General strain theory, key strains, and deviance

Journal of Criminal Justice 37 (2009) 98–106

Contents lists available at ScienceDirect

Journal of Criminal Justice

General strain theory, key strains, and deviance Byongook Moon a,⁎, Kraig Hays b, David Blurton b a b

Department of Criminal Justice, University of Texas-San Antonio, 501 West Durango Boulevard, San Antonio, TX 78207, United States Department of Justice, University of Alaska Fairbanks, 501A Gruening, P. O. Box 756425, Fairbanks, AK 99775-6425, United States

a b s t r a c t The current study, using a sample of 294 university students, examined the effects of key strains, negative emotions, and conditioning factors on various types of deviance. Eight key strains most likely to lead to deviance, but largely ignored in the previous research on the general strain theory (GST), such as teachers' emotional punishment and race/gender discrimination, were measured. Overall, the findings indicated that teachers' emotional punishment and race discrimination were significantly related to deviance, consistent with GST's prediction. Students who were emotionally punished by teachers and/or were racially discriminated against were more likely to engage in deviance. The findings, however, showed that anger had no significant mediating effect linking strains to deviance and that interaction factors between strain and conditioning variables had limited effects on deviance. © 2008 Elsevier Ltd. All rights reserved.

Introduction General strain theory (GST) has gained a significant level of academic attention, since its development in 1992. Criticizing the narrow conception of strain delineated by classical strain theory, Agnew (1992) expanded the concept of strain, identifying additional sources of strain such as removal of positively valued stimuli and presentation of negative stimuli. According to the GST, various types of strains have significant effects on youth by producing negative emotions, especially anger and depression, which in turn lead to deviance (Agnew, 1992, 2001). Furthermore, Agnew (2001, p. 343) identifies the types of strains most likely to result in deviance: failure to achieve core values, parental rejection, negative secondary school experience, abusive peer relations, criminal victimization, and gender/ race discrimination. A considerable number of empirical studies provided support for GST's key propositions (Agnew & Brezina, 1997; Agnew & White, 1992; Aseltine, Gore, & Gordon, 2000; Bao, Haas, & Pi, 2004; Baron, 2004; Broidy, 2001; Jang & Johnson, 2003; Mazerolle & Maahs, 2000; Mazerolle & Piquero, 1997; Moon & Morash, 2004; Piquero & Sealock, 2000). Consistent with GST's prediction, these findings indicated that there was a strong positive relationship between strains and deviance. Regarding the mediating effect of negative emotions linking strain to deviance, there was less consistent evidence. Several studies found that strains predicted negative emotions (i.e., anger), which in turn lead to deviance, while others found that strains had independent effects on deviance, regardless of the experience of negative emotions ⁎ Corresponding author. Tel.: +1 210 458 2622; fax: +1 210 458 2680. E-mail address: [email protected] (B. Moon). 0047-2352/$ – see front matter © 2008 Elsevier Ltd. All rights reserved. doi:10.1016/j.jcrimjus.2008.12.011

(Baron, 2004; Mazerolle, Burton, Cullen, Evans, & Payne, 2000; Mazerolle & Piquero, 1998; Piquero & Sealock, 2004). Most previous research on GST suffered from several limitations. First, previous research often failed to measure strain properly (Agnew, 2001, 2006). Second, a majority of previous empirical studies examined the cumulative effect of strains on deviance, rather than examining the relative impact of these strains on deviance (Agnew, 2001; Baron, 2004; Moon, Blurton, & McCluskey, 2008; Moon & Morash, 2004). Third, few previous studies on GST measured the key strains believed most likely to lead to deviance and/or tested their effects on deviance. The present research, using a sample of 294 university students, attempted to address these limitations. The study measured eight different types of strains (i.e., racism, gender discrimination, and teachers' emotional punishment), most of which were identified by Agnew (2001) as important strains leading to deviance, but had been ignored by previous research. To better understand which types of strains were more likely to lead to deviance, the relative impacts of eight strains on deviance were examined. Characteristics and types of strains most likely to lead to deviance Agnew (1992, p. 48) defines strain as “relationships in which others are not treating the individual as he or she would like to be treated” and organizes strains into three broad categories: (1) situations that block individuals from achieving highly valued goals (i.e., positively valued goal blockages), (2) situations that remove positive stimuli (i.e., death of parent(s) or friends), and (3) situations that produce negative stimuli (i.e., child abuse, homelessness, race/ gender discrimination, and criminal victimization). According to

B. Moon et al. / Journal of Criminal Justice 37 (2009) 98–106

Agnew (1992), strained individuals may engage in deviant behaviors in an attempt to prevent, retrieve, or manage the loss of positively valued stimuli or escape from or terminate the negative stimuli. In response to criticism that the concept of strain delineated by GST is too broad and does not provide any specific guideline for researchers as to what types of strains should be examined, Agnew (2001) delineates four characteristics of strain most likely associated with deviant behaviors: (1) unjustness, (2) magnitude or severity, (3) association with low self-control, and (4) the pressure or incentive to engage in criminal behaviors. Agnew (2001) posits that strain is more likely to lead to deviance when it is perceived as being unjust, consequently provoking negative emotions (i.e., anger). Strains are more likely to be perceived as unjust when individuals believe that strains are not deserved, result in significant harm, and/or fail to ensure the fairness of process. Agnew (2001, p. 333) also claims that strains high in magnitude and severity are more likely to result in deviant behaviors because severe strains have significant influence on individuals' abilities to address strains in a noncriminal manner. Several factors such as frequency, duration, recency, and centrality of strain are assumed to have a considerable influence on the individuals' perceived magnitude/severity of strains. For example, chronically, frequently, and/or recently occurring strains are more likely than resolved and/or older strains to have a significant negative effect on individuals. Agnew (2001, p. 335) also suggests that strains associated with low social-control such as parental abuse and homelessness are more likely to lead to deviance and crime because of the lack of attachment to significant others and the lack of social and financial support. Strains caused by high social-control (i.e., parental supervision and occupational demands), however, are less likely to result in deviance and crime because high social-control may increase the cost of crime, and increase one's emotional and cognitive ability to deal with strain in a noncriminal manner. Lastly, Agnew (2001, p. 336) indicates that strains generating pressure or incentives to utilize crime as alternative means are more likely to result in deviance. For example, individuals experiencing certain types of strain, such as being bullied by peers and/or child abuse, are more likely to develop a positive attitude toward the use of violence, consequently modeling criminal behaviors to manage, escape from, or terminate strains. Agnew (2001) distinguishes the types of strains most likely to be related to crime from the types of strain unrelated or weakly related to crime. Agnew (2001, pp. 340-341) argues that strains caused by accidents/illness of family member(s), parental/teachers' supervision/ discipline, and excessive demands associated with conventional pursuits are not related or at most weakly related to deviance and crime because they are less likely to be perceived as unjust or are more likely to create high levels of social-control. Contrary to aforementioned strains, certain strain (i.e., the failure to achieve core goals, parental abuse, negative secondary school experience, homelessness, criminal victimization, being bullied by peers, and discrimination) are assumed to be significantly related to deviance, because they are more likely to be seen as unjust, high in magnitude, or associated with low social-control (Agnew, 2001, pp. 343–346). Empirical research on GST and limitations As originally suggested by Agnew (1992), most previous research on GST examined the cumulative effect of strains on deviance. Relatively few researchers tested the individual effect of those strains on deviance and/or measured key strains (i.e., excessive parental punishment, criminal victimization, neighborhood problems, and race/gender discrimination) that are most likely to lead to deviance (Baron, 2004; Eitle, 2002; Mazerolle, 1998; Moon et al., 2008; Moon & Morash, 2004; Morash & Moon, 2007). Studies which examined the individual effect of strain on deviance found that strains caused by excessive parental punishment, family conflicts, negative experience in school, gender/race discrimination,

