggested, if there is substantially less gender variation in frequency of offending than there is in participation, the large differences found between males and females in aggregate arrest rates may simply be due to differences that arise from gender-based variation in participation rates (Blumstein et al., 1986:67; see also Fagan, 1990). Whether this hypothesis can be generalized to nonwhites and adult violent offenders also warrants further exploration. As we have seen, a simultaneous consideration of gender and race in aggregate homicide data places the rates of black females closer to those of white males than white females. Because race also appears to be a stronger predictor of participation than frequency (Blumstein et al., 1986:72), attention must be directed toward estimating gender differences by race in violent crime participation rates.⁹ If the violent crime male/female participation ratio is lower among blacks than whites (see Visher and Roth, 1986:251), we may want to rethink some of the etiological gender-based theories of violent offending.

Gangs Finally, one particular type of criminal career, gang membership, also has received considerable attention in criminological research. Although the bulk of this research has focused on male gang members, one can find selected accounts of the activities of female gang members and, more recently, changes in their roles and status over time.¹⁰

Prior to 1970, there was little evidence of any substantial changes in the roles or functions of female gang members; they acted primarily as weapon carriers and decoys. The most common form of female gang involvement remained as auxiliaries or branches of male gangs (Thrasher, 1963; Short, 1968; Hanson, 1964; Miller, 1973). Subsequent to 1970, several ethnographic accounts of gangs implied that females were fighting in more arenas than previously and were increasingly employing the weapons that males use. For example, Quicker's (1974) interviews with 13 female Chicana gang members in East Los Angeles revealed that women would fight

other female auxiliary gang members, as well as carry weapons for males (cited in Bowker, 1978, and Campbell, 1984). Brown (1977), studying black female gangs in Philadelphia, found that in sexually integrated gangs, females were not relegated to peripheral activities but fought alongside the males in gang warfare. Further, his analysis of an all-female gang (the Holly Ho's) suggested a considerable amount of aggression among the women; the gang was reputed to attack both males and females, and owned weapons that ranged from knives to sawed-off shotguns. Male gang members in Los Angeles also reported some female involvement in violent gang activities; nevertheless, females were excluded from most of the economic criminal activity (Bowker et al., 1980).

A contrasting view of the role and activities of female gang members occurs in concurrent and subsequent studies. A 1975 survey of gang activities in six major U.S. cities suggested that any noted changes in female gang member activities does not alter the general position and role of women in gangs. Specifically, although Miller (1975) documented some reports of increased violence by female gang members, the overall distribution of female participation in gang activities did not differ from the past (cited in Bowker, 1978:145): females still acted as weapon carriers, with the most common form of involvement continuing to be as auxiliaries of male gangs. Campbell (1984), completing an ethnographic study of three nonwhite female gang members in New York City, found little change in female gang members' dependence on male gangs but some change in the internal dynamics of the female gang. According to Campbell (1984:32), although it is still the male gang that paves the way for the female affiliate, a girl's status now depends to a larger extent on her female peers (see also Vigil, 1988). Finally, Fagan (1990), studying predominantly nonwhite youth in three U.S. cities, provides perhaps the most systematic assessment of gender variation in gang and nongang criminal activity. He found that self-reported prevalence rates for violent offenses (i.e., the percentage reporting at least one incident of a felony assault, minor assault, or robbery) were significantly lower for female than for male gang members. However, female gang members' violent crime prevalence rates exceeded the rates of nongang males. The context of the gang itself, or its links to criminal opportunities, then may be a particularly important factor in explaining females' initial participation in violent offending, but there is no evidence that it affects their subsequent frequency of violent offending (Fagan, 1990:12-13).

The data on female gang activity at least suggest the possibility

of transition in status and increased violence among women members. However, it is equally probable that these changes represent changes in the attitudes of, and methods used by, researchers studying female offenders as much as they do real qualitative changes in the women's behavior. As Campbell (1984:6) suggests, "It is difficult to separate the true nature of girls' involvement from the particular interpretive stance of the writer (usually male), whose moral and political view most probably reflects the prevailing community standard." Ultimately, answers about changes in female gang activity will depend on the acquisition of longitudinal data from various cities with racially diverse gang populations. In the interim, however, we would do well to concentrate on gender differences both in the etiology of gang participation and in violent criminal careers. The gang appears to be an important source of initiating female violent criminal offending, but at the same time, it remains independent of the development of subsequent rates of offending (cf. Klein and Maxson, 1989; Fagan, 1990). Whether this is due to changes in the nature of gangs themselves or in adolescents' lives in communities where gangs are prominent should be explored.¹¹

The data we have presented suggest that, in most contexts and across various time periods, interpersonal violence is predominantly a male phenomenon. The degree to which these data are valid indicators of the proportional involvement of males and females in criminal violence remains the subject of some controversy.

As we have seen, the primary source of information used in examining gender-related rates of, and trends in, interpersonal violence are police generated arrest statistics that appear annually in the Uniform Crime Reports. Gender differences in officially recorded criminal activity could be biased if citizens fail to notify the police when the offense involves a female or if, once notified, the police are less likely to arrest female suspects. With regard to reporting, there is no evidence of discrimination that would favor female offenders. Hindelang's (1979:151) analysis of NCS data for 1972-1976 reveals that, when the seriousness of the offense is controlled, few differences emerge in the rate of reporting personal

crime victimizations by male and female offenders. In fact, for the most serious victimizations, female offender crimes are slightly more likely to be reported to the police than male offender crimes. Further, a simultaneous consideration of offense seriousness and sex of victim uncovered a counterbalancing effect: female victims were less likely to report female than male offenders to the police, whereas male victims reported female offenders more often than they reported male offenders to the police. This reporting pattern may even be sustained when victimizations involving strangers and nonstrangers are considered separately (Hindelang, 1979:152, note 15).

Research on the effects of gender bias in arrest is limited and generally quite poor; traditionally it has involved only observational impressions, a limited number of offenses, and few, if any, statistical controls (Visher, 1983:7-8). Further, although several scholars have speculated that changes in law enforcement practices may account for notable increases in female arrest rates for property offenses (Rans, 1978; Krohn et al., 1983; Steffensmeier and Streifel, 1989), few have focused on violent offenses. Perhaps the best available basis for assessing the effect of gender on arrest probability is Smith and Visher's (1981) analysis of the determinants of police arrest decisions in three metropolitan areas of the United States. Beginning with a probit analysis of the effects of both legal and extralegal variables on the arrest decision, Smith and Visher (1981) found that, all else constant (including offense severity), gender has no significant effect on probability of arrest. Further exploring the possibility that police treat women preferentially, Visher (1983) used the same data set to estimate separate probit models for males and females (the analysis included a t-test for equality of coefficients across equations). Her results indicate that Uniform Crime Reports may overestimate female involvement in property offenses and underestimate the criminality of older white females. Specifically, violent offenses had a greater impact on males, whereas property offenses were more likely to influence arrest probability for female suspects, relative to the suppressed category of public order violations (Visher, 1983:21). Does this suggest that UCR data underestimate female involvement in violence?

The available evidence does not show police discrimination in favor of violent female offenders. First, Visher's rotation of the

suppressed offense category revealed that, regardless of gender, the probability of arrest is not substantially affected by whether the offender was suspected of a violent or a property offense. Females and males suspected of property offenses are arrested as often as those suspected of violent offenses. Second, and of greater import, is the homogeneity for offenses included in the analyses: most of the cases recorded in these data as ''violent" were domestic assaults or acquaintance fights-not murders, rapes, or robberies. The probability of arrest for men and no arrest for women in such assaults, relative to public order violations, is probably a function of gender-based variations in the severity of the attacks. Lacking analyses for each offense category, with controls for the severity of the crime, such as weapon use and victim injury, we cannot conclude that women are underrepresented in violent crime statistics because of police bias.

A final analysis of this same data set by Smith (1987a) focused only on those cases that involved acts of violence in which the police had contact with both victim and offender (n = 102). Smith used a multinomial logit model to predict the odds of separation, mediation, or arrest based on a variety of legal (including injury and weapon use) and nonlegal (including race and gender of the disputing parties) factors. Although the analysis did not include female offenders, the results point to police bias in cases involving female and nonwhite victims. Violence between males or whites was most likely to be resolved by arrest, whereas violence between a male and a female or two nonwhites was most likely to be handled by separation (Smith, 1987a:776). Again, these results do not pertain to a full range of violent crimes or gender combinations of victims and offenders. Nevertheless, they do imply that the accuracy of violent crime data may depend not only on the severity of the criminal acts but also on the victim's attributes (Smith, 1987a:779-780).

A last look at a wider range of violent crimes can be obtained by referring back to Hindelang's (1979) data. Comparing official arrest statistics, disaggregated by gender, and victimization reports of the perceived sex of the offender for the offenses of rape, robbery, and assault (aggravated and simple) revealed comparable findings: by both criteria, males substantially outnumber females. Notably, in the cases of robbery and aggravated assault, Hindelang also found an overrepresentation of females in arrest statistics by comparison to victimization reports.¹²

Gender differences are uniformly smaller in self-report data of delinquency than in official records. Hindelang and his colleagues (1979), however, maintain that the differences are largely due to discrepancies in content and seriousness between the two instruments. Those items in self-report studies that are even roughly comparable to UCR items (e.g., beat up/assault) produce most similar results. More recently, Bridges and Weis (1989), regressing the magnitude of the gender-violence association on design characteristics of 115 studies, found that sources (e.g., official versus self-report data) and domain (types of crime included) do not create major discrepancies in gender-violence research, in part because of the relative accuracy with which violent crime is measured. However, because the correlations based on social aggregates also produced much stronger associations than those based on analyses of individuals, Bridges and Weis (1989:29-30) recommend disaggregating data to obtain the most precise information on the individual characteristics associated with violent offending. As such, an important subsequent step in research pertaining to gender and interpersonal violence may be focusing on those microsocial contexts (e.g., youth gangs, domestic setting) in which violence is prevalent.

Although differences in item content may explain differences in the gender ratio for crime at one point in time, Smith and Visher (1980:698-699) have suggested that other factors may be more important for explaining why self-report data show a greater convergence in the gender gap than official data. Specifically, they suggest either that the gender convergence is limited to only the less serious criminal acts or that official reactions to changes in female crime are occurring at a slower rate than the actual behavior. For several reasons, we feel that the former explanation seems most plausible. First, recall that a closing of the gender gap, regardless of data type, appears to be least evident in the case of violent crimes (Smith and Visher, 1980:696-697). Second, and relatedly, there is strong evidence that self-report measures are most valid when they are applied to the less serious delinquent populations (Hindelang et al., 1981). Third, and finally, analyses of short-term trends in crime and delinquency suggest that rates of discovery of crime by police and arrest rates are both increasing over time (Menard, 1987); we have no evidence that this increased efficiency on the part of the police systematically excludes female offenders (Steffensmeier and Streifel, 1989).

The preceding analysis has put together pieces of relevant information

on gender and violence from various articles that speak to the issue of the validity of our measures of crime. It is clear that a more adequate assessment would involve (1) progressively examining the factors that influence victim reporting, police response, and offender accountability solely as they pertain to males and females who commit crimes of violence; and (2) constructing adjustment factors to correct for any gender biases found in these measures. At this point, we can only suggest that the biases existing in official statistics either are nonspecific to gender (in the known underreporting of offenses involving nonstrangers) or, in the case of assaults, have a counterbalancing effect on the probability of males being arrested (male offenders with male victims are more likely to be arrested than male offenders with female victims; Smith, 1987a).¹³ The important question before us then is why women are so much less likely than men to be involved in acts of interpersonal violence.

It is now well known that research and policy directed toward female offenders lags far behind that on male offenders (Klein, 1973; Smart, 1976; Datesman and Scarpitti, 1980; Leonard, 1982; Kruttschnitt and Johnson, 1984; Naffine, 1987). Nowhere is this more obvious than in the area of interpersonal violence. Consideration of the substantial gender gap in violent crime has arisen primarily in the context of refuting the hypothesized relationship between the women's movement and changes in female arrest rates for specific property offenses (e.g., see Simon, 1975; Chapman, 1980; Steffensmeier, 1980, 1982). Not surprisingly, sociological and criminological theories offer virtually no explanation for the relative lack of female involvement in crimes of violence (see Siann, 1985; Simon and Baxter, 1989).

Perhaps the largest body of empirical research relevant to this question comes from the work in experimental psychology on aggression. The relevance of examining gender differences in aggression for criminal violence pertains to both the noted association between aggression in youth and adult criminal violence (Robins,

1966; Farrington, 1978; Magnusson et al., 1983) and the etiological issue of the relative importance of biological or cultural/environmental determinants of violent behavior (Wilson and Herrnstein, 1985:117).

Maccoby and Jacklin (1974, 1980) provided compelling evidence for gender differences in aggression in their review of cross-cultural studies of aggression, studies of subhuman primates, experimental manipulation of sex hormones, and studies of "presocialized" children. These data led them to conclude that males may be biologically predisposed toward aggressive behavior; they did not, however, discount environmental influences. Subsequent reviews and critiques of Maccoby and Jacklin have argued that (1) cross-cultural studies of children (which speak more directly to the etiology of aggression than studies of adults) tend to demonstrate the interplay of social environmental factors in children's behavior (Tieger, 1980:945; White, 1983:6-7); (2) there is a great deal of species-specific variability in the degree of sex differentiation in aggression (Tieger, 1980:945-947; White, 1983:5); (3) the organizing and activating functions of hormones in aggression among humans remains unclear (Tieger, 1980:951; White, 1983:6-7; Siann, 1985:33-36; Widom and Ames, 1988:316-322); and (4) although observational research on aggression among young children provides some support for a biological basis for sex differences, developmental research suggests that parental and environmental factors may be equally important determinants of gender differences in aggression (Tieger, 1980:952-959; White, 1983:8-9). Thus, although there may be some evidence for sex differences in aggression, the most challenging issues appear in the area of understanding how social, situational, or cultural factors ameliorate or aggravate aggressive or violent behavior patterns in males and females (see also Brain in Volume 2).

