The Invisible Woman. Joanne Belknap
Ackerman, Schwartz, & Agha, 2006; T. Stevens, Morash, & Chesney-Lind, 2011). According to UCR data, girls and women were 10% of arrests in 1965, 14% in 1980, 16% in 1990, 20% in 2000 (Steffensmeier & Schwartz, 2004), 26% in 2011, and, as we can see in Table 4.1, 30% in 2018. In other words, between 1965 and 2018, women and girls’ percentage of all arrests increased 270%. However, given that this was only a 1% increase in the last decade, the convergence could be leveling into stability. At the same time that we acknowledge indications of gender convergence, it is also necessary to remember that for the 28 offenses listed in Table 4.1, all were solidly male-gender-related except larceny-theft (approaching-male-gender-related for combined ages), embezzlement (gender-neutral for combined ages), prostitution and commercialized vice (female-gender-related for combined ages, approaching-male-gender-related for youths), and liquor law violations (approaching-male-gender-related, but only for youths).
Research Assessing the Three Steps.
The “percentage 10-year change” columns in Table 4.1 indicate that for Step 1, the gender convergence is at least in part due to men and boys’ arrest rates decreasing at a steeper rate than women and girls’ arrest rates are decreasing. This is consistent with many other studies analyzing UCR, NVCS, and other data sets (Kruttschnitt et al., 2008; Lauritsen et al., 2009; Rennison, 2009; J. Schwartz & Beltz, 2018).
Regarding Step 2 since the 1990s, arrest data more consistently support gender convergence; the victimization data, and especially self-report offending data, indicate far more gender stability and in some cases gender divergence, particularly for violent offending (e.g., Brener et al., 1999; Chesney-Lind & Belknap, 2004; Kaufman, 2009; Lauritsen et al., 2009; Marcus, 2009; Rosenfield et al., 2006; J. Schwartz et al., 2009; Stevens et al., 2011). Stevens and her colleagues (2011) found that although UCR data indicated significantly increased official involvement of African American boys and all girls (but particularly African American girls), self-report offending data indicated gender stability. Significantly, in their analysis of 30 years of NIBRS violent offense data, J. Schwartz, Conover-Williams, and Clemons (2015) confirmed their “consistent sex-stratification hypothesis,” that the gender composition of violent crimes has not changed over time. Specifically, the gender gap among the violent crimes stayed consistent over time (1% of homicides, 2%–3% of robberies, and 11% of felony assaults). This impressive study suggests that NIBRS data are preferable to UCR data for assessing gender–crime patterns, which is not unexpected as NIBRS police data are lauded as more accurate than UCR police data (as noted previously).
Recall that Step 3 is accounting for changes in economic, social, and political conditions, including net widening, when analyzing gender–crime patterns. Three important studies published in the 1980s countered the official (police data) data of the time indicating that women were committing more property offenses (gender convergence). First, Box and Hale’s (1983, 1984) studies of crime patterns in England and Wales from 1951 to 1980 found women’s “increasing economic marginalization” (along with less chivalrous police and court treatment, consistent with Step 3) was driving their upward theft crimes (Box & Hale, 1983, p. 43). Second and similarly, Chilton and Datesman’s (1987) analysis of the larceny arrests in the five largest U.S. cities from 1960 to 1980 noted “the most plausible explanation” for gender convergence was “the worsening economic situation of young Black women in older U.S. central cities” (p. 152). This work evolved into the economic marginalization hypothesis and the feminization of poverty, to account for relative disadvantage in assessing crime trends. Campaniello and Gavrilova’s (2018) analysis of 1995 to 2015 NIBRS property crime and robbery data reported that women and girls committed 30% of crimes but controlled for incentives to commit crimes (i.e., criminal earnings and probability of arrest). Women and girls on average, earned 13% less than men and boys and had a 9% lower risk of arrest. Their analysis concluded no gender differences in responses to changing arrest rates, but men and boys “respond more to changes in illegal earnings,” which explained 8% of the gender gap (p. 289).
