The Invisible Woman. Joanne Belknap
under the violent crime index.”
cSex offenses other than rape and prostitution.
Source: U.S. Department of Justice. (2019). Crime in the United States 2018: Uniform Crime Reports. Washington, DC: Federal Bureau of Investigation, U.S. Government Printing Office. https://ucr.fbi.gov/crime-in-the-u.s/2018/crime-in-the-u.s.-2018. Data for this table are directly calculated from Tables 33 and 42.
Now turning to the nature and extent of offending, of 2018 U.S. arrests, the index crimes are divided into violent and property crimes (and even further broken down into the specific violent and property index crimes) and the non-index offenses (Table 4.1). First, women and girls account for about 21% of the composite violent index offense arrests and almost 37% of composite index property crime arrests, making both of these composite measures (overall violent and overall property index crimes) male-gender-related. Second, all individual violent index crimes, are male-gender-related with women and girls constituting 12% of murder/nonnegligent manslaughter, 15% of robbery, 24% of aggravated assaults, and 3% of rape arrests. In addition to the U.S. 2018 index crime arrests, Table 4.1 includes 20 non-index offense arrest rates. In sum, only four individual offenses were not solidly male-gender-related: (1) embezzlement was gender-neutral among combined ages (but male-gender-related among youth); (2) larceny-theft approached male-gender-related for combined ages (but was solidly male-gender-related among youths) and was the only violent or nonviolent index crime that was not solidly male-gender-related; (3) prostitution and commercialized vice were solidly female-gender-related for combined ages but, for the first time, approached male-gender-related among youths; and (4) liquor law violations approached male-gender-related among youths but were solidly male-gender-related among combined ages. It is remarkable that prostitution and commercialized vice have always been female-gender-related and indicates the vulnerability of boys to being sex workers/prostitutes, but it is necessary to remember that arrest rates do not necessarily reflect the actual offending rates.
Documenting and Assessing Gender Patterns in Offending Over Time
Four Options to Describe Gender–Crime Patterns Over Time
The gender patterns of crime rates over time received unprecedented attention with the advent of Adler’s (1975) and R. J. Simon’s (1975) women’s “liberation” emancipation hypothesis (WLEH) (described in Chapter 2). “Moral panic” in the last third of the 20th century (about the time of the WLEH), whereby fears about advancing gender equality resulted in harsher policies targeting women and girls (Kruttschnitt, Gartner, & Hussemann, 2008). Figure 4.2 portrays the possible options in assessing gender patterns of offending over time. Gender stability (Option A) is any pattern where the gender rates are stable over time. Stated alternatively, they covary: rising, falling, and staying flat together1. For example, we might expect that in an era of “get-tough-on-crime” policies or in times of economic depression, men’s and women’s crime rates would be equally affected—unless the policy of economic hardship was likely to affect one gender more than another. With gender divergence (Option B), the gender gap widens over time, meaning gender differences/gaps in crime rates are increasing. Gender convergence (Option C), consistent with WLEH, describes any time that the gender gap is decreasing, where gender–crime rates approach each other. A final possibility, no trend (Option D), is when there is no gendered trend or pattern over time.2
1 O’Brien (1999, p. 100) labels this phenomenon “co-integration,” where the gender trends maintain a linked relationship with each other over time.
2 Option D, “no trend” is from O’Brien (1999).
Figure 4.2 ● Examples of Comparing Gender Patterns Over Time
Three Steps to Assess, Interpret, and Explain Gender Convergence Findings
Defining the Three Steps.
When examining crime patterns over time, particularly gender convergence, it is useful to unpack the data with what other measurement and social, political, economic factors might be affecting any changes found over time or that differ across groups over time (e.g., gender differences, in our case). Step 1 is to examine the gender convergence patterns, even if “only” using official, say UCR, data, to determine whether it is because (1) male rates are decreasing at a faster pace than female rates(are decreasing), or (2) male rates are increasing at a slower rate than female rates (are increasing). Steps 2 and 3 are much more difficult. Step 2 stresses that using police data to assess gender–crime patterns is problematic given how many crimes are unreported and unknown to the police, that reporting to the police varies by the type of crime (e.g., gender-based abuse crimes are some of the most underreported), and arrest data may be a better indication of police disproportionately targeting and privileging individuals (e.g., racial profiling, policing neighborhoods differently, etc.) than the actual offending rates. Most gender gap crime research uses U.S. UCR (arrest) data, although NIBRS is increasingly used. Step 2, then, is to compare arrest rates (and other CLS-generated data, such as court convictions) with self-reported offending and/or victimization data (e.g., NCVS) for the same period. Indeed, most criminologists agree that self-report data are a far more accurate measure of the actual crime rate than arrest data, given the numerous crimes unknown and unreported to the police. Also, people are quite honest and consistent in their self-reports of offending (W. Pollock, Hill, Menard, & Elliott, 2016).
Step 3 is accounting for potential economic, social, and/or policy changes happening that might affect offending patterns and, for the purposes of this chapter, do so in a gendered way. Stated alternatively, when there is gender convergence documented by official statistics, it is important to assess how much is offender-generated (individuals committing more or fewer crimes) and how much is due to economic changes such as the feminization of poverty (increasing women’s “survival” offenses) and to changes in the behavior of social control agents and policy, also referred to as “law in action” (Harmon & O’Brien, 2011; J. Schwartz & Rookey, 2008; J. Schwartz, Steffensmeier, & Feldmeyer, 2009). A key aspect of Step 3 has to do with net widening—CLS policies or practices that define and include more behaviors as offenses. Given that net widening typically involves including more minor behaviors as delinquent or criminal, and women and girls disproportionately commit the more minor offenses, net widening is likely to result in a higher percentage of female than male entanglement in the CLS, indicating gender convergence when actual offending has not changed but responses to it have (Harmon & O’Brien, 2011; Javdani et al., 2011; J. Schwartz & Rookey, 2008).
In contrast with the WLEH, most research addressing gender–crime patterns, until the 1990s, reported a strong tendency toward gender stability, that is, finding women’s crime rates basically stayed the same except in the areas of less serious property crimes and drugs (Boritch & Hagan, 1990; Canter, 1982; Chilton & Datesman, 1987; Giordano, Kerbel, & Dudley, 1981; Kruttschnitt et al., 2008; Leonard, 1982; Naffine, 1987; Steffensmeier, 1993; Steffensmeier & Cobb, 1981; Steffensmeier & Steffensmeier, 1980; Steffensmeier & Streifel, 1992). An exception is a UCR study of youth arrest rates during the 1960s and early 1970s that found the larceny offenses by girls increased by 250%, which the authors attributed to “baby boomers” hitting the high-risk offending ages (Chesney-Lind & Shelden, 1992).
Studies published since the 1990s using arrest data (primarily the UCR) are mostly consistent with gender convergence; that is, evidence showed an increasingly larger percentage of arrests among girls and women (e.g., Brener, Simon, Krug, & Lowy, 1999; Chesney-Lind & Belknap, 2004; Kaufman, 2009; Lauritsen, Heimer, & Lynch, 2009; Marcus, 2009; Rosenfield, Phillips, & White, 2006; J. Schwartz