Fagot and her colleagues address some of the questions raised by Maccoby and Jacklin concerning the development of aggression in "presocialized" children. In several studies involving both home and peer play group settings, Fagot observed white children, primarily from intact families, ranging in age from 12 to 48 months. Beginning with the youngest group (12 to 16 months old), Fagot et al. (1985) found no sex differences in aggression (e.g., hit, push, kick, take objects), but they did find that caregivers reacted very differently to the aggression of boys and girls. When girls aggressed,

their behavior was ignored 80 percent of the time, whereas aggression by boys was attended to 80 percent of the time. By the time these infants were toddlers, there were marked differences in aggression, with boys being consistently more aggressive than girls. Toddler peer response to the aggression of boys and girls also mimicked that of their caregivers; girls' aggression was ignored more than boys', and boys responded more to the acts of other boys than to the acts of girls. Girls generally showed less gender-based differentiation toward acts of aggression (Fagot and Hagan, 1985). Finally, Fagot and her colleagues (1986) also discovered a significant relationship between the acquisition of gender labels and social behavior. Children who understood the boy-girl distinction (usually acquired by 30 months of age) showed gender-typical differences in aggression, whereas those who did not understand it failed to display such differences. As Fagot et al. (1988:92) conclude, these results suggest that the male's use of aggression is built into the child's construction of his gender schema, whereas aggression is not a part of the girl's gender schema; in fact, the typical female gender schema may include avoidance of the expression of aggression.

Extending this work on the determinants of aggression among males and females to the work that has been done on sibling aggression and generally coercive families, Fagot et al. (1988) attempt to explain the developmental determinants of male aggression against females. Specifically, they argue that three conditions must be met to produce a child who will be a consistent aggressor against females. First, the family must be out of control: the parents lack discipline skills, fail to monitor the child, fail to effectively problem solve, and fail to provide positive role models. Second, the child must have female models on which to practice (mothers or siblings), with no effective punishment of the boy's aggressive behavior toward the female. Third, there must be a system of family values that devalues females, so that they are perceived as appropriate targets of male violence (Fagot et al., 1988:103). Partial support for this model can be found in Felson and Russo's (1988) study of sibling aggression (hitting and slapping) among fourth through seventh graders. Here, no gender differences emerged in levels of aggression among siblings, but differential reinforcement appeared for acts of aggression among boys and girls. Parents were particularly likely to punish boys when they fought with their sisters, and subsequent aggression was more frequent when boys were punished than when some other strategy was used.¹⁴

Other scholars have focused on the particular contexts in which boys and girls are most likely to aggress. Caplan (1975), for example, argues that (1) the socialization of boys and girls shapes their responses to tests and experimental situations; (2) the need for achievement may be greater for boys and the need for approval greater for girls; and (3) the previous findings of no gender differences, or elevated levels among boys, in aggression can be explained by facets of the study design that elicit these different needs. Caplan's review of the relevant research does reveal covariation between characteristics of the experiment and the presence or absence of gender differences in aggression. Tasks that result in failure experiences (arousing the need for achievement) and those that involve an adult (arousing the need for approval) are more likely to demonstrate that boys are more aggressive than girls; in contrast, success or no-failure experiences and the absence of an adult are conducive to finding no gender differences in aggression. Barrett's (1979) subsequent observational study of 5- to 8-year-old boys and girls at a summer day camp over a six-week period provides comparable results with regard to the effect of an adult's presence. His data also further our understanding of contextual effects by suggesting that the sex of the target, the degree of structure in a setting, and the type of aggression measured affect the magnitude and direction of gender differences in aggression. Specifically, boys appear to be both more physically and more verbally aggressive than girls only when the target is male and the children are engaged in a moderately structured activity (see White, 1983:13).

Finally, numerous studies have identified parental correlates of aggression and antisocial behavior in middle childhood and adolescence. A meta-analysis of longitudinal and cross-sectional research revealed that the most powerful predictors of juvenile conduct problems, including delinquency and aggression, are lack of parental supervision, parental rejection, and low parent-child involvement; to a lesser extent, background variables, such as parents' marital relations and parental criminality, also predict conduct problems (Loeber and Stouthamer-Loeber, 1986). However, two caveats should be noted. First, although it appears that these findings pertain equally to boys and girls, far fewer analyses that relate parental behavior to delinquency and aggression of boys and girls could be found. Second, these findings are based primarily on samples of white youth (Loeber and Stouthamer-Loeber, 1986:316, 321; see also Lewis et al., 1983; VanVoorhis et al., 1988; Farrington, 1989). The importance of disaggregating these data by

race and gender can be seen at least preliminarily in studies of the developmental consequences of family violence. Abused and neglected males, and nonwhites, seem to be more likely than their female, and white, counterparts to subsequently engage in violent criminal behavior (Kruttschnitt et al., 1986; Rivera and Widom, 1989; Widom, 1989a).

Thus, although observational studies of aggression in very young children can provide compelling evidence for biologically based sex differences, the developmental perspective suggests that the environment, from birth on, also has a strong influence in shaping the nature, extent, and target of aggression among males and females. We cannot ignore the importance of parental variables, both in the microsocial world of preschoolers and in the emergent social life of adolescents. Regardless of a child's inborn or early acquired disposition, parental behaviors appear to have a strong mediating effect (Reid and Patterson, 1989:116). Clearly, however, we need to expand beyond samples of white youth to explore whether, and how, these socialization experiences reappear among nonwhite families.

By comparison to the studies of aggression in children, studies of gender differences in adult aggression reveal much smaller and less consistent findings (Frodi et al., 1977; Eagly and Steffen, 1986). The most robust findings suggest that (1) men aggress more than women when they have the opportunity to aggress physically rather than psychologically; and (2) gender differences are larger when women perceive that aggression produces harm to others or anxiety, guilt, or danger for themselves (Eagly and Steffen, 1986). As was true of the studies of gender differences in aggression among children, these results point to the importance of contextual variables in the magnitude of the noted association. However, there are also design factors that speak to the external validity of the study. As Eagly and Steffen (1986:325) point out, the best-known and most popular methods for studying human aggression (laboratory, Buss-paradigm) happen to be the ones that elicit greater aggression in men than women and greater aggression toward men than women. Unfortunately, the most popular subjects for these studies also tend to be white male college students, and the types of aggression observed (delivering electric shocks or honking horns) are not the most common forms found in daily life (Macaulay, 1985). However, if we assume that the variation in the magnitude

of gender differences in aggression between children and adults results from social roles that channel and regulate aggression, we would do well to concentrate on those environments in which aggression appears (Eagly and Steffen, 1986:326; Macaulay, 1985).

A number of studies of college students' attitudes toward, and experience with, courtship and marital violence report that men with traditional attitudes toward the role of women are more violent than men with more liberal attitudes (Sigelman et al., 1984; Bernard et al., 1985; Finn 1986). These results concur with both experimental studies (White, 1983) and family violence research that finds higher rates of violence in marriages where husbands make the majority of the decisions as opposed to more egalitarian marriages (Frieze and Browne, 1989:191). The relationship between gender role attitudes and violence is less clear, however, for women. Whereas some field research suggests that women who hold traditional gender role attitudes are more likely to endorse marital violence (Finn, 1986) or less likely to report or act upon being assaulted (Bernard et al., 1985; Walker and Browne, 1985), laboratory studies suggest that traditional females are not always less aggressive than liberal females. A woman's gender role attitudes appear to interact with other factors (e.g., level of provocation, sex of instigator, presence of a supportive observer) in producing levels of aggression (White, 1983:14-15). Some support for the hypothesis that aggression levels among females are mediated by other factors also can be found in the work of Giordano (1978). Using self-reports of criminality from a study of high school students and a group of institutionalized delinquent girls, she found that the more delinquent, aggressive girls were receiving reference group support for their nontraditional behavior (see also Marsh and Paton, 1986). More generally, however, self-report studies either find no association between females' gender role attitudes or gender identity and their reported level of involvement in antisocial and criminally violent behavior (Widom, 1978; Norland et al., 1981; Grasmick et al., 1984) or find that females with the most liberal attitudes are the least likely to be involved in aggressive acts of delinquency (James and Thornton, 1980).

Finally, stemming from the hypothesized relationship between the women's movement and increases in female arrest rates, researchers have endeavored to determine whether trends toward gender equality have affected female involvement in crime. Perhaps stated most strongly by Adler (1975:3), "The movement for full equality has a darker side which has been slighted even by the scientific community .... In the same way that women are

demanding equal opportunity in the field of legitimate endeavor, a similar number of determined women are forcing their way into the world of major crimes." Initial examination of this hypothesis with UCR violent crime data suggested that female involvement in criminal violence remains unchanged, despite claims of major shifts in women's involvement in all types of crime (Steffensmeier et al., 1979). More sophisticated analyses of this question have involved regressing the FP/A on various indicators of liberation (e.g., percent of women in labor force relative to percent of men, proportion of female college students, fertility rate) for the years 1960-1985. These analyses generally indicate weak support for the liberation hypothesis. Both female labor force participation and the female-to-male ratio of college enrollments were negatively associated with the FP/A for homicide and assault (Steffensmeier and Streifel, 1989; see also Steffensmeier and Allan, 1988). Finally, INTERPOL data, in combination with various measures of female social role participation (e.g., marriage and fertility rate, female labor force participation, percentage of female university students), have also been used to assess this hypothesis. Although the violent crime analyses are limited to homicides, there is no consistent evidence that the equalization of gender roles, which may accompany social and economic development, increases female crime (cf. Hartnagel, 1982; Widom and Stewart, 1986; Steffensmeier et al., 1989). However, as Widom and Stewart (1986) argue, it may be that gender equality or emancipation is a complex construct and that each of its components (social, political, and biological) affects female crime differently.

A comparison of the results of experimental studies of aggression in children and adults suggests that socialization plays a large part in determining gender-related rates and patterns of aggression. However, at this time, we cannot conclude that gender role socialization is a sufficient explanation for the gender gap in violence. On the one hand, if it were, we should observe some convergence in the violent crime rates of men and women who express/engage in nontraditional gender roles. On the other hand, it may be that particular aspects of gender role socialization, such as the repression of aggression in girls or the child rearing functions performed predominantly by mothers, are maintained across generations despite the adoption of nontraditional roles or attitudes. Still others may argue that the fundamental problem lies in the measurement of gender roles. Although there can be no doubt that gender roles can constrain social behavior, research designs influence our observation and understanding of this process.

Use of the term, and measurement of the concept, "gender roles" have been less than theoretically precise, encompassing types of activities and attitudes (job preferences/roles, profeminist attitudes), gender identity (masculinity and femininity), and gender orientation (expressive and relational). However, there does not appear to be any consensus as to which of these dimensions is most important or which measure is most valid (e.g., see Norland and Shover, 1978; Spence and Sawin, 1985; Gill et al., 1987; Naffine, 1987).

Opportunity Theory and Female Violence In addition to theories involving women's liberation, opportunity theory (Cloward and Ohlin, 1960) has also been invoked to explain changes in female arrest rates. In its most frequent application, it is hypothesized that changes in female labor force participation will create new opportunities for women to commit crimes (e.g. white-collar offenses, embezzlement; see also Steffensmeier, 1980). With regard to violent crime, Simon (1975:2) suggested that as women's employment and educational opportunities expand, their feelings of being victimized and exploited will decrease, and their motivation to kill will become muted. The evidence to support this hypothesis is mixed. Cross-cultural (INTERPOL) data show almost no relationship between women's educational and economic opportunities and female homicide rates (Widom and Stewart, 1986; Simon and Baxter, 1989; Steffensmeier et al., 1989).

However, as we have seen, recent analyses of UCR data do show that decreases in some violent crimes are associated with increased female labor force participation. Steffensmeier and Streifel (1989:24-25) suggest that the latter finding not only supports Simon's original hypothesis but also supplements it in (1) the aspect of family violence theory that points to economic insecurity and financial worry as a proximate cause of domestic violence and (2) the routine activities framework, which would posit that women outside of the home have less opportunity for engaging in violent encounters with children and spouses (see also Steffensmeier and Allan, 1988).

Labeling Theory Changes in the social status of, and opportunities available for, women can affect violent crime rates by affecting the behavior of deviance-processing agents. As previously

noted, official measures of crime are a product of the act of defining and responding to behavior as illegal. Viewed this way, the gender gap in violent crime is a function of differential willingness to report and sanction violent females.

Harris (1977) directed attention to social ''typescripts," or the everyday assumptions we hold about who does what, including deviance, in society to explain gender-based variations in crime. Specifically, Harris argues that the distribution of these typescripts may provide the single strongest causal account of the empirical differences in male and female criminality. With regard to the underrepresentation of women in violent crime, Harris (1977:12, 14) notes that these typescripts are so deeply embedded in society that they impact both motivations for deviance and rates of processing offenders: "It is (still) type-scripted that it is unlikely or 'impossible' for women to attempt assassination, robbery or rape." A comparable set of hypotheses was also developed by Widom (1984). She argued that much of what we know about criminal behavior and psychopathology is influenced by gender role stereotypes because (1) individuals comply with appropriate gender role behavior and (2) expectations about appropriate gender role behavior influence the labeling and diagnosis of deviant acts. She also uses these hypotheses to explain the relatively greater involvement of women in certain types of crimes (prostitution, child abuse, shoplifting) than others (rape, robbery, pedophilia) and the noted associations between certain diagnostic categories and gender role stereotypic behavior (see also Schur, 1984).