99

and neighborhood problems had criminogenic effects (Baron, 2004; Eitle, 2002; Moon & Morash, 2004; Paternoster & Mazerolle, 1994; Simons, Chen, Steward, & Brody, 2003). For example, Paternoster and Mazerolle (1994) found that four types of strains (negative relationships with adults, feelings of dissatisfaction with friends and school life, and the experience of stressful events) had significant positive effects on delinquency. Gender and race discrimination were also found to have significant and positive effects on delinquency/criminal behaviors (Eitle, 2002; Simons et al., 2003). Studies, using samples of non-American youths (i.e., Canada, China, and Korea) also found that teachers' excessive punishment, negative relationship with parents, or victimization experiences were significantly related to criminal/ deviant behaviors. As indicated above, relatively little of the GST research measured key strains that are most likely to lead to deviance and examined their effects on deviance (Baron, 2004; Eitle, 2002; Mazerolle, 1998; Moon et al., 2008; Moon & Morash, 2004; Morash & Moon, 2007). For example, key strains such as being bullied, negative experience in school, and gender/race discrimination were rarely measured, despite GST's specification that these strains have the most significant effects on deviance. Moreover, none of the previous research on GST, sampling American youths, examined the effect of teachers' emotional punishment on students. Previous studies with samples of Korean youths or Chinese youths (Bao et al., 2004; Kim, 2002; Moon et al., 2008; Moon & Morash, 2004), however, indicated that teachers' punishment had the most significant and detrimental effect on deviance among East Asian youths, consistent with Agnew's (2001) specification that severe punishment by school officials is one of the strains most likely to lead to delinquency. Considering the significant effect the teachers' punishment had on delinquency among East Asian youths, it is necessary to examine whether teachers' emotional punishment has a significant effect on delinquency in the United States setting. Another limitation was that most previous research examined the effect of cumulative strain on deviance, instead of testing the relative impact of individual strains on deviance. This approach, however, may combine strains least likely to lead to crime with those most likely to result in crime, and consequently, a cumulative strain has a moderate effect on deviance (Agnew, 2001). It also failed to clarify which types of strains are most likely to lead to deviance. The present study attempted to address these limitations. First, eight key strains, most of which were identified by Agnew and/or empirical findings as important strains leading to deviance, were measured. In order to clarify which types of strain have significant effects on deviance, the individual impact of these strains was examined separately. It was hypothesized that each of eight key strains is positively related to deviance. The study also examined the mediating and moderating effects of negative emotions (focusing on anger) and conditioning factors linking strain to deviance, because of the critical roles of negative emotions and conditioning factors in GST. According to GST, strains generate negative emotions, especially anger, which in turn lead to deviance (Agnew, 1992). Empirical findings, however, had showed mixed results in that several studies showed a medium strength mediating effect of negative emotions in linking strains to deviance, while other studies indicated that strains had significant direct effects on deviance, independent of negative emotions (Bao et al., 2004; Baron, 2004; Broidy, 2001; Jang & Johnson, 2003; Mazerolle & Piquero, 1997; Mazerolle, Piquero, & Capowich, 2003; Moon et al., 2008; Piquero & Sealock, 2004). Therefore, there is need for additional tests, specifically focusing on the mediating effect of negative emotions. It was hypothesized that anger is positively related to deviance and direct effects of individual strain on deviance is signifantly reduced when anger is added. In order to test interaction effects between strains and conditioning factors on deviance, four variables (positive relationship with parents, problem solving ability, perceived legitimacy of violence, and

100

B. Moon et al. / Journal of Criminal Justice 37 (2009) 98–106

association with delinquent peers) were included as conditioning factors. GST (Agnew, 1992) states that youths are more likely to alleviate strains through nondelinquent behaviors when they maintain positive relationship with parents and have a high level of problem solving ability, while youths who associate with delinquent peers and/or have a positive attitude toward the use of violence are more likely to respond to strain with delinquency. It was hypothesized that strains interacted with positive relationship with parents and problem solving ability are less likely to result in deviance, while strain interacted with association with deviant peers and positive attitude toward the use of violence are more likely to lead to deviance. Method Sample Data analyzed in the study were based on questionnaires administered to college students registered in freshman-level courses at a state university in the western United States. The questionnaire was administered during the 2005-2006 academic year. Enrollment data for each semester were examined to identify freshman-level courses with significant enrollment numbers. The selection of courses for administration of the questionnaires was determined by size of enrollments (larger classes were selected), discipline (diversity was the primary consideration), and cooperation of instructors. While the study used a nonrandom, convenience sample, the researchers attempted to achieve diversity of participants with regard to their backgrounds and interests by sampling from various academic disciplines. The courses selected were mainly from social science disciplines. Two of the courses selected satisfied general university baccalaureate requirements, and consequently included students from a broad spectrum of academic interests. Freshman-level classes were targeted because researchers hoped to have captured social interactions occurring during the participants' high school years. The questions explored various strains such as social, family, and student-teacher interactions over a five-year period, a period that corresponds with the high school years of stereotypical college freshmen. The questionnaires were administered in a total of six courses and the researchers personally administered the questionnaires in two of the six courses, and relied upon the course instructors in the other four. Instructors administering the questionnaires in their courses were provided with a brief introduction to read to students. The introductory statements emphasized that participation was to be voluntary, and requested that students not participate if they had already done so in another class. Overall, 343 completed questionnaires were collected. For current analyses, 294 valid cases were used, after listwise deletion of missing values. Among the 294 students in the sample, 52 percent (n = 153) were female and 48 percent (n = 141) were male. Seventy-nine percent (n = 232) were Caucasians and 21 percent were non-Caucasians. The sample had fewer females and more Caucasians, compared with overall university student population (59 percent females and 64 percent Caucasians in the whole student population). For the analyses, respondent's race and gender were used as control variables to better understand the effects of strains, negative emotions, and conditioning factors on deviance. The researchers recognized that sampling college students for social science research involves inherent weaknesses. The population was not representative of the general population, and in fact, it was expected that the college populations would have a relatively low exposure to strain and anger when compared to the general population. This study, however, investigated sources of main strains that had been ignored in previous studies on GST, which necessitated the collection of specific data. This was especially true for the study's attempt to assess the verbal abuse by teachers as a significant source of strain.

