Situational Action Theory’s Self-Control/Morality Interaction Effects and the Moderating Influence of Being Female: A Comparison of Property and Violent Offending Using a Sample of Juvenile Delinquents

Abstract

This study tests the interaction effects between self-control and morality that are proposed by situational action theory (SAT) and examines the ability of biological sex to condition those processes for both property and violent offending. This study employs negative binomial regression to analyze data from the Pathways to Desistance data set (n = 1,354). Results generally support the posited general nature of SAT for violent offending. The results for property offending were also supported; however, some of the results indicated that sex may moderate the associations of SAT’s key variables for this offense type. Our results indicate that the assumption of generalizability across the sexes may be less problematic for SAT than for other theories, but more work is needed to fully articulate how sex affects the processes at work in the theory.

Keywords

situational action theory gender self-control morality criminal propensity

Situational action theory (SAT) is a relative newcomer to the theoretical landscape in criminology. Tests of the theory have generally supported its main propositions (Antonaccio & Tittle, 2008; Wikström, Oberwittler, Treiber, & Hardie, 2012; Wikström & Svensson, 2010), but scholars are just beginning to evaluate the gendered nature of some of the key variables, most notably self-control and morality, and the possibility that the theory may work differently for males and females. This is an important area of inquiry because initial presentations of SAT include no explicit discussion of potential sex differences relevant to the theory. Early tests of the theory done by Wikström et al. (2012) use data containing an evenly divided sample of males and females, but stop short of exploring whether the theory has different implications for males than for females.

Mainstream criminological theories are frequently criticized for assumptions of sex neutrality, or generalizability (Arditti & Few, 2006; Reisig, Holtfreter, & Morash, 2006). Indeed, traditional theories overwhelmingly posit that their models explain both male and female criminality, leading to conclusions about female offending that are based on findings from data sets comprised solely of male subjects. For example, both social bonding and social learning theories have been criticized on these grounds (Holtfreter & Cupp, 2007; Morash, 1999). Such omissions often are acknowledged and dismissed, as illustrated by Hirschi (1969), who, in his original test of social bonding theory, did not assess the role of sex. Indeed, while he indicated that “ . . . girls have been neglected for too long by students of delinquency . . .,” he ultimately excluded them from his analyses, even while admitting that “the exclusion of them is difficult to justify.”

Only recently conceived, it is notable that SAT has followed the path of early criminological theories, as it did not initially address the potential for differences between boys and girls. Nor has most research drawing from or testing SAT considered gendered aspects of the model. Notably, Weerman, Bernasco, Bruinsma, and Pauwels (2016), as well as a study done by Hirtenlehner and Treiber (2017), provide important exceptions. These two studies shed light on the gendered aspects of SAT; however, Weerman et al. (2016) did not evaluate the interaction effects proposed by the theory, and Hirtenlehner and Treiber (2017) only explored the propensity-criminogenic exposure interaction as it related to shoplifting. This study expands the SAT literature by providing a test of the self-control–morality interaction effects, exploring these processes as they relate to both property and violent offending, and exploring the possibility of biological sex conditioning these effects. We argue that a number of elements of SAT are inherently gendered in nature, and as such, identify SAT as a theoretical framework that may be useful in expanding the understanding of the gendered nature of crime.

Literature Review

SAT

The basic propositions of SAT are depicted in Figure 1. The primary proposition of SAT is that the likelihood of criminal behavior is determined by the convergence of different kinds of people (i.e., differing criminal propensities) with different kinds of settings (Wikström et al., 2012). Wikström et al. (2012) argue that a person’s moral rules and level of self-control will be determined by background variables such as race, age, sex, and socio-economic status (SES). (They refer to these variables as the causes of the causes.) The interaction between morality and self-control yields a person’s criminal propensity¹ (Wikström & Svensson, 2010). When a person with a high criminal propensity finds themselves in a setting with criminogenic characteristics, crime becomes a likely outcome. SAT further specifies that for crime to happen, a person must first see crime as an actionable alternative. How a person evaluates alternatives that are actionable or not is conditioned by their criminal propensity. Given two settings in which criminal propensities are equal, the likelihood of criminal behavior is reduced in settings with fewer criminogenic characteristics (Hirtenlehner & Treiber, 2017; Wikström et al., 2012; Wikström & Treiber, 2009).

Figure 1.

Graphical representation of SAT’s primary propositions.

The incorporation of morality is one of the unique aspects of the SAT model. While findings that higher levels of morality can inhibit criminal and delinquent behavior (Antonaccio & Tittle, 2008; Bachman, Paternoster, & Ward, 1992; Hannon, De Fronzo, & Prochnow, 2001; Mears, Ploeger, & Warr, 1998; Paternoster & Simpson, 1996; Rogers, Smoak, & Jia, 2006; Thurman, 1984) are common in the criminological literature, the concept of morality is rarely explicitly incorporated into theoretical structures. Morality is most often inferred in theorizing. One example is when Braithwaite (1989) talks about the “pangs of conscience” a person feels after doing something that they feel makes them an object of shame (p. 74).

Wikström et al. (2012) posit that potential offenders undergo a perception-choice process. Certain people see crime as a valid alternative in certain settings while others do not. Whether or not a person sees crime as an actionable alternative is dependent on their criminal propensity. Thus, only a certain portion of the population will engage in the rational choice process (weighing costs and benefits) when deciding whether to commit a crime. This process is bypassed for those who do not see crime as a viable alternative. In sum, SAT recognizes that not everyone weighs costs and benefits in deciding whether to commit a crime because a substantial portion of the population does not see crime as a viable alternative to begin with, rendering the weighing process irrelevant (Braithwaite, 1989; Kroneberg, Heintze, & Mehlkop, 2010).²

SAT also deals with criminogenic exposure and lifestyle risk that are reminiscent of some ideas related to routine activity theory (Cohen & Felson, 1979; Svensson & Pauwels, 2010). Wikström et al. (2012) assert that crime should be analyzed as a situational phenomenon. They further argue that there are no personal or environmental influences that are sufficient on their own to cause crime, rather it is the interaction between personal factors and environmental factors that yields the probability of criminal behavior. While routine activity theory (Cohen & Felson, 1979) is one of the only theories to address crime at the situational level, it does not, due to the theoretical assumptions it makes about human nature, try to directly explain why an offender is motivated to commit crime in the way other theories do, nor does it explicitly define what crime is (Wikström, Ceccato, Hardie, & Treiber, 2010).

SAT also contains elements reminiscent of Gottfredson and Hirschi (1990), general theory of crime, which identifies low self-control as the primary explanatory variable in the causation of crime. SAT uses the concept of self-control, positing that it interacts with morality to create a person’s criminal propensity. A difference in the treatment of self-control lies in the fact that SAT is more about a person’s ability to exercise self-control, not a measured level of self-control per se. While it is not entirely clear if these are truly two distinct concepts, it is worth noting that in the original data collection, Wikström et al. (2012) used a modified version of the Grasmick scale as a measure of a person’s ability to exercise self-control. This would seem to make it likely that the conception of self-control used by Wikström et al. (2012) has more similarities to that used by Gottfredson and Hirschi (1990) than differences.

The interaction of morality and self-control has been previously examined by some researchers. Antonaccio and Tittle (2008) found that morality and self-control were both significant predictors of crime, but morality emerged as a more important factor and withstood the addition of control variables better than self-control in their regression models. In addition, Svensson, Pauwels, and Weerman (2010) assessed the interaction between morality and self-control in predicting offending, reporting support for the SAT position that self-control is a more powerful predictor of offending for adolescents with low morality compared to those with higher morality. Wikström and Svensson (2010) reported additional support, concluding that self-control was important only as a predictor for lower morality groups. They attributed this to the assumption that the high morality group did not see crime as an actionable alternative. Finally, Kroneberg et al. (2010) provided further support, determining that incentives to commit shoplifting or tax fraud were far less influential for individuals with high levels of morality than those with low morality levels.

