Strain,Negative Emotions,and Juvenile Delinquency

Abstract

General strain theory (GST) is an established criminological theory. Although the theory has been examined by many and enjoys empirical support, some limitations of previous studies need to be addressed. Many previous studies rely heavily on samples from Western countries, mostly the United States; thus, possible cultural influences are ignored. Although a few studies have moved forward by using subjects from Asia (e.g., China, Korea), these studies only provide empirical results regarding whether GST is applicable in other cultures. However, these studies do not directly compare Western and Eastern countries. The present study used two samples from the United States and Taiwan to directly compare and contrast central GST propositions. Although most of the GST propositions are found to be similar between the U.S. and Taiwanese juveniles, some differences were also discovered. Explanation of these similarities and differences from their cultural perspectives are offered in this study.

Keywords

strain anger depression juvenile delinquency

Introduction

Psychologists and sociologists often refer to adolescence as a period of storm and turmoil, a reference that points out not only the high risk of involvement in various antisocial behaviors during this period (Gottfredson & Hirschi, 1990), but also the occurrence of increased stress and high levels of negative emotions. DuRant, Getts, Cadenhead, Emans, and Wood (1995, p. 233) suggest that “[L]ife stress can have a deleterious impact on the psychological adjustment of adolescents,” and relate the impact of such stress or strain to various negative outcomes, including delinquency and negative emotions (Lin, Cochran, & Mieczkowski, 2011).

Agnew’s (1992, 2001, 2006) general strain theory (GST) provides a rich framework for analyzing the underlying mechanisms that connect strain, negative emotions, and delinquency in adolescents. Although GST seems to be an important criminological theory, the tests so far carried out on GST have relied almost exclusively on samples drawn from the United States. Only a few studies have applied this theory in other countries (e.g., Canada, China, Greece, Korea, and Russia), and studies that focus on similarities and differences between the test results of the GST across nations are virtually nonexistent (see Botchkovar, Tittle, & Antonaccio, 2009; Sigfusdottir, Kristjansson, & Agnew, 2012, for exceptions).That explains why Froggio (2007) questions the utility of GST in explaining juvenile delinquency in other countries, because cultural attitudes and values may influence one to define events and conditions as either typical or stressful; hence, what is seen as a strain in the United States may not be seen as such in other cultures, and may not lead to negative emotions and delinquency (Chun, Moons, & Cronkite, 2005). Consequently, Adler (1996) argues for testing and applying criminological theories in other countries, and cross-cultural studies help in refining a theory to accommodate cultural differences (Kim, Triandis, Kagitcibasi, Choi, & Yoon, 1994; Kohn, 1987).

To test GST, the present study employs two samples, one from the United States (Western culture) and the other from Taiwan (Eastern culture). Both samples recruit similar subjects (12- to 19-year-old adolescents) and use identical survey instruments. Hence, similarities in results should not only support the core theoretical propositions of GST but also validate the GST model in both cultures, whereas differences could help to expand and revise GST to incorporate variations among cultures.

Review of GST

Agnew’s GST defines strain as the product of negative relationships with others: “relationships in which the individual is not treated as he or she wants to be treated” (Agnew, 1992, p. 48). This negative treatment refers to different types of strain, among which Agnew identified three general types. First, strain may arise because the individual fails to achieve positively valued goals, such as good grades in school. Second, strain may be generated by the removal of positively valued stimuli, that is, the loss of something valued, such as breakup with a boyfriend or girlfriend. Finally, strain may arise from the presence of noxious stimuli, such as the experience of victimization. More recently, Agnew (2001, 2006) adds four characteristics of strain that help to define its criminogenicity. According to these characteristics, a strain is criminogenic if it is perceived as unjust, high in magnitude, low in social control, and likely to create an incentive to commit crime. Hence, whether a strain is related to delinquency is dependent on these four characteristics.

According to GST, each type of strain can lead an individual to experience an array of negative emotions, including anger, fear, and depression. Among these negative emotions, anger is the most important to GST (e.g., by increasing an individual’s level of outward attribution toward the source of strain). Although anger is the most criminogenic and an important negative emotion in GST, other negative emotions, such as depression, might also be related to delinquency. In fact, some studies found that depressed juveniles are highly likely to display problem behaviors, including general delinquency (Beyers & Loeber, 2003), substance use (Ford & Schroeder, 2009), as well as violent behavior (Ostrowsky & Messner, 2005).

As the foregoing discussion shows, the basic premise of GST is that strain generates negative feelings, which, in turn, make an individual want to do something about the bad feeling and/or stressful situation. Thus, strain has not only direct effects on delinquency, but also indirect effects through negative emotions. The general finding of empirical research is that a strain–delinquency relationship exists (Baron, 2004; Broidy, 2001; Eitle & Turner, 2003; Hay, 2003; Slocum, Simpson, & Smith, 2005). However, with regard to the strain–negative emotions and the mediating effects of negative emotions, the results are mixed (Ford & Schroeder, 2009). For example, Broidy (2001) finds that strain is significantly related to anger, which in turn is related to crime, whereas Baron and Hartnagel (1997) find no relationship between anger and delinquency.

Besides the general theoretical pattern, GST also recognizes differences between groups. Agnew (1992) believes that different groups (e.g., adults vs. adolescents) may experience different strains and develop different strain experiences. For example, family stressors are said to be influential among young adolescents, while peer strain affects middle adolescents (Compas & Phares, 1991), but strains from peers may exert a lesser effect on adults. Broidy and Agnew (1997) use group and cultural influences and extend GST to explain gender differences in offending. They suggest that gender differences might be due to qualitative differences in strain and negative emotions (e.g., males and females might encounter different type of strains) but not quantitative differences (e.g., males experience more strains than females). For example, they argue that females experience anger along with other negative emotions (e.g., distress), and as a result females commit fewer outer-directed delinquent acts (e.g., violence) than males (Jang & Johnson, 2003). Similarly, social class and/or race, age, and living environment may create different life experiences and culture; hence, differences may exist in each component of GST or in the whole GST process. Indeed, Agnew (2006) extends GST to accommodate these differences by showing, for example, different criminogenic strains for different age groups.

Cultural Influences and Cross-Cultural Studies

While GST can theoretically explain group differences, these extensions are limited to the subculture of a particular country (e.g., males and females in the United States). The impact of culture on the relationships between strain and consequent outcomes (e.g., emotions and delinquency) in different countries with dramatically different cultures is unknown (e.g., Chinese culture). Most of the published studies employed samples from the United States (Froggio, 2007) or other countries with fairly similar cultural backgrounds (e.g., Canada or England), or cultures that are very different from the East (e.g., Russia). This may hinder the generalizability of GST to other dramatically different cultures, and limit the possibility of further understanding cultural influences on GST. Piquero and Sealock (2000) argue that researchers need to adapt and test the theory in other countries to increase generalizability and foster empirical development of the theory.

To extend GST’s generalizability and continue recognizing its sensitivity to cultural influences, applying GST to countries with different cultural backgrounds is important. The case in hand is Eastern culture, particularly Chinese culture, which is very different from Western culture such as that of the United States. The major difference between the Western or U.S. culture and the Eastern or Chinese culture is in their positions on the continuum of individualism and collectivism. The United States has an individualistic culture, as documented in various historical accounts and other scholarly writings (Bellah, Madsen, Sullivan, Swidler, & Tipton, 1996; Hsu, 1983). For example, Bellah et al. (1996) write that “individualism lies at the very core of American culture” (p. 142). In contrast to this, Chinese culture has been described as particularly collectivistic (Leung & Bond, 1982).

