The Importance of Social Support and Coercion to Risk of Impulsivity and Juvenile Offending

Abstract

The current study provides a comprehensive test of differential social support and coercion (DSSC) theory of crime as proposed by Colvin, Cullen, and Vander Ven. DSSC suggests that social interactions are either coercive or supportive in nature and that these interactions figure prominently into the development of self-control and delinquent behavior. Data drawn from the Evaluation of the Gang Resistance Education and Training (GREAT) assess four DSSC research hypotheses. Ordinary least squares (OLS) regression suggests that parental support reduces impulsive behavior whereas coercive relationships correlate with increased impulsivity. Logistic regression indicates that parental support associates with reduced violent offending odds and that interpersonal coercion and low self-control increase violent offending. Beyond testing micro-level DSSC theory, these findings have important implications for criminal justice responses to juvenile offending.

Keywords

support coercion strain adolescent offending juvenile delinquency

Differential social support and coercion (DSSC) theory is an emerging concept in criminology that connects several important theories—most notably general strain (Agnew, 1992) and self-control theories (Gottfredson & Hirschi, 1990)—while further developing aspects of the relationship between criminal behavior and various types of social support and coercive elements of society. At the core, differential support and coercion builds upon the strain tradition in criminology as it explores potential macro- and micro-level social processes important to the development of criminal behavior. Drawing on Merton’s (1938) anomie strain theory, which argues that stress or strain encourages individuals to commit criminal or delinquent behaviors, DSSC further explores negative coercive elements in society.

Merton argued that individuals unable to obtain wealth through legitimate means could turn to criminal opportunities to achieve their goals (Merton, 1938). Building on earlier versions of strain theory, Agnew (1992) extended Merton’s work with general strain theory (GST). In particular, Agnew expanded the sources of strain to include three categories of social-psychological strain: (a) the failure to achieve positively valued goals, (b) the possible or actual loss of positively valued stimuli, and (c) the presentation of noxious stimuli. Experiencing any of these conditions could create anger or frustration that, in return, leads to criminal or deviant behaviors. Put together, individuals are more likely to commit criminal or deviant behaviors when experiencing negative emotions, particularly when he or she perceives the treatment as unjust (Agnew, 2006).

GST posits that negative interpersonal relationships can lead a person to criminal behavior. According to Agnew, the greater the strain, the greater the chances for crime because negative emotions produce pressure for corrective action. In essence, individuals feel pressure from strain and want to respond. Four situational factors can increase the likelihood that strain will prompt criminal behaviors: (a) the strain is seen as unjust, (b) strain is high in magnitude, (c) the strain is caused by or associated with low self-control, and (d) the strain creates pressure or provides incentive to engage in criminal behavior. However, the availability of other goals and individual coping resources may reduce or diminish the risk of criminal adaptation (Agnew, 1992, 2001, 2006). Over the past two decades, a considerable body of empirical research supports the basic propositions of strain theories (Agnew, Brezina, Wright, & Cullen, 2002; Broidy, 2001; Capowich, Mazerolle, & Piquero, 2001; Hoffmann & Miller, 1998; Paternoster & Mazerolle, 1994). Similar to strain, additional studies found that involvement in crime and delinquency is heighted with elevated levels of anger and frustration (Brezina, 1998; Leeper Piquero & Sealock, 2000; Mazerolle, Piquero, & Capowich, 2003).

Differential Support and Coercion Theory

In development of DSSC, Colvin, Cullen, and Vander Ven (2002) refined existing theory on social support, coercion, and strain in an attempt to account for chronic criminal behavior and, to a certain degree, poor mental health. For the theory, coercion is described as “a force that compels or intimidates an individual to act because of fear or anxiety” (Colvin et al., 2002, p. 19). The theory holds that coercive interactions can be employed consistently or erratically and that individuals who receive consistent coercion will develop strong self-directed anger, high externalized self-control, and attenuated social bonds. Although likely to have minimal legal difficulties, these individuals may suffer persistent mental health problems (Colvin et al., 2002). Consistent coercion might be rare and the authors note that total institutions such as prisons or social contexts such as patriarchal households may, perhaps, produce these conditions.

The theory argues individuals disciplined in a coercive, but erratic manner will tend to develop strong anger directed toward others, low self-control, and weak alienated social bonds. In this way, the theory attempts to account for low self-control as envisioned by Gottfredson and Hirschi (1990). Erratically disciplined children become self-centered, pursue pleasure, and fail to develop self-control that mediates criminal impulses (Gottfredson & Hirschi, 1990). The degrees to which parents recognize and reprimand deviant behavior correspond to the child’s capacity to self-regulate behaviors (Gottfredson & Hirschi, 1990; Grasmick, Tittle, Bursik, & Arneklev, 1993; Hay, 2001; Vazsonyi & Crosswhite, 2004), and erratic coercion may aggravate or generate low self-control. This integrates support and coercion theory with self-control, which remains one of the most tested and supported criminological theories to date (Pratt & Cullen, 2000). In addition, if coercion is erratic, arbitrary, or unjust, it will heighten outward directed anger (Agnew, 2001, 2006; Spohn & Kurtz, 2011), thus increasing the risk of involvement in violent and predatory behavior (Colvin et al., 2002). Accordingly, those facing erratic coercion will develop deficits that may lead to chronic offending.

Colvin et al. (2002) further differentiate interpersonal and impersonal forms of coercion. Impersonal coercion relates to negative structural elements within the broader macro-level construction of society and more or less corresponds to Merton’s original version of strain theory. This particular form of coercion may have implications for broader societal behavior and also criminal justice practices of law enforcement and corrections. Interpersonal coercion is most applicable to family relations and other micro-social processes, and involves intimidation or threatening personal relationships intended to force compliance. The authors believe that coercive interpersonal relationships are probably the most aversive form and likely result in anger on the victim’s part. This is particularly true if the individual perceives the treatment as unwarranted (Colvin et al., 2002). This line of reasoning is consistent with aspects of Agnew’s (2006) strain theory as he suggests that strains seen as “unjust” are more likely to result in delinquency than strain perceived as just.

Colvin et al. (2002) described social supports to include “the delivery (or perceived delivery) of assistance from communities, social networks, and confiding partners in meeting the instrumental and expressive needs of individuals” (p. 20). Expressive social supports include sources of emotional support and confirmation of an individual’s importance and worth. Instrumental social support includes a wide range of items from material needs and financial backing to informal guidance and connection to pro-social networks within society. Both expressive and instrumental forms of support can originate from informal sources such as friends and family or larger social structures such as formal networks and social institutions (Colvin et al., 2002).