The key proposition of SAT is that a person’s criminal propensity (i.e., the interaction of morality and self-control) will interact with their environment to predict the likelihood of criminal involvement (Wikström et al., 2012). Exposure to a criminogenic setting will be more influential for those with a higher criminal propensity than those with a lower criminal propensity. Studies have found, consistent with the theory, that subjects with higher criminal propensity were more likely to consider criminal involvement when exposed to a higher risk setting than those with a lower criminal propensity exposed to similar conditions (Svensson & Pauwels, 2010; Wikström et al., 2010; Wikström & Svensson, 2008; Wikström & Treiber, 2009).

To date, several studies have tested SAT’s primary propositions (Antonaccio & Tittle, 2008; Kroneberg et al., 2010; Pauwels & Svensson, 2009, 2010; Svensson & Pauwels, 2010; Svensson et al., 2010; Wikström et al., 2010; Wikström, Oberwittler, Treiber, & Hardie, 2012; Wikström & Svensson, 2008, 2010; Wikström & Treiber, 2007, 2009; Wikström, Tseloni, & Karlis, 2011). Nearly all of these studies use the data Wikström collected as part of the Peterborough Adolescent and Young Adult Development Study (PADS). These data include equal numbers of males and females, but, as noted previously, Wikström et al. (2012) did not explore the generalizability of SAT for both males and females in their initial published tests.

SAT and the effects of biological sex

Social systemic factors and processes, such as race, sex, SES, and family structure, are described by SAT as “causes of the causes” (Wikström et al., 2012, p. 12). Wikström et al. (2012) assert that such systemic variables have indirect effects on crime through self-control, morality, and exposure to criminogenic settings; however, these factors are not characterized as causes of crime. Pauwels and Svensson (2009) found weak direct effects for family structure on lifestyle risk (criminogenic exposure), which may only partially support SAT’s assertion that these background variables are not directly causal for crime. Wikström et al. (2012) argue that systemic background variables affect the development of one’s ability to exert self-control and the type of moral rules to which one adheres. These variables are antecedent to the posited direct causes of criminal behavior: criminal propensity and exposure to criminogenic settings (Wikström et al., 2012). As SAT relates to biological sex, the theory would assert that sex differences in self-control and moral rules would be observed, thus explaining differences in offending patterns.

While Wikstrom and colleagues acknowledge that these social systemic factors, including biological sex, are important, they are seldom treated in any detail in the existing SAT literature. This state of affairs has limited our understanding of the importance of sex as a driving factor in criminal offending, as it relates to SAT. Belknap and Holsinger (2006) noted that sex can intersect with race, sexual identity, and even age in explaining the risk factors associated with delinquency. A detailed examination of the effects of sex is also important because research indicates that several of the key SAT constructs are themselves gendered. Scholars have noted that low self-control has similar effects on the likelihood of offending across males and females, although females generally exhibit significantly higher levels of self-control than males do (Gibbs, Giever, & Martin, 1998; Harrison, Jones, & Sullivan, 2008; Lagrange & Silverman, 1999; Özbay, 2008; Tittle, Ward, & Grasmick, 2003). Lagrange and Silverman (1999) found in their study sample that girls had higher levels of self-control than boys. Offending was driven primarily by self-control and opportunities to offend, but sex remained a significant predictor net of the effects of those two constructs. The differences between males and females that explain the gender gap in offending are not entirely clear, but it is clear that self-control is an important predictor and that girls tend to have more of it than boys.

There are also qualitative differences in how males and females evaluate moral questions. Gangl (2010) has highlighted the fact that unchangeable characteristics of people, such as biological sex, place them in certain socialization patterns and opportunity structures. This in turn affects the type and extent of moral education that people receive during adolescence. This gives reason to believe that sex-based differences would be found when examining the moral rules by which boys and girls evaluate moral questions. Other scholars have observed that females tend to evaluate moral questions from a caring perspective that evaluates who will get hurt by certain decisions whiles males tend to operate from a rule-based perspective when evaluating moral issues (Gilligan, 1982; Jackson, Zhao, Witt, Fitzgerald, & von Eye, 2009; Mitchell, 2002). Although morality is predictive of delinquency for boys and girls, it is possible that sex influences the evaluation of moral questions and could affect the ability of any theory incorporating morality to predict offending. Because morality and self-control are key variables in SAT, such findings suggest that a gendered assessment of the theory is important to determine if it is equally predictive for males and females and whether these differences in self-control and methods of evaluating moral questions are likely to lead to differences in criminal propensity between males and females (Weerman et al., 2016).

Associations with delinquent peers are another consideration that may affect SAT in light of documented differences between boys and girls in this area. Research indicates that females tend to have fewer delinquent friends than do boys (Mears et al., 1998; O’Donnell, Richards, Pearce, & Romero, 2012; Weerman et al., 2016; Weerman & Hoeve, 2012) which is one possible explanation for differential opportunities for offending. For adolescents, the presence of delinquent friends can be one important characteristic of a criminogenic setting. If girls have fewer delinquent friends, they may be exposed to fewer criminogenic settings than boys. Since the characteristics of particular settings are at the heart of SAT, differences between boys and girls in exposure to criminogenic settings are important to consider. Examining this aspect of SAT may shed some light on explanations for the gender gap in offending.

A number of SAT studies that included sex as a control variable reported that it was a significant predictor (Antonaccio & Tittle, 2008; Pauwels & Svensson, 2010; Svensson & Pauwels, 2010); however, these studies did not elaborate on this particular finding. In contrast, Pauwels and Svensson (2009) focused on sex differences to predict lifestyle risk using family structure, social control, and criminal propensity as independent variables. Pauwels and Svensson (2009) found that their independent variables acted as significant predictors of lifestyle risk for both boys and girls, but the magnitude of these effects differed significantly between the groups. Examinations of why such differences existed were not further explored.

Most popular criminological theories are constructed under the assumption of generality. In other words, the mechanisms and hypothesized effects are thought to operate the same for everybody. Research (Gibson, Ward, Wright, Beaver, & Delisi, 2010; Holtfreter & Cupp, 2007; McCarthy, Felmlee, & Hagan, 2004; Morash, 1999; Nofziger, 2010; Özbay & Özcna, 2008) has identified that this assumption may not hold up in a variety of theories when it comes to the effects of biological sex. A review of the literature on SAT reveals that some work has examined the theory’s implications for boys and girls separately (Pauwels, 2012; Schils & Pauwels, 2014). A notable effort was recently conducted by Weerman et al. (2016). In their study, they used four variables as key indicators for SAT: morality, self-control, unsupervised activities with peers, and the presence of rule-breaking peers. All four of these variables were significant predictors of delinquency for both boys and girls in the initial analyses presented. Additional analyses, which took lagged effects into account, revealed that “boys and girls differ in which effects are statistically significant, but there are no significant differences between boys and girls in the magnitude of the effects” (Weerman et al., 2016, p. 1202). Given these findings, Weerman and colleagues concluded that there was support for the assertion that SAT is a general theory. Even in light of this conclusion, Weerman et al. (2016) note that their test of the theory is only a partial test because they did not investigate the interaction effects proposed by SAT.

Another recent study examining male/female differences related to SAT was conducted by Hirtenlehner and Treiber (2017). This study examined the ability of SAT to explain the differential involvement in shoplifting of boys and girls. The authors concluded that observed differences in the criminal propensity-criminogenic exposure interaction were sufficient to explain involvement in shoplifting for both sexes. This led to the conclusion that SAT appears to generalize well to both boys and girls.

The Current Study

This study expands on the work of Weerman et al. (2016) and Hirtenlehner and Treiber (2017) in two ways. First, this study tests the self-control–morality interaction effects posited by SAT and examines whether or not those interaction effects are conditioned by the influence of sex. Second, this study examines violent and property offending separately. The study conducted by Weerman et al. (2016) used total delinquency frequency as the dependent variable. Their variable included a range of property and violent offenses, both minor and serious. The study done by Hirtenlehner and Treiber (2017) only examined shoplifting. By analyzing property offending and violent offending separately, it is possible that nuanced differences as to the etiology of these types of offenses, concerning SAT constructs, may manifest themselves. This study will provide a test of whether or not the interaction of self-control and morality significantly predict delinquency. In addition, a three-way interaction between self-control, morality, and sex will be used to determine if biological sex influences these associations.