An individualistic society (e.g., the United States) places higher priority on self and emphasizes “I” consciousness. As such, the individual is the central unit of society, and its members give priority to personal goals over those of others (Hofstede, 2001; Triandis, 1995). Therefore, the individualistic society promotes the “independent self” (Markus & Kitayama, 1991), such that people place high value on independence, individual freedom, and personal achievement. In contrast, collectivistic societies are oriented toward groups (e.g., family and nation) as the central units of society. Hence, obligation, interdependence, and fulfillment of social roles are the focal points. Collectivistic societies stress the feeling of “we,” and their members choose to give priority to collective goals and to emphasize group solidarity (Hofstede, 2001; Triandis, 1995). Collectivism, then, cultivates an “interdependent self,” so that individuals place high value on cooperation, mutual support, and maintenance of group harmony.

Because self-achievement, personal rights, and autonomy are the primary values in an individualistic society, struggles in these areas in life may be stressful in that society (Chun et al., 2005). In contrast, individuals from a collectivistic society, where relational harmony and interdependence are more important, may see problems related to pleasing parents and fulfilling family goals as more stressful. Compared with American college students, Asian students (e.g., Hong Kong Chinese) have reported higher needs for affiliation and higher sensitivity to social rejection (Yamaguchi, Kuhlman, & Sugimori, 1995). Heine and Lehman (1995) find that Japanese college students, members of a collectivistic culture, considered interdependent events (e.g., “Sometimes in the future you will do something that makes your family ashamed of you”) more stressful, and independent events (e.g., “After growing old, you will find that you never realized your most important dreams”) less stressful, compared with Euro-Canadian college students.

Cultural differences may affect emotional responses to strain as well. As mentioned earlier, the self is an important unit in an individualistic culture such as the United States, and expressing anger to some extent might be considered appropriate when this expression of anger is related to one’s personal goal or identity. Markus and Kitayama (1994),when discussing expressing anger, state that in the United States, “it is the emotional states that have the individual’s internal attributes (his or her needs, goals, desires or ability) as the primary referent that are most commonly manifest” (p. 101). Hence, one would expect that strains related to an individual’s goals or identity are more likely to arouse anger in the United States. A recent study (Horton, Rice, Piquero, & Piquero, 2012) found that American adults, in comparison to their Tibetan counterparts, are more likely to feel anger after experiencing injustice and think that expressing anger is sometimes a good thing. In Chinese culture, which focuses on harmony within relationships and interdependence (Markus & Kitayama, 1991), negative emotions may be caused by failing to maintain relationships or meeting others’ expectations; consequently, one would expect anger to arise from negative interpersonal relationships, rather than from failure to achieve a goal. From an extension of this point, it follows that because the focus of each culture is different, the types of criminogenic strain and the negative emotions in response may also be different.¹

Because of these potential differences, applying GST in Eastern countries seems important. Several previous studies have directly applied GST in some Asian countries: China (Bao, Haas, & Pi, 2007), Hong Kong (Cheung & Cheung, 2010), South Korea (Moon, Blurton, & McCluskey, 2008; Moon & Morash, 2004; Moon, Morash, McCluskey, & Hwang, 2009; Morash & Moon, 2007), Taiwan (Lin & Mieczkowski, 2011), and the Philippines (Maxwell, 2001). The basic conclusion from these studies is that the strain–delinquency relationship does exist in these collectivistic cultures. However, they also found some differences that suggest a need for further cross-cultural investigations. For example, Chinese culture emphasizes educational attainment and relational harmony, which creates or increases strains emanating from these areas in an adolescent’s life (Bao et al., 2007; Moon & Morash, 2004). Similarly, in a collectivistic culture, parent-related strain may have a lesser effect on delinquency because of the importance of obeying parents and authorities in the culture (Cheung & Cheung, 2008).

Although the above studies employed GST to explain juvenile delinquency in some Asian countries, they had certain limitations. First, the sample employed by those studies was not similar enough to the one from United States for a valid comparison; consequently, similarities and differences were discussed only on the basis of different samples and research instruments. Second, those studies did not apply the full GST model, which includes all major types of strain and negative emotions (see Moon et al., 2009, for an exception). Therefore, what is needed is a study that can overcome these two limitations and shed light on the influence of culture on GST.

The Present Study and Research Question

Based on the foregoing discussion, the present study attempted to compare the basic GST model between a Western and an Eastern country. For this, two samples, one of U.S. adolescents and the other of Taiwanese adolescents, were used. Most importantly, subjects from both countries are very similar in many respects (e.g., age) and the instrument used to collect information was the same for both groups. This is important because the comparison between the effects of a particular strain on delinquency across countries would be meaningful only if the samples and their methods of study are identical. A recent study conducted by Sigfusdottir et al. (2012) also points out that previous studies could not afford this comparison because their results were based on “differences,” that is, different measures on the “same” theoretical concepts and different analytical techniques (e.g., control different variables).

This study addresses two research questions: First, whether strain increases delinquency in the United States and Taiwan, and second, whether the strain–delinquency relationships in the United States and Taiwan are similar. From the literature review presented earlier, one would expect that unjust strain, victimization, and negative life-events increase the likelihood of delinquency, because previous studies found that these strains increase delinquency in the United States (Agnew, 2006; Hay, 2003; Hoffmann, Cerbone, & Su, 2000) and Eastern countries (Cheung & Cheung, 2010). They further show that goal-related strain had no impact on delinquency in the United States and Russia (Botchkovar & Broidy, 2010; Broidy, 2001); its influence on Taiwanese students, however, remains to be tested. In addition, the influences of different cultures on the strain–delinquency relationship will have to be distinguished, because culture helps members of a society define a strain (Sanchez, Spector, & Copper, 2005). In regard to the question of whether strain increases anger and depression similarly or differently in the two countries, it has already been mentioned that expressing anger might threaten relationship harmony; hence, it is discouraged in Chinese culture. On the contrary, expressing anger that is related to some strains is sometimes regarded as acceptable in the United States (Horton et al., 2012; Markus & Kitayama, 1991). Hence, one would expect that the influences of strain on anger would be different. Expression of depression, however, has no such harmful outcomes. A recent report from the World Health Organization (2001) indicates that depression is becoming epidemic among youths around the world; hence, the strain–depression relationship might be quite similar in both countries.

Method

Sample

This study used two samples:² one from the United States and the other from Taiwan. The U.S. sample is from an existing cross-sectional data set compiled for studying juvenile delinquency among middle and high school students. All the subjects of that sample were recruited from a middle school and a high school—both public—in a medium-sized southwest city. The subjects of Taiwanese sample were from a public junior and a public senior high school in one of the school districts of Taiwan’s second largest city in Southern Taiwan.

U.S. sample

The U.S. data were collected in 1998 from a metropolitan area. The middle school selected for the study was one of the two middle schools (Grades 6-8) in that area; the school enrolled 1,294 students during the 1998-1999 school year and had an average class size of 25 students. Students from all Social Studies classes were invited to participate in the survey. Before the actual survey, a passive parental consent form was distributed to all the students (see Verrill, 2008, for details). On the day of the survey, a researcher explained them the purpose of the study, reminded them that their participation was voluntary, and that the confidentiality of the information they gave would be ensured. The researcher remained on-site thereafter to answer questions related to the survey. The final response rate was 81% (N = 1,049).