In terms of the net-widening component of Step 3, Schwartz and colleagues’ (2015) analysis of 30 years of NIBRS data (referred to in the previous paragraph) found only one indication of gender convergence, “a small but steady 5% cumulative increase over 13 years in the share of all-female simple assault incidents (15%-20%),” which they noted “ coincide with implementation of mandatory and prearrest policies for minor violence in which women typically have engaged” (p. 75). Another example of Step 3 net widening can be seen when accounting for the practice “charging up” or “up-criming”—policies that relabeled some status crimes such as “child in need of supervision” or a simple assault as a “delinquent act,” including an aggravated assault (Acoca, 1999; L. M. Brown, Chesney-Lind, & Stein, 2007; Cauffman, 2008; Chesney-Lind & Belknap, 2004; Javdani et al., 2011; Kruttschnitt et al., 2008; Steffensmeier et al., 2005; T. Stevens et al., 2011). Examples include mothers calling the police because their daughters threw a Barbie doll or cookie at them. Parents are more likely to call the police on their daughters than their sons, and sometimes for actions such as the girl running away because of the mother’s boyfriend’s inappropriate and/or illegal behavior (Lederman & Brown, 2000).
Up-criming can also be seen in many school “no tolerance” practices where less serious infractions, as well as youth-on-youth fighting that was formerly handled by school officials, now involve the police and courts (Lederman & Brown 2000; Stevens et al., 2011). Moreover, up-criming policies are hardest hitting in communities of Color and economically marginalized communities where youths are already more heavily monitored by the police and other officials (Annamma et al., 2019; L. M. Brown et al., 2007; Gion et al., 2018; M. W. Morris, 2015; Stevens et al., 2011). Recall Stevens et al.’s (2011) study lauded in reporting on Step 2, where self-report data indicated gender stability when UCR did not. Consistent with Step 3, they further attributed the disproportionately harsher CLS responses to racist and sexist enforcement of policy changes (e.g., domestic violence policies and in-school “zero tolerance” policies), whereby African American boys and all girls, but particularly African American girls, were treated disproportionately harshly. Thus, feminist scholars have been identifying the complicated and often sexist, racist, classist, and homophobic responses to the identification of girls and women’s use of violence as a result of victimization (e.g., Belknap, Holsinger, et al., 2012; L. M. Brown et al., 2007; Henriksen & Miller, 2012; N. Jones, 2008; Kruttschnitt et al., 2008; J. Miller, 2001, 2008; S. L. Miller, 2001; Potter, 2008; Richie, 2012; J. Schwartz et al., 2009; Stevens et al., 2011).
A final example of the net widening in terms of Step 3 is DUI arrests. Although men’s driving under the influence (DUI) rates outpace women’s by a “sizeable” gap, where the more serious the DUI, the wider the gender gap (women were 25% of the DUI arrests but 18% of DUI traffic fatalities) (J. Schwartz & Beltz, 2018, p. 10), DUI arrest data since the 1980s support gender convergence (A. A. Robertson, Liew, & Gardner, 2011; J. Schwartz & Beltz, 2018; J. Schwartz & Rookey, 2008). Three studies attribute the DUI gender convergence, at least in part, to states’ changes in DUI laws and law enforcement, primarily hightened police enforcement and lowering the DUI blood alcohol content (BAC) maximum from 0.10% to 0.08%. Such BAC-lowering laws disproportionately affect women because they generally weigh less than men, so they can more easily “hit” the lower BAC level and thus are more likely then men to get snared in this net-widening policy (A. A. Robertson et al., 2011; J. Schwartz & Beltz 2018; J. Schwartz & Rookey 2008).
The Most Recent UCR Data and the Gender–Crime Gap 2009–2018.
Table 4.1 includes a gendered account of U.S. arrest (using UCR data) trends over a decade, from 2009 through 2018. Total females’ (women and girls’) and males’ (men and boys’) percentage changes over this time period are reported, as well as girls’ and boys’ (under the age of 18) percentage changes, separately. “Eyeballing” these data suggests that, consistent with most of the gender–crime gap research since the 1990s, gender convergence is the most common pattern. Again, keep in mind the limitations of UCR data and that almost all offending, measured by UCR data (arrests), were male-gender-related. Table 4.1 also indicates that 24 of 28 individual offense types, as well as the composite violent and property index crimes, were male-gender-related.