Social supports can be erratic or consistent in nature. The delivery of erratic support “means than an individual cannot depend on receiving assistance from others or from social institutions” (Colvin et al., 2002, p. 25). These individuals may not be strained or feel duress, but they must fend for themselves without appropriate encouragement and may feel disconnected to communities. Individuals lacking pro-social forms of support may seek out support from criminal associates, role models, and illegitimate social networks in a manner consistent with aspects of differential association (Colvin et al., 2002). According to the theory, erratic social supports will generally result in low self-control, moderate anger, and intermediate social bonds. These individuals may engage in criminal behavior, and depending on the access to illegitimate social supports offered by criminals, they may continue with disorganized criminal action or skilled and organized crime. Individuals who receive consistent support will have low anger, highly internalized self-control, and strong social bonds. These individuals are likely to display few criminal behaviors and possess higher levels of pro-social beliefs (Colvin et al., 2002). Although support and coercion are inversely related, they are not simply “polar opposites,” but rather substantively different experiences that form four ideal types of control and support as a theoretical continuum (for a detailed diagram of this typology, see Colvin et al., 2002, p. 26).

Criticisms and Empirical Test

Pointed critiques and a dearth of empirical validation leave DSSC theory’s place in criminology unsettled. Alexander and Bernard (2002) argued that Colvin and others fail to appropriately define ways to empirically test the theory, and these authors further believe that DSSC is not an integrated theory of criminality and more aptly represents a specific version of strain theory. While acknowledging the contributions and similarities to strain theory, Unnever, Colvin, and Cullen (2004) outright reject beliefs that support coercion theory fits uniquely within the strain paradigm, and argue that it integrates a number of criminological theories and more specifically defines coercion, as opposed to the innumerable forms of strain currently found in the literature (Unnever et al., 2004).

Although peer-reviewed tests of the differential support and coercion theory remain limited, a certain degree of empirical validation rests in test of foundation principles, particularly those within the strain tradition (Agnew, 2006; Agnew et al., 2002; Broidy, 2001; Hoffmann & Miller, 1998; Spohn & Kurtz, 2011) and studies documenting the important influence of social support (Jones, Cauffman, & Piquero, 2007; Kim & Pridemore, 2005; Pratt & Godsey, 2002). Although still limited, the last decade included a number of important tests of DSSC theory.

In an attempt to apply coercion and social support to an organizational setting, Colvin (2007) explored the theory in relationship to historical behavioral practices of the Penitentiary of New Mexico. He argues that during certain periods of time, the practices of the prison were, in essence, a test of differential support and coercion theory. For example, from 1956 to 1967, the institution employed consistent coercion resulting in limited violent behavior, but demoralized inmates. In contrast, he argues that from 1968 to 1972, the organization experienced low violence and pro-social behavior among inmates because of the consistent support employed during this period. During other time frames, the prison’s practices were erratically supportive or coercive with varied outcomes, which are in line with the theory’s tenants. Although this research offers a point of support for the theory, it is difficult to generalize from this specific setting and broader social situations (Colvin, 2007).

In a more empirical test, Unnever et al. (2004) find general support for the link between different forms of coercion and criminal behavior among a sample of middle school students. These researchers tested the influence of parental, peer, school, and neighborhood coercion on criminal behavior. They found all but peer coercion statistically significant and linked to offending, even when controlling for a number of social-psychological deficits and other controls. The parental coercion measure was clearly linked to delinquent offending in this research and included parental threats and abuse. The authors argue that the results “offered fairly consistent support” (p. 257) for differential support and coercion theory.

More recently, DSSC theory has received fairly consistent empirical support across multiple research populations. For example, Baron (2009a) tested the theory with a sample of street youths residing in Toronto, Canada, and found coercion was positively correlated with violent offending. More specific, he found that physical abuse, street victimization, being incarcerated, and being unemployed (all types of coercion) were related to violent offending. In another study of homeless youths, Baron (2015) provided evidence that social support decreases criminal behavior by reducing anger and increasing self-control. This study also showed that coercion increases anger, lowered self-control, and serves as a catalyst to seek illegitimate social support. A study by Kurtz, Linnemann, and Green (2014) analyzed data from the National Survey of Adolescents and found that interpersonal coercion (physical child maltreatment) predicted delinquency and violent offending, whereas social support reduced the odds of offending. Finally, Zavala and Kurtz (2016) analyzed data from a sample of police officers and found that coercive forces predicted officer’s perpetration of intimate partner violence. However, their measures of social support were not significant in their paper.

Despite the modest expansion of research on DSSC, additional research is warranted and this article endeavors to test multiple aspects of the theory. DSSC suggests that various forms of coercion should decrease social control and increase risk of many forms of delinquent behavior. Inversely, social supports should increase social control and decrease involvement in criminal behavior. Based on previous studies of DSSC, we test the following four hypotheses in the current study.

Hypothesis 1: Higher levels of coercion will associate with higher impulsivity (lower self-control) among a sample of eighth-grade students.

Hypothesis 2: Higher levels of coercion will associate with increased odds of self-reported violent behavior among a sample of eighth-grade students.

Hypothesis 3: Higher levels of social support will associate with lower impulsivity (higher self-control) among a sample of eighth-grade students.

Hypothesis 4: Higher levels of social support will associate with decreased odds of self-reported violent behavior among a sample of eighth-grade students.

Method

Data

Data for the current study were obtained from the Evaluation of the Gang Resistance Education and Training (GREAT) Program in the United States, 1995-1999 (see Esbensen & Winfree, 1998) . The “GREAT” program was established to educate youth about gang involvement and to reduce gang activity among program participants. The survey employed a multisite/multistate cross-sectional design completed during the spring of 1995. Survey respondents consisted of eighth-grade students and the primary investigators selected cities and research sites to maximize geographic and demographic diversity. In total, 11 cities were chosen to conduct the survey: Torrance, California; Pocatello, Idaho; Providence, Rhode Island; Will County, Illinois; Orlando, Florida; Milwaukee, Wisconsin; Kansas City, Missouri; Philadelphia, Pennsylvania; Phoenix, Arizona; Omaha, Nebraska; and Las Cruces, New Mexico. The data are open to academic use and accessible for downloading at the Inter-University Consortium for Political and Social Research (ICPSR) website and some elements of the data are frequently used by published scholars to explore crime prevention, educational programs, gang violence, and general criminal behavior (Agnew et al., 2011; Childs, Sullivan, & Gulledge, 2010; Esbensen & Osgood, 1999; Turanovic & Pratt, 2013; Watkins & Melde, 2007).