The high school selected was one among the many schools (Grades 9-12) in that area with about 1,848 students enrolled during the 1998-1999 school year; the average class size was 33 students. Students from a random sample of 30 third-period school classes were asked to participate in the survey. A passive parental consent procedure was followed, similar to the one followed in the case of middle school students. On the day of administration, a researcher apprised the participants of the purpose of the study, explained that participation was voluntary, and assured maintenance of confidentiality for the information they gave. Furthermore, the researcher remained available to answer questions. The final response rate was 79% (n = 625).

Taiwanese sample

Additional data for this study were collected from a sample of junior and senior high school students in a large city in the southern part of Taiwan, which spreads over an area of about 59.3 square miles and has a population of about 1.5 million. Two public schools were selected from the same school district for this study. One of the three large junior high schools in the district, similar in size to the American public middle school, was selected. The age range of junior high school students in Taiwan, as a whole, was 12 to 15 years. The total number of students in the selected junior high school was 2,265 in 2010; among these, male students constituted about 51% (n = 1,153), and the average class size was 37 students (Education Bureau, Kaohsiung City, 2010). The public senior high school selected for the present study had 1,789 students in 2010, with 48% male students (n = 867) and an average class size of 35 students (Education Bureau, Kaohsiung City, 2010). Compared with the public high school in the U.S. sample, the Taiwanese senior high school had slightly fewer students.

The selected school principals agreed to provide teachers help in administering the survey by distributing the survey questionnaire and by remaining on-site to answer questions. All the students who were present on the day of the survey were given the opportunity to participate, but because participation was voluntary, not all students participated. The participation rates were 91% and 96% for the junior high and the senior high school, respectively.

Measurement of Variables³

Delinquency

Self-reported delinquent behaviors included in the present study are fairly commonly reported in the criminological literature (e.g., drug use, stealing). However, in both countries, for many items, less than 1% of the subjects answered in the affirmative; this was especially true in Taiwan. Because of such lack of variance to be explained, all those items were discarded and finally only three delinquent acts were retained as the outcome variables. These variables not only had relatively high percentages of students reporting that they had done the acts in the past 12 months, but also were theoretically relevant to and representative of the general delinquency categories⁴ (e.g., property crime, violent crime). The three delinquent acts were “alcohol use,” “purposely damage property,” and “hit someone with intention to hurt.”

Individuals reporting that they had not committed a particular act received a score of 0, and those reporting that they had committed the act received a score of 1. Each of these three delinquent acts was examined separately for two reasons: first, because they represent different domains of acts, and the second, because Agnew (2006) advocates separation of coping behaviors by the type of act. This analysis helps to identify specific strain–delinquency relationships.

Strain

Agnew (1992) outline three major types of strain:⁵ failure to achieve positively valued goals, removal of positively valued stimuli, and presentation of noxious stimuli. For the present study, the first type of strain was measured by (a) the discrepancy between desired goals and actual outcomes and (b) experience with unjust outcomes; the second type by (c) stressful life-events involving loss; and the third type by (d) victimization.

Failure to achieve positively valued goals—disjunction between desired and actual outcomes

To measure goal blockage, the present study covered various goals that the youth might find important in his or her current life (e.g., relationship, autonomy). Specifically, students were asked to state whether they strongly agree, disagree, or strongly disagree with several statements regarding receiving respect from parents and teachers (three items), relationships with others (two items), and autonomy (two items).

Each item was divided by its own standard deviation (SD) rather than creating z score before summing them for two reasons. First, dividing the raw score by its SD prevents assigning unduly greater weight to items that have great variance; there is no theoretical reason to believe that some items should have greater weight than others. Second, the z score was not used, because it would potentially eliminate the mean differences across countries. Higher scores indicated that students did not achieve or were unsatisfied with their situation regarding these goals and hence presumably experienced greater strain.

Failure to achieve positively valued goals—unjust outcomes

Agnew (1992) argues that a relationship is stressful when it is unjust, in other words, when the outcome/input ratio is not one, that is, when individuals feel that they have been under-rewarded (Hegtvedt, 1990). To capture these feelings of unjust strain, seven statements about unequal relationships were made in which the students would have been involved and the students were asked to report whether they agree or disagree with those statements. For example, in response to the statement, “Many students don’t study as hard as I do, but they still make better grades,” the students can choose from “strongly agree” to “strongly disagree.” The theme used for creating this scale is similar to the one used for goal strain. Higher scores indicated greater unjust strain.

Loss of positive stimuli—negative life-events

Strain can result from losing a positively valued goal (Agnew, 1992). The most widely used instrument to capture this type of strain whether in the stress literature or in studies that examine GST is the negative life-event scale. Although many negative life-event scales exist, Turner and Wheaton (1995) argue that there is no advantage in using one particular scale rather than the other and that the scales should be tailored to fit the population under study. They further suggest that the 1-year time frame should be used and that unweighted indices are as useful as any. The negative life-event scale (10 items) used in the present study has all the four recommended features, and all events are related to loss of positive stimuli (e.g., death of a relative, loss of a friendship). Students were asked to report whether such event had occurred in the past 12 months (yes = 1, no = 0). Because each item had only two categories, the z score was created for each item and the summation of the z scores gave the negative life-event scale. Higher scores indicated that a student had experienced many stressful life-events and consequently had a higher level of strain.

Presentation of noxious stimuli—victimization

We used criminal victimization to capture this strain because it is one of the most severe noxious stimuli and types of strain. Six types of victimization were used to study this strain, such as “having been forced to give up money or possessions” or “having been physically attacked by others.” Students were asked to indicate whether they experienced any such incident in the past 12 months (yes = 1, no = 0). Summation of z scores created this scale, as the one for negative life-events. Higher scores indicated a higher level of strain.

Negative emotions

Anger

Eight items were used to measure anger; five of them were adapted from Spielberger’s (1988) State-Trait Anger Expression Inventory (STAXI), which examines anger as a personality trait that is situational (Wareham, Cochran, Dembo, & Sellers, 2005). For example, one of the five items goes thus: “When I get frustrated, I feel like hitting others.” The other three items capture angry feelings or reactions that are more situational too (e.g., “It makes me mad when I don’t get the respect from others that I deserve”). The eight items used here appear to be primarily situational, which is in conformity with the suggestions of Agnew (2006). Response categories were so coded that a higher score indicates a higher level of anger.

Agnew (2006) and others (Horton et al., 2012) contend that a state-like measure is more closely related to GST’s perspective than a trait-like measure. However, the results from some studies that used state-like measures (Botchkovar & Broidy, 2010; Ganem, 2008; Jang, 2007) show results similar to those from studies that used a trait-like measure (Piquero & Sealock, 2000; Robertson, Stein, & Schaefer-Rohleder, 2010). Hence, while using a state-like measure is to be encouraged, it does not mean that trait-like measures provide no insight into GST. Besides this general point, one related issue is that the state-like measure should be related to the strain measured in the study. Although our anger measure is situational in nature, each item was not linked to a specific strain situation measured in the present study. However, some of the items represent an anger response to strain situations that bear a meaning similar to that of the strain measure of this study. For example, the item “It makes me mad that others are able to spend more money than I can,” is related to unjust feelings or failure to achieve a goal. Although some items are related to a similar strain situation, they are still limited because each item does not directly indicate the strain situation measured in this study. Hence, our measure provides only a conservative measure.