As we have seen, it is difficult to confirm empirically that gender role stereotypes provide the motivation for specific kinds of delinquency or crime. However, there is some evidence that violent behavior may be a more common adaptation to stress in males than it is in females. For example, Horwitz and White (1987) found that gender identity was associated with different types and rates of pathology. Among a large sample of adolescents, females displayed the greatest amount of psychological distress whereas males showed more self-reported delinquency (including crimes of violence). Further, over the course of adolescence, masculine gender identity was increasingly associated with high rates of delinquency for males and low rates of psychological distress among both males and females. Widom's (1989a) findings from an analysis of the cycle of violence hypothesis also speak to the question of gender roles and styles of pathology. Specifically, she found that although abused and neglected males had substantially higher arrest rates for violent crimes than a matched group

of controls, the same pattern did not appear for females. Instead, abused and neglected females were at increased risk for property, drug, and order offenses. However, it is unclear whether these different developmental patterns are due to an internalization of stereotypes about gender-appropriate behaviors or simply to the difference in types of stress/abuse suffered by males and females.

Research pertaining to the hypothesized influence of gender role stereotypes on the labeling and diagnosis of deviant acts is most pronounced at the latter stages of the criminal justice system and rarely focuses on violent offenders (e.g., see Kruttschnitt and Johnson, 1984; Erez, 1989). However, some evidence concerning changes in police behavior that may affect the probability of female arrests for minor acts of violence can be found in the recent work of Steffensmeier and Streifel (1989). Specifically, Steffensmeier and Streifel found that, for the years 1960-1985, indicators of the formalization of social control/social labeling (i.e., annual changes in police per capita and the percentage of civilians among law enforcement personnel) were significantly associated with female arrests for property offenses and other assaults but not for the more serious violent crimes (aggravated assault and homicide). It should be clear, however, that these data do not provide measures of either police officers' gender role stereotypes or their attitudes toward female offenders; instead they measure changes in police officers' capacity to make arrests (see also Steffensmeier and Allan, 1988, for a similar hypothesis concerning the EP/A for assault among black females).

In summary, the most prominent explanatory variables in current research on gender and interpersonal violence are gender role socialization and, to a lesser extent, gender identity. Gender role socialization is used to account for observed differences in the aggression of infants and toddlers, the incongruent results of gender differences in laboratory studies of aggression in children and adults, and national and cross-cultural data on violent crime among females. However, the data are not always congruent with the hypothesized effects of gender roles on crime. We have little evidence that objective indicators of changes in gender roles (such as the female percentage of the labor force) correlate with changes in levels of interpersonal violence, and we have no evidence that gender role identity or gender role attitudes are associated with violent crime for females. Thus, as a number of feminists argue, we may need to reexamine this explanatory framework and restructure it with both inductive and deductive data.

Although it is important to acknowledge that there are numerous feminist perspectives in criminology, a dominant theme emerging from them is that the study of crime has been a male phenomenon (Daly and Chesney-Lind, 1988; Simpson, 1989). As a result, it is argued, explanations for female crime are really explanations of male criminality (Klein, 1973; Smart, 1976; Leonard, 1982; Naffine, 1987). Role theory was acknowledged as one of the first theories that explained female criminality in other than psychological or biological terms, and that addressed the significance of such factors as differential socialization, differential opportunity, and differential social reaction in understanding female crime (Smart, 1976:66). However, despite its contributions, role theory is criticized both for failing to take account of the historical and political derivation of gender roles (Smart, 1976; Leonard, 1982) and for employing the culturally variable and empirically unreliable concepts of masculinity and femininity (Carlen, 1985; Naffine, 1987). Further, its clear application to questions such as, Why are women less likely than men to be involved in crime? and What explains gender differences in rates of arrest and types of criminal activity? places it squarely in what is termed the "gender ratio" paradigm.

Daly and Chesney-Lind (1988:515-516) argue that theories of gender and crime can be built in several ways. The most common approaches focus on either the generalizability problem or the gender ratio problem. The former concerns the degree to which theories of men's crime apply to women, and the latter focuses on what explains gender differences in rates and types of criminal activity. Work on the gender ratio problem, they believe, is dominated by male scholars who rely on statistical evidence and elements of existing theory to develop new theoretical formulations that, in reality, are only variations of old theories of male criminality (e.g, social control and conflict theory as recently applied to the issue of gender and delinquency by Hagan et al., 1987). More generally, a number of feminist scholars believe that a more fruitful approach to answering questions about both inter-and intragender variability in crime is to give equal weight to different methodological approaches to studying crime. Our understanding of the causes of crime among both men and women could be improved by emphasizing qualitative, historical, and subjectivist approaches in addition to the dominant, quantitative, research paradigm (Daly and Chesney-Lind, 1988; Simpson, 1989; for similar arguments pertaining to research in other fields, see Macaulay,

1985; Harding, 1986). Such approaches might provide more accurate portrayals of how women become involved in crime, what keeps them involved, and what factors influence their desistance. None of these scholars have formulated their own theory to explain the role of gender in crime or in violent crime; nevertheless, their work is significant in pointing out how the reliance on one type of methodology or style of work can limit our understanding of whether, and how, the processes that lead to violence differ for men and women and, within gender, for whites and nonwhites (Simpson, 1989).

Research on the demographic and social correlates of violent crime suggests considerable gender variation. However, the quantity of the research is limited and the quality is varied (e.g, based on institutional populations, uneven coverage of all types of violent crime and, within crime categories, of the demographic correlates). The largest body of data comes from studies of homicide. As previously noted, these data indicate that black women are more often involved in homicides (as both victims and offenders) than white men or women; regardless of gender, this offense appears to be primarily intraracial. However, although the data on gender and race also suggest that women's homicides are more often intersexual and intrafamilial than men's (Benedek, 1982; Mann, 1987; Riedel, 1988a), there is additional evidence that black women tend to kill friends, acquaintances, and other females more often than do their white counterparts. Women appear to be infrequently involved in felony homicides (Mann, 1987:182).

It is much more difficult to obtain a portrait of women involved in other types of serious violent crime. Analysis of NCS data, as well as probation and prison data, pertaining to females involved in assaultive offenses reaffirms at least one of the patterns found in homicide data: the victims of these women are frequently known to or involved with them (Ward et al., 1969; Kruttschnitt, 1985a), especially when the women are acting as lone offenders (Young, 1979). Parisi (1982) examined NCS data and PROMIS (Prosecutor Information System) data from Washington, D.C. for 1974 and 1975 and found that, overall, females were more likely to commit crimes alone, rather than with an accomplice, and were less likely to use lethal weapons than males. Unfortunately, however, she did not specify the types of offenses included in these data. The study by Ward and his colleagues

(1969) of female inmates in California indicated that women often commit robberies with a partner, and only infrequently with a weapon, but act alone in the commission of assaults and homicides where lethal force is much more likely to occur. With regard to robbery, some comparable findings appeared in a recent study of 33 female robbers incarcerated in Florida: only one-third of these women committed their offense alone, but most used a firearm (Fortune et al., 1980).

Certainly more exploration is needed about the patterns and correlates of female violence. Further analyses might address some of the following omissions in our knowledge of violent female offenders: (1) Do the race and gender patterns observed for homicide (i.e., intraracial and, generally, intersexual) reoccur for robbery and assault? (2) If a women is more likely to commit a robbery (as opposed to homicide or assault) with an accomplice, what is her relationship to the accomplice? (3) Are violent women offenders more or less likely than their male counterparts to have a substance abuse problem or to use drugs or alcohol before committing a crime, or does substance use vary with the type of crime (e.g., is it more likely to occur in domestic assaults and relational homicides than in robberies)? (4) Do female career criminals specialize in certain types of offending? (5) Do the women who commit felony homicides and aggravated assaults against strangers present a very different social and demographic profile from those who commit domestic homicides and nonstranger or relational assaults?

Because a sizable proportion of women's violence appears in encounters with nonstrangers (Ward et al., 1969; Mann, 1987; Riedel, 1988a), special attention should be directed toward whether, and how, women who engage in relational violence differ from those who direct their violence toward a stranger. As Fagan and Wexler (1987:7) point out, incorporating family violence into the study of criminal violence may enhance the explanatory power of existing theory derived from the separate disciplines. Mann's (1988) analysis of women who kill in domestic and nondomestic encounters in six U.S. cities provides one significant attempt to bridge this gap. Although many of her findings confirm previous research (e.g, homicides committed by women are generally intersexual, intraracial, and intrafamilial, and offenders are usually older members of low-income minority groups), others suggest that this type of violence may not be so much a product of self-defense or a violent intimate relationship as it is a continuation of a violent background. Women involved in domestic homicides (58.3%) were just as likely

to premeditate their crimes as women involved in nondomestic homicides (58%). Further, the domestic homicide offenders who claimed self-defense had more prior arrests than the nondomestic offenders claiming self-defense, and there was little difference in violent offense histories between the two groups (30% of the domestic homicide offenders and 37.5 percent of the nondomestic homicide offenders). As is true of men, then, it may be that women who are extremely violent in the home are also the ones who are more likely to be violent outside the home; women's choice of victims may also not be as static as it is commonly believed to be (Fagan and Wexler, 1987:12; Hotaling et al., 1989).

We are now at a point where it is possible to offer some speculative propositions about gender and interpersonal violence. Three predominant patterns emerge from the data we have presented: (1) female participation in crimes of violence is significantly lower than that of males; (2) race may represent an important variation in the gender-violent crime paradigm since black females have higher rates of homicide, and possibly other violent crimes, than white males; and (3) by comparison to male violence, the target of female violence is more often relational. Because a good theory both parsimoniously explains known facts and accurately generates predictions, we begin our explanatory model with these three factors of participation, racial variation, and target. We then integrate psychological, social, and situational/contextual variables to suggest the following preliminary hypotheses.

First, from infancy forward, aggression in boys appears to be attended to, and reinforced, more than it is in girls. Although most boys gradually learn to repress their aggression, physical aggression is always in the male's background as a gender-acceptable behavior; girls typically drop aggression from their behavioral repertoire at a very early age (Fagot et al., 1988). Second, contributing to this initial difference in the place of aggression according to gender schema is the role of parenting in middle childhood and adolescence. A lack of supervision and parental involvement appears to predict antisocial and violent behavior among both boys and girls (Loeber and Stouthamer-Loeber, 1986). However, poor parenting may be especially detrimental for boys both because of their known propensity for being more vulnerable than girls to stressful life events (Hetherington, 1981; Rutter, 1982; Eme, 1984) and because of the traditional pattern of gender role

socialization in which the activities of boys are often monitored less carefully than those of girls (Hagan et al., 1987). Further, because there is also evidence that parental involvement and supervision vary by race as well as gender, with white girls being the most protected group (Matsueda and Heimer, 1987; Jensen and Thompson, 1990), we should expect to see higher rates of violence among nonwhite females than among white females. Finally, for the relatively few girls who do engage in violence, their greater propensity to select relational targets may be due to a combination of childhood exposure to family violence and specific contextual stressors such as current victimization, isolation, and economic stress (e.g., see Widom, 1987). Because it is now well known that prior exposure to family violence is not a sufficient explanation for engaging in subsequent familial violence (Widom, 1989b) and that the presence of a number of stressful events increases the probability of subsequent violence (Kruttschnitt et al., 1987), it seems most likely that women who engage in relational violence are those who are also currently experiencing family-related stressors.

In sketching these propositions, we do not mean to exclude other social and cultural factors (e.g., criminal opportunities, community norms toward crime and violence) that may either reinforce or counterbalance existing developmental patterns (e.g., see Fagan and Wexler, 1987; Fagot et al., 1988). In fact, our ability to extend a theory of gender and interpersonal violence into a broader explanation of the major variations in gender and crime will probably require the inclusion of these variables. Again, gender role socialization may be a necessary, but not sufficient, explanatory variable for both the absolute gender gap in crime and the greater concentration of females in selected, "gender-appropriate" offense categories. However, to explain the relative narrowing of the gender gap in larceny, fraud, and forgery (and perhaps robbery), we need to look more closely at the role of other social and cultural factors.

Steffensmeier and Streifel (1989:22) suggest, for example, that larger and more bureaucratized police forces result in greater diligence in counting and pursuing minor, instead of only serious, offenders. Because women commit relatively more minor than serious offenses, these larger and more formalized police departments have had a greater effect on female levels of reported crime. It is also possible, as they suggest (Steffensmeier and Streifel, 1989:23,35-36), that the growth of consumer products, illicit drug use, and female economic marginality has created more opportunities

and more motivation for female involvement in crime (see also Fagan, 1990).

Drug use may be particularly important in explaining observed increases in female property offenses. The use of addictive drugs, such as cocaine and heroin, appears to be at least as prevalent, if not more so, among female as among male arrestees (Wish and Gropper, 1990:371). Further, although the association between drug use and crime is strongest for high-frequency drug users regardless of their gender, the criminal activities in which these users engage, is not gender neutral. Females who use drugs frequently are less likely than their male counterparts to engage in predatory crimes; instead, they resort to prostitution, shoplifting, and other covert, nonviolent, crimes at high rates (Chaiken and Chaiken, 1990:212-213). Increasing rates of female drug addiction then could well be contributing to the rise in female property crimes (see also Anglin et al., 1987). Relatedly, the rise in female-headed households and the attending increases in poverty (Center for the Study of Social Policy, 1986) may motivate more women to commit property crimes, but the evidence for this association has yet to be established. Multivariate analyses of the effect of female economic marginality on property crime are rare and do not provide particularly strong support for this hypothesis (cf. Chapman, 1980; Steffensmeier, 1980; Giordano et al., 1981; and Steffensmeier and Streifel, 1989). More generally, however, an assessment of the full range of individual and situational conditions (e.g., frequency of drug use, peer group influences, and expanded criminal opportunities) that could impact the levels of female property crime, including economic marginality, has yet to be undertaken.