Data and Methods

The data used for this study were from the Pathways to Desistance project, a multi-site, longitudinal study of serious adolescent offenders. Data collection occurred between November, 2000, and January, 2003, in Phoenix, Arizona and Philadelphia, Pennsylvania. Respondents were between 14 and 18 years old at the time of their offense and had been adjudicated guilty of a serious offense (predominantly felonies, with a few exceptions for some misdemeanor property offenses, sexual assault, or weapons offenses). Each respondent was followed for a period of 7 years, yielding multiple waves of data over that time period (Mulvey, 2012). In the current study, independent variables are drawn from the first wave of data, while dependent variables are drawn from the second wave of data, which was collected 6 months later. This approach aids in establishing causal order in the relationships observed in the analysis. The sample comprised 1,170 males and 184 females (n = 1,354), reflecting the reality that far fewer females are adjudicated delinquent in the juvenile courts than males.

Dependent Variables

The Pathways data set contains self-reported offending items on 24 illegal behaviors. Exploratory factor analysis was used to identify groups of variables that accurately represent the constructs of violent offending and property offending. Detailed information regarding coding of the independent and dependent variables used in the analyses are given in Table 1.

Table 1.

Descriptive Statistics for Dependent and Independent Variables Used in Analysis.

Variable	Coding	M	SD
Self-Control Scale	(1) False, (2) Somewhat False, (3) Not Sure, (4) Somewhat True, (5) True (indicates highest self-control)	3.03	0.72
Moral Disengagement	(1) Disagree, (2) Neither agree nor disagree, (3) Agree (highest moral detachment)	1.62	0.35
Neighborhood Conditions	(1) Never, (2) Rarely, (3) Sometimes, (4) Often (highest disorder)	2.35	0.75
Unsupervised Routine Activities	5-point Likert-type scale, 1 = never, 5 = almost every day (most unstructured)	3.83	0.85
Religiosity/Spirituality Scale	1 = not at all true, 2 = not very true, 3 = somewhat true, 4 = pretty true, 5 = completely true (highest religiosity)	3.07	1.04
Lives with Bio. Father	0 = father absent 1 = father present	0.23	0.42
Parents Education Level	1 = grade school or less, 2 = some high school, 3 = high school diploma, 4 = trade school/some college/2-year degree, 5 = college graduate, 6 = graduate school	4.30	0.95
Age	Age is truncated to whole numbers R = 14-18	16.04	1.14
Race/Ethnicity	White (20.2%), Non-White (79.8%)
Female	Sex coded as: Male = 0 (86.4%), Female = 1 (13.6%)
Violent Offending	(Time 1) Count Variable	2.49	14.33
Property Offending	(Time 1) Count Variable	6.56	35.98
Violent Offending (time 2)	Scale includes shooting someone, shooting at someone, robbery with a weapon, and beating someone.R = 0-32	0.45	1.87
Property Offending (time 2)	Scale includes burglary, shoplifting, and illegal use of checks or credit cards.R = 0-195	1.07	8.30

Violent offending (time 2)

This count variable comprised items identifying how many times during the 6-month recall period a respondent shot at someone (whether the bullet struck its target or not), committed a robbery with a weapon, or beat up a person to the point that they required medical attention. Scores ranged from 0 to 32 incidents (M = 0.45, SD = 1.87).

Property offending (time 2)

This variable taps how many times during the 6-month recall period a respondent committed a burglary (entered a building with intent to steal), shoplifted, or illegally used checks or credit cards. Scores ranged from 0 to 195 incidents (M = 1.07, SD = 8.3).

Independent Variables

Independent variables in this study were drawn from the first wave of the Pathways data. Key theoretical variables include self-control, moral disengagement, neighborhood conditions, unsupervised routine activity, and religiosity/spirituality. In addition, a number of control variables are included in these models.

Self-control scale (time 1)

The Pathways data do not contain a direct measure of self-control, but do include four separate measures for the scales included in the Weinberger Adjustment Inventory (WAI; Weinberger & Schwartz, 1990). The WAI was developed as an indicator of self-restraint. Self-restraint is conceptualized as a person’s ability to “inhibit immediate self-focused desires in the interest of promoting long-term goals and positive relations with others” (Feldman & Weinberger, 1994, p. 196). These scales measure this construct by asking subjects to rate how true or false certain statements are about their own behavior. Included items tap the constructs of impulse control (e.g., “I say the first thing that comes into my mind without thinking enough about it”), suppression of aggression (e.g., “People who get me angry better watch out”), consideration of others (e.g., “Doing things to help other people is more important to me than almost anything else”), and temperance (combines items from impulse control and suppression of aggression). While perhaps not an ideal measure of self-control, these constructs are consistent with other measures that include considerations of impulsivity, forward thinking, and a general concern for others. For the purposes of this study, the four items representing those constructs were combined into a scale as a proxy for a direct measure of self-control. Exploratory factor analysis confirmed that these four items load onto a single factor. Reliability analysis also produced satisfactory results (α = .80).

Moral disengagement (time 1)

The Pathways data set includes a measure of moral disengagement based on the Mechanisms of Moral Disengagement instrument (Bandura, Barbaranelli, Caprara, & Pastorelli, 1996). This instrument has 32 items, with responses coded on a 3-point Likert-type scale ranging from disagree to agree, with agree representing higher levels of moral disengagement. This measure taps the constructs of moral justification, euphemistic language, advantageous comparison, displacement of responsibility, diffusion of responsibility, distorting consequences, attribution of blame, and dehumanization (Mulvey, 2012). While this variable is not a direct measure of a person’s moral beliefs, because of the dimensions it taps, it was deemed an appropriate measure to operationalize the concept of morality. Higher scores on this scale represent higher levels of moral disengagement (which would predict worse behavior), and a reliability analysis produced favorable results (α = .978).

Neighborhood conditions (time 1)

The neighborhood conditions variable is an adaptation developed for the Pathways data set from a study regarding neighborhood disorder done by Sampson and Raudenbush (1999). This measure captures the degree of physical disorder (e.g., litter, graffiti) and social disorder (e.g., adults fighting or arguing loudly, visible drug use) surrounding the respondent’s home. Responses were coded using a 4-point Likert-type scale, with higher numbers representing higher levels of disorder. This measure was included to partially tap the construct of criminogenic exposure.

Unsupervised routine activity (time 1)

This variable is an additional measure included to operationalize the construct of criminogenic exposure. The Pathways data draw on the “Monitoring the Future” questionnaire developed by Osgood, Wilson, O’Malley, Bachman, and Johnston (1996), which included three items tapping the quantity of time a juvenile spends in unstructured socializing with friends away from any authority figures. Although these items were not originally combined by Osgood et al. (1996), they were combined into a single scale (α = .623) by the researchers who compiled Pathways data (Mulvey, 2012).

Religiosity/Spirituality Scale (time 1)

Religiosity was identified by Antonaccio and Tittle (2008), along with SES and family structure, as being important considerations in a test of SAT because these constructs are likely antecedent to self-control and morality and most likely have effects on their development. The measures used in the Pathways data for this construct were originally developed by Maton (1989) and are recognized as valid and reliable measures (Maton et al., 1996). Five items are included in the Pathways data set. Respondents were asked to indicate the extent to which they agreed or disagreed with the following statements: “I experience God’s love and caring on a regular basis,” “I experience a close personal relationship to God,” “and religion helps me to deal with my problems.” Respondents also were asked to indicate how often they attended church in the past year and how important religion was in their life. Responses to all five items were recorded using a 5-point Likert-type scale, with higher numbers indicating a higher level of religiosity and spirituality. For this study, these five items were combined into single scale. Exploratory factor analysis confirmed that these five items loaded onto a single factor. A reliability analysis also produced satisfactory results (α = .84).