Depression

Four items were used to measure depression. In contrast to the anger measures, which are more situational, the depression items appear to be trait like or symptoms of depression. The four depression items, adapted from the Beck Depression Inventory–Second Edition (BDI-II; Beck, Steer, & Brown, 1996), asked the students to indicate to what extent the following statements described them: (a) I do not look forward to things as much as I used to; (b) I find it hard to keep my mind on schoolwork; (c) I sleep very well (reverse coded); and (d) I have lots of energy (reverse coded). The first two items include an affective component of depression (Storch, Roberti, & Roth, 2004) and the last two items a somatic component. The response categories are identical to those of the anger measure.

The exploratory factor analysis results indicate that a one-factor solution is acceptable in the U.S. sample, because the loadings of all the four items were more than .4 except for the first item, which was .349. On the contrary, these four items did not load evenly in the Taiwanese sample, as the first two items had very low loadings. However, the eigenvalue (sum of squared loadings) was more than 1 for the Taiwanese sample. Thus, this study still treated depression as one factor in Taiwan, so that the model construct was similar across the two countries. One possible explanation is that Chinese people complain of physical illness readily, but report psychological discomfort relatively seldom. This phenomenon, called somatization, refers to people who emphasize physical symptoms of depression more than they emphasize the psychological ones (Heine, 2008). Moreover, in Chinese society, psychological illness carries enormous stigma; hence, the results obtained in the Taiwan sample are on the expected lines.

Demographic variables

The present study included two important demographic variables, gender and age, which have been shown to influence the strain–delinquency relationship. Males and females may have different reactions and coping strategies under strain/stress (Mirowsky & Ross, 1995). Gender was, therefore, included in the survey, with male coded 1 and female coded 0. Age (students’ age on the date of the survey) was also included, because as individuals grow older, they may accumulate different experiences and develop different responses to strain and different coping skills (Thoits, 1995). Race was not considered a control variable, because, in Taiwan, there is virtually no “minority group,” except for some aboriginal individuals who constitute only 2% of the total population (Directorate-General of Budget, Accounting, and Statistics, Executive Yuan, Taiwan, 2011; see the descriptive statistics for all variables in Table 1).

Table 1.

Descriptive Statistics of Strain and Negative Emotion Variables.

	United States (N = 1,516)			Taiwan (N = 1,717)
Variable	M (α)^a	SD ^b	Minimum/Maximum	M (α)	SD	Minimum/Maximum
Goal strain	15.03 (.69)	3.70	7.22/28.88	14.57 (.80)	4.13	7.45/29.80
Unjust strain	16.84 (.65)	3.96	7.47/28.81	19.10 (.68)	4.10	8.03/32.13
Life-event^c	0 (.77)	4.97	−7.35/15.12	0 (.68)	5.25	−4.56/24.33
Victimization^c	0 (.45)	3.21	−2.29/22.07	0 (.40)	3.20	−1.83/35.24
Anger	19.37 (.77)	4.92	8.18/32.73	17.58 (.80)	5.14	9.03/36.13
Depression	8.7 (.54)	2.60	4.01/16.03	9.34 (.51)	2.54	3.98/15.90

The reported mean and α is the average across the five data sets.

The reported SD is the average across the five data sets.

The means of the life-event and victimization scales are 0 because of standardization.

Analytical Strategies

The present study used path analysis to test the GST model in U.S. and Taiwanese samples. The regular standard error of the mean (SEM) and path analysis models employ maximum likelihood (ML) estimation and assume multivariate normality. However, the delinquent acts covered by this study are all dichotomized variables, which violate the normality assumption. In such a situation, the proper alternative estimator is the weighted least squares mean and variance (WLSMV), which is used to account for non-normality. Flora and Curran (2004) show that WLSMV is an accurate estimator when sample sizes vary from 100 to 1,000 with various degrees of non-normality and model complexity. Each sample in the present study had more than 1,500 subjects; hence, employing WLSMV as the estimator was considered appropriate.

For the further understanding of the GST model, by comparing and contrasting it across the two countries, the SEM multiple-groups approach was used. In contrast to these conventional approaches, multiple-groups analysis in the SEM framework provides greater flexibility, offering direct comparisons across different groups on various parameters (e.g., path coefficients or group means), which reveal not only differences but also similarities. For example, to test for similarities, one can impose the equality constraint on paths across samples. If such constraint does not worsen the fit of the model, then similarity is confirmed. In addition, this approach utilizes all the subjects from all the groups and thus provides greater statistical power. Maruyama (1997) argues that “modeling groups as multiple populations is a superior alternative to dummy coding,” when the samples are sufficiently large (p. 259).

The present study employed two steps to handle missing data. First, students who did not report either their gender or age and those who omitted many items in reporting were excluded from the analysis. After such exclusion, which left the U.S. sample with 1,516 subjects and the Taiwanese sample with 1,717, multiple imputation (MI; Rubin, 1987) with the Markov chain Monte Carlo (MCMC) method was used to replace the missing values for all items, thereby generating five complete data sets (Allison, 2003) for subsequent analyses.

Results

The first analysis tested the GST model (strain → negative emotions → delinquency) in the U.S. and Taiwan samples. As can be seen in Table 2, unjust strain and goal strain have no direct effects on delinquency in either sample. In contrast, negative life-events and victimization exerted large, direct, and statistically significant effects on delinquency in both countries. For example, the students in both countries whose victimization was 1 SD above the mean, other things being equal, were about 55% more likely to use alcohol than the students at the mean level. In addition, anger and depression remained potent risk factors in delinquency in both countries. These results support only partially Agnew’s (2006) assertion that anger should be related to outer-directed crime (e.g., violent crime) and depression to inner-directed crime (e.g., substance use), because both anger and depression are related to hitting someone and alcohol use. Besides the significant effects, gender and age exerted significant impacts on negative emotions and delinquency.⁶

Table 2.

The Multiple-Groups Analysis of the GST Model.

	Anger	Depression	Damaging property	Hitting someone	Alcohol use
Model for the United States (N = 1,516)
Goal strain	.237 (.032)**	.122 (.018)**	−.021 (−.014)	−.017 (.013)	−.010 (.012)
Unjust strain	.328 (.031)**	.159 (.018)**	−.023 (−.013)	0 (.012)	−.018 (.011)
Negative life-event	.053 (.023)*	.028 (.012)*	.020 (.009)*	.013 (.008)	.017 (.007)*
Victimization	.212 (.036)**	.051 (.019)**	.051 (.012)**	.091 (.011)**	.063 (.012)**
Age	.298 (.065)	.305 (.030)**	−.027 (.024)	−.068 (.023)**	.255 (.020)**
Gender	.065 (.230)**	−.567 (−.122)**	.204 (.083)*	.352 (.075)**	−.148 (.071)*
Anger			.060 (.009)**	.047 (.009)**	.038 (.008)**
Depression			.082 (.017)**	.067 (.017)**	.075 (.015)**
Model for Taiwan (N = 1,717)
Goal strain	.138 (.025)**	.133 (.014)**	.014 (.011)	−.003 (.012)	−.005 (.009)
Unjust strain	.204 (.027)**	.098 (.015)**	.014 (.011)	.019 (.011)	.009 (.009)
Negative life-event	.097 (.020)*	.032 (.012)**	.026 (.007)**	.033 (.008)**	.023 (.006)**
Victimization	.129 (.026)**	.035 (.017)*	.026 (.009)**	.027 (.010)**	.008 (.009)
Age	.322 (.080)**	.110 (.036)**	−.010(.027)	−.084 (.032)**	.047 (.021)*
Gender	.652 (.242)**	−.143 (−.118)	.432 (.085)**	.381 (.099)**	.033 (.066)
Anger			.029 (.008)**	.043 (.009)**	.022 (.006)**
Depression			.046 (.016)**	.037 (.017)*	.059 (.013)**

Note. Unstandardized coefficients are shown, with standard errors in parentheses, and the coefficients are the averaged estimation from the five complete data sets. All paths are free to be estimated. The model is estimated with anger correlated with depression (United States, r = .262; Taiwan, r = .186).