Thus, at this time, we know too little about the factors that elicit and sustain female crime and about the range, or degree of specialization, of women's criminality to offer more than this speculative explanation. At the very least, however, the data we have reviewed do suggest that any theory that fails to consider past and current differences in socialization and opportunities for women and men will be incomplete (see also Heidensohn, 1985).

Research on patterns and trends in violent offending by females has generally proceeded as if it had no relevance to sanctions and sentencing. Deterrence research, for example, virtually ignores potential gender variations in analyzing both the deterrent

influence of threats of criminal sanctions and the effectiveness of the criminal justice system's activities in controlling crime rates (e.g., see Cook, 1980; Simpson, 1989:620-621). Although in large part this omission can be attributed to the infrequent involvement of women in violent crime, two important questions remain: do gender-based differences in criminal court processing exist and, if so, are they related to subsequent levels of female involvement in crimes of violence? In this section we are particularly concerned with the ways in which the processing of women through the criminal justice system could either aggravate or mitigate their subsequent involvement in acts of interpersonal violence.¹⁵ We review what little literature there is on sanctioning of both male and female violent offenders; what, if any, conclusions can be drawn; and what questions remain. We then suggest additional areas of inquiry that could affect criminal justice policy pertaining to interpersonal violence.

The vast majority of both quantitative and qualitative analyses of gender variations in the treatment of juveniles addresses the issue of differential selection and processing of male and female status offenders (see Chesney-Lind, 1973, 1974, 1977, 1988; Boisvert and Wells, 1980; Shelden, 1981). Among the few studies that include a broader array of juvenile offenders, no clear association among gender, offense type, and severity of disposition emerges. The relationships among these variables appear to vary by study and stage of criminal court processing. For example, in deciding who will be formally adjudicated in juvenile court, Cohen and Kluegal (1979a) find that violent offenders, regardless of gender, are less likely to be treated informally than juveniles charged with other types of offenses. However, in the case of pretrial detention, violent female offenders are treated more leniently than their male counterparts (Cohen and Kluegal, 1979b). Similar results were reported for the decision to incarcerate by Staples (1984). Nevertheless, the contrary conclusion-that females are treated more severely than males in these latter two stages of criminal

justice processing-was supported by evidence from a different jurisdiction (Feld, 1989).

One study suggests that, once the decision to incarcerate has been made, violent females are much less likely than violent males to be sent to a correctional school. Lewis and her colleagues (1981) compared the psychiatric symptoms, violent behaviors, and medical histories of an entire one-year sample of adolescents from the same community who were sent either to the correctional school or to the only state hospital adolescent psychiatric unit serving the area. They found that although the degree of violence and psychiatric symptomatology were not associated with placement, gender was. Aggressive behaviors in adolescent boys were treated as the deliberate acts of healthy youngsters whereas the same acts were considered as psychologically aberrant when performed by girls (Lewis et al., 1981:518). Certainly, replicating this study in other jurisdictions and following these youngsters forward to obtain criminal career data would provide much-needed information on the consequences of mental health sanctions, as opposed to penal sanctions, for youthful violent offenders, as well as the potential interactions among gender, treatment mode, and recidivism.

As with juveniles, few analyses of the effects of gender on adult criminal justice processing decisions focus on violent crime. Accordingly, the following review relies heavily on studies including felony offenses and either controlling for or analyzing offense severity in interaction with the defendant's gender.¹⁶

Information on the relationship between gender and various presentencing adjudicatory decisions (e.g., the decision to prosecute, pretrial release, charge bargaining, and severity of final conviction) is both sparse and inconsistent. Women may actually be treated more severely than men in the decision to prosecute and plea negotiations (Figueira-McDonough, 1985; Ghali and Chesney-Lind, 1986) but less severely in the pretrial release and final conviction decisions (Stryker et al., 1983; Farrell and Swigert, 1986; Wilbanks, 1986).

However, analyses of sentence severity, the most studied aspect of adjudication, repeatedly find that women are less likely to be recipients of prison sentences, and receive shorter prison sentences when incarcerated, than men (Frazier and Bock, 1982; Curran, 1983; Zingraff and Thomson, 1985; Spohn and Welch, 1987). Additionally,

although attention has been directed to the interaction of race and gender in sentencing outcomes, there are too few studies-and the outcomes of these studies are too disparate-to draw any firm conclusions (see Daly, 1989b).

The noted gender differences in sentence severity can be explained in two ways. One explanation pertains to the methodological variations and omissions in this research. For example, does either a failure to consider qualitative differences in the violence of men and women, or a failure to model the outcome of prior deviance processing decisions, produce leniency toward females (Kruttschnitt and Green, 1984; Zatz and Hagan, 1985; Miethe, 1987)? The other explanation pertains to the gender role attitudes and stereotypes of the judiciary, as encompassed by the chivalry and paternalism these (see Nagel and Hagan, 1982, for a review of these theories; and Kritzer and Uhlman, 1977; Gruhl et al., 1981; Kruttschnitt, 1985b, for analyses of the effects of the gender of deviance-processing agents). Most recently, the paternalism thesis has been extended/reformulated to include a specific discussion of the way in which family status interacts with gender to produce leniency for female offenders (Daly, 1987, 1989a,b). Daly's (1989a:15) analysis provides some evidence that family ties have a stronger mitigating effect on the sanctions accorded women, regardless of offense severity, than men (see also Kruttschnitt, 1984). For our purposes, however, the important question is whether the apparent leniency extended to female offenders subsequently affects their probability of recidivating.

Probation If women are more likely to be given probation than men, are they also more likely to complete probation successfully? Norland and Mann (1984), in perhaps the only study focusing on this question, examined rates of probation violations for both men and women. The authors found that men were charged with committing new crimes while on probation in much greater proportion than women. New charges for women were disproportionately for technical violations of probation. In fact, probation officers regarded these women as troublesome not because of their criminality but because they made time-consuming demands that tended to be organizationally disruptive.

Prison Over the past decade the issue of inmate violence and victimization in institutions for adult males has dominated the

field of penology. The vast majority of research on female inmates, however, focuses on (1) female patterns of adaptation to prison life, including the role of homosexuality and the development of pseudofamilies in this process; (2) the effects of incarceration on mothers and their children; and (3) historical accounts of the development of women's prisons (see Kruttschnitt and Krmpotich, 1990). Despite evidence of violence and degradation in women's institutions, empirical research documenting this phenomenon is scarce. What little we do know suggests that both the extent and the nature of aggression in women's prison facilities differ from the aggression and violence found in men's prison facilities. Although women may be punished and put ''on report" more frequently than men for offenses against prison discipline (Dobash et al., 1986; Mandaraka-Sheppard, 1986), their proportional involvement in prison violence appears to be considerably lower than men's. There is little evidence of either predatory sexual behavior or collective violence in women's prisons (Bowker, 1980; Kruttschnitt and Krmpotich, 1990).

The correlates of inmate violence also vary significantly by gender. Whereas a minority racial status, a conviction for violent offenses, a lengthy current sentence, or a previous history of incarceration is each associated with a higher rate of violence in male correctional facilities (Carroll, 1974; Ellis et al., 1974; Bennett, 1976; Sylvester et al., 1977; Bowker, 1980; Irwin, 1980), none of these variables seems to be associated with prison violence among females. Instead, age and family composition (both family of origin and procreative family) appear to be the most significant correlates of female prison violence. The most aggressive women are young, single, and childless and, more often than not, were raised in a traditional two-parent household (Mandaraka-Sheppard, 1986; Kruttschnitt and Krmpotich, 1990).

These conclusions about female inmate aggression should be treated cautiously. They are based largely on data obtained from only two studies, one conducted in the United States (Kruttschnitt and Krmpotich, 1990) and the other in Great Britain (Mandaraka-Sheppard, 1986). Nevertheless, the significant gender variations that emerge from them suggest that this area of research may improve our understanding of the processes that affect rates and trends of interpersonal violence by men and women. By replicating and extending these studies in other correctional facilities for women, questions such as the following, relevant to both theory and policy, could be addressed: (1) Has violence among female inmates increased over time? (2) Do institutional characteristics

predict rates of inmate violence, regardless of gender (e.g., cf. Ekland-Olson, 1986; Mandaraka-Sheppard, 1986)? (3) Upon release, do violent inmates' rates of offending vary by gender? (4) Do the effects of violent prison victimization differ by gender?

Parole Finally, we examine whether men and women released from prison on parole have different success rates. Unfortunately, there appears to be little information on gender variation in the parole performance of violent offenders. However, in one study of female parolees, violent offenders (assaultive women) were no more likely to violate parole than women convicted of other offenses (Spencer and Berocochea, 1972). Among the few studies that compare male and female parolees, regardless of the offense of conviction, there is evidence that women have a greater probability of parole success than men: women are less likely than men to be returned to prison either for technical violations or for new major convictions (Moseley and Gerould, 1975; Mowbray, 1982). Women with drug histories, however, may be an important exception to this pattern (Moseley and Gerould, 1975; Simon, 1979; Mowbray, 1982).

We began by asking whether violent female offenders are treated differently from their male counterparts in the criminal justice system and, if so, whether this affects their probability of engaging in subsequent violence. The extant research offers no definitive answer to this question. Although the available data suggest that, at least in the sentencing phase, women receive preferential treatment, there is no evidence that this leniency amplifies their violence. Both during and after periods of incarceration, women exhibit less violence and subsequent criminality than men. Is it because, relative to gender socialization and informal social control, criminal justice interventions have little influence on behavior? As Heidensohn (1985:198) suggests, women are the one segment of society whose policing has already been "privatized," even though public means of control are still employed. Or is it simply that our knowledge of the careers of violent female offenders is so limited that we are unable at this time to detect the influence of sanctions on women's behavior?

Although the leniency extended to women has been attributed primarily to their family status/ties, the interaction between gender and violent offense type, with family status being controlled for, needs to be assessed more carefully (cf. Bernstein et al., 1979). Certainly, a systematic examination of race and gender variations in these outcomes is warranted. The general omission of race in these analyses is surprising, given that at least half of all incarcerated

women are black and that the social locations of and pressures for criminal involvement may differ for black women as opposed to women of other cultures or races (Lewis, 1981; Young, 1986; Daly, 1989b).

Up to this point, the analyses of gender and interpersonal violence largely exclude the crime of rape. Because of the virtual dominance of male offenders, rape is not a useful offense for studying gender variations in violent offending, victimization, and adjudication. However, when we turn to the questions of whether and how the criminal justice system's treatment of violent crime victims impacts subsequent violence, rape-and more recently domestic violence (see Fagan and Browne, in this volume)-take center stage. Concern over the level of female violent crime victimizations has drawn increasing attention to the ways in which official treatment of female assault victims may encourage subsequent victimizations (e.g., see Brownmiller, 1975; Dobash and Dobash, 1979; Klein, 1982; Russell, 1984). Specifically, the diminution of rape victims, or of their criminal cases, may result in lowering the perceived costs of committing a rape. Although child sexual abuse cases are also relevant to our discussion, at this time, too few cases proceed to the point of prosecution to make a systematic assessment possible (Stewart, 1985). Thus, the following discussion focuses on adult victims of sexual assaults/rape.¹⁷

Reporting Simply stated, women are generally reluctant to report sexual assaults to the police. Factors that appear to affect the likelihood of reporting include peer support, a woman's conviction that a rape has occurred, and the evidentiary strength of the case. For example, Holmstrom and Burgess's (1978) study of 94 rape victims admitted to a city hospital revealed that only 23 percent of these women reported the rape on their own initiative; 10 percent of the cases were reported by a stranger; and in 38 percent a police report resulted from friends and family members prompting the victim to report the crime (see also Feldman-Summers and Ashworth, 1981). Empirical analyses of NCS data also suggest that evidentiary variables influence reporting decisions: a greater amount of property stolen, serious injury, an unknown offender, and a married victim all increased the likelihood that a

rape will be reported to the police (Lizotte, 1985). A study of women in rape crisis centers (admittedly not highly representative of the population) produced comparable results (Williams, 1984).

Finally, Russell's (1984) interviews with 930 randomly selected adult females in San Francisco showed that black victims and offenses involving black offenders were most likely to be reported. However, because the race of the offender was confounded with the relational distance between the victim and the offender (i.e., black rapists were overrepresented among stranger rapists), this latter finding pertaining to the offender's race must be treated with caution.

Police Processing Just as there is a significant amount of self-selection in the willingness of victims to invoke legal action in sexual assault cases, so also is there selection bias from the responses of the police to these victims. LaFree (1989) analyzed 904 sexual assault cases that occurred in 1970, 1973, and 1975 in Indianapolis, focusing on the participation of police in arrest, charge seriousness, and felony screening decisions (which were strongly influenced by the detectives). The arrest decision was significantly influenced by such evidentiary concerns as suspect identification, victim willingness to testify, prompt reporting, and presence of a weapon. The probability of arrest also increased when the offense involved an acquaintance, rather than a stranger. LaFree (1989:78) suggests that this may be due, as in domestic assaults, to the use of arrest by the police to protect the victim from further attack. Finally, consistent with feminist arguments, arrest was less likely when the victim engaged in "nonconformist" behaviors (e.g., hitchhiking, using alcohol at the time of the offense, willingly going with the suspect).

The same evidentiary factors (weapon, reporting promptness, and victim testimony) also predicted the more serious charges. Additionally, if (1) the offense included sexual penetration, (2) the victim was both white and older, and (3) the suspect was black, the case resulted in more serious charges. The best predictors of the decision to file a case as a felony were charge seriousness, the presence of multiple offenders, and victims over 18 years of age. This latter finding was attributed to the increased likelihood that incidents involving younger suspects included characteristics that usually make the case less serious to the detectives (e.g., "date rapes"; LaFree, 1989:79).