Independent variables Strains Eight types of strain most likely to lead to deviance were measured: desired goal blockage, family conflict, parental punishment, teachers' emotional punishment, gender discrimination, race discrimination, criminal victimization, and negative community environment (see Table 1 and Appendix A for all the items used to measure strains, negative emotions, conditioning factors, and deviance). All the scales were coded so that a higher score indicated a higher level of each strain, negative emotion, or deviance. The goal blockage scale was partially adopted from the study of Broidy (2001) and consisted of four items. These items measured whether respondents were successful at achieving socially desirable goals such as academic/career goals, financial goals, and health goals. The response options for each item ranged from 1 (very successful) to 4 (not at all successful). It was coded so that a higher score indicated a higher level of goal blockage (Cronbach's alpha = .65). The family conflict scale consisted of three items, measuring the extent of experience of argument and tension among family members (Cronbach's alpha = .60). These items were partially adapted from Aseltine et al. (2000) and the response options ranged from 0 (never) to 3 (always). The parental punishment scale measured the frequency of parents' emotional and physical punishment of the respondents. Adapting from Piquero and Sealock (2000), six items (i.e., pushing, slapping, and name-calling) were used to create the parental punishment scale (Cronbach's alpha = .88). The response options for each item ranged from 0 (never) to 4 (ten or more times). The teachers' punishment scale (Cronbach's alpha = .89) consisted of five items, measuring the frequency of teachers' emotional punishment of the respondents (i.e., isolating, embarrassing, and ignoring). These items were adopted from Moon and Morash (2004) and the response options ranged from 0 (never) to 4 (ten or more times). The racial discrimination scale (Cronbach's alpha = .88) was adopted from a study of Landrine and Klonoff (1996) and consisted of ten items. These items measured the experience of racial discrimination of respondents by others such as teachers, police officers,

Table 1 Descriptive statistics of key strains, negative emotions, conditioning factors, and deviance Mean (SD)

Range

Strains Goal blockage Family conflict Parental punishment Teachers' emotional punishment Racial discrimination Gender discrimination Criminal victimization Negative community environment

9.1 (2.2) 5.0 (1.8) 3.0 (4.4) 2.5 (3.9) 3.2 (5.1) 3.8 (4.3) 5.2 (1.9) 1.4 (1.9)

4 ~ 15 0~9 0 ~ 24 0 ~ 20 0 ~ 36 0 ~ 20 4 ~ 16 0 ~ 11

Negative emotions Anger

2.7 (2.1)

0~9

Conditioning variables Positive relationship with parents Deviant peer association Problem solving ability Attitude toward the use of violence Deviant variables General deviancy Violent deviancy Nonviolent deviancy Total number = 294

9.9 6.5 12.9 10.5

(2.3) (5.0) (2.3) (3.6)

3 ~ 12 0 ~ 21 4 ~ 16 5 ~ 20

3.8 (5.6) 1.6 (2.9) 2.2 (3.5)

0 ~ 36 0 ~ 21 0 ~ 20

B. Moon et al. / Journal of Criminal Justice 37 (2009) 98–106

and neighbors. The response options of each item ranged from 0 (none) to 4 (ten or more times). The gender discrimination scale (Cronbach's alpha = .88) was created by summing five items, partially adopted from Landrine and Klonoff (1996). The original items (intended to measure racial discrimination) were slightly rephrased to measure the experience of gender discrimination (i.e., treated with less courtesy and respect, insulted, and threatened because of gender). It was coded so that a higher score indicated a higher experience of gender discrimination. The victimization scale (Cronbach's alpha = .56) consisted of five items, measuring respondents' and/or their family's victimization experiences. The items included various victimization experiences such as theft, property damage, and physical assault. The response options ranged from 0 (never) to 4 (four or more times) and the scale was coded so that a higher score indicated more victimization experiences. The negative neighborhood environment scale (Cronbach's alpha = .81) was created by summing six items, measuring various types of problems in a respondent's neighborhood such as vandalism, abandoned houses, burglaries, and assaults. The response options ranged from 0 (not a problem) to 2 (big problem) and the scale was coded so that a higher score indicated a high level of perceived community problems. Negative emotion To examine the mediating effect of negative emotion linking strains to deviance, anger, a key negative emotion in GST, was measured. The anger scale (Cronbach's alpha = .79) was partially adapted from a study of Derogatis (1977) and consisted of three items. These items measured whether a respondent had felt uncontrollable outbursts of temper, urges to beat and harm someone, or urges to break things during the last five years. The response options for each item ranged from 0 (never) to 3 (always) and the scale was coded so that a higher score indicated a high level of anger. Conditioning factors As mentioned above, GST recognizes the importance of conditioning factors on the relationship between strains and delinquency. To examine the interaction effects between strains and conditioning factors on deviance, the present study included four conditioning factors: supportive family, deviant peer association, problem solving ability, and legitimacy of violence. The supportive family scale (Cronbach's alpha =.84) was created by summing three items, measuring the degree of a respondent's parent(s) or his/her family members' understanding, interest, and closeness toward a respondent during the last five years. The response options of each item ranged from 1 (strongly disagree) to 4 (strongly agree) and the scale was coded so that a higher score indicated a high level of family support. The deviant peer association scale (Cronbach's alpha = .87) was partially adopted from a study by Mazerolle and Maahs (2000) and consisted of seven items. These items measured whether a respondent's close friends engaged in various deviant behaviors (i.e., purposely damaging or destroying property, hitting someone, and carrying a hidden weapon) during the last five years. The response options of each item ranged from 0 (none of them) to 4 (all of them) and the scale was coded so that a higher score indicated a high level of association with deviant peers. The problem solving ability scale (Cronbach's alpha = .71) consisted of four items (i.e., there are lots of ways around any problem that I am facing now and I can think of many ways to reach my current goals). These items measured whether a respondent had the ability to find ways to solve problems in his/her life during the last five years. The response options of each item ranged from 1 (strongly disagree) to 4 (strongly agree) and the scale was coded so that a higher score indicated a high level of problem solving ability.