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

All these results thus indicate that, up to this point, the GST model is reasonably similar across cultural boundaries. This conclusion is important, because the data were collected from almost identical survey items. This conclusion confirms the results of previous studies that applied GST models in other Asian countries (Bao et al., 2007; Lin & Mieczkowski, 2011; Maxwell, 2001; Moon & Morash, 2004). However, more can be gained if the differences in path coefficients are directly examined. To this end, multiple-groups analysis was conducted across these two samples. ⁷

The results of Table 2 provided the basis for imposing constraints on paths, which facilitates further testing of similar or different paths.⁸ The most interesting differences related to damaging property are as follows: goal strain → damaging property, unjust strain → damaging property, and negative life-event → damaging property. Consequently, constraints were imposed on these three paths when conducting multiple-groups analysis. Table 3 sums up the results of this analysis. It can be seen that students who experienced negative life-events were more likely to experience depression and anger in both the United States and Taiwan, and the magnitude was the same for these relationships between strain and negative emotions. However, students who experienced victimization were more likely to experience depression in both countries and at a similar magnitude than students who did not have the experience. The same is applied to the link between goal strain and depression. However, while both unjust strain and goal strain increased students’ anger, the magnitude was not the same across countries; it was always stronger in the United States than in Taiwan. Looking at depression, the only cultural difference that can be seen was in the unjust strain–depression relationship, which was stronger in the U.S. sample. The other three strain–depression relationships were similar between the United States and Taiwan.

Table 3.

Multiple-Groups Analysis With Constraints (Damaging Property).

	Anger		Depression		Damaging property
	United States	Taiwan	United States	Taiwan	United States	Taiwan
Goal strain	.230 (.032)**	.140 (.025)**	.129 (.011)**		.001 (.005)
Unjust strain	.328 (.031)**	.204 (.027)**	.159 (.018)**	.098 (.015)**	−.022 (.013)	.014 (.011)
Negative life-event	.077 (.015)**		.030 (.009)**		.023 (.005)**
Victimization	.211 (.036)**	.129 (.011)**	.042 (.013)**		.052 (.013)**	.025 (.009)**
Age	.298 (.059)**	.322 (.080)**	.305 (.030)**	.111 (.036)**	−.025 (.024)	−.011 (.027)
Gender	.066 (.230)	.651 (.242)**	−.567 (.122)**	−.143 (.118)	.202 (.083)**	.432 (.085)**
Anger					.058 (.009)**	.030 (.008)**
Depression					.079 (.017)**	.049 (.016)**

Note. Unstandardized coefficients are shown, with standard errors in parentheses, and the coefficients are the averaged estimation from the five complete data sets (United States, N = 1,516; Taiwan, N = 1,717). The model is estimated with anger correlated with depression (United States, r = .268; Taiwan, r = .185). The model fits the data well: χ ² (6) = 7.21, ns. comparative fit index (CFI) = .999; Tucker Lewis Index (TLI) = .992; root mean square error of approximation (RMSEA) = .011. For some paths, we present one coefficient for both countries because we put constraint on this particular path.

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

Of the three paths on which constraints were imposed (goal strain, unjust strain, and negative life-events → damaging property), one was rejected: unjust strain → damaging property, χ²(1) = 5.44, p < .05. Although the magnitude and the sign of this path were different between the United States and Taiwan, the path had an insignificant effect on damaging property in both countries. The other two imposed constraints did not worsen the model fit. Hence, goal strain did not have any influence on damaging property in either country, whereas the occurrence of negative life-events exerted a significant and similar effect on damaging property in both countries. That is, students who experienced more negative life-events were about 53% more likely to damage property than other students in both countries. Victimization, anger, and depression continued to have effects on damaging property in both countries, although the magnitude of the effects was always stronger in the United States than in Taiwan.

When looking at hitting someone, besides the constraints imposed on the strain–negative emotion relationships, four other constraints were imposed: unjust strain → hitting someone, negative life-events → hitting someone, victimization → hitting someone, and anger → hitting someone. Of these, one was rejected (victimization → hitting someone), χ²(1) = 21.78, p < .01. Hence, although victimization is a risk factor for violent delinquency (e.g., hitting someone) in both countries, the effect of this stressor on hitting someone was significantly stronger in the United States than in Taiwan. Specifically, American students who experienced more victimization incidents were about 57% more likely to hit others than other students, but this difference was only 52% in Taiwan. In addition, U.S. students (β = .065) were more likely to soothe their feeling of depression by hitting someone than were the students in Taiwan (β = .039).

Negative life-events are also an important risk factor for hitting someone in both countries, but the influence was about the same. That is, students from both countries who encountered more negative life-events were about 53% more likely to cope with this strain by hitting someone than their counterparts who experienced fewer events of that type. The most interesting similarity was that of the anger–hitting someone relationship, because one would expect that relationship to be stronger in the United States than in Taiwan. However, the imposed constraint did not worsen the model fit, which indicates that relationship was the same across cultures. One possible counterexplanation is the recent surge in violent crime and campus violence in Taiwan, which might make students more likely to vent anger through violent acts.

Finally, five paths deserve particular attention with regard to alcohol use: goal strain → alcohol use, unjust strain → alcohol use, victimization → alcohol use, anger → alcohol use, and depression → alcohol use. Consequently, constraints were imposed on these five paths in addition to the constraints imposed on the strain–negative emotions paths. The results are fairly similar to those in Tables 3 and 4, insofar as the strain link with negative emotion is concerned (Table 5). Of these five paths, only one was rejected: victimization → alcohol use, χ²(1) = 14.93, p < .01. Hence, the effects of victimization on alcohol use had dramatically different effects between the two countries. Students in the United States were about 55% more likely to drink alcohol to cope with victimization, whereas those in Taiwan did not resort to such a coping strategy. Other imposed paths did not worsen the model fit. The similarities of the four paths were confirmed. Hence, students in both countries who experienced anger and depression were more likely to use alcohol, but goal strain and unjust strain had no effects on drinking alcohol in either the United States or Taiwan. In contrast, the occurrence of negative life-events exerted a significant effect on alcohol use in both countries, although the magnitude was significantly different.

Table 4.

Multiple-Groups Analysis With Constraints (Hitting Someone).

	Anger		Depression		Hitting Someone
	United States	Taiwan	United States	Taiwan	United States	Taiwan
Goal strain	.237 (.033)**	.138 (.025)**	.129 (.011)**		−.017 (.013)	−.003 (.012)
Unjust strain	.325 (.031)**	.207 (.028)**	.157 (.018)**	.098 (.015)**	.010 (.008)
Negative life-event	.077 (.015)**		.030 (.008)**		.023 (.005)**
Victimization	.211 (.036)**	.129 (.026)**	.042 (.013)**		.092 (.011)**	.027 (.010)**
Age	.298 (.059)**	.323 (.080)**	.305 (.030)**	.110 (.036)**	−.067 (.023)**	−.084 (.032)**
Gender	.064 (.230)	.651 (.242)**	−.567 (.122)**	−.143 (.118)	.351 (.075)**	.380 (.099)**
Anger					.045 (.006)**
Depression					.065 (.017)**	.039 (.017)*

Note. Unstandardized coefficients are shown, with standard errors in parentheses, and the coefficients are the averaged estimation from the five complete data sets (United States, N = 1,516; Taiwan, N = 1,717). The model is estimated with anger correlated with depression (United States, r = .263; Taiwan, r = .186). The model fits the data well: χ²(7) = 8.37, ns. comparative fit index (CFI) = .999; Tucker Lewis Index (TLI) = .993; root mean square error of approximation (RMSEA) = .011. For some paths, we present one coefficient for both countries because we put constraint on this particular path.