As in the reporting of sexual assaults, these data indicate that

there is no easy answer to the question of whether sexual assault cases are affected more by the behavior of the victim or by the behavior of the offender. Subsequent analysis of the processing of these cases through the court system, however, sheds further light on this issue and speaks directly to the role of victim credibility in the processing of sexual assault cases in criminal court (e.g., see Randall and Rose, 1981).

Criminal Court Processing Although few systematic analyses of the court processing of rape cases exist, the available data implicate victim characteristics and behaviors in both pretrial and final case dispositions. PROMIS data from the District of Columbia, for example, revealed that in forcible sex cases, victims who know the offender and who abuse alcohol are more likely to have their cases dropped by prosecutors at the initial screening (Williams, 1976). LaFree (1989:91-113) subsequently determined that once charges have been filed, legally relevant factors (e.g., weapon use, type, and seriousness of offense; victim's willingness to testify and to identify the suspect) and victim credibility both influence court decisions. Specifically, victims who exhibited "nonconforming" behaviors (e.g., drinking in a bar unescorted), who failed to report the assault promptly to the police, and who were black were less likely to obtain guilty verdicts or convictions of guilt (by trial or plea). It is important to note that these cases with black victims refer primarily to intraracial offenses because there were too few white offenders with black victims to be included in the analyses.

Feminist concerns about the selectivity with which rape cases are handled by the criminal justice system and the treatment accorded victims seem to be at least partially justified (cf. Myers and LaFree, 1982). The data suggest that both the legal seriousness of the case and factors pertaining to the victim's credibility/reputation influence a victim's willingness to report being assaulted, police response, and criminal prosecution of the case. It may not be surprising, then, that over the past two decades we have seen a concerted effort by feminists to change state laws that contribute to low conviction rates in rape cases.

Changes in Legal Procedures in Rape Cases Statutory changes in the processing of rape cases aim at reducing the role of victim attributes/credibility and increasing conviction rates. These include redefining rape in a broader offense category, relaxing proof requirements, and restricting testimony concerning the complainant's sexual history (Gates, 1978; Polk, 1985). Evaluations of the effects

of these statutory reforms on prosecution and conviction rates are, however, rare.

Some states, such as Michigan and Minnesota, placed rape in the broader category of assaults. California, on the other hand, adopted a multifaceted approach that included (1) providing enhancements for the use of weapons or violence, (2) including a rape shield law (to constrain the use of prior sexual history in proving consent), (3) establishing gender-neutral language in defining rape, and (4) removing the spousal exception in cases of rape (Polk, 1985:193).

Polk (1985) examined California's Bureau of Criminal Statistics data, for 1977-1982, to determine whether these statutory changes affected the processing (i.e., arrests, the filing of felony complaints, felony convictions, superior convictions, and state institutional sentences) of rape cases relative to other predatory crimes. The changes he found were not specific to rape and were limited to the tail end of the criminal justice system. There was no evidence that police were more likely to clear a reported incidence of rape with an arrest, and although prosecutors showed some tendency to file more rape cases as felonies, the felony conviction rates for rapes did not systematically increase over the years. Even at the sentencing level, the increase in incarceration rates was not unique to rape cases but instead appeared to be part of California's general shift toward increased penalties for all serious felonies. An early analysis of legal reform in the state of Washington also showed a "tail end" effect that was consistent with the penal philosophy of "treating" rather than incarcerating those convicted of sexual assaults (Loh, 1980).

Caringella-MacDonald (1985) examined similar data for Michigan five years after the enactment of the model Michigan rape law. She found that sexual assault cases were more likely to be authorized for prosecution than nonsexual assault cases, but it is unclear whether this is an artifact of offense severity. The nonsexual assault cases included simple as well as aggravated assaults and case characteristics were not controlled. Further, among the cases authorized for prosecution, victim credibility problems (e.g., inconsistent statements, implausible accounts) were noted in a significantly higher proportion of the criminal sexual assaults, perhaps because of the absence of other strong case characteristics (corroborative evidence, presence of weapon, stranger assailant; see also Reskin and Visher, 1986). At subsequent stages of processing, Caringella-MacDonald uncovered no significant differences in the proportion of cases terminated with conviction or plea bargained.

However, criminal sexual assaults were reduced to a greater extent than other assaults. The objective of treating criminal sexual assaults like nonsexual assaults did not then appear to have been fully realized in Michigan.

Indiana's attempts to improve rape conviction rates and lessen the importance of victim characteristics also seem somewhat unsuccessful. LaFree (1989) gathered extensive data (e.g., on victim, defendant, courtroom characteristics and testimony, interviews with jurors, and questionnaires from judges) from 38 trials three years after the passage of the Indiana rape shield law. Additionally, he developed operational definitions of traditional and nontraditional victim behavior; the latter included (1) alcohol or drug use in general and at the time of the incident, (2) extramarital sexual activity, (3) having illegitimate children, and (4) having a reputation as a "partyer."

First, LaFree (1989:203) found that the rape shield law was invoked in only one-third of the trials observed, but that in virtually all of these trials the women had allegedly engaged in nontraditional behavior (LaFree, 1989:203). Because the point of this law was to ensure that information about the victim's nontraditional behavior was not introduced in the courtroom and courtroom observers coded only the evidence of nontraditional behavior that was presented in the court while the jury was present, the rape shield law appeared to be totally ineffective.

Second, LaFree (1989:201-208) also examined the interrelationships among the victim's behavior, defense strategy, and verdict.¹⁸ This analysis revealed that allegations of nontraditional behavior were always made when the issue was whether the victim consented to the sexual act but never when it was the defendant's diminished responsibility . Further, those cases that used the consent defense were least likely to result in a guilty verdict, whereas those that employed diminished responsibility were most likely to obtain the guilty verdict.

Third, and finally, LaFree (1989:208-233) also analyzed what factors influenced jurors' perceptions of the defendant's guilt. The victim's behavior was important only when consent and no-sex defenses were mounted, and in these cases, it was more important than measures of physical evidence and seriousness of offense. Women who engaged in extramarital sex, used alcohol and drugs regularly, knew the assailant, or were black were "blamed" for the offense. Notably, LaFree's (1989:220) interviews with jurors led him to conclude that this latter finding is due to the stereotypes that white middle-class jurors hold about black women,

such as "they are more likely to consent to sex or they are more sexually experienced and hence less harmed by the assault."

Legal reforms aimed at increasing conviction rates and reducing the role of victim attributes in the processing of sexual assault cases have been, at best, only partially successful. Despite the passage of rape shield laws in more than 45 states and the elimination of corroboration of the victim's testimony as a prerequisite for conviction in most states, the victim's credibility remains a central issue in the processing of sexual assault cases. Attending to the victim's behavior can be understood, in part, when relevant evidentiary concerns (e.g., weapon use, injury, eyewitness) are missing from a case, but as we have seen, the victim's behavior can be weighted more heavily than the legal seriousness of the cases when a no-consent or no-sex defense is mounted. An important policy question arises here that should be considered a priority for future research: Do attorneys select their mode of defense based on the victim's credibility, or does the defense selected dictate whether the victim's behavior will be raised as an issue? An additional and related priority for future research pertains to the role of the victim's race in the processing of sexual assaults. The clear message from LaFree's (1989) findings is that only sexual assaults involving white victims are considered seriously by the criminal justice system. We need to determine whether the victim's race has a similar effect in other jurisdictions and in other violent crimes; recall that at least one empirical analysis of police responses to assaults revealed that the probability of arrest was least likely when the assault involved nonwhites.

Admittedly, our knowledge of the criminal court processing of sexual assault cases is limited to specific jurisdictions and time periods. Perhaps evaluation data from a broader array of states, agencies dealing with victims of sexual assaults (hospitals, social agencies, crisis center), and postlegislative reform years, may reveal cause for a more optimistic outlook. Popular and scientific interest in rape victims has generated a considerable amount of change in the reporting rates, treatment, and even some aspects of processing offenders. At this time, however, it appears that we are not doing enough to raise the ante for the perpetrators of sexual assault.

Almost every review of women and crime ends with a litany of topics for future research. This has been useful because it has

generated considerably more information on, for example, patterns and trends in female offending and factors that influence the processing of female offenders. It has not, however, generated much interest in violent female offending or, more generally, the problem of explaining the gender gap in crimes of violence. As such, the plan for future research begins by suggesting that if we want to advance our understanding of the relationship between gender and crime we cannot exclude violent offending. We will not significantly change our knowledge about the etiology and careers of violent female offenders by continuing to study violent male offenders with the excuse that the smaller sample sizes of seriously violent females make firm conclusions more difficult.

It may be that there are specific contexts in which women are as violent as men. Hints of this potential phenomenon appear in several areas of the literature that we have reviewed. For example, the data on domestic violence suggest that for the less serious types of interpersonal violence, female rates may equal male rates. Fagan's (1990) research on self-reported violence among gang and nongang members also suggests that participation rates of female gang members exceeded those of nongang males. Even laboratory studies of gender differences in aggression among adults conclude that contextual factors, such as the presence of supportive observers, the level of provocation, and the level/nature of the aggression (e.g., verbal versus physical retaliation), influence the magnitude of the gender ratio. In the aggregate, then, although women do appear to be less violent than men, it may be that in specific supportive contexts where violence is encouraged or viewed as appropriate, women's levels of violence equal those of men. A closer contextual analysis may also reveal more variation in the targets of female violence than is frequently assumed. For example, the homicide data gathered by Mann (1987) suggested that black women tend to kill friends and acquaintances (including other females) more often than white women. Thus, as recently suggested by both Bridges and Weis (1989) and a number of feminist criminologists, in order to create a more valid profile of the gender/violence paradigm, we need to disaggregate the data and examine both the nature and the extent of violence in different contexts.

Certainly, the racial variations we observe in the gender ratio for homicides and, more generally, in self-report studies provide one of the most important avenues for future research. We need to determine whether the narrower gender gap for blacks, by comparison to whites, applies to all types of interpersonal violence (e.g., see Visher and Roth, 1986:251) and to both participation rates and frequency of offending. These data may lead to a more informed approach to biological and sociological explanations for gender and interpersonal violence. With regard to biology, for example, we might expect that less attention would be focused on hormonal antecedents of violence and more attention would be directed toward prenatal and peri-or postnatal antecedents (e.g., see Shanok and Lewis, 1981). With regard to the latter, a greater emphasis might be directed to understanding the contribution of race, independent of social class and other familial or social constraints. A limited number of studies that control for social class report black-white differences in serious crime (Rutter and Giller, 1983:154-155; Visher and Roth, 1986:257). However, as Rutter and Giller (1983:154-155) point out, these data are inadequate for determining whether this race difference is explicable in terms of family circumstances, living conditions, or area of residence.

With the exception of gang research, virtually all of our knowledge about the relationships among gender, race, and crime come from aggregate data that dichotomize race into white and black, or white and "nonwhite." In the first instance, all other ethnic groups are ignored; in the second, all other ethnic groups are assumed to be equal (Simpson, 1989:618). Criminologists are only beginning to explore the relevance of etiological theory for crime among blacks; their preliminary work suggests that although the mechanisms that produce crime may be racially invariant (e.g., differential association, gender-and class-linked mechanisms of social control, poor parenting), the structural and cultural contexts (e.g., presence of, and opportunities for involvement in, illegal markets; underlying norms and values) in which these etiological factors emerge are not (e.g, see Matsueda and Heimer, 1987). Future research may suggest that these larger structural and cultural variables are indispensable for etiological theories pertaining to violence among men and women of different ethnic origins.

A large body of research has accumulated on the relationship between gender role and pathology. For our purposes the most important aspect of this research is whether men and women develop different styles of, or responses to, distress. The Dohrenwends (1976), for example, suggest that each sex has a distinct style of expressing mental disorder: for males, this involves acting-out behaviors such as personality disorders, drug and alcohol problems; for females, it entails internalization, with symptoms of depression, anxiety, and other neuroses. Although there is some controversy over whether drug, alcohol, and antisocial behaviors stem from different causal factors than disorders involving subjective distress (Gove, 1978), it could well be, as Horwitz and White (1987:159) argue, that some underlying variable such as gender identity or gender role socialization leads men and women to express pathology in fundamentally different ways. Another important avenue for future inquiry, then, is to explore whether women are less likely to engage in violence than men but more likely to exhibit other forms of psychological distress. We envision this research tackling the large bodies of data on vulnerable populations: abused, neglected, and economically deprived children, as well as children of alcoholic and divorced parents. It would also be important to consider the effects of gender roles themselves on the diagnosis and labeling of these populations. As Lewis and her colleagues (1981) found, among boys and girls with the same violent history and symptomatology, girls were more likely to be sent to a state hospital and boys to a correctional school. Thus, gender may be important not only in the etiology of specific kinds of pathology but also in the responses to pathology.

Finally, we also want to direct attention to the role of victim attributes in the processing of violent crimes. Both Smith (1987a) and LaFree (1989) found that cases involving female or nonwhite victims were treated less seriously by the police and the courts. Although these data pertain to only two types of violent crimes-assault and rape-they suggest that we need to examine more carefully whether victim attributes influence responses to all types of interpersonal violence. Our ability to deter violence, and to assist those who provide services to victims, will depend on whether

we can apply sanctions for violence without regard to the victim's race or gender.

A thorough knowledge of the relationship between gender and interpersonal violence is imperative for implementing new policies and interventions that may curb the incidence of this phenomenon. We hope that this paper will assist in that endeavor, as much perhaps by the acknowledgment of what remains to be uncovered by future research as by what the current data reveal.

Adler, F. 1975 Sisters in Crime. New York: McGraw-Hill.

1981 The Incidence of Female Criminality in the Contemporary World . New York: New York University Press.

Ageton, S. 1983 The dynamics of female delinquency, 1976-1980. Criminology 21(4):555-584.

Aizenman, M., and G. Kelley 1988 The incidence of violence and acquaintance rape in dating relationships among college men and women. Journal of College Student Development 29(4):305-311.