101

The legitimacy of violence scale (Cronbach's alpha = .82) consisted of five items measuring a respondent's attitude toward the use of violence to defend one's right, obtain fair treatment, resist exploitation, and/or avoid appearing weak. The response options of each item ranged from 1 (strongly disagree) to 4 (strongly agree) and the scale was coded so that a higher score indicated a high level of legitimacy of violence. Dependent variables Respondents were asked how often they engaged in various types of deviant behaviors during the last five years. Examples of these deviant behaviors were: fighting in a group, joining a gang, hitting others, using force to get money from others, stealing something, and using or selling illegal drugs (see Appendix A). The response options ranged from 0 (never) to 4 (nine or more times). A general deviance scale (Cronbach's alpha = .85) was created by combining these items and was coded so that a higher score indicated a higher level of general deviant behaviors. In addition to examining the effects of strains on general deviance, the current study also examined the distinctive effect of strains on different types of deviance, since the previous studies (Mazerolle & Piquero, 1998; Piquero & Sealock, 2000, 2004) found that strains had unique effects on different types of deviance. Two different types of deviance (violent deviance and nonviolent deviance) were created. A violent deviance scale was created by summing seven items, measuring the frequency of committing violent deviant behaviors (i.e., hitting others, throwing objects at others, and using force to get money from others). The scale was coded so that a higher score indicated a higher level of violent deviancy (Cronbach's alpha = .80). A nonviolent deviance scale (Cronbach's alpha = .77) was created by summing five items, measuring the frequency of engaging in nonviolent deviant behaviors such as property and/or drug-related deviance (i.e., breaking into a building to steal; knowingly buying, selling, or holding stolen goods; and using/selling illegal drugs). It was also coded so that a higher score indicated a higher level of nonviolent deviance during the last five years. Natural log transformation for each deviance index was used for multiple analyses, because the deviance indexes were positively skewed. Control variables To better understand relationships between key variables in the GST and deviance, several demographic variables such as gender, race, and age, which are known to have significant effects on deviance, were used as control variables (Bao et al., 2004; Brier, 1995; Moon & Morash, 2004; Morash & Moon, 2007). A respondent's gender was dichotomized coding female as 0 and male as 1. A respondent's race was also dichotomized coding non-White as 0 and White as 1. A respondent's age was a continuous variable. Findings Table 2 presents the results of zero-order correlations between eight key strains, anger, conditioning factors, and deviant behaviors. As expected, a majority of strains (i.e., goal blockage, family conflicts, parental and teachers' punishment, and racial/gender discrimination) were significantly and positively related to negative emotions (see Table 2). The results also indicated that there was positive and significant relationship between most strains and various types of deviance. Anger was significantly and positively related to three types of deviance. Consistent with GST's prediction, association with deviant peers and legitimacy of violence were positively related to deviance, while problem solving ability was negatively related to all three types of deviance. A series of stepwise ordinary least squares regression analyses was performed to examine the predictors of various types of deviance. In

102

B. Moon et al. / Journal of Criminal Justice 37 (2009) 98–106

Table 2 Correlation matrix among strains, negative emotions, conditioning factors, and deviance

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

1.00 .15⁎⁎ .17⁎⁎ .11 .01 - .01 .22⁎⁎⁎ .09 .27⁎⁎⁎ - .22⁎⁎⁎ .14⁎ - .30⁎⁎⁎ .13⁎ .17⁎⁎ .09 .20⁎⁎

1.00 .45⁎⁎⁎ .14⁎ .14⁎ .20⁎⁎ .20⁎⁎ .15⁎⁎ .25⁎⁎⁎ - .31⁎⁎⁎ .19⁎⁎⁎ - .11 .10 .09 .10 .06

1.00 .29⁎⁎⁎ .21⁎⁎⁎ .18⁎⁎ .21⁎⁎⁎ .11 .22⁎⁎⁎ - .39⁎⁎⁎ .26⁎⁎⁎ - .07 .10 .17⁎⁎ .17⁎⁎ .13⁎

1.00 .44⁎⁎⁎ .33⁎⁎⁎ .34⁎⁎⁎ .11 .26⁎⁎⁎ - .18⁎⁎ .25⁎⁎⁎ - .07 .09 .36⁎⁎⁎ .34⁎⁎⁎ .30⁎⁎⁎

1.00 .35⁎⁎⁎ .45⁎⁎⁎ .18⁎⁎ .19⁎⁎ - .15⁎ .23⁎⁎⁎ - .06 .09 .29⁎⁎⁎ .37⁎⁎⁎ .16⁎⁎

1.00 .26⁎⁎⁎ .17⁎⁎ .13⁎ - .14⁎ .18⁎⁎ .01 .05 .09 .09 .06

1.00 .09 .22⁎⁎⁎ - .23⁎⁎⁎ .30⁎⁎⁎ - .18⁎⁎ .15⁎ .26⁎⁎⁎ .27⁎⁎⁎ .19⁎⁎

1.00 .08 - .10 .22⁎⁎⁎ - .12⁎ .05 .13⁎ .07 .14⁎

1.00 - .12⁎ .35⁎⁎⁎ - .22⁎⁎⁎ .28⁎⁎⁎ .42⁎⁎⁎ .47⁎⁎⁎ .29⁎⁎⁎

1.00 - .07 .17⁎⁎ - .10 - .06 - .10 - .02

1.00 -.14⁎ .27⁎⁎⁎ .54⁎⁎⁎ .43⁎⁎⁎ .50⁎⁎⁎

1.00 - .13⁎ - .26⁎⁎⁎ - .21⁎⁎ - .23⁎⁎⁎

1.00 .26⁎⁎⁎ .32⁎⁎⁎ .14⁎

1.00 .84⁎⁎⁎ .90⁎⁎⁎

1.00 .51⁎⁎⁎

1.00

Note: 1 = goal blockage, 2 = family conflict, 3 = parental punishment, 4 = teachers' punishment, 5 = racial discrimination, 6 = gender discrimination, 7 = criminal victimization, 8 = negative community environment, 9 = anger, 10 = supportive family, 11 = deviant peer association, 12 = problem solving ability, 13 = legitimacy of violence, 14 = general deviance, 15 = violent deviance, and 16 = nonviolent deviance. ⁎ = p b .05. ⁎⁎ = p b .01. ⁎⁎⁎ = p b .001.