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

Table 5.

Model—Multiple-Groups Analysis With Constraints (Alcohol Use).

	Anger		Depression		Alcohol Use
	United States	Taiwan	United States	Taiwan	United States	Taiwan
Goal strain	.237 (.033)**	.138 (.025)**	.129 (.011)**		−.007 (.007)
Unjust strain	.325 (.031)**	.207 (.027)**	.157 (.018)**	.099 (.015)**	−.001 (.007)
Negative life-event	.077 (.015)**		.030 (.008)**		.017 (.007)*	.023 (.006)**
Victimization	.212 (.036)**	.129 (.026)**	.042(.013)**		.066 (.012)**	.007 (.009)
Age	.298 (.059)**	.322 (.080)**	.306 (.030)**	.110 (.036)**	.260 (.020)**	.044 (.021)*
Gender	.065 (.230)	.651 (.242)**	−.567 (.122)**	−.143(.118)	−.152 (.070)*	.030 (.066)
Anger					.029 (.005)**
Depression					.067 (.010)**

Note. Unstandardized coefficients are shown, with standard errors in parentheses, and the coefficients are the averaged estimation from the five complete data sets (United States, N = 1,516; Taiwan, N = 1,717). The model is estimated with anger correlated with depression (United States, r = .263; Taiwan, r = .186). The model fits the data well: χ²(8) = 9.31, ns. comparative fit index (CFI) = .999; Tucker Lewis Index (TLI) = .995; root mean square error of approximation (RMSEA) = 0. For some paths, we present one coefficient for both countries because we put constraint on this particular path.

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

Discussion and Conclusion

The present study set out to examine the GST model in the United States and Taiwan. The main objective was to assess whether there is any difference in GST processes between the two countries. It needs to be stressed here that the cultural backgrounds of the United States and Taiwan are very different. Thus, this study provides insight into the issue of applying criminological theory to a cultural setting that is different from typical Western countries.

The general finding is that the tested paths of strain to delinquency were fairly similar and significant in both countries. For example, students who experienced more negative life-events and victimization incidents are more likely to hit someone and damage property, regardless of the country they live in. These results are consistent with those of previous researchers who applied GST in other Asian countries (Bao et al., 2007; Maxwell, 2001; Moon & Morash, 2004). The results also confirmed that both negative life-events and victimization have a negative impact on the youth. This is important because the results were validated through a cross-cultural comparison.

Interestingly, the link between victimization and alcohol use was significant only in the United States. Two possible explanations can be offered here. First, victimization directly damages wellness of self, which is detrimental to individuals in the individualistic culture. Hence, it is perceived to be high in magnitude, and the resultant strain, according to Agnew (2006), becomes criminogenic. Consequently, the U.S. students respond to victimization with alcohol use to soothe their bad feelings. Second, in Taiwan, students are under close supervision of family members and others because of the limited land area and crowding; hence, opportunities for students’ deviant behavior are greatly reduced. In addition, the drinking norms of Chinese culture demand that individuals drink in the company of others during feasts or at meal times (Harrell, 1981). Although students’ drinking with other students or peers may be considered norm abiding, their drinking without adult supervision is still considered deviant by most Taiwanese. Therefore, Taiwanese students may be aware of this cultural restriction and hence reject alcohol use.

With regard to the strain–negative emotion relationship, strain–anger relationships were different between the United States and Taiwan, the negative life-event → anger being the sole exception. As a result, most students in the United States were more likely than Taiwanese students to react to the four strains with anger. Markus and Kitayama (1994) write that, in an individualistic culture, negative emotion which is related to self (e.g., goal, self-identity) is more likely to be felt and expressed. Goal strain, unjust strain, and victimization threaten an individual in various ways (e.g., blocking goals, or harming oneself) and increase anger. In addition, Chinese students often attribute their failures to themselves, but successes to the group (Heine, 2008; Yu, 1996); this self-attribution might make one attribute the strain experience to oneself, which in turn could lower anger, because anger is more likely when one attributes failure to external conditions. Another explanation is that maintaining relational harmony is of high priority in a collectivistic culture, and the Confucian ethos regard the expression of anger as immature. Therefore, Taiwanese students may be reluctant to express anger, especially when the source of anger is goal related or relationship related.

In contrast, strain–depression relationships were similar between the two samples, except unjust strain–depression relationships. Strains were significantly related to depressive feelings of students in both countries, which echoes a World Health Organisation (WHO) report that revealed that depression is a problem in youth around the world. Although we found similar results for strain–depression relationship, the difference in the magnitude and combining with the results in strain–anger relationship showed that negative emotional responses to strain were stronger in the U.S. than in Taiwan students. One explanation for the difference may be that the Chinese start socializing at an early age to control affective display (Wu, 1996). In a collectivistic culture, enormous effort is made to maintain social harmony; hence, expressing negative emotions, especially anger, may be prohibited because of its potential to damage interpersonal relationships. Consequently, Taiwanese students may not prefer to disclose their anger, although they are, perhaps, just as likely to experience it as their counterparts in the United States. The reluctance to express negative emotions is even greater when these emotions are considered a serious social stigma, as in the case of depression (Russell & Yik, 1996).

Although the above differences are interesting, we discover one unexpected similarity. That is, the link between anger and hitting someone is similar in magnitude in both cultures. This is quite puzzling, because we would expect this relationship to be stronger in the United States than in Taiwan. One explanation is that the anger–hitting someone relationship indicates that angered students are more likely to be involved in violent delinquency, which does not mean that the students are equally angry or equally violent in both the countries. In Table 1, we can see that the mean scores of hitting someone and anger are always higher in the United States than in Taiwan.

The differences noticed in this study may become fewer with gradual fading of cultural differences. Previous researchers documented that, as a country moves toward greater development, individuals become more individualistic (Hofstede, 2001; Triandis, 1995). However, similarities may mask the differences underneath. For example, in a collectivistic society, a verbal insult from an outside group member may be sanctioned, based on the “mind your own business” rule (Bond, Wan, Leung, & Giacalone, 1985), but the insult is likely to cause conflict and strain. A similar verbal insult in a more individualistic society invades personal identity, which is treasured by such societies, and results in strain and anger. Thus, the same incident in different cultural settings may have the same results, although the mechanisms or underlying meanings may be different. Data limitations preclude unearthing of the underlying mechanism. Future studies may need to go a step further to disentangle the underlying mechanisms of the relationships between strain, negative emotions, and delinquency.

This study makes a valuable contribution to the literature on GST; however, it has several limitations needing to be addressed. First, because the data are cross-sectional, this study cannot firmly establish the causal relationships between the variables. Second, the samples collected in both countries are convenience samples; hence, their generalizability may be limited. Convenience samples help to discover and lay foundation for further large-scale and systematic studies, which is consistent with the purpose of the present study. Finally, the U.S. sample was collected more than a decade ago. Comparing those data with the recent data of Taiwan might pose some challenges. However, the rapid development of the Internet, together with globalization, makes these two cultures more similar than they ever were. Hence, one can find more similarities than differences, which is consistent with the findings of other studies that applied GST in other Asian countries.