Sample

Esbensen’s research team distributed self-administrated questionnaires to eighth graders with appropriate parental consent at schools in the GREAT program. Only the schools in the selected cities that offered the GREAT program in the previous 2 years were selected for the survey and the program was administered during the school year of 1993-1994 during respondents’ seventh-grade year. In 1995, the survey was circulated to students on a specific school day and the sample includes all eighth-grade students in attendance that day. This sampling technique yielded 5,935 eighth-grade students encompassing 315 classrooms in 42 different middle schools, and the attendance rates on the survey day varied from a low of 75% to a high of 93%.

Measures

For this study, we utilize several sets of variables drawn from the GREAT data set including demographic characteristics/common control variables, interpersonal coercion and victimization, self-control impulsivity, measures of family social support, and self-reported violent behavior. The research design employs comparative cross-tabulations and quantitative analysis with these variable groupings.

Demographic and control variables

For the analysis, a number of demographic and control variables were constructed from existing data. To examine the influence of gender, a dummy variable was created for males and, therefore, girls serve as the reference category. Age is measured as a continuous variable ranging from 12 to 17, although it must be noted that the sampling procedure limited to only eighth graders restricted the range for this variable and roughly 90% of respondents were either aged 13 or 14 years (M = 13.82 years). Race was coded using dummy variables for the represented groups. In the current analyses, Caucasian serve as the reference group for all regression models and African Americans and Hispanics are included as unique variables.

Coercion

Interpersonal coercion represents one of the most applicable micro-social process envisioned by the theory and involves intimidation or threatening behavior intended to force control. This form of coercion is probably the most aversive and likely to result in anger on the victim’s part (Colvin et al., 2002). Although the broader theory posits multiple forms of coercion including many structural elements, we are only testing interpersonal aspects of coercion. Thus, our research focus pivots from structural version of coercion in a similar fashion to the way Agnew’s GST focuses more on negative interpersonal relationships or micro-level sources of noxious stimuli than Merton’s more structural version of strain (Agnew, 1992). The coercive variables included in the current study examine victimization and the perceptions of interpersonal coercion at school.

The victimization variable captures whether the respondent has ever been the victim of a violent crime and was constructed from the following three questions: “Have you ever been hit by someone trying to hurt you?” “Had someone use a weapon or force to get money or things from you?” and “Been attacked by someone with a weapon or by someone trying to seriously hurt or kill you?” The victimization variable was constructed by combining these questions and coding responses as 1 for a “yes” to any of these questions and 0 for “no” to all the questions. School coercion is an indexed variable constructed from responses to five statements. Respondents were asked the level of agreement with the following five statements: “There are a lot of fights between different groups at my school,” “Students beat up teachers,” “There is a lot of racial conflict between students at my school,” “There is a lot of pressure to join gangs at my school,” and “There are gang fights at my school.” Participants were allowed to answer on a 5-point, Likert-type, ordinal scale with answers ranging from strongly disagree to strongly agree. Reliability analysis revealed a Cronbach’s alpha of .654 and indicated that removing any variable from the index wound not significantly increase or decrease the alpha level. Scores range from 5 to 25 on this index, with higher scores indicating a more coercive school environment.

Delinquent peers

To measure delinquent peers, respondents were asked how many of their current friends committed various criminal/deviant acts. This indexed variable was constructed from the following items that asked how many of their current friends “skipped school without an excuse”; “lied, disobeyed, or talked back to adults such as parents, teachers, or others”; “purposely damaged or destroyed property that did not belong to them”; “stolen something worth less than US$50.00”; “stolen something worth more than US$50.00”; “gone into or tried to go to a building to steal something”; “stolen or tried to steal a motor vehicle”; “hit someone with the idea of hurting them”; “attacked someone with a weapon”; “used a weapon or force to get money or things from people”; “sold marijuana”; “sold illegal drugs such as heroin, cocaine, crack or LSD”; “used tobacco products”; “used alcohol”; “used marijuana”; “used other illegal drugs such as heroin, cocaine, crack or LSD.” The possible responses were on the following ordinal scale: 1) = none of them; 2) = few of them, 3) = half of them, 4) = most of them, and 5) = all of them. A reliability analysis revealed a Cronbach’s alpha of .94 and that removing any of the variables from the index wound not significantly increase the alpha. Scores range from 16 to 80, with higher scores indicating peers with more significant involvement in delinquent behaviors.

Social support variables

Four variables were constructed to examine family-related structure, social supports, and supervision. The survey asked respondents whether they were currently living with their mother only, father only, with their biological parents, or other. The influence of family structure is explored in regression models using the intact family variable, and it was dichotomously coded with living with both biological parents as one and all other responses as zero.

Several questions assessed the level of expressive and emotional social support from parents, and two unique indexed variables were constructed from these items. Expressive social supports include sources of emotional support and confirmation of an individual’s importance and worth. Respondents answered six ordinal scaled measures that independently assessed beliefs about support and attachment to their mother/mother figure and father/father figure. These six measures included “can talk to mother/father about anything,” “mother/father trusts you,” “mother/father knows your friends,” “mother/father understands you,” “can ask for mother/father advice,” and “mother/father praises me.” Responses to each question utilized an ordinal scale from 1 to 7, with lower scores indicating less support and levels and higher scores indicative of greater social support. The mother’s support index scores range from 6 to 42 with higher scores indicating higher support. A reliability analysis revealed a Cronbach’s alpha of .844 for this variable. Likewise, the father’s support variable was constructed from the scaled responses measuring the degree of father’s support. The range for this variable is from 6 to 42 and the Cronbach’s alpha for this index is .883. Reliability analyses indicated that removing any of the variables from either index wound not significantly increase or decrease the alpha.

Supervision represents the final support variable examined in the current research, and this indexed variable is constructed by four questions that assess the degree of informal parenting. The four original measures included “When I go someplace, I leave a note for my parents or call them to tell them where I am”; “My parents know where I am when I am not at home or at school”; “I know how to get in touch with my parents if they are not at home”; and “My parents know who I am with if I am not at home.” Participants responded on a 5-point, Likert-type, ordinal scale with answers ranging from strongly disagree to strongly agree. Responses range from 5 to 20 on this index, with higher scores indicating greater levels of supervision. Reliability analysis generated a Cronbach’s alpha of .732 and also indicated that removing any variable from the index wound not significantly increase or decrease the alpha.