each table, four different models were used to examine the unique effects of individual strain, anger, conditioning factors, and interaction terms between strains and conditioning factors on the dependent variables. At every step, control variables (gender, race, and age) were included in the model. Table 3 presents the results of OLS regression analyses of the independent variables on the log transformed general deviance. The results in Model 1 show that only two of eight strains were significantly related to general deviance in the expected direction (see Table 3). Students who experienced teachers' emotional punishment or racial discrimination were more likely to commit general deviance. Criminal victimization and negative community environment also had marginally significant positive effects on general deviance. In Model 2, anger was added to examine their mediating effects linking strains to deviance. As expected, anger had a significant positive effect on general deviance in that students with higher levels of anger personality were more likely to engage in deviance. The mediating effects of anger linking strains to deviance were marginal in that teachers' emotional punishment continued to exert a significant effect on general deviance. Criminal victimization was significantly related to general deviance after the inclusion of the negative emotion. In Model 3, four conditioning factors were added to the baseline model. The results show that three of four conditionings factors had significant effects on general deviance in expected directions. Students who associated with delinquent peers or had positive attitudes toward the use of violence were more likely to engage in deviance, while those with high levels of problem solving ability were less likely to commit deviance. In Model 4, interaction terms between strains and conditioning factors were created and included for a more precise and theoretically appropriate estimate of the interaction effects between strains and conditioning factors on deviance. First, thirty-two interaction terms between eight individual strains and four conditioning were created in the final analysis. The VIFs and the condition index scores, however, indicated a significant level of multicollinearity in the multivariate analyses. In order to address a multicollinearity issue, a composite measurement of total strain was created and used to produce four interaction terms. To further reduce the potential of a multicollinearity problem, total strain and each of four conditioning variables were standardized before they were multiplied to create interaction terms. The examination of the VIF, condition index, and variance proportion values showed no significant problem of multicollinearity in the final model. The results in Model 4 indicated that only one of four

interaction terms was significantly related to general deviance in the expected direction. Strained students were less likely to engage in deviance, when they had higher levels of problem solving ability. Table 3 Effects of strains, negative emotions, conditioning factors on general deviance

Control variables Gender (male = 1) Race (non-White = 1) Age Key strains Goal blockage Family conflict Parental punishment Teacher emotional punishment Racial discrimination Gender discrimination Criminal victimization Negative community environment Total strain

Model 1

Model 2

Model 3

Model 4

B (Beta)

B (Beta)

B (Beta)

B (Beta)

.41 (.24)⁎⁎⁎ - .09 (- .04) - .01 (- .04)

.28 (.17)⁎⁎ - .12 (- .06) - .00 (- .01)

.15 (.09)a -.12 (- .06) .01 (.07)

.28 (.17)⁎⁎⁎ - .19 (- .09)⁎ .01 (.04)

.01 (.03) .03 (.06) .01 (.06) .04 (.18)⁎⁎

- .02 (- .05) - .00 (- .01) .01 (.04) .03 (.13)⁎

-.03 (- .08) -.00 (- .01) .00 (.02) .03 (.14)⁎⁎

.02 (.11)a - .01 (- .06) .05 (.11)⁎ .04 (.09)a

.02 (.12)⁎ -.02 (- .09)a .03 (.06) .01 (.03)

.02 (.14)⁎ - .01 (- .03) .05 (.12)a .04 (.10)a

.14 (.16)⁎⁎

Negative emotions Anger

.17 (.41)⁎⁎⁎

Conditioning factors Supportive family Deviant peer association Problem solving ability Legitimacy of violence

.12 (.29)⁎⁎⁎

.11 (.28)⁎⁎⁎

.03 (.07) .06 (.34)⁎⁎⁎ -.04 (- .11)⁎ .03 (.13)⁎⁎

.07 (.08)a .28 (.33)⁎⁎⁎ - .07 (- .08)a .10 (.11)⁎

Interaction variables Total strain × supportive family Total strain × deviant peer association Total strain × problem solving ability Total strain × legitimacy of violence R2 a

= b .10. ⁎ = p b .05. ⁎⁎ = p b .01. ⁎⁎⁎ = p b .001.

.04 (.05) .04 (.06) - .08 (- .10)⁎ .00 (.00)

.26

.39

.52

.51

B. Moon et al. / Journal of Criminal Justice 37 (2009) 98–106

Table 4 presents the results of OLS regression analyses of the independent variables on the log transformed violent deviance. The same procedure used in Table 3 was employed. Consistent with the findings in Table 3, the results in Model 1 showed that two of eight strains (teachers' emotional punishment and racial discrimination) were significantly and positively related to violent deviance (see Table 4). In Model 2, anger had a significant positive effect on violent deviance as expected. Almost identical with the results in Table 3, teachers' emotional punishment and racial discrimination continued to have significant effects on violent deviance, even after the negative emotion was added. Three of the conditioning variables had significant effects on violent deviance in the expected directions. In Model 4, none of the interaction terms were significantly related to violent deviance, though interaction terms between strain and problem solving ability and strain and legitimacy of violence had marginal effects on violent deviance in the expected directions. Table 5 presents the results of OLS regression analyses of the independent variables on the log transformed nonviolent deviance. In Model 1, two of eight strains (goal blockage and teachers' emotional punishment) had significant effects on nonviolent deviance (see Table 5). Students experiencing desired goals blockage or teachers' emotional punishment were more likely to commit nonviolent deviance. Similar to findings in Tables 3 and 4, anger had a significant positive effect on nonviolent deviance. When the measure of anger was included, teachers' emotional punishment continued to exert

Table 4 Effects of strains, negative emotions, conditioning factors on violent deviance

Control variables Gender (male = 1) Race (non-White = 1) Age Key strains Goal blockage Family conflict Parental punishment Teacher emotional punishment Racial discrimination Gender discrimination Criminal victimization Negative community environment Total strain

Model 1

Model 2

Model 3

Model 4

B (Beta)

B (Beta)

B (Beta)

B (Beta)

.37 (.24)⁎⁎⁎ - .10 (- .05) - .00 (- .01)

.25 (.16)⁎⁎ - .13 (- .07) .00 (.02)

a

.14 (.09) - .12 (- .07) .02 (.09)a

.00 (.01) .03 (.07) .01 (.06) .04 (.19)⁎⁎ .02 (.16)⁎ - .01 (- .04) .04 (.09) .02 (.05)

- .03 (- .07) .00 (.00) .01 (.04) .03 (.14)⁎

- .04 (- .10)⁎ - .00 (- .00) .00 (.02) .03 (.15)⁎⁎

.02 (.14)⁎ - .01 (- .07) .04 (.09)a .02 (.04)