Lives with biological father (time 1)

This measure is included to control for the effects of family structure on the development of self-control and morality (Antonaccio & Tittle, 2008). The Pathways data report whether or not the respondent lives with their biological father. Although this is not an ideal measure of family structure, it does tap some of the broadest aspects this variable is attempting to capture. It does have the limitation, however, of not being able to tease out differences between a wide variety of possible family situations.

Parents’ education level (time 1)

Parents’ education levels were included as a proxy for SES. The Pathways data contain measures for the education level of a juvenile’s mother and father separately. This study uses a combined measure, taken directly from the Pathways data, which represents a mean of the two individual measures. Responses included six choices ranging from a grade school education up to and including attendance in graduate school (see Table 1 for details). The lowest education level was coded as 1, and the highest level was coded as 6.

Self-control/morality interaction term (time 1)

To test the hypothesized interaction between self-control and morality, a multiplicative interaction term was computed by multiplying the self-control scale variable with the moral disengagement scale variable. Before carrying out this computation, the self-control and morality variables were mean centered to reduce the effects of multicollinearity in the regression models (Aiken & West, 1991).

Age (time 1)

Controls for age are included in this analysis. The ages of the Pathways study participants ranged from 14 to 18 years old, with an average age of 16.04 (SD = 1.14).

Race/ethnicity

Controls were also included for race/ethnicity. Within these data, race is coded in four separate categories: White, Black, Hispanic, and other. These categories were dummy coded zero and one; the White category was the reference group for interpretation purposes.

Sex

The sex of the respondents is included as a control variable in the analysis (male = 0, female = 1).

Additional controls

To control for prior offending, variables mirroring the dependent variables were used from the first wave of data: violent offending at time 1 and property offending at time 1. A control was also added to account for differences between the two collection sites (Philadelphia, PA and Phoenix, AZ).

Analytic Strategy

Negative binomial regression is used to estimate the effects of the study variables on violent offending and property offending. This type of regression is well suited to count-based dependent variables with skewed distributions. To test the interaction of self-control and morality specified by SAT, two-way multiplicative interaction terms were computed and included in the statistical models.³ Three-way interaction terms were used to assess the influence of gender on SAT’s posited interactions.⁴

In terms of criminogenic exposure, SAT argues that crime patterns can be explained by the fact that “the social dynamics created by the differentiated urban environment, and related processes of social and self-selection, help create varying intersections of [different] kinds of people in [different] kinds of settings” (Wikström et al., 2012, p. 43). Some of these intersections of people and places are more likely to result in criminal activity than others. This study employed the use of two control variables intended to capture these dynamics as specified in the theory: neighborhood conditions and unsupervised routine activity.

Results

Prior to conducting the multivariate analyses, independent sample t-tests were performed to see how the males and females compared along the key variables being assessed. Those results are detailed in Table 2.

Table 2.

Means of Key Variables and t Test Results for Males and Females.

	Male mean	SE	Female mean	SE	Mean difference	p
Violent Offending (T2)	0.51	0.060	0.11	0.041	0.39**	<.001
Property Offending (T2)	1.06	0.257	1.15	0.545	−0.09	.897
Self-Control	3.01	0.021	3.10	0.057	−0.09	.110
Moral Disengagement	1.63	0.010	1.55	0.026	0.08**	.005
Neighborhood Disorder	2.35	0.022	2.35	0.83	0.00	.989
Unsupervised Socializing	3.85	0.024	3.67	0.072	0.18*	.019
Religiosity Scale	3.07	0.031	3.09	0.078	−0.02	.785
Lives with Bio. Father	0.23	0.012	0.28	0.033	0.05	.145
Parents Education Level	4.31	0.028	4.28	0.069	0.03	.667

p < .05. **p < .01.

Statistically significant differences were noted between the boys and girls for three of the key study variables: violent offending, moral disengagement, and unsupervised socializing. Girls reported lower levels of violent offending compared to boys as well as lower levels of moral disengagement compared to the boys in the sample. Boys reported higher levels of unsupervised socializing than the girls.

Negative binomial regression models were constructed in a stepwise fashion to see how the addition of key variables affected the outcome before assessing the final model. The regression results for violent offending are presented in Table 3, and the results for property offending are presented in Table 4.

Table 3.

Negative Binomial Regression Predicting Violent Offending (T2), Combined Sample (n = 1,204).

	Violent offending
	Model 1		Model 2		Model 3		Model 4
	b (SE)	IRR	b (SE)	IRR	b (SE)	IRR	b (SE)	IRR
Key theoretical variables
Female	−1.63*** (0.256)	0.20	−1.66*** (0.266)	0.19	−1.61*** (0.266)	0.20	−2.42*** (0.449)	0.09
Self-Control			−0.67*** (0.098)	0.51	−0.70*** (0.099)	0.50	−0.73*** (0.110)	0.48
Moral Disengagement			0.58*** (0.166)	1.78	0.87*** (0.197)	2.39	0.80** (0.234)	2.22
Neighborhood Disorder			0.08 (0.082)	1.08	0.06 (0.082)	1.07	0.08 (0.082)	1.08
Unsupervised Socializing			0.32*** (0.077)	1.38	0.31*** (0.077)	1.37	0.31*** (0.078)	1.36
SC × MD					0.68** (0.248)	1.98	0.48 (0.269)	1.62
SC × Sex							−0.44 (0.486)	0.65
MD × Sex							−0.81 (1.147)	0.45
SC × MD × Sex							−2.83* (1.137)	0.06
Control variables
Religiosity	−0.05 (0.053)	0.95	0.10 (0.057)	1.11	0.10 (0.057)	1.11	0.10 (0.057)	1.10
Lives with Bio. Father	−0.19 (0.130)	0.83	−0.04 (0.136)	0.96	−0.02 (0.137)	0.98	−0.05 (0.138)	0.95
Race (White)	−0.46** (0.151)	0.63	−0.46** (0.160)	0.63	−0.44** (0.160)	0.64	−0.44** (0.161)	0.65
Age	−0.04 (0.047)	0.96	−0.01 (0.050)	0.99	−0.01 (0.050)	0.99	−0.02 (0.051)	0.98
Parents Education (SES)	−0.16** (0.057)	0.82	−0.16** (0.060)	0.82	−0.16** (0.061)	0.83	−0.17** (0.061)	0.81
Violent Offending (Time 1)	0.04*** (0.009)	1.04	0.02 (0.008)	1.02	0.02* (0.008)	1.02	0.02* (0.008)	1.02
Philadelphia Site	−0.50*** (0.114)	0.61	−0.42** (0.124)	0.66	−0.40** (0.124)	0.67	−0.42** (0.124)	0.66

Note. SES = socio-economic status. IRR=incidence rate ratio. SC=self-control. MD=moral disengagement.

p < .05. **p < .01. ***p < .001.

Table 4.

Negative Binomial Regression Predicting Property Offending (T2), Combined Sample (n = 1,224).