Impulsivity/social control

Gottfredson and Hirschi (1990) did not originally specify how to quantify self-control and, thus, measurement of the concept is mostly left to researchers, and perhaps no criminological theory has generated more debate about measurement than self-control (Delisi, Hochstetler, & Murphy, 2003; Grasmick et al., 1993; Higgins, 2007; Hirschi & Gottfredson, 1993; Longshore, Stein, & Turner, 1998; Longshore, Turner, & Stein, 1996; Marcus, 2003, 2004; Nofziger, 2008; Piquero & Rosay, 1998; Ward, Gibson, Boman, & Leite, 2010). The current research assesses impulsivity and risk-taking behavior using measures derived from Grasmick et al. (1993) and numerous studies demonstrated that the index is a valid and reliable measure (Arneklev et al., 1999; Longshore et al., 1996; Piquero & Rosay, 1998; Piquero & Tibbetts, 1996). The Grasmick index remains perhaps the most widely utilized measure of self-control (Pratt & Cullen, 2000), but some criticized the index for exclusively focusing on attitudinal measures and failing to capture behavioral dimensions of self-control (Marcus, 2003, 2004). However, it should be noted that other criminologists reject behavioral measures of self-control, believing such measures are tautological in nature because they mostly assess criminal or deviant acts themselves (Akers, 1991). The debate regarding the most appropriate measure of self-control will surely continue (see Ward et al., 2010).

In the current research, the impulsivity/self-control variable was constructed from responses to the following eight questions: “I often act on the spur of the moment without stopping to think.” “I don’t devote much thought and effort to preparing for the future.” “I often do whatever brings me pleasure here and now, even at the cost of some distant goal.” “I’m more concerned with what happens to me in the short run than in the long run.” “I like to test myself every now and then by doing something a little risky.” “Sometimes I will take a risk just for the fun of it.” “I sometimes find it exciting to do things for which I might get in trouble.” “Excitement and adventure are more important to me than security.” Respondents answered using a 5-point, Likert-type, ordinal scale with answers ranging from strongly disagree to strongly agree. Scores on the index range from 8 to 40, with higher scores on the measure indicating higher impulsivity and, thus, lower self-control. Reliability analysis generated a Cronbach’s alpha of .800 and removing any variable from the index wound not significantly increase or decrease the alpha.

Violent behavior

The current research uses the violent behavior variable to gauge delinquent conduct, and this binary variable was constructed from affirmative responses to a subset of violent offending questions in the original survey. The survey items for violent offending are as follows: “hit someone with the idea of hurting them, attacked someone with a weapon, used a weapon or force to get money or things from people, been involved in gang fights, shot at someone because you were told to by someone else.” Affirmative responses to any of these violent crimes are coded as 1 for “yes” and “no” responses are coded as 0.

Analytic Strategy

The chief analytical strategy includes cross-tabulations and multivariate analysis to test key aspects of support and coercion theory on both self-control and violent behavior. DSSC theory posits that levels of support and coercion influence both self-control and delinquency. As such, the current research utilizes impulsivity/self-control as both a dependent variable in an ordinary least squares (OLS) regression model and as an independent variable in a logistic regression for violent behavior.¹ We utilize logistic regression analysis for the violent acts variable as it is the most appropriate analytic technique given the binary nature of this variable.² The primary goal of the logistic regression model was to determine whether support, interpersonal coercion, and self-control, controlling for demographic and control variables, influence the odds of respondents reporting violent behavior. In regard to variable measurement and distribution, the independent and control variables can take any level of measurement, and logistic regression has no assumptions regarding normal distribution variables.

Findings

Table 1 displays the mean scores for each of the parental attachment variables and the impulsivity/self-control variable differentiated by types of self-reported violent behavior. We utilize t test to evaluate statistical significant between groups with the first non-violence subgroup serving as the reference category and all t tests are statistically significant. The first subgroup represents respondents indicating involvement in none of the five measured types of delinquency in this sample, and 40% of survey respondents reported no involvement with violent behavior. Consistent with DSSC theoretical assumptions, the non-violent subgroup displays the lowest mean impulsivity (21.66) and the highest means for both mother’s support (31.05) and father’s support (28.44).

Table 1.

Relationship Between Self-Reported Violence, Parental Support, and Self-Control.

Violent behavior	Impulsivity/self-control			Mother’s support			Father’s support
Violent behavior	M	n	SD	M	n	SD	M	n	SD
No reported violence	21.66	2,424	5.52	31.05	2,496	7.45	28.44	2,289	8.90
Hit someone	25.17***	2,909	5.64	27.57***	2,993	8.12	25.36***	2,741	9.51
Gang fight	27.19***	1,058	5.46	26.37***	1,081	8.85	24.34***	976	10.10
Attack with weapon	27.45***	759	5.56	25.55***	779	8.63	23.64***	696	10.34
Armed robbery	28.60***	329	5.77	24.02***	333	9.21	23.28***	306	10.56
Shot at someone	28.26***	280	6.01	25.04***	274	9.88	23.09***	255	10.79
All forms of violence	30.58***	117	5.99	22.28***	111	9.89	20.72***	105	10.60

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

As predicted by DSSC theory, mean scores for the impulsivity/self-control increases with the escalating forms of violent behavior and numbers of youths involved in more aggressive types, expectedly associates with higher impulsivity. The impulsivity means for hitting someone is 25.17, for involvement in a gang fight is 27.19, for attacking someone with a weapon is 27.45, for armed robbery is 28.60, and for shooting at someone is 28.26. The mean impulsivity score is 30.58 among those reporting engagement with all types of violent behavior and this is nearly 10 points greater than the non-violence group. Involvement in various forms of violent behavior is also associated with decreases in both the parental social support variables, with mother’s support slightly higher across all forms of delinquency compared with father’s social support. The escalating types of violent behavior each associates with statistically significant declines in both support measures. A rather sizable change in means is noted when comparing youth reporting no violent behavior to those respondents indicating involvement in all measured forms of violence . Mother’s support among those reporting all forms of violence is nearly nine units lower than the non-violence subgroup, 22.28 compared with 31.05. The same pattern is observed for father’s support with mean support scores of 28.44 for the non-violent group and 20.72 for those reporting involvement in all forms of violent. Across each sub-population, mean difference in both support variables and level of impulsivity/self-control is clearly observable.³