.02 (.14)⁎ - .02 (- .10)a .02 (.04) - .01 (- .01)

.15 (.43)⁎⁎⁎

Key strains Goal blockage Family conflict Parental punishment Teacher emotional punishment Racial discrimination Gender discrimination Criminal victimization Negative community environment Total strain

Model 2

Model 3

Model 4

B (Beta)

B (Beta)

B (Beta)

B (Beta)

.24 (.14)⁎ .01 (.01) - .01 (- .05)

.18 (.11) - .00 (- .00) - .01 (- .04)

.06 (.04) .00 (- .01) .01 (.04)

.06 (.15)⁎ .02 (.04) - .00 (- .01) .04 (.19)⁎⁎

.04 (.11)a .01 (.01) - .00 (- .02) .04 (.17)⁎⁎

.04 (.09)a .01 (.02) - .01 (- .04) .04 (.17)⁎⁎

- .01 (- .04) .01 (.07) .02 (.05) .04 (.10)a

- .01 (- .05) .01 (.05) .02 (.05) .04 (.09)

- .01 (- .05) .00 (.02) .01 (- .02) .00 (.01)

.10 (.06) .01 (.00) .01 (.03)

.11(.13)⁎

Negative emotions Anger

.08 (.19)⁎⁎

Conditioning factors Supportive family Deviant peer association Problem solving ability Legitimacy of violence

.03 (.07)

.04 (.10)a .08 (.49)⁎⁎⁎ - .05 (- .13)⁎ - .01 (- .03)

Interaction variables Total strain × supportive family Total strain × deviant peer association Total strain × problem solving ability Total strain × legitimacy of violence

.03 (.07)

.06 (.07) .41 (.48)⁎⁎⁎ - .12 (- .14)⁎⁎ - .02 (- .02)

.10 (.14)⁎ - .00 (- .01) - .04 (- .05) - .04 (- .06)

.14

.17

.37

.37

= b .10. ⁎ = p b .05. ⁎⁎ = p b .01. ⁎⁎⁎ = p b .001.

.11 (.31)⁎⁎⁎

.11 (.31)⁎⁎⁎

.02 (.05) .04 (.26)⁎⁎⁎ - .04 (- .11)⁎ .04 (.17)⁎⁎⁎

.06 (.08) .19 (.25)⁎⁎⁎ - .06 (- .08)a .11 (.15)⁎⁎

significant effect on nonviolent deviance, but the significant effect of goal blockage disappeared. Two of the conditioning factors (deviant peer association and problem solving ability) were significantly related to nonviolent deviance in that students associating with deviant peers or having low levels of problem solving ability were more likely to commit nonviolent deviance. For the effects of interaction terms on nonviolent deviance, the results showed that an interaction term between total strain and family support had a significant effect on nonviolent deviance. Strained students, however, were more likely to engage in nonviolent deviance when they had a higher level of family support, contrary to the GST's prediction. Discussion and conclusion

Interaction variables Total strain × supportive family Total strain × deviant peer association Total strain × problem solving ability Total strain × legitimacy of violence

= b.10. ⁎ = p b .05. ⁎⁎ = p b .01. ⁎⁎⁎ = p b .001.

Control variables Gender (male = 1) Race (non-White = 1) Age

Model 1

a

.09 (.12)⁎

Conditioning factors Supportive family Deviant peer association Problem solving ability Legitimacy of violence

a

Table 5 Effects of strains, negative emotions, conditioning factors on nonviolent deviance

R2

Negative emotions Anger

R2

.25 (.17)⁎⁎⁎ - .20 (- .11)⁎ .01 (.06)

103

- .01 (- .01) .04 (.06) - .06 (- .08)a .06 (.08)a

.25

.40

.50

.48

The results indicated that only three of the eight strains were significantly related to general deviance, violent deviance, or nonviolent deviance. Students who experienced desired goal blockage, teachers' emotional punishment, or racial discrimination were more likely to engage in deviance. Family related strains (family conflict and parental punishment) and gender discrimination, contrary to Agnew's predictions, were not significantly related to any types of deviance. With regard to family generated strains, it must be questioned whether the age and relative educational success of the convenience sample masked any relationship with deviance. In general, a significant number of participants in the study were

104

B. Moon et al. / Journal of Criminal Justice 37 (2009) 98–106

unlikely to have recent contact with their families since they were attending college. Agnew has suggested that recent strain will be more significantly related to deviance than older sources of strain. Consequently, the lack of a significant relationship between family strain and deviance may be a reflection of the absence of recent family contact, a product of the convenience sample. As expected, the results showed that racial discrimination had a significant effect on violent deviance; students experiencing racial discrimination were more likely to engage in violent deviant behaviors. Agnew (2001) argues that delinquent behavior is most likely to be the result of strain when it is seen as unjust, when it threatens identities, and when it leaves few options for coping other than delinquency. It is more likely that victims perceive racial discrimination as intentional, aggressive, undeserved, and an unfair assault which could threaten their self-esteem and identity. An interesting and surprising finding was that teachers' emotional punishment had a significant effect on various types of deviance. As mentioned above, teachers' punishment was ignored as a source of strain in empirical studies on GST, using American youths, probably because corporal punishment is largely banned in a majority of states and the relationship between students and teachers is relatively positive. The limited studies in the education literature (Brendgen, Bukowski, Wanner, Vitaro, & Tremblay, 2007; Brendgen, Wanner, & Vitaro, 2006; Casarjian, 2000), however, demonstrated that a considerable number of students experienced teachers' verbal abuse in Western societies, and teachers' verbal punishment had a shortterm as well as long-term negative effects on students. For example, Brendgen et al. (2007), using a longitudinal sample of 231 Canadian children over eight years (starting in kindergarten) and again at age twenty-three, examined the relationship between teachers' verbal abuse and children' behavioral and emotional problems. The results indicated that teachers' verbal abuse had a detrimental effect on children's self-esteem as well as scholastic and behavioral competence (Brendgen et al., 2007). The findings also showed that high levels of teacher abuse led to behavior problems at age twenty-three even when controlling for sex, SES, and other potential childhood risk factors. Overall, these results were consistent with previous findings with samples of Korean and Chinese youths (in both countries corporal punishment as well as emotional punishment is common) in that youths who were emotionally and/or physically punished by teachers were more likely to engage in deviance (see Moon & Morash, 2004). These results also suggested that teachers' emotional punishment/abuse may be as significant of a source of strain as physical punishment and future research testing GST should further examine the prevalence of teachers' emotional punishment and its effect on deviant behaviors in the U.S. setting. The findings showed that anger had a significant positive effect on deviance. Regarding the mediating effect of anger linking strains to deviance, these findings indicated a minimal mediating effect of anger on the relationship between strains and deviance, contrary to GST's prediction. Several strains (i.e., teacher emotional punishment and racial discrimination) which were significant predictors of deviance before the inclusion of the negative emotion, continued to exert significant effects on deviance, even after the inclusion of anger. These findings were consistent with some of the previous research on the mediating role of negative emotions between strain and deviance, indicating that strains may have direct and independent effects on deviance, regardless of experience of negative emotions (Baron, 2004; Mazerolle et al., 2000; Mazerolle & Piquero, 1998; Moon et al., 2008; Piquero & Sealock, 2004). This research provided partial support for the interaction effects between total strain and conditioning factors on deviance. Consistent with GST's prediction, the results showed that strained individuals were less likely to engage in general deviance, when they had high levels of problem solving ability. Strained individuals were more likely to commit violent deviance when they have positive attitudes toward