Anglin, D., Y. Hser, and W. McGlothin 1987 Sex differences in addict careers. American Journal of Drug and Alcohol Abuse 13:59-71.

Archer, D., and R. Gartner 1984 Violence and Crime in Cross-National Perspective. New Haven, Conn.: Yale University Press.

Barrett, D.E. 1979 A naturalistic study of sex differences in children's aggression. Merrill-Palmer Quarterly 25:193-207.

Bender, L. 1988 A lawyer's primer on feminist theory and tort. Journal of Legal Education 38(1 & 2):3-36.

Benedek, E.P. 1982 Women and homicides. Pp. 150-164 in B.L. Canto, J. Bruhns, and A.H. Kutscher, eds., The Human Side of Homicide. New York: Columbia University Press.

Bennett, L.A. 1976 The study of violence in California prisons: A review with policy implications. Pp. 149-168 in A.K. Cohen, G.F. Cole, and R.G. Bailey, eds., Prison Violence. Lexington, Mass.: Lexington Books.

Bennie, A.B., and A.B. Sclare 1969 The battered child syndrome. American Journal of Psychiatry 125(Jan.):975-979.

Berk, S.F., and D. Loseke 1981 "Handling" family violence: Situational determinants of police

arrest in domestic disturbances. Law & Society Review 15(2):317-346.

Berk, R., S.F. Berk, D. Loseke, and D. Rauma 1983 Mutual combat and other family violence myths. Pp. 197-212 in D. Finkelhor, R.J. Gelles, G.T. Hotaling, and M.A. Straus, eds., The Dark Side of Families: Current Family Violence Research. Beverly Hills, Calif.: Sage Publications.

Bernard, J.L., S.L. Bernard, and M.L. Bernard 1985 Courtship violence and sex-typing. Family Relations: Journal of Applied Family and Child Studies 34(4):573-576.

Bernstein, I.N., J. Cardascia, and C.E. Ross 1979 Defendant's sex and criminal court decisions. Pp. 329-354 in R. Alvarez, K.G. Lutterman et al., eds., Discrimination in Organizations . San Francisco: Jossey-Bass.

Black, D. 1976 The Behavior of Law. New York: Academic Press.

1980 The Manners and Customs of the Police. New York: Academic Press.

Blumstein, A., J. Cohen, J.A. Roth, and C.A. Visher, eds. 1986 Criminal Careers and "Career Criminals", Vol. 1. Panel on Research on Criminal Careers, Committee on Research on Law Enforcement and the Administration of Justice, National Research Council. Washington, D.C.: National Academy Press.

Blumstein, A., J. Cohen, and D.P. Farrington 1988a Criminal career research: Its value for criminology. Criminology 26(1):1-35.

1988b Longitudinal and criminal career research: Further clarifications. Criminology 26(1):57-74.

Boisvert, M.J., and R. Wells 1980 Toward a rational policy on status offenders. Social Work 25(3):230-234.

Bowker, L.H. 1978 Gangs and prostitutes: Two case studies of female crime. Pp. 143-169 in L.H. Bowker, ed., Women, Crime and the Criminal Justice System. Lexington, Mass.: Lexington Books.

1980 Prison Victimization. New York: Elsevier.

1981 Women as victims: An examination of the results of LEAA's National Crime Survey Program. Pp. 158-179 in L.H. Bowker, ed., Women and Crime in America. New York: Macmillan.

Bowker, L.H., H.S. Gross, and M.W. Klein 1980 Female participation in delinquent gang activities. Adolescence 15:509-515.

Bridges, G.S., and J.G. Weis 1989 Measuring violent behavior. Pp. 14-34 in N.A. Weiner and M.E. Wolfgang , eds., Violent Crime, Violent Criminals. Newbury Park, Calif.: Sage Publications.

Brown, W.K. 1977 Black female gangs in Philadelphia. International Journal of Offender Therapy and Comparative Criminology 21(3):221-228.

Brownmiller, S. 1975 Against Our Will: Men, Women and Rape. New York: Simon and Schuster.

Bureau of the Census 1989 Projections of the Population of the United States, by Age, Sex and Race: 1988 to 2080. Current Population Reports, Series P-25, No. 1018. Washington, D.C.: U.S. Department of Commerce.

Bureau of the Census (jointly with the Department of Agriculture) 1989 Rural and Rural Farm Population: 1988. Current Population Reports, Series P-20, NO. 439. Washington, D.C.: U.S. Department of Commerce.

Campbell, A. 1984 The Girls in the Gang. Oxford, England: Basil Blackwell.

Caplan, P.J. 1975 Sex differences in antisocial behavior: Does research methodology produce or abolish them? Human Development 18:444-460.

Caringella-MacDonald, S. 1985 The comparability in sexual and nonsexual assault case treatment: Did statute change meet the objective? Crime and Delinquency 31(2):206-222.

Carlen, P. 1985 Criminal Women. Cambridge: Policy Press.

Carroll, L. 1974 Hacks, Blacks, and Cons: Race Relations in a Maximum Security Prison. Lexington, Mass.: D.C. Heath and Company.

Center for the Study of Social Policy 1986 The "flip-side" of black families headed by women: The economic status of black men. Pp. 232-238 in R. Staples, ed., The Black Family: Essays and Studies. Belmont, Calif.: Wadsworth.

Chaiken, J.M., and M.R. Chaiken 1990 Drugs and predatory crime. Pp. 203-239 in M. Tonry and J.Q. Wilson, eds., Drugs and Crime. Chicago: University of Chicago Press.

Chapman, J.R. 1980 Economic Realities and the Female Offender. Lexington, Mass.: Lexington Books.

Chappell, D. 1989 Sexual criminal violence. Pp. 68-108 in N.A. Weiner and M.E. Wolfgang, eds., Pathways to Criminal Violence. Newbury Park, Calif.: Sage Publications.

Chesney-Lind, M. 1973 Judicial enforcement of the female sex role: The family court and the female delinquent. Issues in Criminology 8(Fall):51-70.

1974 Juvenile delinquency: The sexualization of female crime. Psychology Today 7:43-46.

1977 Judicial paternalism and the female status offender. Crime and Delinquency 23(2):121-130.

1988 Girls in jail. Crime and Delinquency 34:150-168.

Cloward, R.A., and L.E. Ohlin 1960 Delinquency and Opportunity. A Theory of Delinquent Gangs. New York: Free Press.

Cohen, L.E., and J.R. Kluegal 1979a Selecting delinquents for adjudication. Journal of Research in Crime and Delinquency 16:143-163.

1979b The detention decision: A study of the impact of social characteristics and legal factors in two metropolitan courts. Social Forces 58(1):146-161.

Cook, P.J. 1980 Research in criminal deterrence: Laying the groundwork for the second decade. Pp. 211-268 in N. Morris and M. Tonry, eds., Crime and Justice, An Annual Review of Research, Vol. 2. Chicago: University of Chicago Press.

Corrado, R.R. 1980 Female terrorists: Competing perspectives. Pp. 37-50 in C.T. Griffiths and M. Nance, eds., The Female Offender. Selected Papers from an International Symposium, Vancouver, Canada. Simon Fraser University: Criminology Research Centre.

Cressey, D. 1970 Measuring crime rates. Pp. 55-59 in A.L. Guenther, ed., Criminal Behavior and Social Systems. Chicago: Rand McNally.

Curran, D.A. 1983 Judicial discretion and defendant's sex. Criminology 21(1):41-58.

Daly, K. 1987 Discrimination in the criminal courts: Family, gender, and the problem of equal treatment. Social Forces 66(1):152-175.

1989a Rethinking judicial paternalism: Gender, work-family relations, and sentencing. Gender and Society 3(1):9-36.

1989b Neither conflict nor labeling nor paternalism will suffice: Intersections of race, ethnicity, gender and family in criminal court decisions. Crime and Delinquency 35(1):136-168.

Daly, K., and M. Chesney-Lind 1988 Feminism and criminology. Justice Quarterly 5(4):497-535.

Datesman, S.K., and F.R. Scarpitti 1980 The extent and nature of female crime. Pp. 3-64 in S.K. Datesman and F.R. Scarpitti, eds., Women, Crime, and Justice. New York: Oxford University Press.

Dobash, R.E., and R. Dobash 1979 Violence Against Wives: A Case Against the Patriarchy. New York: Free Press.

Dobash, R.P., R. Emerson, and S. Gutteridge 1986 The Imprisonment of Women. Totowa, N.J.: Basil & Blackwell.

Dohrenwend, B.P., and B.S. Dohrenwend 1976 Sex differences in psychiatric disorders. American Journal of Sociology 81:1447-1454.

Eagly, A., and V.J. Steffen 1986 Gender and aggressive behavior: A meta-analytic review of the social psychological literature. Psychological Bulletin 100:309-330.

Ekland-Olson, S. 1986 Crowding, social control and prison violence: Evidence from the post-Ruiz years in Texas. Law & Society Review 20(3):389-421.

Elliott, D., S. Ageton., D. Huizinga, B. Knowles, and R. Canter 1983 The Prevalence and Incidence of Delinquent Behavior: 1976-1980 . National Youth Survey, Report No. 26. Boulder, Colo.: Behavioral Research Institute.

Elliott, D.S., D. Huizinga, and B. Morse 1986 Self-reported violent offending: A descriptive analysis of juvenile violent offenders and their offending careers . Journal of Interpersonal Violence 1:472-514.

Ellis, D., H.G. Grasmick, and B. Gilman 1974 Violence in prisons: A sociological analysis. American Journal of Sociology 80:16-43.

Eme, R. 1984 Sex roles, sex differences and child psychopathology. Pp. 279-316 in C.S. Widom, eds., Sex Roles and Psychopathology. New York: Plenum.

Epstein, C.F. 1988 Deceptive Distinctions: Sex, Gender and the Social Order. New Haven, Conn.: Yale University Press.

Erez, E. 1989 Gender rehabilitation and probation decisions. Criminology 27(2):307-327.

Erickson, M.L. 1975 Delinquency in a birth cohort: A new direction in criminological research? Journal of Criminal Law and Criminology 66:362-367.

Fagan, J. 1990 Social processes of delinquency and drug use among urban gangs. In C.R. Huff, ed., Gangs in America. Newbury Park, Calif.: Sage Publications.

Fagan, J., and S. Wexler 1987 Crime at home and crime in the streets: The relationship between

family and stranger violence. Violence and Victims 2(1):5-23.

Fagot, B.I., and R. Hagan 1985 Aggression in toddlers: Responses to the assertive acts of boys and girls. Sex Roles 12:341-351.

Fagot, B.I., R. Hagan, M.B. Leinbach, and S. Kronsberg 1985 Differential reactions to assertive and communicative acts of toddler boys and girls. Child Development 56:1499-1505.

Fagot, B.I., M.D. Leinbach, and R. Hagan 1986 Gender labeling and adoption of gender role behaviors. Developmental Psychology 22:440-443.

Fagot, B.I., R. Loeber, and J.B. Reid 1988 Developmental determinants of male-to-female aggression. Pp. 91-105 in G. W. Russell, ed., Violence in Intimate Relationships. New York: PMA Publishing Corp.

Farrell, R.A., and V.L. Swigert 1986 Adjudication in homicide: An interpretative analysis of the effects of defendant and victim social characteristics. Journal of Research in Crime and Delinquency 23(4):349-369.

Farrington, D.P. 1978 The family background of aggressive youths. Pp. 73-93 in L.A. Hersov, M. Berger, and D. Shaffer, eds., Aggression and Antisocial Behavior in Childhood and Adolescence. Oxford: Pergamon.

1982 Longitudinal analyses of criminal violence. Pp. 171-201 in M.E. Wolfgang and N.A. Weiner, eds., Criminal Violence. Beverly Hills, Calif.: Sage Publications.

1986 Age and crime. Pp. 189-250 in M. Tonry and N. Morris, eds., Crime and Justice, Vol. 7. Chicago: University of Chicago Press.

1989 Early predictors of adolescent aggression and adult violence. Violence and Victims 4(2):79-100.

1990 Age, period, cohort, and offending. Pp. 51-75 in D.M. Gottfredson and R.V. Clarke, eds., Policy and Theory in Criminal Justice: Contributions in Honour of Leslie T. Wilkins. Aldershot, England: Avebury.

Feld, B.C. 1989 The right to counsel in juvenile court: An empirical study of when lawyers appear and the difference they make. Journal of Criminal Law and Criminology 79(4):1185-1346.

Feldman-Summers, S., and C.D. Ashworth 1981 Factors related to intentions to report a rape. Journal of Social Issues 37:53-70.

Felson, R.B., and N. Russo 1988 Parental punishment and sibling aggression. Social Psychology Quarterly 51(1):11-18.

Figueira-McDonough, J. 1985 Gender differences in informal processing: A look at charge

bargaining and sentence reduction in Washington D.C. Journal of Research in Crime and Delinquency 22(2):101-133.

Finn, J. 1986 The relationship between sex role attitudes and attitudes supporting marital violence. Sex Roles 14(5-6):235-244.

Fortune, E.P., M. Vega, and I.J. Silverman 1980 A study of female robbers in a southern correctional institution. Journal of Criminal Justice 8:324.

Frazier, C.E., and E.W. Bock 1982 Effects of court officials on sentence severity. Criminology 20(2):257-272.

Frieze, I.H., and A. Browne 1989 Violence in marriage. Pp. 163-218 in L. Ohlin and M. Tonry, eds., Family Violence. Chicago: University of Chicago Press.

Frodi, A., J. Macaulay, and P.R. Thome 1977 Are women always less aggressive than men? A review of the experimental literature. Psychological Bulletin 84:634-660.

Gartner, R., K. Baker, and F.C. Pampel 1987 The Sex Differential in Homicide Victimization. Paper presented at a meeting of the American Society of Criminology, Montreal, Canada.