Table 2 displays the OLS regression model with impulsivity/self-control as the dependent variable and findings generally support the theoretical framing of DSSC theory. Among the control variables in this model, boys displayed a statistically significant increase (b = 0.621) in impulsivity compared with girls in this sample. This is consistent with a significant body of research that finds girls to have higher levels of self-control than boys (Chapple & Johnson, 2007; Hayslett-McCall & Bernard, 2002; Hope & Chapple, 2005; Nofziger, 2010; Tittle, Ward, & Grasmick, 2003) . African Americans associate with statistically significant decrease (b = −1.196) in impulsivity/low self-control compared with the Whites reference group. The difference in degree of self-control remains a complicated element in regard to differentiation by race, with most research finding a degree of support for links between self-control and crime across all racial and ethnic groups (Chan & Chui, 2015; Shekarkhar & Gibson, 2011). However, most studies do not explore the relative levels of self-control among divergent groups and such levels may dramatically differ. Although social control may vary by group, it does not mean that the key causal process linking self-control and deviance may significantly differ by group (Vazsonyi & Crosswhite, 2004). Two additional control variables, Hispanic and age, are not statistically significant in this model. The restricted variance of the age variable in this sample likely results in this finding as nearly 90% of the sample are within a year in age because the sampling technique only included eighth graders.

Table 2.

OLS Regression Analysis for Impulsivity/Self-Control Level.

Variables	b	SE	β
Male	0.621	.158	.053***
Age	0.190	.124	.020
African American	−1.196	.201	−.083***
Hispanic	−0.011	.205	.001
Delinquent peers	0.187	.011	.260***
Victimization	0.723	.252	.039**
School coercion	0.182	.023	.108***
Intact family	−0.500	.177	−.038**
Mother’s support	−0.103	.012	−.139***
Father’s support	−0.041	.009	−.065***
Parental supervision	−0.339	.028	−.184***
F significance .000
Adjusted R² = .305

Note. OLS = ordinary least squares.

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

The impulsivity/self-control OLS regression model assesses a number of core theoretical elements of DSSC theory by including measures of delinquent peer relationships, multiple types of coercion, and various forms of social support. The level of peer delinquency (b = 0.187) is statistically significant and linked to increased levels of impulsivity. Both school coercion (b = 0.182) and interpersonal victimization (b = 0.732) associate with increased impulsivity. Each of the four social support variables included in this model—an intact family (b = −0.500), mother’s support (b = −0.103), father’s support (b = −0.041), and parental supervision (b = −0.339)—are statistically significant and display the expected negative relationship with impulsivity/self-control. An exploration of the beta weights of this model indicates that delinquent peer variable has the strongest influence on impulsivity followed in order by parental supervision, mother’s support, and school coercion among the top four. Findings of this model support Hypotheses 1 and 3 and the general assumptions of DSSC theory.

Table 3 displays the findings of a logistic regression model with violent behavior as the dependent variable. This model also includes a number of demographic and control variables, and findings indicate that boys (odds ratio [OR] = 1.466) are associated with a 46% increase in odds of delinquency. We anticipated the importance of gender as a great deal of prior research finds girls are still less likely to engage in deviance, particularly violent crime (Heimer & De Coster, 1999; Mears, Ploeger, & Warr, 1998; Miller, 1998; Nofziger & Kurtz, 2005). Prior research indicates that offense rates of men/boys are higher across all crime categories with the noted exception of prostitution; however, the so-called “gender gap” is most prevalent for violent offending and less noted among minor offense (Steffensmeier & Allan, 1996). African Americans (OR = 1.541) in this sample are 54% more likely than Whites to participate in violence. Research validates the importance of race in regard to delinquent behavior (Matsueda & Heimer, 1987), but race and ethnicity remains a complicated factor in the study of crime. Some research suggests that observed difference in offending by race and ethnicity is far from a proven fact. Clearly, findings based on Uniform Crime Report (UCR) data support differential rates of crime arrest by race. Yet, research utilizing some national representative samples, such as the National Youth Surveys, and other self-reported delinquency finds inconclusive links between race and delinquent behavior (Elliott, Huizinga, & Morse, 1986; Hawkins, Laub, Lauritsen, & Cothern, 2000). As with the impulsivity model, the age and Hispanic variables are not statistically significant in this research.

Table 3.

Logistic Regression for Violent Behavior.

Variables	b	SE	Wald	OR
Male	0.383	.074	26.758	1.466***
Age	0.022	.060	0.127	1.022
African American	0.433	.097	19.911	1.541***
Hispanic	−0.170	.098	3.041	0.844
Delinquent peers	0.113	.007	243.81	1.119***
Victimization	1.170	.158	54.693	3.223***
School coercion	0.012	.011	1.165	1.012
Impulsivity/self-control	0.049	.007	46.203	1.050***
Intact family	−0.111	.085	1.717	0.895*
Mother’s support	−0.015	.006	6.773	0.985**
Father’s support	−0.008	.005	2.672	0.992**
Parental supervision	−0.028	.014	3.961	0.972*

Note. n = 4,317. Nagelkerke R² = .314. OR = odds ratio.

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

The majority of variables testing DSSC propositions display a noteworthy association with self-reported violent behavior. The variable measuring the amount of peer delinquency (OR = 1.119) is significant and associated with an 11% increase in the odds of reported violence with each unit increase in this variable. Among the coercion variables, school coercion (OR = 1.012) is not statistically significant, rather weak, and directionally appropriate. The victimization variable (OR = 3.223) is rather robust, statistically significant, and associated with a 223% increase in odds of reporting engagement violence behavior. The impulsivity/self-control index (OR = 1.05) is included as a dependent variable in this model, and this variable is statistically significant and indicates that each unit increase in impulsivity associates with a 5% increase in violent offending.

Controlling for other factors associated with violent offending, findings indicate the importance of interpersonal supports as each of the four support variables are statistically significant and associate with decreased odds of violent behavior. Living with an intact family (OR = 0.895) appears a slight protective factor against delinquency with the odds of delinquency 11% less for those living in this family structure. Mother’s support (OR = 0.985) is associated with a 15% reduction in the odds of violent offending with each unit increase in this indexed variable. Although weaker, father’s support (OR = 0.992) also associates with reduced odds of violent behavior with an 8% reduction with each unit increase in support. Finally, the parental supervision variable (OR = 0.972) indicates that each unit increase in supervision correlates with a 28% reduction in odds of violent behavior. Overall, results of the violent behavior model support Hypotheses 2 and 4 and the general assumptions of DSSC theory.