the use of violence, though it was marginally significant. Strained students, however, were more likely to engage in nonviolent deviance, when they had high levels of family support, contrary to the prediction of GST. The current study had several limitations. First, this research could not properly examine causations of strains, negative emotions, and deviance because of the use of the cross-sectional data. For example, it may be possible that students' deviant behaviors preceded any emotional punishment by teachers, rather than teachers' emotional punishment causing students' deviance. To better understand the sequences of strains, negative emotions, and deviance, longitudinal research is necessary. Second, the study used a convenience sample of university students. Therefore, there was a limitation of generalizing the current findings to a large general population. Another limitation of the current study was the use of a retrospective survey which may compromise recall accuracy of life events that happened during the last five years. Fourth, the current research measured trait-based negative emotions rather than situational negative emotions in direct response to strains, based on the assumption that individuals with high levels of persistent negative emotions (or called “emotional traits” by Agnew) are more likely to respond to strains with negative emotions (see Mazerolle et al., 2003). To adequately test the mediating effect of negative emotions on strains and deviance, future research is necessary to measure situational negative emotions in response to strain. The current research contributed to the empirical development of the general strain theory, measuring an extensive list of strains most likely to lead to deviance. As predicted by Agnew (2001), some strains such as teachers' emotional punishment and racial discrimination, largely ignored in previous empirical studies on GST, are significantly related to deviance. Considering the significant effect of teachers' emotional punishment on deviance among college students, future research needs to replicate these findings in the context of general strain theory, especially sampling U.S. adolescents. Acknowledgments The authors thank anonymous reviewers for their helpful comments. Appendix A. Scales used in the analysis Scales

Items and response items

Goal blockage How successful have you been at achieving the following (Cronbach's alpha = .65) goals over the past five years? 1. Academic/career goals 2. Social/family life goals 3. Financial goals 4. Health/appearance goals Response options: 1 (very successful) to 4 (not at all successful) The following questions ask about interaction among Family conflict (Cronbach's alpha = .60) family members during the past five years. 1. Did your parents argue with each other? 2. Did your family members (other than your parents) argue or disagree with each other? 3. Did you frequently argue or disagree with your parents? Response options: 0 (never) to 3 (always) Parental punishment The following items are about behaviors your parents (Cronbach's alpha = .88) may have engaged in to punish or discipline you during the last five years. 1. Pushed, grabbed, or shoved 2. Slapped 3. Hit or tried to hit with something

B. Moon et al. / Journal of Criminal Justice 37 (2009) 98–106 Appendix A (continued)

Appendix A (continued) Scales

Items and response items

Scales

4. Punished you unfairly 5. Called you names, embarrassing you 6. Negatively compared you to others

Response options: 0 (never) to 3 (always) Supportive family The following statements are about your relationship with (Cronbach's alpha = .84) your parents during the last five years. 1. My parents try to understand my situation 2. My parents show love and interest for me 3. My family members are close emotionally to each other

Response options: 0 (none) to 4 (ten or more times) Teacher emotional punishment (Cronbach's alpha = .89)

105

The following items are about behaviors your teachers may have engaged in to punish or discipline you during the last five years. 1. Isolated you 2. Embarrassed you 3. Negatively compared you to others 4. Ignored you 5. Made negative comments about you Response options: 0 (none) to 4 (ten or more times)

Racial discrimination Because of your race, how many times (Cronbach's alpha = .88) 1. Have you been treated unfairly by teachers? 2. Have you been treated unfairly by your fellow students? 3. Have you been treated unfairly by people in service jobs (store clerks, waiters, or others)? 4. Have you been treated unfairly by strangers? 5. Have you been treated unfairly by people in helping jobs (social workers, doctors, school counselors, or others)? 6. Have you been treated unfairly by neighbors? 7. Have you been treated unfairly by the police? 8. Have you been accused or suspected of doing something wrong (stealing, cheating, or breaking the law)? 9. Have people misunderstood your intentions and motives? 10. Have you been called a racist name or slur? Response options: 0 (never) to 4 (ten or more times) Gender discrimination Because of your gender, how many times (Cronbach's alpha = .88) 1. Have you been treated with less courtesy than other people? 2. Have you been called names or insulted? 3. Have you been threatened or harassed? 4. Have you been treated with less respect than you deserve? 5. Have people acted as if they are better than you? Response options: 0 (none) to 4 (ten or more times) Criminal victimization How many times have you and your family members been (Cronbach's alpha = .56) victimized (including attempts) by the following criminal conduct? 1. Theft 2. Robbery 3. Burglary 4. Sexual assault 5. Physical assault

Items and response items

Response options: 1 (strongly disagree) to 4 (strongly agree) Deviant peer association How many of your close friends (Cronbach's alpha = .87) 1. Purposely damaged or destroyed property that did not belong to them? 2. Used inhalant or illegal drugs? 3. Hit or threatened to hit someone? 4. Stole or tried to steal something of worth? 5. Sold drugs? 6. Carried a hidden weapon? 7. Injured someone? Response options: 0 (none of them) to 4 (all of them) Problem solving ability The following statements are about your personality during (Cronbach's alpha = .71) the last five years. 1. If I should find myself in a jam, I can think of many ways to get out of it. 2. My values and beliefs help me to meet daily challenges. 3. I can think of many ways to reach my current goals. 4. There are lots of ways around any problem that I am facing. Response options: 1 (strongly disagree) to 4 (strongly agree) Legitimacy of violence Do you believe that violence is often justified to (Cronbach's alpha = .82) 1. Defend one's right? 2. Achieve respect? 3. Obtain fair treatment? 4. Resist exploitation? 5. Avoid appearing weak? Response options: 1 (strongly disagree) to 4 (strongly agree) Violent deviance How many times in the past five years, have you (Cronbach's alpha = .80) 1. Fought in a group with others? 2. Carried a hidden weapon? 3. Joined a gang? 4. Hit or threatened to hit others? 5. Threw objects such as rocks or bottles at people? 6. Used force to get money (things) from other people? 7. Damaged, destroyed, marked up, or tagged somebody else's property on purpose? Response options: 0 (none) to 4 (nine or more times)