Gates, M. 1978 Victims of rape and wife abuse. Pp. 176-201 in W.L. Hepperle and L. Crites, eds., Women in the Courts. Williamsburg, Va.: National Center for State Courts.

Gelles, R.J. 1979 Family Violence. Beverly Hills, Calif.: Sage Publications.

1982 Domestic criminal violence. Pp. 201-235 in M.E. Wolfgang and N.A. Weiner, eds., Criminal Violence. Beverly Hills, Calif.: Sage Publications.

Gelles, R.J., and M.A. Straus 1988 Intimate Violence. New York: Touchstone.

Ghali, M.A., and M. Chesney-Lind 1986 Sex bias and the criminal justice system: An empirical investigation. Sociology and Social Research 70(2):164-171.

Gil, D. 1970 Violence Against Children: Physical Child Abuse in the United States. Cambridge, Mass.: Harvard University Press.

Gill, S., J. Stockard, M. Johnson, and S. Williams 1987 Measuring gender differences: The expressive dimension and critique of androgyny scales. Sex Roles 17(7-8):375-400.

Gilligan, C. 1982 In a Different Voice: Psychological Theory and Women's Development . Cambridge, Mass.: Harvard University Press.

Giordano, P. 1978 Research note: Girls, guys and gangs: The changing social

context of female delinquency. Journal of Criminal Law and Criminology 69(1):126-132.

Giordano, P., S. Kerbel, and S. Dudley 1981 The economics of female criminality: An analysis of police blotters, 1890-1976. Pp. 65-82 in L. Bowker, ed., Women and Crime in America . New York: Macmillan.

Gottfredson, M., and M. Hindelang 1979 A study of the behavior of law. American Sociological Review 44:3-18.

Gove, W.R. 1978 Sex differences in mental illness among adult men and women. Social Science and Medicine 12B:187-198.

Gove, W.R., M. Hughes, and M. Geerken 1985 Are Uniform Crime Reports a valid indicator of the index crimes? An affirmative answer with minor qualifications. Criminology 23:451-501.

Grasmick, H.G., N.J. Finley, and D.L. Glaser 1984 Labor force participation, sex-role attitudes and female crime. Social Science Quarterly 65(3):703-718.

Gruhl, J., C. Spohn, and S. Welch 1981 Women as policymakers: The case of trial judges. American Journal of Political Science 25(2):308-322.

Hagan, J., A.R. Gillis, and J. Simpson 1985 The class structure of gender and delinquency: Toward a power-control theory of common delinquent behavior. American Journal of Sociology 90(6):1151-1178.

Hagan, J., J. Simpson, and A.R. Gillis 1987 Class in the household: A power-control theory of gender and delinquency. American Journal of Sociology 92(4):788-816.

Hamparian, D.M., R. Schuster, S. Dinitz, and J.P. Conrad 1978 The Violent Few: A Study of Dangerous Juvenile Offenders. Lexington, Mass. : Lexington.

Hanson, K. 1964 Rebels in the Streets: The Story of New York's Girl Gangs. New York: Prentice Hall.

Harding, S. 1986 The Science Question in Feminism. Ithaca, N.Y.: Cornell University Press.

Harris, A. 1977 Sex and theories of deviance. American Sociology Review 42:3-16.

Hartnagel, T.F. 1982 Modernization, female social roles and female crime: A cross-national investigation. Sociological Quarterly 23(4):477-490.

Heidensohn, F.M. 1985 Women and Crime. New York: New York University Press.

Hetherington, E.M. 1981 Children and divorce. Pp. 33-58 in R. Henderson, ed., Parent-Child Interaction: Theory, Research and Prospect. New York: Academic Press.

Hindelang, M.J. 1979 Sex differences in criminal activity. Social Problems 27(2):143-156.

Hindelang, M., and M. Gottfredson 1976 The victim's decision not to invoke the criminal process. Pp. 57-78 in W. McDonald, ed., The Victim and the Criminal Justice System . Beverly Hills, Calif.: Sage Publications.

Hindelang, M.J., T. Hirschi, and J.G. Weis 1979 Correlates of delinquency: The illusion of discrepancy between self-report and official measures. American Sociological Review 44:995-1014.

1981 Measuring Delinquency. Beverly Hills., Calif.: Sage Publications.

Holmstrom, L.L., and A.W. Burgess 1978 The Victim of Rape: Institutional Reactions. New York: Wiley-Interscience.

Horwitz, A.V., and H.R. White 1987 Gender role orientations and styles of pathology among adolescents. Journal of Health and Social Behavior 28(2):158-170.

Hotaling, G.T., M.A. Straus, and A.J. Lincoln 1989 Intrafamily violence, and crime and violence outside the family. Pp. 315-375 in L. Ohlin and M. Tonry, eds., Family Violence. Chicago: University of Chicago Press.

Inciardi, J.A. 1979 Heroin use and street crime. Crime and Delinquency 25:335-346.

Irwin, J. 1980 Prisons in Turmoil. Boston: Little Brown & Co.

James, J., and W. Thornton 1980 Women's liberation and the female delinquent. Journal of Research in Crime and Delinquency 17(2):230-244.

Jensen, G.F., and K. Thompson 1990 What's class got to do with it? A further examination of power-control theory. American Journal of Sociology 95(4):1009-1023.

Klein, D. 1973 The etiology of female crime: A review of the literature. Crime and Social Justice: Issues in Criminology (Fall):3-30.

1982 Violence against women: Some considerations regarding its cause and eliminations. Pp. 203-222 in B.R. Price and N.J. Sokologg, eds., The Criminal Justice System and Women. New York: Clark Boardman Co.

Klein, M.W., and C.L. Maxson 1989 Street gang violence. Pp. 198-234 in N.A. Weiner and M.E.

Wolfgang, eds., Violent Crime, Violent Criminals. Newbury Park, Calif.: Sage Publications.

Kratcoski, P.C. 1987 Families who kill. Marriage and Family Review 12(1-2):47-70.

Kritzer, H.M., and T.M. Uhlman 1977 Sisterhood in the courtroom: Sex of judge and defendant in criminal case disposition. Social Science Journal 14(2):77-78.

Krohn, M.D., J.P. Curry, and S. Nelson-Kilger 1983 Is chivalry dead: An analysis of changes in police dispositions of males and females. Criminology 21:417-437.

Kruttschnitt, C. 1984 Sex and criminal court dispositions: The unresolved controversy. Journal of Research in Crime and Delinquency 21(3):213-232.

1985a ''Female crimes" or legal labels? The effect of deviance processing agents on our understanding of female criminality. Pp. 76-94 in I.L. Moyer, ed., The Changing Roles of Women in the Criminal Justice System . Prospect Heights, Ill.: Waveland.

1985b Legal outcomes and legal agents. Adding another dimension to the sex-sentencing controversy. Law and Human Behavior 9(3):287-303.

Kruttschnitt, C., and D. Green 1984 The sex-sanctioning issue: Is it history? American Sociology Review 21:213-232.

Kruttschnitt, C., and C. Johnson 1984 Sentencing recommendations and women offenders: The biopsychological model and the treatment of female offenders. Law and Inequality 2(1):97-120.

Kruttschnitt, C., and S. Krmpotich 1990 Aggressive behavior among female inmates: An exploratory study. Justice Quarterly 7(2):371-389.

Kruttschnitt, C., L. Heath, and D. Ward 1986 Family violence, television viewing habits, and other adolescent experiences related to violent criminal behavior. Criminology 24:235-267.

Kruttschnitt, C., D. Ward, and M. Sheble 1987 Abuse-resistant youth: Some factors that may inhibit violent criminal behavior. Social Forces 66(2):501-519.

Kurz, D. 1989 Social science perspectives on wife abuse: Current debates and future directions. Gender and Society 3(4):489-505.

LaFree, G.D. 1989 Rape and Criminal Justice. The Social Construction of Sexual Assault. Belmont, Calif.: Wadsworth.

Laub, J.H., and M.J. McDermott 1985 An analysis of serious crime by young black women. Criminology 23(1):81-99.

Law Enforcement Assistance Administration 1972 San Jose Methods Test of Known Crime Victims. Statistics Technical Report No. 1, National Institute of Law Enforcement and Criminal Justice, Statistics Division, U.S. Department of Justice, Washington, D.C.

Leonard, E.G. 1982 Women, Crime and Society. New York: Longmans.

Letcher, M. 1979 Black women and homicide. In H.M. Rose, ed., Lethal Aspects of Urban Violence. Lexington, Mass.: D.C. Heath.

Lewis, D.K. 1981 Black women offenders and criminal justice: Some theoretical considerations. Pp. 89-105 in M.Q. Warren, ed., Comparing Female and Male Offenders. Research Progress Series in Criminology, Vol. 21. Beverly Hills, Calif.: Sage Publications.

Lewis, D.O., S.S. Shanok, R.J. Cohen, M. Kligfeld, and G. Frisone 1981 Race bias in the diagnosis and disposition of violent adolescents. Annual Progress in Child Psychiatry and Child Development 14:508-520.

Lewis, D.O., S.S. Shanok, M. Grant, and E. Rita 1983 Homicidally aggressive young children: Neuropsychiatric and experiential correlates. American Journal of Psychiatry 140(2):148-153.

Lizotte, A.J. 1985 The uniqueness of rape: Reporting assaultive violence to the police. Crime and Delinquency 31(2):169-190.

Loeber, R., and M. Stouthamer-Loeber 1986 Family factors as correlates and predictors of juvenile conduct problems and delinquency. Pp. 219-339 in M. Tonry and N. Morris, eds., Crime and Justice, Vol. 7. Chicago: University of Chicago Press.

Loh, W.D. 1980 The impact of common law reform and rape statutes on prosecution: An empirical study. Washington Law Review 55:543-652.

Macaulay, J. 1985 Adding gender to aggression research: Incremental or revolutionary change? Pp. 191-224 in V.E. O'Leary, R.K. Unger and B.S. Wallston, eds., Women, Gender and Social Psychology. Hillsdale, N.J.: Lawrence Erlbaum.

Maccoby, E.E., and C.N. Jacklin 1974 The Psychology of Sex Differences. Stanford, Calif.: Stanford University Press.

1980 Sex differences in aggression: A rejoinder and reprise. Child Development 51:964-980.

Magnusson, D., H. Stottin, and A. Duner 1983 Aggression and criminality in a longitudinal perspective. Pp.

277-301 in K.T. Van Dusen and S.A. Mednick, eds., Antecedents of Aggression and Antisocial Behavior. Boston: Kluwer-Nijhoff.

Makepeace, J.M. 1983 Life events stress and courtship violence. Family Relations: Journal of Applied Family and Child Studies 32(1):101-109.

1986 Gender differences in courtship violence victimization. Family Relations: Journal of Applied Family and Child Studies 35(3):383-388.

Mandaraka-Sheppard, A. 1986 The Dynamics of Aggression in Women's Prisons in England. Brookfield, Vt.: Gower.

Mann, C.R. 1984 Female Crime and Delinquency. Huntsville: University of Alabama Press.

1987 Black women who kill. Pp. 157-186 in R.L. Hampton, ed., Violence in the Black Family. Lexington, Mass.: Lexington Books.

1988 Getting even? Women who kill in domestic encounters. Justice Quarterly 5(1):33-51.

Marsh, P., and R. Paton 1986 Gender, social class and conceptual schemes of aggression. Pp. 59-86 in A. Campbell and J. Gibbs, eds., Violent Transactions. The Limits of Personality. New York: Basil Blackwell.

Matsueda, R.L., and K. Heimer 1987 Race, family structure and delinquency: A test of differential association and social control theories. American Sociological Review 52:826-846.

Menard, S. 1987 Short-term trends in crime and delinquency: A comparison of UCR, NCS and self-report data. Justice Quarterly 4(3):455-474.

Miethe, T.D. 1987 Stereotypical conceptions and criminal processing: The case of the victim-offender relationship. Justice Quarterly 4(4):571-593.

Miller, W.B. 1973 The Molls. Society 11:32-35.

1975 Violence by Youth Gangs and Youth Groups as a Crime Problem in Major American Cities. Washington, D.C.: U.S. Government Printing Office.

Moseley, W.H., and M.H. Gerould 1975 Sex and parole: A comparison of male and female parolees. Journal of Criminal Justice 3:47-57.

Mowbray, E.J. 1982 Parole Prediction and Gender. Paper presented at the annual meeting of the American Society of Criminology, Toronto, Canada.

Myers, M.A., and G.D. LaFree 1982 The uniqueness of sexual assault: A comparison with other

crimes. Journal of Criminal Law and Criminology 73:1282-1305.

Naffine, N. 1987 Female Crime: The Construction of Women in Criminology. Sydney, Australia: Allen and Unwin.

Nagel, I.H., and J. Hagan 1982 Gender and crime: Offense patterns and criminal court sanctions. Pp. 91-144 in N. Morris and M. Tonry, eds., Crime and Justice, Vol. 4. Chicago: University of Chicago Press.

Norland, S., and P.J. Mann 1984 Being troublesome-Women on probation. Criminal Justice and Behavior 11(1):115-135.

Norland, S., and N. Shover 1978 Gender roles and female criminality. Criminology 15:87-104.

Norland, S., R.C. Wessel, and N. Shover 1981 Masculinity and delinquency. Criminology 19(3):421-433.

Pagelow, M.D. 1984 Family Violence. New York: Praeger.

Parisi, N. 1982 Exploring female crime patterns. Pp. 111-129 in N.H. Rafter and E.A. Stanko, eds., Judge, Lawyer, Victim, Thief. Boston, Mass.: Northeastern University Press.

Parke, R.D., and C.W. Collmer 1975 Child abuse: An interdisciplinary analysis. Pp. 1-102 in M. Hetherington, ed., Review of Child Development Research, Vol. 5. Chicago: University of Chicago Press.