	Property offending
	Model 1		Model 2		Model 3		Model 4
	b (SE)	IRR	b (SE)	IRR	b (SE)	IRR	b (SE)	IRR
Key theoretical variables
Female	−0.63*** (0.136)	0.54	−0.58*** (0.144)	0.56	−0.54*** (0.147)	0.58	−0.83*** (0.180)	0.44
Self-Control			−0.80*** (0.094)	0.45	−0.84*** (0.099)	0.43	−0.89*** (0.102)	0.42
Moral Disengagement			0.91*** (0.168)	2.49	1.53*** (0.202)	4.60	1.67*** (0.208)	5.30
Neighborhood Disorder			0.07 (0.078)	1.08	0.03 (0.080)	1.03	0.03 (0.081)	1.03
Unsupervised Socializing			−0.004 (0.066)	1.00	−0.01 (0.067)	0.99	−0.04 (0.067)	0.97
SC × MD					1.67*** (0.272)	5.31	1.76*** (0.280)	5.83
SC × Sex							0.42 (0.228)	1.53
MD × Sex							0.50 (0.515)	1.65
SC × MD × Sex							−2.62*** (0.670)	0.07
Control variables
Religiosity	−0.24*** (0.045)	0.79	−0.05 (0.050)	0.95	−0.05 (0.050)	0.96	−0.03 (0.052)	0.97
Lives with Bio. Father	−0.83*** (0.123)	0.44	−0.71*** (0.129)	0.49	−0.65*** (0.131)	0.52	−0.68*** (0.132)	0.51
Race (White)	0.73*** (0.103)	2.08	0.74*** (0.117)	2.10	0.81*** (0.119)	2.25	0.81*** (0.120)	2.25
Age	−0.19*** (0.039)	0.83	−0.14** (0.044)	0.87	−0.12** (0.044)	0.89	−0.11* (0.044)	.90
Parents Education (SES)	−0.43*** (0.052)	0.46	−0.54*** (0.057)	0.28	−0.54*** (0.058)	0.29	−0.55*** (0.059)	0.27
Property Offending (Time 1)	0.01*** (0.002)	1.01	0.01*** (0.002)	1.01	0.01*** (0.002)	1.01	0.01*** (0.002)	1.01
Philadelphia Site	−1.59*** (0.105)	0.20	−1.65*** (0.114)	0.19	−1.63*** (0.115)	0.20	−1.65*** (0.116)	0.19

Note. SES = socio-economic status. IRR=incidence rate ratio. SC=self-control. MD=moral disengagement.

p < .05. **p < .01. ***p < .001.

Violent Offending Results

Model 1 in Table 3 includes only the control variables and the female (sex) variable. As expected, being female is a significant predictor of reduced violent offending (b = −1.63, p < .001) as females committed far fewer violent offenses than their male counterparts. This is consistent with past research which indicates that the gender gap is the largest for violent offense types (Koons-Witt & Schram, 2003). Model 2 adds self-control, moral disengagement, neighborhood disorder, and unsupervised socializing. Three of these variables significantly predicted violent offending in the expected direction. Specifically, higher levels of self-control reduced violent offending (b = −0.67, p < .001) and higher levels of moral disengagement increased violent offending (b = 0.58, p < .001). Neighborhood disorder was not a significant predictor, but higher levels of unsupervised socializing significantly predicted higher levels of violent offending (b = 0.32, p < .001).

Model 3 adds the two-way interaction term between self-control and moral disengagement. As posited by SAT, this interaction is statistically significant in predicting violent offending. The significance of the interaction term indicates that the influence of either self-control or morality individually will depend on the value of the other. Notably, the self-control and morality variables remained significant, indicating that they are still contributing independent effects.

Model 4 includes the three-way interaction between self-control, moral disengagement, and sex. This interaction indicates whether the interaction between self-control and moral disengagement is working differently for boys than it is for girls. The interaction term was significant (b = −2.83, p < .05), indicating the possibility that the previously mentioned interaction is working differently across the sexes. At this stage, caution is warranted in drawing firm conclusions based on the significant three-way interaction terms since the interpretation of these interactions is not as straightforward as the evaluation of non-interaction terms in a regression model. A graphical representation of these interactions is provided later to allow for a closer examination of these effects.

Other noteworthy results include the analysis of the race and SES (parents’ education level) variables. Being White predicted a significant reduction in violent offending (b = −0.44, p < .01). Higher SES was also associated with lower levels of violent offending (b = −0.17, p < .01).

Property Offending Results

In many ways, the results for property offending are similar to the results obtained for violent offending. One difference lies with the unsupervised socializing measure, which significantly predicted violent offending but was not significant in the final property offending model.

In Model 3 in Table 4, the two-way interaction between self-control and moral disengagement is significant (b = 1.67, p < .001), consistent with the predictions of SAT. Sex, self-control, and moral disengagement also were significant in the expected directions. In the final model for property offending, the three-way interaction term was significant (b = −2.62, p < .001) net of all other effects in this model, just as for the violence model.

The results for SES were similar to the violence analysis; higher status yields lower levels of property offending (b = −0.55, p < .001). The results for race, however, were somewhat unexpected. Being White predicted significantly higher levels of property offending as compared to minorities (b = 0.81, p < .001).

Further Exploration of the Interaction Results

Statistical tests comparing males and females were not done as part of this study due to a lack of statistical power associated with the small size of the female sample (n = 184). Instead, graphs of the interaction effects were created to help visualize the effects more clearly and to stimulate scholarly thought on the processes at work in SAT. What follows is not intended to be a definitive statement about precise differences between males and females within the SAT framework. Instead, the figures provided are conceptual devices intended to help visualize the interaction effects tested in this study. The interaction graphs and the accompanying slope difference tests were constructed and computed in accordance with the methodology specified by Aiken and West (1991), Dawson and Richter (2006), and Dawson (2014).

Two-way interactions

Figures 2 and 3 graphically illustrate the two-way interactions tested in this study.

Figure 2.

Interaction effects of moral disengagement and self-control on violent offending.

Figure 3.

Interaction effects of moral disengagement and self-control on property offending.

Figure 2 shows the interaction effects taking place between self-control and moral disengagement as they relate to violent offending. As expected, the low self-control group offends at a higher rate than the high self-control group across all levels of moral disengagement. The more interesting result is that as moral disengagement increases, offending increases for both groups, but the effects appear to be more pronounced for the high self-control group, indicated by a steeper slope of the line.

Figure 3 details the analysis for the same two-way interaction between self-control and moral disengagement for property offending. The results for this dependent variable are similar with one key difference. Offending across different levels of moral disengagement for the juveniles with low self-control was largely unchanged. For juveniles with high self-control, however, an increase in offending is seen as moral disengagement increases. This is important because this shows that for property offending, which is much more common than violent offending, moral beliefs may not have much influence on behavior for those with low self-control. The influence of morality appears to be more salient for those juveniles with high self-control.

Three-way interactions

Figure 4 illustrates the three-way interaction between sex, moral disengagement, and self-control for violent offending.

Figure 4.

Three-way interaction effects of sex, moral disengagement, and self-control on violent offending.

As would be expected, females are represented by the bottom two lines on the graph that represent the lowest rate offenders, and boys are represented by the top two lines which represent the higher rate offenders. The rate of offending for girls with higher levels of self-control remains relatively stable across varying levels of moral disengagement. Boys with higher levels of self-control, on the other hand, exhibit a marked increase in offending as moral disengagement increases. Turning to the juveniles with lower levels of self-control, girls showed an increase in offending as moral disengagement increased, but boys’ levels of offending remained high and stable across varying levels of moral disengagement.

In examining the results of the slope difference tests, the only significant difference was seen in a comparison of males with high self-control and males with low self-control. There are no significant differences that emerged in any of the comparisons between males and females for violent offending. This is an unexpected result given the fact that the three-way interaction term in the regression model was significant. A possible reason for this may be that although there are significant intra-sex differences for males in the interaction between self-control and morality, the significant three-way interaction with sex may simply be a reflection of the gender gap in violent offending. If this is the case, the conclusion that SAT predicts violent offending for boys and girls effectively would be reasonable. The differences observed in this study may simply be reflective of the fact that boys commit violent offenses at a significantly higher rate than girls do. From the results presented here, it seems best to conclude that, for this study sample, SAT’s assumption of gender neutrality held up reasonably well as far as violent offending was concerned.

Figure 5 illustrates the three-way interactions between sex, moral disengagement, and self-control for property offending. The results here are somewhat different than those from violent offending in that the two groups with the lowest overall offense rate are boys and girls with higher levels of self-control (represented by the bottom two lines on the graph). For these two groups, the interaction between self-control and moral disengagement works similarly with an increase in offending as moral disengagement increases, and as indicated by the slope difference test, the strength of the graphed interactions for these groups are not significantly different from each other.