In conclusion, the results show that GST is useful in explaining juvenile delinquency in both the United States and Taiwan, and combining this result with the results of previous studies suggests that GST is a potential candidate for a general theory because of its usefulness in explaining delinquency across varied cultures (e.g., West vs. East, male vs. female). In addition, this study corroborates Agnew’s assertion that victimization and negative life-events are criminogenic to adolescents regardless of their culture. Anger and depression are detrimental to the youth in both countries, because these negative emotions not only affect adolescents’ well-being but also lead to delinquency. The parents and schools in the United States, for instance, can learn from Asian cultures by using “shame” to educate their students and by emphasizing relationship harmony. If this is done, students may have lower levels of anger when experiencing strain, which, according to GST, could reduce delinquency (Agnew, 1995). However, parents of Chinese culture may learn from their counterparts in the United States to encourage their students to explore themselves, so that they may have less strain and depression from parents and social expectations.

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

References

Adler

(1996). Our American Society of Criminology, the world, and the state of the art—The American Society of Criminology 1995 presidential address. Criminology, 34, 1-9.

Agnew

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

Agnew

(1995). Controlling delinquency: Recommendations from general strain theory. In Barlow

(Ed.), Crime and public policy: Putting theory to work (pp. 43-70). Boulder, CO: Westview.

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 Publishing.

Allison

P. D.

(2003). Missing data techniques for structural equation modeling. Journal of Abnormal Psychology, 112, 545-557.

Bao

W.-N.

Haas

(2007). Life strain, coping, and delinquency in the People’s Republic of China: An empirical test of general strain theory from a matching perspective in social support. International Journal of Offender Therapy and Comparative Criminology, 51, 9-24.

Baron

S. W.

(2004). General strain, street youth and crime: A test of Agnew’s revised theory. Criminology, 42, 457-483.

Baron

S. W.

Hartnagel

T. F.

(1997). Attributions, affect, and crime: Street youths’ reactions to unemployment. Criminology, 30, 519-533.

10.

Beck

A. T.

Steer

R. A.

Brown

G. K.

(1996). Manual for the Beck Depression Inventory–II. San Antonio, TX: Psychological Corporation.

11.

Bellah

R. N.

Madsen

Sullivan

W. M.

Swidler

Tipton

S. M.

(1996). Habits of the heart: Individualism and commitment in American life. Los Angeles, CA: University of California Press.

12.

Beyers

J. M.

Loeber

(2003). Untangling developmental relations between depressed mood and delinquency in male adolescents. Journal of Abnormal Child Psychology, 31, 247-266.

13.

Bond

M. H.

Wan

K. C.

Leung

Giacalone

R. A.

(1985). How are responses to verbal insult related to cultural collectivism and power distance? Journal of Cross-Cultural Psychology, 16, 111-127.

14.

Botchkovar

E. V.

Broidy

(2010). Accumulated strain, negative emotions, and crime: A test of general strain theory in Russia. Crime & Delinquency. Advance online publication.

15.

Botchkovar

E. V.

Tittle

C. R.

Antonaccio

(2009). General strain theory: Additional evidence using cross-cultural data. Criminology, 47, 131-176.

16.

Broidy

L. M.

(2001). A test of general strain theory. Criminology, 39, 9-35.

17.

Broidy

L. M.

Agnew

(1997). Gender and crime: A general strain theory perspective. Journal of Research in Crime & Delinquency, 34, 275-306.

18.

Cheung

N. W. T.

Cheung

Y. W.

(2008). Self-control, social factors, and delinquency: A test of the general theory of crime among adolescents in Hong Kong. Journal of Youth and Adolescence, 37, 412-430.

19.

Cheung

N. W. T.

Cheung

Y. W.

(2010). Strain, self-control, and gender differences in delinquency among Chinese adolescents: Extending general strain theory. Sociological Perspectives, 53, 321-345.

20.

Chun

C.-A.

Moons

R. H.

Cronkite

R. C.

(2005). Culture: A fundamental context for the stress and coping paradigm. In Wang

P. T. P.

Wong

L. C. J.

(Eds.), Handbook of multicultural perspectives on stress and coping (pp. 29-53). New York, NY: Springer.

21.

Compas

B. E.

Phares

(1991). Stress during childhood and adolescence: Source of risk and vulnerability. In Cummings

E. M.

Greene

A. L.

Karraker

K. H.

(Eds.), Life-span developmental psychology: Perspectives on stress and coping (pp. 111-130). Hillsdale, NJ: Lawrence-Erlbaum.

22.

Directorate-General of Budget, Accounting, and Statistics, Executive Yuan. (2011). Statistics of population. Retrieved from http://www.dgbas.gov.tw/ct.asp?xItem=15408&CtNode=4594

23.

DuRant

R. H.

Getts

Cadenhead

Emans

S. J.

Woods

E. R.

(1995). Exposure to violence and victimization and depression, hopelessness, and purpose in life among adolescents living in and around public housing. Developmental & Behavioral Pediatrics, 16, 233-237.

24.

Education Bureau, Kaohsiung City Government. (2010). Current educational statistics. Retrieved from http://163.32.250.1:8000/Members/grp07/new_page_3.htm

25.

Eitle

D. J.

Turner

R. J.

(2003). Stress exposure, race, and young adult crime. Sociological Quarterly, 44, 243-269.

26.

Flora

D. B.

Curran

P. J.

(2004). An empirical evaluation of alternative methods of estimation for confirmatory factor analysis with ordinal data. Psychological Methods, 9, 466-491.

27.

Ford

J. A.

Schroeder

R. D.

(2009). Academic strain and non-medical use of prescription stimulants among college students. Deviant Behavior, 30, 26-53.

28.

Froggio

(2007). Strain and juvenile delinquency: A critical review of Agnew’s general strain theory. Journal of Loss & Trauma, 12, 383-418.

29.

Ganem

N. M.

(2008). The role of negative emotion in general strain theory: An empirical study. La Vergne, TN: Lightning Source.

30.

Gottfredson

M. R.

Hirschi

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

31.

Harrell

(1981). Normal and deviant drinking in rural Taiwan. In Kleinman

Lin

T.-Y.

(Eds.), Normal and abnormal behavior in Chinese culture (pp. 49-59). Dordrecht, Netherlands: Reidel.

32.

Hay

(2003). Family strain, gender, and delinquency. Sociological Perspectives, 46, 107-135.

33.

Hegtvedt

K. A.

(1990). The effects of relationship structure on emotional response to inequity. Social Psychology Quarterly, 53, 214-228.

34.

Heine

S. J.

(2008). Cultural psychology. New York, NY: Norton.

35.

Heine

S. J.

Lehman

D. R.

(1995). Social desirability among Canadian and Japanese students. Journal of Social Psychology, 135, 777-779.

36.

Hoffmann

J. P.

Cerbone

F. G.

S. S.

(2000). A growth curve analysis of stress and adolescent drug use. Substance Use & Misuse, 35, 687-716.

37.

Hofstede

(2001). Culture’s consequences (2nd ed.).Thousand Oaks, CA: Sage.

38.

Horton

Rice

S. K.

Piquero

N. L.

Piquero

A. R.

(2012). On the variability of anger cross-culturally: An assessment of general strain theory’s primary mediator. Deviant Behavior, 33, 260-281.

39.