Response options: 0 (none) to 4 (four or more times) Negative community environment (Cronbach's = .81)

The following items ask how much of a problem each of the following was in your neighborhood during the last five years. 1. Vandalism (buildings and personal belongings broken and torn up) 2. Winos and junkies 3. Abandoned houses 4. Burglaries and thefts 5. Run-down and poorly kept buildings 6. Assaults and muggings

Nonviolent deviance How many times in the past five years, have you (Cronbach's alpha = .77) 1. Broke or tried to break into a building to steal something or just to look around? 2. Stole (tried to steal) something not belonging to you? 3. Knowingly bought, sold, or held stolen goods, or tried to do any of these things? 4. Used illegal drugs (marijuana, crack, or others)? 5. Sold illegal drugs (marijuana, crack, or others)? Response options: 0 (none) to 4 (nine or more times)

Response options: 0 (not a problem) to 2 (big problem) Anger The following questions ask about emotional feelings (Cronbach's alpha = .79) during the time that you were having troubles with problems mentioned above (i.e., family conflict, discrimination, or victimization) during the last five years. 1. Uncontrollable outbursts of temper 2. Urges to beat or harm someone 3. Urges to break things

References Agnew, R. (1992). Foundation for a general strain theory of crime and delinquency. Criminology, 30, 47−87. Agnew, R. (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 and Delinquency, 38, 319−361. Agnew, R. (2006). Pressured into crime: An overview of general strain theory. Los Angeles: Roxbury.

106

B. Moon et al. / Journal of Criminal Justice 37 (2009) 98–106

Agnew, R., & Brezina, T. (1997). Relational problems with peers, gender, and delinquency. Youth and Society, 29, 84−111. Agnew, R., & White, H. R. (1992). An empirical test of general strain theory. Criminology, 30, 475−499. Aseltine, R. H., Gore, S., & Gordon, J. (2000). Life stress, anger and anxiety, and delinquency: An empirical test of general strain theory. Journal of Health and Social Behavior, 41, 256−275. Bao, W., Haas, A., & Pi, Y. (2004). Life strain, negative emotions, and delinquency: An empirical test of general strain theory in the People's Republic of China. International Journal of Offender Therapy and Comparative Criminology, 48, 281−297. Baron, S. W. (2004). General strain, street youth and crime: A test of Agnew's revised theory. Criminology, 42, 457−483. Brendgen, M., Bukowski, W. M., Wanner, B., Vitaro, F., & Tremblay, R. (2007). Verbal abuse by the teacher during childhood and academic, behavioral, and emotional adjustment in young adulthood. Journal of Educational Psychology, 99, 26−38. Brendgen, M., Wanner, B., & Vitaro, F. (2006). Verbal abuse by the teacher and child adjustment from kindergarten through grade 6. Pediatrics, 117, 1585−1598. Brier, N. (1995). Predicting antisocial behavior in youngsters displaying poor academic achievement: A review of risk factors. Journal of Developmental and Behavioral Pediatrics, 16, 271−276. Broidy, L. (2001). A test of general strain theory. Criminology, 39, 9−35. Casarjian, B. E. (2000). Teacher psychological maltreatment and students' schoolrelated functioning. Dissertation Abstracts International, 60, 4314A. (UMI No. AAT 9956340) Derogatis, L. R. (1977). SCL 90: Administration, scoring, and procedures manual for the revised version. Baltimore: Johns Hopkins University. Eitle, D. J. (2002). Exploring a source of deviance-producing strain for females: Perceived discrimination and general strain theory. Journal of Criminal Justice, 30, 429−455. Jang, S. J., & Johnson, B. R. (2003). Strain, negative emotions, and deviant coping among African Americans: A test of general strain theory. Journal of Quantitative Criminology, 19, 79−105. Kim, J. (2002). An application of general strain theory to Korean adolescents: A test of the relations between negative stimuli at school and delinquency. Unpublished master's thesis, Michigan State University, East Lansing.

Landrine, H., & Klonoff, E. A. (1996). The schedule of racist events: A measure of racial discrimination and a study of its negative physical and mental health consequences. Journal of Black Psychology, 22, 144−168. Mazerolle, P. (1998). Gender, general strain, and delinquency: An empirical examination. Justice Quarterly, 15, 65−91. Mazerolle, P., Burton, V. S., Cullen, F. T., Evans, D., & Payne, G. L. (2000). Strain, anger and delinquent adaptations: Specifying general strain theory. Journal of Criminal Justice, 28, 89−101. Mazerolle, P., & Maahs, J. (2000). General strain and delinquency: An alternative examination of conditioning influences. Justice Quarterly, 17, 753−778. Mazerolle, P., & Piquero, A. (1997). Violent responses to strain: An examination of conditioning influences. Violence and Victims, 12, 323−343. Mazerolle, P., & Piquero, A. (1998). Linking exposure to strain with anger: An investigation of deviant adaptations. Journal of Criminal Justice, 26, 195−211. Mazerolle, P., Piquero, A., & 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. Moon, B., Blurton, D., & McCluskey, J. (2008). General strain theory and delinquency: Focusing on the influences of key strain characteristics on delinquency. Crime and Delinquency, 54, 582−613. Moon, B., & Morash, M. (2004). Adaptation of theory for alternative cultural contexts: Agnew's general strain theory in South Korea. Journal of International and Comparative Criminal Justice, 28, 77−104. Morash, M., & Moon, B. (2007). Gender differences in the effects of strain on three types of delinquency: A test of theory outside of the United States. Youth and Society, 38, 300−321. Paternoster, R., & Mazerolle, P. (1994). General strain theory and delinquency: A replication and extension. Journal of Research in Crime and Delinquency, 31, 235−263. Piquero, N. L., & Sealock, M. D. (2000). Generalizing general strain theory: An examination of an offending population. Justice Quarterly, 17, 449−484. Piquero, N. L., & Sealock, M. D. (2004). Gender and general strain theory: A preliminary test of Broidy and Agnew's gender/GST hypotheses. Justice Quarterly, 21, 125−158. Simons, R. L., Chen, Y., Steward, E. A., & Brody, G. H. (2003). Incidents of discrimination and risk for delinquency: A longitudinal test of strain theory with an African American sample. Justice Quarterly, 20, 827−854.