Figure 5.

Three-way interaction effects of sex, moral disengagement, and self-control on property offending.

Looking at the juveniles with lower self-control, girls exhibited an increase in offending as moral disengagement increased, but the boys actually exhibited a moderate decrease in offending as moral disengagement increased. This result for the boys from these data was unexpected because the expectation for all groups at higher levels of moral disengagement would be to exhibit a higher rate of offending as compared to the levels of offending associated with lower levels of moral disengagement.

Up to this point, SAT’s assumption of gender neutrality has been generally supported by the results of this study, but in examining the differences between male and female property offenders with low self-control, significant differences are indicated by the slope difference test. (Note that all slope difference tests were significant except for the comparison between boys with high self-control and girls with high self-control.) This result indicates, in part, that sex appears to significantly condition the interaction between morality and self-control for property offending when dealing with low self-control individuals.

Discussion

One of the primary hypotheses of this study was that biological sex would moderate the interaction of self-control and morality. Stated another way, the interaction of self-control and morality could yield different results when comparing boys and girls. To a certain extent, this may be what some refer to as a black box argument where the focus is only on inputs and outputs. Admittedly, this study is focused more on what is happening in the interaction of sex, self-control, and morality while less attention is given to why it is happening. Other researchers have examined this issue. Nofziger (2010) found that feminine traits predicted higher levels of self-control and reduced deviance while masculinity had no effect on deviant behavior. While Nofziger was dealing with the more complex issue of gender identity, as opposed to sex, these findings may shed light on some of the processes occurring in the “black box” of the three-way interactions explored in the current study. In addition, some of the studies already cited have dealt with the gendered nature of self-control and moral development and reasoning.

As originally formulated, SAT would explain differences in levels of offending in boys and girls as stemming from differing levels of criminal propensity (self-control and law relevant morality) and criminogenic exposure. In light of this, one may ask, “What effect does sex have on offending, over and above criminal propensity and criminogenic exposure?”⁵ In both of the full statistical models presented, being female was still a significant predictor of property and violent offending, even after accounting for the three-way interaction between sex, self-control, and morality. While SAT does seem to provide a sensible explanation for the gender gap in offending in predicting differential levels of self-control, morality, and criminogenic exposure between boys and girls, it seems clear that there is still more work to be done in determining all of the causal mechanisms behind this phenomenon.

Results of the regression analyses are generally supportive of SAT’s key propositions. Consistent with the assertions made by SAT is the persistent strength of the morality variable as a predictor of delinquency. Indeed, past research has found a person’s morality to be the more salient factor, as compared to self-control, when it comes to delinquency (Wikström & Svensson, 2010). Also consistent with SAT were the findings regarding the interaction term for self-control and moral disengagement. In Model 3, for both violent and property offending, this term was a significant predictor, the figures provided with graphs of this interaction indicate that it is working in the way SAT predicts. This means that the independent influence of either self-control or morality will depend on the value of the other. This is consistent with the conception of criminal propensity put forward by Wikström et al. (2012).

A key piece of analysis for the present study was the test of a three-way interaction between self-control, morality, and sex. The three-way interaction between these measures was a significant predictor in the models for both violent and property offending. When the three-way interaction with sex was added to the model, however, the two-way interaction between self-control and moral disengagement only remained significant in the property offending model. This suggests that the interaction specified by SAT between these two variables is not only conditioned by the effects of sex, but by the type offending as well.

The literature provides some cause to believe that sex influences the way self-control and morality might interact. In a study involving the gendered effects of self-control on delinquency, Nofziger (2010) found that feminine character traits were associated with higher self-control and lower rates of deviant behavior; however, that particular study did not assess whether self-control had equal predictive power for boys and girls. In a study examining the incidence of institutional misconduct among incarcerated juveniles, low self-control was only found to be predictive of increased misconduct among males but not females (DeLisi et al., 2010). A full review of the literature on the gendered nature of self-control is beyond the scope of this article, but these ideas taken together at least raise the possibility that self-control as a distinct measure may not predict female delinquency in the same way that it does for males. Although the end result may be similar, there may be other mechanisms at work. Further research is needed to determine what these mechanisms are, if they indeed exist.

Morality also has some nuance that should be considered. Wikström et al. (2012) argue as part of SAT that when a person intersects with a potentially criminogenic setting, there is a choice-perception process that occurs. In short, some people, due to differing levels of moral commitment to obeying the law, see criminal behavior as a legitimate possible course of action, and some do not. Wikström et al. (2012) further argue that for those who do not see criminal activity as a viable option in any circumstance, variables such as self-control become irrelevant to a certain degree because self-control is only required in the face of temptation to do something that one knows is wrong or illegal (Wikström & Svensson, 2010; Wikström & Treiber, 2007).

Another idea to consider, as it applies to biological sex conditioning the effects of SAT variables, is the finding of McCarthy et al. (2004) that female-dominated peer networks tend to act as a protective factor against delinquency. This has implications for the situational aspect of SAT. Although the data used for this study were not suitable for rigorous tests of differences between males and females due to the large disparity in sample size, one possibility to consider is that girls tend to be more likely to engage with female-dominated peer groups than boys (Benenson, 1990). In the present study, higher levels of unsupervised socializing predicted higher levels of violent offending, but this was not the case for property offending. Future research should examine whether or not the composition of peer networks influence the interaction effects specified by SAT.

A finding that seems to be contrary to the predictions made by SAT is seen with regard to the control variables. Wikström et al. (2012) argue that background variables (like race, sex, and age) are not directly causal for crime or delinquency. This means that these types of variables should not have any predictive power after taking a person’s criminal propensity and the characteristics of their present setting into account. In Model 4 for property offending, all of our control variables remained significant, net of SAT construct effects, with the exception of religiosity. For violent offending, religiosity, living with a biological father, and age were reduced to non-significance, while race, SES, and previous offending remained significant predictors. While the two models differ somewhat, it can be said that not all background variables were reduced to non-significance for either type of offending. This leaves room for further investigation of the question as to whether or not SAT constructs fully account for these “causes of the causes,” but in the analysis done for the present study, they did not.

The results of this study indicate that being female had a negative association with all types of offending net of SAT constructs. Sex was a significant predictor across all statistical models which highlights the need for future research to continue to explore the ways in which sex affects offending and delinquent behavior.

The idea that the interaction between self-control and morality would significantly predict violent and property offending net of the interactive effects with sex received partial support. The two-way interaction term was significant for both offense types prior to the addition of the three-way interaction with sex. When the three-way interaction term was added, the interaction between self-control and morality was reduced to non-significance in the violence model (see Table 3).

One research question in this study was whether a three-way interaction between self-control, morality, and sex would be statistically significant in the regression models. Although the results indicated statistical significance for our three-way interaction term, caution is warranted in the interpretation of this result. Although the three-way interaction term (Self-control × Moral Disengagement × Sex) was statistically significant in both the violent and property offending models, difference of slope tests indicated that there were no significant differences in the strength of the moderating effect of sex with regard to violent offending. For property offending, many differences did emerge with the exception of the comparison between boys and girls with high self-control. This indicates that although SAT’s assumption of generality appears to hold up well under many circumstances, there may be some nuanced between the effects observed for boys and girls in examining different types of criminal offenses.

Limitations of the Study

One limitation of this study lies in the nature of the sample used for analysis. Using a sample of juvenile delinquents may yield different results than using a community sample. While a delinquent sample could potentially exhibit less variation in some of the key independent variables as compared to a community sample, the variation observed in the data was sufficient for a meaningful analysis. Caution should be exercised, however, as to the generalizability of our results. We also recognize that although separate analyses were done for violent and property offending, these categories do not represent two distinct groups of offenders. Many of the individuals in the data set were involved in both types of offending. Another limitation of this study is that it was not able to make direct comparisons between males and females. This type of analysis would have been problematic given the disparity in the number of males (n = 1,170) and females (n = 184) in the data set. While this disparity in the number of males and females in the data present challenges for analysis and interpretation of results, these numbers reflect the reality that fewer females are adjudicated delinquent than males.