Discussion and Limitations

The central undertaking of this current research is to offer the most comprehensive test of DSSC theory to date, and the findings are supportive of the theory’s major propositions across each of the research models. In this study, we included multiple measures of social support, coercion, and test integrated and intermediate elements of the theory by examining the influences of impulsivity/self-control. In general, the findings indicated consistent support for the tested elements of this rather complex theory with the four measures of interpersonal social support displaying statistically significant outcomes across each of the included research models. In addition, two elements of coercion also displayed expected outcomes, with victimization strongly associated with the odds of violent offending. Finally, findings support the DSSC propositions that delinquent peer relationships and self-control may serve important intermediate processes between support and coercion and various forms of delinquent behavior (Colvin et al., 2002).

The four types of social control measures included in this research display results supportive of Hypotheses 3 and 4 and importance of social supports for reducing impulsivity and decreasing offending. The degree of parental support, parental supervision, and the family structure variables all link to reduced impulsivity/low self-control and link to reduced odds of violent offending. The theory argues that individuals who receive consistent support will have low anger, highly internalized self-control, and thus display few criminal behaviors (Colvin et al., 2002). The assumption of parental support also fits with Gottfredson and Hirschi’s (1990) belief that low self-control relates to parenting practices, including failure to recognize, monitor, and correct criminal behavior. We believe that higher scores on the parental support variables are indicative of more consistent social support, although we acknowledge that parents represent one, all-be-it extremely important, form of social support. Thus, the current research does not assess some important macro-level forms of social supports that are likely linked to reduced delinquent behavior.

DSSC theory posits that those lacking consistent support may seek out illegitimate supports offered by delinquent peers and that such youths may have chronic and skilled criminal careers if the social supports are offered by illegitimate means (Colvin et al., 2002). We find that delinquent peers are associated with increased impulsivity and the odds of violent offending in a manner supportive of this major proposition of DSSC theory. However, one could also argue that impulsive and delinquent youth are more likely to seek out delinquent peers than those with higher self-control and less inclination toward offending. At a minimum, the delinquent peer variable in the current study controls for general research on the influence of delinquent peers on criminal offending and/or aspects of differential association (Agnew, 1991; Jensen, 1972; Warr, 1991).

The two interpersonal coercion variables included in this research are statistically significant in the impulsivity/low self-control model, and the victimization variable is strongly linked to increased odds of violent offending. Forms of interpersonal coercive, such as the victimization variable identified in the current study, are recognized as the most destructive and likely to produce a strong sense of anger and a coercive orientation (Colvin et al., 2002).

The coercion outcomes are also rather consistent with those expected from GST perspective, and these findings are an important point of discussion given a prior criticism that DSSC merely represents a version of strain (Alexander & Bernard, 2002). Certainly, any type of interpersonal coercion could be envisioned as a specific type of strain. Although this may seem to complicate the theory, Unnever et al. (2004) argued that it is not uncommon for variables to measure elements from different theoretical perspectives and no rules exist to clearly define particular applications of theories to variables. We believe that the current victimization and school coercion represent valid measures of coercion as outlined by DSSC theory, and also argue that the interpersonal victimization variable used in this sample represents a form of coercion most adverse and likely to result in aggression (Colvin et al., 2002; Unnever et al., 2004).

Beyond testing micro-level DSSC theory, we believe these findings have important implications for criminal justice responses to juvenile offending. Our data seem to indicate that criminogenic origins frequently arise from a combination of inconsistent or weak supports and/or coercive interpersonal relationships. Thus, the more punitive juvenile justice practices of the last three decades appear ill-suited to promote change in those lacking appropriate supports and already subjected to coercive social environments. These retributive juvenile justice practices are deliberately constructed as coercive mechanisms, and aggressive net-widening criminal justice behaviors are easily envisioned as community coercion. Furthermore, interaction with the criminal justice system likely reduces various types of social supports, and criminologists have long argued that system involvement drastically changes the life-course of the offender and effectively limits opportunities (Sampson & Laub, 1997).

Policy changes should target reducing the negative impacts of system coercion and the expansion of more supportive elements. An example may be recent, albeit, modest growth in restorative justice programs. A wide range of juvenile justice practices fall under the umbrella of restorative justice programs and they generally focus on healing interpersonal and community relations as opposed to legalistic and coercive harms directed at the offender (Bazemore & Walgrave, 1999; Braithwaite, 1989). Some research indicates that restorative programs have better outcomes for both the offender and the victims of crimes (Latimer, Dowden, & Muise, 2005). Clearly, findings such as these indicate the need to reevaluate juvenile justice practices, and recently, many states have returned to the early and more progressive responses to juvenile offending (Childress, 2016).

Limitations

Admittedly, this study does not assess a few aspects of the theory, primarily because of data limitations and the sample. First, the cross-sectional nature of data means it is not possible to determine the casual ordering for some of the theoretically important variables and juvenile offending. Therefore, we use caution when interpreting these findings as it is not possible to undeniably establish the temporal ordering of study variables. The current study is, therefore, restricted in only discussing the correlates of impulsivity/self-control and violent offending. Second, the sample only includes measures of interpersonal and school coercion and did not allow for test the structural aspects of the theory such as high levels of unemployment, group competition, and community poverty. Third, it is necessary to point out the limitations that plague most of school-based surveys, and these data are no exception. The sample did not capture private schools or students not in attendance on the day of the study. The sampling technique also likely excluded a number high-risk youths such as the runaways, the homeless, and those confined in correctional centers. Finally, future studies should try to further delineate coercion from strain by measuring various levels and types of social support and coercive aspects of society, and including measures of the consistency of coercion and supports. In spite of these limitations, our study provides one of the most comprehensive tests of DSSC theory to date.

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

Don L. Kurtz is an associate professor of social work and a criminologist in the Department of Sociology, Anthropology and Social Work at Kansas State University. His research interests include police stress, youth violence, and the influence of support and coercion on delinquency. He has published in the Journal of Research in Crime and Delinquency, Feminist Criminology, Deviant Behavior, Criminal Justice Review, Critical Criminology, and Women and Criminal Justice among many others. Prior to pursuing an academic career, he was employed in the juvenile justice system.