Hsu

F. L. K.

(1983). Rugged individualism reconsidered: Essay in psychological anthropology. Knoxville: University of Tennessee Press.

40.

Jang

S. J.

(2007). Gender differences in strain, negative emotions, and coping behaviors: A general strain theory approach. Justice Quarterly, 24, 523-553.

41.

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.

42.

Kim

Triandis

H. C.

Kagitcibasi

Choi

S.-C.

Yoon

(1994). Introduction. In Kim

Triandis

H. C.

Kagitcibasi

Choi

S.-C.

Yoon

(Eds.), Individualism and collectivism: Theory, method, and applications (pp. 1-11). Thousand Oaks, CA: Sage.

43.

Kohn

M. L.

(1987). Cross-national research as an analytic strategy: American Sociological Association, 1987 presidential address. American Sociological Review, 52, 713-731.

44.

Leung

Bond

M. H.

(1982). How Chinese and Americans reward task-related contributions: A preliminary study. Psychologia, 25, 32-39.

45.

Lin

W.-H.

Cochran

J. K.

Mieczkowski

(2011). Direct and vicarious violent victimization and juvenile delinquency: An application of general strain theory. Sociological Inquiry, 81, 195-222.

46.

Lin

W.-H.

Mieczkowski

(2011). Subjective strains, conditioning factors, and juvenile delinquency: General strain theory in Taiwan. Asian Journal of Criminology, 6, 69-87.

47.

Markus

H. R.

Kitayama

(1991). Culture and the self: Implications for cognition, emotion, and motivation. Psychological Review, 20, 568-579.

48.

Markus

H. R.

Kitayama

(1994). The cultural construction of self and emotion: Implications for social behavior. In Kitayama

Markus

H. R.

(Eds.), Emotion and cultural: Empirical studies of mutual influence (pp. 89-130). Washington, DC: American Psychology Association.

49.

Maruyama

G. M.

(1997). Basics of structural equation modeling. Thousand Oaks, CA: Sage.

50.

Maxwell

(2001). A focus on familial strain: Antisocial behavior and delinquency in Filipino society. Sociological Inquiry, 71, 265-292.

51.

Mirowsky

Ross

C. E.

(1995). Sex difference in distress: Real or artifact? American Sociological Review, 60, 449-468.

52.

Moon

Blurton

McCluskey

J. D.

(2008). General strain theory and delinquency: Focusing on the influences of key strain characteristics on delinquency. Crime & Delinquency, 54, 582-613.

53.

Moon

Morash

(2004). Adaptation of theory for alternative cultural contexts: Agnew’s general strain theory in South Korea. International Journal of Comparative and Applied Criminal Justice, 28, 77-103.

54.

Moon

Morash

McCluskey

C. P.

Hwang

H. W.

(2009). A comprehensive test of general strain theory: Key strains, situational- and trait-based negative emotions, conditioning factors, and delinquency. Journal of Research in Crime & Delinquency, 46, 182-212.

55.

Morash

Moon

(2007). Gender differences in the effects of strain on the delinquency of South Korean youth. Youth & Society, 38, 300-321.

56.

Ostrowsky

M. K.

Messner

S. F.

(2005). Explaining crime for a young adult population: An application of general strain theory. Journal of Criminal Justice, 33, 463-476.

57.

Piquero

N. L.

Sealock

M. D.

(2000). Gender and general strain theory: A preliminary test of Broidy and Agnew’s gender/GST hypotheses. Justice Quarterly, 21, 125-158.

58.

Robertson

A. R.

Stein

J. A.

Schaefer-Rohleder

(2010). Effects of hurricane Katrina and other adverse life events on adolescent female offenders: A test of general strain theory. Journal of Research in Crime & Delinquency, 47, 469-495.

59.

Rubin

D. B.

(1987). Multiple imputation for nonresponse in surveys. New York, NY: Wiley.

60.

Russell

J. A.

Yik

M. S. M.

(1996). Emotions among the Chinese. In Bond

M. H.

(Ed.), The handbook of Chinese psychology (pp. 166-188). New York, NY: Oxford University Press.

61.

Sanchez

J. I.

Spector

P. E.

Copper

C. L.

(2005). Frequently ignored methodological issues in cross-cultural stress research. In Wang

P. T. P.

Wong

L. C. J.

(Eds.), Handbook of multicultural perspectives on stress and coping (pp. 29-53). New York, NY: Springer.

62.

Sigfusdottir

I.-D.

Kristjansson

A. L.

Agnew

(2012). A comparative analysis of general strain theory. Journal of Criminal Justice, 40, 117-127.

63.

Slocum

L. A.

Simpson

S. S.

Smith

D. A.

(2005). Strained lives and crime: Examining intra-individual variation in strain and offending in a sample of incarcerated women. Criminology, 43, 1067-1110.

64.

Spielberger

C. D.

(1988). Manual for the State-Trait Anger Expression Inventory (Research ed.). Odessa, FL: Psychological Assessment Resource.

65.

Storch

E. A.

Roberti

J. W.

Roth

D. A.

(2004). Factor structure, concurrent validity, and internal consistency of the Beck Depression Inventory–Second Edition in a sample of college students. Depression and Anxiety, 19, 187-189.

66.

Thoits

(1995). Stress, coping, and social support process: Where are we? What next? Journal of Health and Social Behavior, Extra Issue, 53-79.

67.

Triandis

H. C.

(1995). Individualism & collectivism. Boulder, CO: Westview.

68.

Turner

R. J.

Wheaton

(1995). Checklist measurement of stressful life event. In Cohen

Kessler

R. C.

Gordon

L. U.

(Eds.), Measuring stress: A guide for health and social scientists (pp. 29-58). New York, NY: Oxford University Press.

69.

Verrill

S. W.

(2008). Social structure and social learning: Mediating or moderating? El Paso, TX: LFB Scholarly.

70.

Wareham

Cochran

J. K.

Dembo

Sellers

C. S.

(2005). Community, strain, and delinquency: A test of a multi-level model of general strain theory. Western Criminological Review, 6, 117-133.

71.

World Health Organization. (2001). The World Health Report 2001. Mental health: New understanding new hope. Retrieved from http://www.who.int/whr/2001/en/whr01_en.pdf

72.

D. Y. H.

(1996). Chinese childhood socialization. In Bond

M. H.

(Ed.), The handbook of Chinese psychology (pp. 143-154). New York, NY: Oxford University Press.

73.

Yamaguchi

Kuhlman

D. M.

Sugimori

(1995). Personality correlates of allocentric tendencies in individualist and collectivist cultures. Journal of Cross-Cultural Psychology, 25, 146-158.

74.

A. B.

(1996). Ultimate life concerns, self, and Chinese achievement motivation. In Bond

M. H.

(Ed.), The handbook of Chinese psychology (pp. 227-246). New York, NY: Oxford University Press.

Strain,Negative Emotions,and Juvenile Delinquency

Abstract

Keywords

Introduction

Review of GST

Cultural Influences and Cross-Cultural Studies

The Present Study and Research Question

Method

Sample

U.S. sample

Taiwanese sample

Measurement of Variables 3

Delinquency

Strain

Failure to achieve positively valued goals—disjunction between desired and actual outcomes

Failure to achieve positively valued goals—unjust outcomes

Loss of positive stimuli—negative life-events

Presentation of noxious stimuli—victimization

Negative emotions

Anger

Depression

Demographic variables

Analytical Strategies

Results

Discussion and Conclusion

Footnotes

Declaration of Conflicting Interests

Funding

Notes

References

Measurement of Variables³