This study only tested the interactions between self-control and morality that are specified by SAT. While independent variables were included to account for criminogenic exposure, this study did not test the interaction between criminal propensity and criminogenic exposure. There are reasons to believe that sex may impact this association as well. This is another key element of SAT that should be examined in future research.

As with any study, the validity and reliability of measurements can be a concern. For example, the data employed in this study did not contain a direct measure of self-control, so a scale was constructed that was used as a self-control measure. The authors believe this issue has been mitigated to the extent possible in this project, but we acknowledge the possibility that some of the effects observed could be artifacts of measurement issues.

It is also worth noting that although our graphs depicting interaction effects are labeled with “low self-control” and “high self-control,” no specific cutoffs were determined. The graphs are merely conceptual tools that represent the variation within our study sample and may be different from that found in other samples. Another limitation of the study is that we did not examine how SAT constructs affect behavior over time with a detailed longitudinal analysis. This was beyond the scope of this article and remains for future researchers to address.

Theoretical Implications

The results of this study indicate that the assumption of sex neutrality may be somewhat less problematic for SAT than for some other criminological theories. The results of this study are generally supportive of the notion that SAT can serve as a general theory when it comes to predicting delinquency for both sexes. That said, some of the results suggest that the interaction of morality and self-control does not always manifest itself in an identical fashion for male and female juvenile offenders. Although some conditioning effects of sex were observed, the results may also depend somewhat on the type of criminal offenses that are the subject of analysis. The three-way interaction analysis showed that the difference in the way self-control interacts with moral disengagement was most pronounced when comparing boys and girls with low self-control who had committed property offenses. This seems particularly problematic for predicting crime as this involves the people who are likely to commit the most crime (those with low self-control) and the most common offense type (property crime). Whether or not these differences observed here are large enough to warrant a conclusion that SAT is not a general theory is an empirical question that is beyond the scope of this study. These results do suggest, however, that care should be taken in constructing dependent variables that are measures of criminal involvement. The differences found in this study for violent versus property crime suggest that these categories of crime may have different causes and may be impacted differentially by the key constructs measured in this study. We must, however, also acknowledge the possibility that the sex-based differences between what was observed in this study and what was hypothesized could be attributed to measurement error, the nature of the offender only sample, and/or the time lagged nature of the analytic strategy. Given this possibility, we would characterize the results of our study as being largely supportive of the general nature of SAT.

Although the findings of this study generally support SAT as a general theory, we would posit that more work remains to be done in exploring how biological sex affects the theory’s key variables of morality, self-control, criminogenic settings, and the interaction effects that exist between these constructs. We note, that biological sex or gender (as attributes) cannot be a cause of action but are merely markers of something else that could be.⁶ Future research should pursue this line of inquiry to refine the propositions of SAT so that they more fully account for the differences between the sexes documented in the literature.

Footnotes

Declaration of Conflicting Interests

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

Funding

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

Notes

Author Biographies

Glen A. Ishoy is an assistant professor in the Department of Criminology and Criminal Justice at Indiana University of Pennsylvania. His research interests include policing, decision-making processes in the criminal justice system, family processes and delinquency, and criminological theory. His work has been published in outlets such as Crime and Delinquency, Criminal Justice Ethics, and The Journal of Criminal Justice Studies. Prior to entering academia, he spent nearly 10 years as a police officer in a major metropolitan area.

Brenda Sims Blackwell is the chair of the Department of Criminal Justice and Criminology at Georgia Southern University. She has served on the faculties of Pennsylvania State University and Georgia State University. Her research broadly focuses on women and crime, with an emphasis on gender differences in offending and differences in criminal justice responses to male and female offenders.

References

Aiken

L. S.

West

S. G.

(1991). Multiple regression: Testing and interpreting interactions. Newbury Park, CA: Sage.

Antonaccio

Tittle

C. R.

(2008). Morality, self-control, and crime. Criminology, 46, 479-510. doi:10.1111/j.1745-9125.2008.00116.x

Arditti

J. A.

Few

A. L.

(2006). Mothers’ reentry into family life following incarceration. Criminal Justice Policy Review, 17, 103-123.

Bachman

Paternoster

Ward

(1992). The rationality of sexual offending: Testing a deterrence/rational choice conception of sexual assault. Law & Society Review, 26, 343-372. doi:10.2307/3053901

Bandura

Barbaranelli

Caprara

G. V.

Pastorelli

(1996). Mechanisms of moral disengagement in the exercise of moral agency. Journal of Personality and Social Psychology, 71, 364-374. doi:10.1037/0022-3514.71.2.364

Belknap

Holsinger

(2006). The gendered nature of risk factors for delinquency. Feminist Criminology, 1, 48-71.

Benenson

J. F.

(1990). Gender differences in social networks. The Journal of Early Adolescence, 10, 472-495.

Braithwaite

(1989). Crime, shame, and reintegration. New York, NY: Cambridge University Press.

Cohen

L. E.

Felson

(1979). Social change and crime rate trends: A routine activity approach. American Sociological Review, 44, 588-608.

10.

Dawson

J. F.

(2014). Moderation in management research: What, why, when, and how. Journal of Business and Psychology, 29, 1-19.

11.

Dawson

J. F.

Richter

A. W.

(2006). Probing three-way interactions in moderated multiple regression: Development and application of a slope difference test. Journal of Applied Psychology, 91, 917-926.

12.

DeLisi

Beaver

K. M.

Vaughn

M. G.

Trulson

C. R.

Kosloski

A. E.

Drury

A. J.

Wright

J. P.

(2010). Personality, gender, and self-control theory revisited: Results from a sample of institutionalized juvenile delinquents. Applied Psychology in Criminal Justice, 6, 31-46.

13.

Feldman

S. S.

Weinberger

D. A.

(1994). Self-restraint as a mediator of family influences on boys’ delinquent behavior: A longitudinal study. Child Development, 65, 195-211. doi:10.1111/1467-8624.ep9406130689

14.

Gangl

(2010). Causal inference in sociological research. Annual Review of Sociology, 36, 21-47. doi:10.1146/annurev.soc.012809.102702

15.

Gibbs

J. J.

Giever

Martin

J. S.

(1998). Parental management and self-control: An empirical test of Gottfredson and Hirschi’s general theory. Journal of Research in Crime & Delinquency, 35, 40-71.

16.

Gibson

C. L.

Ward

J. T.

Wright

J. P.

Beaver

K. M.

Delisi

(2010). Where does gender fit in the measurement of self-control? Criminal Justice & Behavior, 37, 883-903. doi:10.1177/0093854810369082

17.

Gilligan

(1982). In a different voice. Cambridge, MA: Harvard University Press.

18.

Gottfredson

M. R.

Hirschi

(1990). A general theory of crime. Stanford, CA: Stanford University Press.

19.

Hannon

De Fronzo

Prochnow

(2001). Moral commitment and the effects of social influences on violent delinquency. Violence & Victims, 16, 427-439.

20.

Harrison

M. L.

Jones

Sullivan

(2008). The gendered expressions of self-control: Manifestations of noncriminal deviance among females. Deviant Behavior, 29, 18-42. doi:10.1080/01639620701382907

21.

Hirschi

(1969). Causes of delinquency. Berkley: University of California Press.

22.

Hirtenlehner

Treiber

(2017). Can situational action theory explain the gender gap in adolescent shoplifting? Results from Austria. International Criminal Justice Review, 27, 165-187. doi:10.1177/1057567717690199

23.

Holtfreter

Cupp

(2007). Gender and risk assessment: The empirical status of the LSI-R for women. Journal of Contemporary Criminal Justice, 23, 363-382.

24.