Egbert Zavala is an assistant professor in the Department of Criminal Justice at The University of Texas at El Paso, where he has been since receiving his PhD in sociology from Kansas State University. His area of research includes intimate partner violence, testing criminological theories, homicide, deviance, and criminal behavior by police officers. His publications appear in Journal of Family Violence, American Journal of Criminal Justice, Deviant Behavior, Journal of Crime and Justice, and others.

References

Agnew

(1991). The interactive effects of peer variables on delinquency. Criminology, 29, 47-72.

Agnew

(1992). Foundation for a general strain theory of crime and delinquency. Criminology, 30, 47-87.

Agnew

(2001). Building on the foundation of General Strain Theory: Specifying the types of strain most likely to lead to crime and delinquency. Journal of Research in Crime & Delinquency, 38, 319-361.

Agnew

(2006). Pressured into crime: An overview of general strain theory. Los Angeles, CA: Roxbury.

Agnew

Brezina

Wright

J. P.

Cullen

F. T.

(2002). Strain personality traits, and delinquency: Extending general strain theory. Criminology, 40, 43-71.

Agnew

Scheuerman

Grosholz

Isom

Watson

Thaxton

(2011). Does victimization reduce self-control? A longitudinal analysis. Journal of Criminal Justice, 39, 169-174.

Akers

R. L.

(1991). Self-control as a general theory of crime. Journal of Quantitative Criminology, 7, 201-211.

Alexander

D. A.

Bernard

T. J.

(2002). A critique of Mark Colvin’s, crime and coercion: An integrated theory of chronic criminality. Crime, Law and Social Change, 38, 389-398.

Arneklev

B. J.

Grasmick

H. G.

Bursik

R. J.

(1999). Evaluating the dimensionality and invariance of low self-control. Journal of Quntitative Criminology, 15, 307-331.

10.

Baron

S. W.

(2009a). Differential coercion, street youth, and violent crime. Criminology, 47, 239-268.

11.

Baron

S. W.

(2009b). Street youths’ violent responses to violent personal, vicarious, and anticipated strain. Journal of Criminal Justice, 37, 442-451.

12.

Baron

S. W.

(2015). Differential social support, differential coercion, and organized criminal activities. Justice Quarterly, 32, 1089-1117. doi:10.1080/07418825.2014.887760

13.

Bazemore

Walgrave

(1999). Restorative juvenile justice: Repairing the harm of youth crime. Monsey, NY: Criminal Justice Press.

14.

Braithwaite

(1989). Crime, shame and reintegration. Cambridge, UK: Cambridge University Press.

15.

Brezina

(1998). Adolescent maltreatment and delinquency: The question of intervening processes. Journal of Research in Crime & Delinquency, 35, 71-99.

16.

Broidy

L. M.

(2001). A test of general strain theory. Criminology, 39, 9-36.

17.

Capowich

G. E.

Mazerolle

Piquero

(2001). General strain theory, situational anger, and social networks: An assessment of conditioning influences. Journal of Criminal Justice, 29, 445-461.

18.

Chan

H. C.

Chui

W. H.

(2015). The influence of low self-control on violent and nonviolent delinquencies: A study of male adolescents from two Chinese societies. The Journal of Forensic Psychiatry & Psychology, 28, 599-619.

19.

Chapple

C. L.

Johnson

K. A.

(2007). Gender differences in impulsivity. Youth Violence and Juvenile Justice, 5, 221-234.

20.

Childress

(2016, June 2). More states consider raising the age for juvenile crime. Frontline Digital Reports. Retrieved from http://www.pbs.org/wgbh/frontline/article/more-states-consider-raising-the-age-for-juvenile-crime/

21.

Childs

K. K.

Sullivan

C. J.

Gulledge

L. M.

(2010). Delinquent behavior across adolescence: Investigating the shifting salience of key criminological predictors. Deviant Behavior, 32, 64-100.

22.

Colvin

(2007). Applying differential coercion and social support theory to prison organizations: The case of the penitentiary of New Mexico. The Prison Journal, 87, 367-387.

23.

Colvin

Cullen

F. T.

Vander Ven

(2002). Coercion, social support, and crime: An emerging theoretical consensus. Criminology, 40, 19-42.

24.

Delisi

Hochstetler

Murphy

D. S.

(2003). Self-control behind bars: A validation study of the Grasmick et al. scale. Justice Quarterly, 20, 241-263.

25.

Elliott

Huizinga

Morse

(1986). Self-reported violent offending: A descriptive analysis of juvenile violent offenders and their offending careers. Journal of Interpersonal Violence, 1, 472-514.

26.

Esbensen

F. A.

Winfree

L. T.

(1998). Race and gender difference between gang and non-gang youths: Results from a multisite survey. Justice Quarterly, 15, 505-526.

27.

Esbensen

F. A.

Osgood

D. W.

(1999). Gang Resistance Education and Training (GREAT): Results from the national evaluation. Journal of Research in Crime & Delinquency, 36, 194-225.

28.

Fox

(1991). Regression Diagnostics. Newbury Park, CA: Sage.

29.

Gottfredson

M. R.

Hirschi

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

30.

Grasmick

H. G.

Tittle

C. R.

Bursik

R. J.

Arneklev

B. J.

(1993). Testing the core empirical implications of Gottfredson and Hirschi’s general theory of crime. Journal of Research in Crime & Delinquency, 30, 5-29.

31.

Hawkins

Laub

J. L.

Lauritsen

J. L.

Cothern

(2000). Race, ethnicity, and serious and violent juvenile offending (U.S. Department of Justice: Office of Justice Programs, Office of Juvenile Justice and Delinquency Prevention). Washington, DC: Juvenile Justice Bulletin.

32.

Hay

(2001). Parenting, self-control, and delinquency: A test of self-control theory. Criminology, 39, 707-736.

33.

Hayslett-Mccall

Bernard

T. J.

(2002). Attachment, masculinity, and self-control: A theory of male crime rates. Theoretical Criminology, 6, 5-33.

34.

Heimer

De Coster

(1999). The gendering of violent delinquency. Criminology, 37, 277-281.

35.

Higgins

G. E.

(2007). Examining the original Grasmick scale: A Rasch model approach. Criminal Justice and Behavior, 34, 157-178.

36.

Hirschi

Gottfredson

M. R.

(1993). Commentary: Testing the general theory of crime. Journal of Research in Crime & Delinquency, 30, 47-54.

37.

Hoffmann

J. P.

Miller

A. S.

(1998). A latent variable analysis of general strain theory. Journal of Quantitative Criminology, 14, 83-110.