Polk, K. 1985 Rape reform and criminal justice processing. Crime and Delinquency 31(2):191-205

President's Commission on Law Enforcement and Administration of Justice 1967 The Challenge of Crime in a Free Society. Washington, D.C.: U.S. Government Printing Office.

Quicker, J.C. 1974 The Chicana Gang: A Preliminary Description. Paper presented at the annual meeting of the Pacific Sociological Association, San Jose, Calif.

Randall, S.C., and V.M. Rose 1981 Barriers to becoming a "successful" rape victim. Pp. 336-354 in L.H. Bowker, ed., Women and Crime in America. New York: Macmillan.

Rans, L. 1978 Women's crime: Much ado about . ...? Federal Probation 42:45-49.

Reid, J.B., and G.R. Patterson 1989 The development of antisocial behavior patterns in childhood and adolescence. European Journal of Personality 3:107-119.

Reiss, A.J. 1981 Problems in developing statistical indicators of crime. Presentation for the XXXI Cours International de Criminologie, Connaître la Criminalité: Le Dernier État de la Question, Aix-en-Provence, December 7-11.

Reskin, B.F., and C.A. Visher 1986 The impact of evidence and extralegal factors in jurors' decisions. Law & Society Review 20:423-438.

Riedel, M. 1988a Black Women and Homicide: Rates, Patterns, and Perspectives. Paper presented at the annual meeting of the American Society of Criminology.

1988b Murder, Race and Gender: A Test of Hagan's Hypotheses. Paper presented at the annual meeting of the American Society of Criminology, Chicago, Ill.

Rivera, B., and C.S. Widom 1989 Childhood Victimization and Violent Offending. Unpublished paper. Departments of Criminal Justice and Psychology, Indiana University.

Robins, L.N. 1966 Deviant Children Grow Up: A Sociological and Psychiatric Study of Sociopathic Personality. Baltimore, Md.: Williams and Wilkins.

Russell, D.E.H. 1984 Sexual Exploitation. Rape, Child Sexual Abuse and Workplace Harassment. Beverly Hills, Calif.: Sage Publications.

Rutter, M. 1982 Epidemiological-longitudinal approaches to the study of development. Pp. 105-144 in W.A. Collins, ed., The Concept of Development: Minnesota Symposia on Child Psychology, Vol. 15. Hillsdale, N.J.: Lawrence Erlbaum.

Rutter, M., and H. Giller 1983 Juvenile Delinquency. Trends and Perspectives New York: The Guilford Press.

Schur, E.M. 1984 Labeling Women Deviant: Gender, Stigma and Social Control. New York: Random.

Shannon, L.W. 1988 Criminal Career Continuity. Its Social Context. New York: Human Sciences Press.

Shanok, S.S., and D.O. Lewis 1981 Medical histories of female delinquents. Archives of General Psychiatry 38:211-213.

Shelden, R.G. 1981 Sex discrimination in the juvenile justice system: Memphis, Tennessee, 1900-1917. Pp. 55-72 in M.Q. Warren, ed., Comparing

Female and Male Offenders. Beverly Hills, Calif.: Sage Publications.

Short, J.F., Jr. 1968 Gang Delinquency and Delinquent Subcultures. New York: Harper & Row.

Siann, G. 1985 Accounting of Aggression: Perspectives on Aggression and Violence . Winchester, Mass.: Allen and Unwin.

Sigelman, C.K., C.J. Berry, and K.A. Wiles 1984 Violence in college students' dating relationships. Journal of Applied Social Psychology 14(6):530-548.

Simon, R.J. 1975 Women and Crime. Lexington, Mass.: D.C. Health.

1979 The parole system: How women fare. Pp. 380-385 in F. Adler and R.J. Simon, eds., The Criminology of Deviant Women. Boston: Houghton Mifflin.

Simon, R.J., and S. Baxter 1989 Gender and violent crime. Pp. 171-197 in N.A. Weiner and M.E. Wolfgang, eds., Violent Crime, Violent Criminals. Newbury Park, Calif.: Sage Publications.

Simpson, S.S. 1989 Feminist theory, crime and justice. Criminology 27(4):605-632.

Smart, C. 1976 Women, Crime and Criminology. London: Routledge & Kegan Paul.

Smith, D.A. 1987a Police response to interpersonal violence: Defining the parameters of legal control. Social Forces 65(3):767-782.

Smith, D.A., and C.A. Visher 1980 Sex and involvement in deviance/crime: A quantitative review of the empirical literature. American Sociological Review 45(4):691-701.

1981 Street-level justice: Situational determinants of police arrest decisions. Social Problems 29(2):167-177.

Smith, D.M. 1987b Changes in the victimization of women: Is there a "new female victim"? Journal of Research in Crime and Delinquency 24(4):291-301.

Spence, J.T., and L.L. Sawin 1985 Images of masculinity and femininity: A reconceptualization. Pp. 35-66 in V.E. O'Leary, R.K. Unger and B.S. Wallston, eds., Women, Gender and Social Psychology. Hillsdale, N.J.: Lawrence Erlbaum.

Spencer, C., and J.E. Berocochea 1972 Recidivism Among Women Parolees: A Long Term Survey. Report No. 47, California Research Division, Department of Corrections.

Spohn, C., and S. Welch 1987 The effect of prior record in sentencing research: An examination of the assumption that any measure is adequate. Justice Quarterly 4(2):287-302.

Staples, W.G. 1984 Toward a structural perspective on gender bias in the juvenile court. Sociological Perspectives 27(3):349-367.

Stark, E., A. Flitcraft, and W. Frazier 1979 Medicine and patriarchal violence: The social construction of a "private" event. International Journal of Health Services 98:461-491.

Steele, B.F., and C.A. Pollock 1974 A psychiatric study of parents who abuse infants and small children. Pp. 89-134 in R.E. Helfer and C.H. Kempe, eds., The Battered Child , 2nd ed. Chicago: University of Chicago Press.

Steffensmeier, D.J. 1980 Sex differences in patterns of adult crimes, 1965-77: A review and assessment. Social Forces 58(4):1080-1108.

1982 Trends in female crime. It's still a man's world. Pp. 117-129 in R.R. Price and N.J. Solkoloff, eds., The Criminal Justice System and Women. New York: Clark Boardman Co.

Steffensmeier, D.J., and E.A. Allan 1988 Sex disparities in arrests by residence, race and age: An assessment of the gender convergence/crime hypothesis. Justice Quarterly 5(1):53-80.

Steffensmeier, D.J., and C. Jordan 1978 Changing patterns of female crime in rural America, 1962-1975. Rural Sociology 43(1):87-102.

Steffensmeier, D.J., and R.H. Steffensmeier 1980 Trends in female delinquency. Criminology 18(1):62-85.

Steffensmeier, D.J., and C. Streifel 1989 Women's Status and the Female Share of Offending, 1960-1985: A Time Series Analyses of Sex Differences in Person and Property Crime Involvement. Unpublished paper, Department of Sociology, Pennsylvania State University.

Steffensmeier, D.J., R.H. Steffensmeier, and A. Rosenthal 1979 Trends in female violence, 1960-1977. Sociological Focus 12(3):217-227.

Steffensmeier, D.J., E.A. Allan, and C. Streifel 1989 Development and female crime: A cross-national test of alternative explanations. Social Forces 68(1):262-283.

Stets, J.E., and M.A. Pirog-Good 1987 Violence in dating relationships. Social Psychology Quarterly 50(3):237-246.

Stewart, J.K. 1985 Prosecution of child sexual abuse: Innovations in practice. National

Institute of Justice. Research in Brief. Washington, D.C.: U.S. Department of Justice.

Straus, M.A., R.J. Gelles, and S.K. Steinmetz 1980 Behind Closed Doors: Violence in the American Family. New York: Doubleday/Anchor.

Stryker, R., I.H. Nagel, and J. Hagan 1983 Methodological issues in criminal court research: Pretrial release decisions for federal defendants. Sociological Methods and Research 11(4):469-500.

Sylvester, S., J. Reed, and D. Nelson 1977 Prison Homicide. New York: Spectrum.

Thornton, W.E., and J. James 1979 Masculinity and delinquency revisited. British Journal of Criminology 19(3):225-241.

Thrasher, F.M. 1963 The Gang. Chicago: University of Chicago Press.

Tieger, T. 1980 On the biological basis of sex differences in aggression. Child Development 51:943-963.

U.S. Department of Health and Human Services 1981 National Study of the Incidence and Severity of Child Abuse and Neglect. Washington, D.C.: U.S. Department of Health and Human Services.

U.S. Department of Justice, Bureau of Justice Statistics 1989 Criminal Victimization in the United States, 1987. A National Crime Survey Report, NCJ-115524. Washington, D.C.: U.S. Department of Justice.

U.S. Department of Justice, Federal Bureau of Investigation 1989 Uniform Crime Reports for the United States, 1988. Washington, D.C.: U.S. Department of Justice.

VanVoorhis, P., F.T. Cullen, R.A. Mathers, and C.C. Garner 1988 The impact of family structure and quality on delinquency: A comparative assessment of structural and functional factors. Criminology 26(2):235-261.

Vigil, J.D. 1988 Barrio Gangs. Austin: University of Texas Press.

Visher, C.A. 1983 Gender, police arrest decisions and notions of chivalry. Criminology 21(1):5-28.

Visher, C.A., and J.A. Roth 1986 Participation in criminal careers. Appendix A. Pp. 211-291 in A. Blumstein, J. Cohen, J.A. Roth, and C.A. Visher, eds., Criminal Careers and "Career Criminals", Vol. 1. Washington, D.C.: National Academy Press.

Walker, L.E., and A. Browne 1985 Gender and victimization by intimates. Special issue: Conceptualizing

gender in personality theory and research. Journal of Personality 53(2):179-195.

Ward, D.A., M. Jackson, and R.E. Ward 1969 Crimes of violence by women. Pp. 843-909 in D.J. Mulvihill and M.M. Tumin, eds., Crimes of Violence, Vol. 13. Washington, D.C.: U.S. Government Printing Office.

Weinberg, L., and W.L. Eubank 1987 Italian women terrorists. Terrorism 9(3):241-262.

Weiner, N.A. 1989 Violent criminal careers and "violent career criminals." Pp. 35-138 in N.A. Weiner and M.E. Wolfgang, eds., Violent Crime, Violent Criminals. Newbury Park, Calif.: Sage Publications.

Weis, J.G. 1986 Issues in the measurement of criminal careers. Pp. 1-51 in A. Blumstein, J. Cohen, J.A. Roth and C.A. Visher, eds., Criminal Careers and "Career Criminals", Vol. 2. Washington, D.C.: National Academy Press.

White, J.W. 1983 Sex and gender issues in aggression research. Pp. 1-26 in R.G. Geen and E.I. Donnerstein, eds., Aggression: Theoretical and Empirical Reviews. Vol. 2. Issues in Research. New York: Academic Press.

Widom, C.S. 1978 Toward an understanding of female criminality. Pp. 245-308 in A. Mafier, ed., Progress in Experimental Personality Research, Vol. 8. New York: Academic Press.

1984 Sex roles, criminality and psychopathology. Pp. 183-217 in C.S. Widom, ed., Sex Roles and Psychopathology. New York: Plenum.

1987 Family Violence and Infanticide. Paper presented at the Pennsylvania State University Conference on Perspectives on Postpartum Depression and Criminal Responsibility .

1989a Child abuse, neglect and violent criminal behavior. Criminology 27:251-271.

1989b The intergenerational transmission of violence. Pp. 137-201 in N.A. Weiner and M.E. Wolfgang, eds., Pathways to Criminal Violence . Newbury Park, Calif.: Sage Publications.

Widom, C.S., and A. Ames 1988 Biology and female crime. Pp. 308-331 in T.E. Moffitt and S.A. Medinick, eds., Biological Contributions to Crime Causation. Dordrecht: Martinus Nijhoff.

Widom, C.S., and A.J. Stewart 1986 Female criminality and the status of women. International Annals of Criminology 24:137-162.

Wilbanks, W. 1982 Murdered women and women who murder: A critique of the literature. Pp. 151-180 in N.H. Rafter and E.A. Sanko, eds.,

Judge, Lawyer, Victim, Thief: Women, Gender Roles and Criminal Justice . Boston: Northeastern University Press.

1986 Are female felons treated more leniently by the criminal justice system. Justice Quarterly 3(4):517-529.

Williams, K.M. 1976 The effects of victim characteristics on the disposition of violent crimes. Pp. 177-213 in W.F. McDonald, ed., Criminal Justice and the Victim. Beverly Hills, Calif.: Sage Publications.

Williams, L.S. 1984 The classic rape: When do victims report? Social Problems 31(4):459-467.

Wilson, J.Q., and R.J. Herrnstein 1985 Crime and Human Nature. New York: Simon and Schuster.

Wish, E.D., and B.A. Gropper 1990 Drug testing by the criminal justice system: Methods, research and applications. Pp. 321-391 in M. Tonry and J.Q. Wilson, eds., Drugs and Crime. Chicago: University of Chicago Press.

Wolfgang, M.E. 1975 Patterns in Criminal Homicide. Montclair, N.J.: Patterson Smith.

Young, V.D. 1979 Victims of female offenders. Pp. 72-87 in W.H. Parsonage, ed., Perspectives on Victimology. Beverly Hills, Calif.: Sage Publications.

1986 Gender expectations and their impact on black female offender and victims. Justice Quarterly 3(2):305-327.

Zatz, M.J., and J. Hagan 1985 Crime, time and punishment: An exploration of selection bias in sentencing research . Journal of Quantitative Criminology 1(1):103-126.

Zingraff, M.T., and R.J. Thomson 1985 Differential sentencing of women and men in the U.S.A. International Journal of the Sociology of Law 12:401-413.