Jackson

L. A.

Zhao

Witt

E. A.

Fitzgerald

H. E.

von Eye

(2009). Gender, race, and morality in the virtual world and its relationship to morality in the real world. Sex Roles, 60, 859-869. doi:10.1007/s11199-009-9589-5

25.

Koons-Witt

B. A.

Schram

P. J.

(2003). The prevalence and nature of violent offending by females. Journal of Criminal Justice, 31, 361-671. doi:10.1016/S0047-2352(03)00028-X

26.

Kroneberg

Heintze

Mehlkop

(2010). The interplay of moral norms and instrumental incentives in crime causation. Criminology, 48, 259-294. doi:10.1111/j.1745-9125.2010.00187.x

27.

Lagrange

T. C.

Silverman

R. A.

(1999). Low self-control and opportunity: Testing the general theory of crime as an explanation for gender differences in delinquency. Criminology, 37, 41-72.

28.

Maton

K. I.

(1989). The stress-buffering role of spiritual support: Cross-sectional and prospective investigations. Journal for the Scientific Study of Religion, 28, 310-323.

29.

Maton

K. I.

Teti

D. M.

Corns

K. M.

Vieira-Baker

C. C.

Lavine

J. R.

Gouze

K. R.

Keating

D. P.

(1996). Cultural specificity of support sources, correlates and contexts: Three studies of African-American and Caucasian youth. American Journal of Community Psychology, 24, 551-587.

30.

McCarthy

Felmlee

Hagan

(2004). Girl friends are better: Gender, friends, and crime among school and street youth. Criminology, 42, 805-835.

31.

Mears

D. P.

Ploeger

Warr

(1998). Explaining the gender gap in delinquency: Peer influence and moral evaluations of behavior. Journal of Research in Crime & Delinquency, 35, 251-266.

32.

Mitchell

(2002). Women’s morality: A test of Carol Gilligan’s theory. Journal of Social Distress & the Homeless, 11, 81-110. doi:10.1023/A:1013336206400

33.

Morash

(1999). A consideration of gender in relation to social learning and social structure: A general theory of crime and deviance. Theoretical Criminology, 3, 451-462.

34.

Mulvey

E. P.

(2012). Research on pathways to desistance [Maricopa County, AZ and Philadelphia County, PA]: Subject measures, 2000-2010. Retrieved from https://dx-doi-org.web.bisu.edu.cn/10.3886/ICPSR29961.v1

35.

Nofziger

(2010). A gendered perspective on the relationship between self-control and deviance. Feminist Criminology, 5, 29-50. doi:10.1177/1557085109353892

36.

O’Donnell

Richards

Pearce

Romero

(2012). Gender differences in monitoring and deviant peers as predictors of delinquent behavior among low-income urban African American youth. Journal of Early Adolescence, 32, 431-459.

37.

Osgood

D. W.

Wilson

J. K.

O’Malley

P. M.

Bachman

J. G.

Johnston

L. D.

(1996). Routine activities and individual deviant behavior. American Sociological Review, 61, 635-655.

38.

Özbay

. (2008). Self-control, gender, and deviance among Turkish university students. Journal of Criminal Justice, 36, 72-80. doi:10.1016/j.jcrimjus.2007.12.009

39.

Özbay

Özcna

Y. Z

. (2008). A test of Hirschi’s social bonding theory: A comparison of male and female delinquency. International Journal of Offender Therapy & Comparative Criminology, 52, 134-157. doi:10.1177/0306624X07309182

40.

Paternoster

Simpson

(1996). Sanction threats and appeals to morality: Testing a rational choice model of corporate crime. Law & Society Review, 30, 549-583.

41.

Pauwels

(2012). How similar is the interaction between low self-control and deviant moral beliefs in the explanation of adolescent offending? An inquiry in subgroups by gender and immigrant background. In Fruili

A. S.

Veneto

L. D.

(Eds.), Psychology of morality (pp. 141-153). New York, NY: Nova.

42.

Pauwels

Svensson

(2009). Adolescent lifestyle risk by gender and ethnic background. European Journal of Criminology, 6, 5-23. doi:10.1177/1477370808098102

43.

Pauwels

Svensson

(2010). Informal controls and the explanation of propensity to offend: A test in two urban samples. European Journal on Criminal Policy & Research, 16, 15-27. doi:10.1007/s10610-009-9115-x

44.

Reisig

M. D.

Holtfreter

Morash

(2006). Assessing recidivism risk across female pathways to crime. Justice Quarterly, 23, 384-405. doi:10.1080/07418820600869152

45.

Rogers

Smoak

N. D.

Jia

(2006). Self-reported deviant computer behavior: A big-5, moral choice, and manipulative exploitive behavior analysis. Deviant Behavior, 27, 245-268. doi:10.1080/01639620600605333

46.

Sampson

R. J.

Raudenbush

S. W.

(1999). Systematic social observation of public spaces: A new look at disorder in urban neighborhoods. American Journal of Sociology, 105, 603-651. doi:10.1086/210356

47.

Schils

Pauwels

(2014). Explaining violent extremism for subgroups by gender and immigrant background: Using SAT as a framework. Journal of Strategic Security, 7, 27-47.

48.

Svensson

Pauwels

(2010). Always “risky”? The interaction between individual propensity and lifestyle risk in adolescent offending: A test in two urban samples. Crime and Delinquency, 56, 608-626. doi:10.1177/0011128708324290

49.

Svensson

Pauwels

Weerman

F. M.

(2010). Does the effect of self-control on adolescent offending vary by level of morality? Criminal Justice & Behavior, 37, 732-743. doi:10.1177/0093854810366542

50.

Thurman

Q. C.

(1984). Deviance and the neutralization of moral commitment: An empirical analysis. Deviant Behavior, 5, 291-304.

51.

Tittle

C. R.

Ward

D. A.

Grasmick

H. G.

(2003). Gender, age, and crime/deviance: A challenge to self-control theory. Journal of Research in Crime & Delinquency, 40, 426-453. doi:10.1177/0022427803256074

52.

Weerman

F. M.

Bernasco

Bruinsma

G. J. N.

Pauwels

L. J. R.

(2016). Gender differences in delinquency and situational action theory: A partial test. Justice Quarterly, 33, 1182-1209. doi:10.1080/07418825.2015.1064987

53.

Weerman

F. M.

Hoeve

(2012). Peers and delinquency among girls and boys: Are sex differences in delinquency explained by peer factors? European Journal of Criminology, 9, 228-244. doi:10.1177/1477370811435736

54.

Weinberger

D. A.

Schwartz

G. E.

(1990). Distress and restraint as superordinate dimensions of self-reported adjustment: A typological perspective. Journal of Personality, 58, 381-417. doi:10.1111/j.1467-6494.1990.tb00235.x

55.

Wikström

P. H.

Ceccato

Hardie

Treiber

(2010). Activity fields and the dynamics of crime. Journal of Quantitative Criminology, 26, 55-87. doi:10.1007/s10940-009-9083-9

56.

Wikström

P. H.

Oberwittler

Treiber

Hardie

(2012). Breaking rules: The social and situational dynamics of young people’s urban crime. Oxford, UK: Oxford University Press.

57.

Wikström

P. H.

Svensson

(2008). Why are English youths more violent than Swedish youths? European Journal of Criminology, 5, 309-330. doi:10.1177/1477370808090835

58.

Wikström

P. H.

Svensson

(2010). When does self-control matter? The interaction between morality and self-control in crime causation. European Journal of Criminology, 7, 395-410. doi:10.1177/1477370810372132

59.

Wikström

P. H.

Treiber

(2007). The role of self-control in crime causation. European Journal of Criminology, 4, 237-264. doi:10.1177/1477370807074858

60.

Wikström

P. H.

Treiber

K. H.

(2009). Violence as situational action. International Journal of Conflict & Violence, 3, 75-96.

61.

Wikström

P. H.

Tseloni

Karlis

(2011). Do people comply with the law because they fear getting caught? European Journal of Criminology, 8, 401-420. doi:10.1177/1477370811416415