38.

Hope

T. L.

Chapple

C. L.

(2005). Maternal characteristics, parenting, and adolescent sexual behavior: The role of self-control. Deviant Behavior, 26, 25-45.

39.

Hosmer

Lemeshow

(2000). Applied logistic regression (2nd ed.). New York, NY: John Wiley.

40.

Jensen

G. F.

(1972). Parents, peers, and delinquent action: A test of the differential association perspective. American Journal of Sociology, 78, 562-575.

41.

Jones

Cauffman

Piquero

A. R.

(2007). The influence of parental support among incarcerated adolescent offenders: The moderating effects of self-control. Criminal Justice and Behavior, 34, 229-245.

42.

Kim

S.-W.

Pridemore

W. A.

(2005). Social support and homicide in transitional Russia. Journal of Criminal Justice, 33, 561-572.

43.

Kurtz

D. L.

Linnemann

Green

(2014). Support, coercion, and delinquency: Testing aspects of an emerging theory. Journal of Crime and Justice, 37, 309-326.

44.

Latimer

Dowden

Muise

(2005). The effectiveness of restorative justice practices: A meta-analysis. The Prison Journal, 85, 127-144.

45.

Leeper Piquero

Sealock

M. D

. (2000). Generalizing general strain theory: An examination of an offending population. Justice Quarterly, 17, 449-484.

46.

Longshore

Stein

J. A.

Turner

(1998). Reliability and validity of a self-control measure: Rejoinder. Criminology, 36, 175-182.

47.

Longshore

Turner

Stein

J. A.

(1996). Self-control in a criminal sample: An examination of construct validity. Criminology, 34, 209-228.

48.

McClendon

M. J.

(1994). Multiple regression and causal analysis. Itasca, IL: F.E. Peacock Publishers.

49.

Marcus

(2003). An empirical examination of the construct validity of two alternative self-control measures. Educational and Psychological Measurement, 63, 674-706.

50.

Marcus

(2004). Self-control in the general theory of crime: Theoretical implications of a measurement problem. Theoretical Criminology, 8, 33-55.

51.

Matsueda

R. L.

Heimer

(1987). Race, family structure, and delinquency: A test of differential association and social control theories. American Sociological Review, 52, 826-840.

52.

Mazerolle

Piquero

A. R.

Capowich

G. E.

(2003). Examining the links between strain, situational and dispositional anger, and crime: Further specifying and testing general strain theory. Youth and Society, 35, 131-157.

53.

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.

54.

Menard

(1995). Applied logistic regression analysis. Thousand Oaks, CA: Sage.

55.

Merton

R. K.

(1938). Social structure and anomie. American Sociological Review, 3, 672-682.

56.

Miller

(1998). Up it up: Gender and the accomplishment of street robbery. Criminology, 36(1), 37-66.

57.

Nofziger

(2008). The “cause” of low self-control: The influence of maternal self-control. Journal of Research in Crime & Delinquency, 45, 191-224.

58.

Nofziger

(2010). A gendered perspective on the relationships between self-control and deviance. Feminist Criminology, 5, 29-50.

59.

Nofziger

Kurtz

D. L.

(2005). Violent lives: A lifestyle model linking exposure to violence to juvenile violent offending. Journal of Research in Crime & Delinquency, 42, 3-26.

60.

Paternoster

Mazerolle

(1994). General strain theory and delinquency: A replication and extension. Journal of Research in Crime & Delinquency, 31, 235-263.

61.

Piquero

A. R.

Rosay

A. B.

(1998). Reliability and validity of Grasmick et al.’s Self-Control Scale: A comment on Longshore et al. Criminology, 36, 157-174.

62.

Piquero

A. R.

Tibbetts

(1996). Specifying the direct and indirect effects of low self-control and situational factors in offenders’ decision making: Toward a more complete model of rational offending. Justice Quarterly, 13, 481-510.

63.

Pratt

T. C.

Cullen

F. T.

(2000). The empirical status of Gottfredson and Hirschi’s general theory of crime: A meta-analysis. Criminology, 38, 931-964.

64.

Pratt

T. C.

Godsey

T. W.

(2002). Social support and homicide: A cross-national test of an emerging criminological theory. Journal of Criminal Justice, 30, 589-601.

65.

Sampson

Laub

(1997). Urban poverty and the family context: A new look at structure and process in a classic study. Child Development, 65, 523-540.

66.

Shekarkhar

Gibson

C. L.

(2011). Gender, self-control, and offending behaviors among Latino youth. Journal of Contemporary Criminal Justice, 27, 63-80.

67.

Spohn

Kurtz

D. L.

(2011). Family structure as a social context for family conflict: Unjust strain and serious delinquency. Criminal Justice Review, 36, 332-356.

68.

Steffensmeier

Allan

(1996). Gender and crime: Toward a gendered theory of female offending. Annual Review of Sociology, 22, 459-487.

69.

Tittle

C. R.

Ward

D. A.

Grasmick

H. G.

(2003). Self-control and crime/deviance: Cognitive vs. behavioral measures. Journal of Quantitative Criminology, 19, 333-365.

70.

Turanovic

J. J.

Pratt

T. C.

(2013). The consequences of maladaptive coping: Integrating general strain and self-control theories to specify a causal pathway between victimization and offending. Journal of Quantitative Criminology, 29, 321-345.

71.

Unnever

Colvin

Cullen

(2004). Crime and coercion: A test of core theoretical propositions. Journal of Research in Crime & Delinquency, 41, 244-268.

72.

Vazsonyi

A. T.

Crosswhite

J. M.

(2004). A test of Gottfredson and Hirschi’s general theory of crime in African American adolescents. Journal of Research in Crime & Delinquency, 41, 407-432.

73.

Ward

J. T.

Gibson

C. L.

Boman

Leite

W. L.

(2010). Assessing the validity of the retrospective behavioral self-control index: Is the general theory of crime stronger than the evidence suggests? Criminal Justice and Behavior, 37, 336-357.

74.

Warr

(1991). The influence of delinquent peers: What they think or what they do? Criminology, 29, 851-866.

75.

Watkins

A. M.

Melde

(2007). The effect of self-control on unit and item nonresponse in an adolescent sample. Journal of Research in Crime & Delinquency, 44, 267-294.

76.

Zavala

Kurtz

D. L.

(2016). Applying differential coercion and social support theory to police officers’ misconduct. Deviant Behavior, 37, 877-892. doi:10.1080/01639625.2016.1153365