Abstract
This study attempts to explore the potential extension of general strain theory (GST) by applying the hypotheses proposed by Broidy and Agnew to a sample of South Korean youths. Specifically, this study examines whether particular forms of strain and stressors have differential effects across genders on various deviant outcomes, delinquency, and suicidal thoughts. Multiple regression analyses using longitudinal data of 3,125 South Korean youths revealed mixed support for the GST proposition. Females experienced a higher level of both anger and depression than males. However, the experience of negative emotions is partly gendered in general. This study also found that different negative emotions and strain/stress factors are important and demonstrate gendered pathways in the case of delinquency. However, it also revealed that similar types of strains and stressors and negative emotions were significant and positive for suicidal thoughts for both males and females. Furthermore, a model examining the impacts of conditioning variables on suicidal thoughts highlighted that depression is particularly important in females. These findings indicate that various types of deviant outcomes and strain–stressors provide a fuller understanding of both similarities and differences by gender.
Introduction
Strain is likely to cause various problematic behaviors (Cullen & Agnew, 2006). However, a pathway from a strain toward a specific problematic behavior is not uniform but diverse. Several factors may affect such a pathway, and thus produce different outcomes. Some problematic behaviors are relatively outward, hurting others, while others are relatively inward, hurting the self. For instance, delinquency is other-destructive behavior, whereas suicide is self-destructive. Gender and cultural differences are well-known factors significantly affecting the pathway and producing different outcomes. The prevalence, process, and type of delinquency differ between males and females (Denno, 1997). Some cultures, such as the Japanese culture, demonstrate less crime (hurting others) and more suicide (hurting themselves) than Western cultures (Komiya, 1999; Wingfield-Hayes, 2015). This article focuses on the interrelations between gender, culture, and problematic outcomes—other-hurting or self-hurting—to understand different pathways from strains toward problematic behaviors.
Criminologists are aware that gender is a primary correlate of varying degrees of crime and delinquency. However, feminist criminologists suggest that common criminological theories do not adequately explain female criminality, and that this shortcoming is due to their heavily skewed focus on male-oriented criminality (see Messerschmidt, 1993). Classic strain theory (Cloward & Ohlin, 1960; Cohen, 1955; Merton, 1938) is one such example, having drawn criticism from feminist criminologists (Chesney-Lind, 1989; Daly & Chesney-Lind, 1988). However, general strain theory (GST), a recent revision of classic strain theory by Agnew (1992, 2001, 2006), was extended to explain the gender differences in crime rates across diverging types of crime (Agnew, 2001; Broidy & Agnew, 1997). Nonetheless, to this day, gender remains a relatively neglected topic within the GST literature (Hay, 2003; Jang, 2007). The propositions of Broidy and Agnew (1997) have yet to be fully examined, and studies that have addressed them yielded mixed results (e.g., Jang, 2007). The current study further investigates the issue highlighted in GST. Specifically, it focuses on gender differences in the criminogenic process. It is assumed that a gender difference may be evident in problematic behavior responses to strains.
The current study also explores the cultural aspect of GST, which has been supported as a sound criminological theory mostly in the American and Western context. As such, this study attempts to expand the GST tradition beyond Western cultural settings by exploring whether gendered pathways to delinquency and suicidal thoughts, as proposed by Broidy and Agnew (1997), hold for the youth in South Korea. Findings drawn from the South Korean context could strongly contribute to GST in the offending literature because of predictable cultural and societal differences between South Korea and other Western industrialized nations (i.e., the United States and Western European nations). South Korea has been undergoing a process of Westernization for the last three decades. It now has a distinct cultural and social context of duality, showcasing the coexistence of Western and traditional cultures (Kim, Kwak, & Yun, 2010). Because of this duality of cultural tension, South Korean youths are equally affected by both peers and parents, unlike their Western counterparts, who are mostly affected by peers (Akers & Cochran, 1985; Erickson, Crosnoe, & Dornbusch, 2000; Neff & Waite, 2007).
Indeed, South Korean society demonstrates the distinct dynamics of problematic behaviors affected by strains different from its Western counterparts but similar to those in Japan. Noteworthy is the high suicide rate for Koreans, which peaked in the last two decades. For example, Organization for Economic Co-Operation and Development (OECD) health statistics revealed that since 2003, the average annual rate of people per 100,000 who commit suicide has been found to be 28.1. In addition, South Korea had the highest suicide rates among developed nations. According to statistics for South Korea for 2014 and 2015, suicide is the number-one cause of death among individuals aged 10 to 39. Specifically, the highest suicide rate for Koreans was found among students (Korea Herald, 2015; Wingfield-Hayes, 2015). By contrast, both juvenile delinquency and the crime rate are moderate compared with other Western OECD countries (Civitas, 2012; OECD, 2009). This feature of the South Korean context differs from that of the country’s Western counterparts but is similar to that of Japan. This could be attributed to the dual cultural nature of contemporary South Korean society. As such, South Korean youths might be caught between traditional parental control and Western cultural forces through peer influence. The current study explores this question by looking at the interaction between cultural settings and types of problematic behaviors.
Briefly, the current study investigates pathways toward problematic behaviors generated by strains. In doing so, the study examines how gender differences shape different procedural types of pathways and how the South Korean cultural context may affect the different outcomes of problematic behaviors. The findings of this study may enrich the GST tradition.
Theoretical Framework: GST and Gender
GST (Agnew, 1992, 2001) proposes three major sources of strain increasing the likelihood of crime and delinquency: (a) failure to achieve positively valued goals, (b) removal of positively valued stimuli, and (c) presentation of noxious/negative stimuli. More recently, Agnew (2001, 2006) suggested four major criteria of key strains and stressors increasing the possibility of crime and delinquency. These criteria indicate that strains must be “seen as unjust, high in magnitude, associated with low social control, and related to some pressure or incentive to engage in criminal copying” (Agnew, 2001, p. 326, 2006; Yun, Kim, & Morris, 2014). Supporting the arguments of the theory, empirical research revealed that certain strains and stressors meeting the four criteria tend to increase or decrease the possibility of crime and delinquency, such as experience of childhood abuse, experience of prejudice and discrimination, being a victim of crime, and being unemployed (Agnew, 2006).
According to GST, the negative effects resulting from strains (e.g., anger, depression, guilt, and frustration) mediate or condition the impact of strains on an individual’s legitimate or illegitimate cognitive and behavioral coping strategies. Agnew (2001) suggested that anger, as a negative emotion, is most likely to lead an individual to illegitimate coping strategies. Empirical studies showed that anger tends to increase the likelihood of interpersonal violence, whereas depression increases the likelihood of self-destructive behavior such as drug and substance abuse—see Agnew (2006) for a review. Agnew also argued that individuals cope with strains through various cognitive, behavioral, and emotional coping strategies, which can be either conventional behaviors or criminal and deviant acts. Whether strain leads to crime is determined by internal (i.e., self-esteem) or external conditioning factors (i.e., social support) associated with the individual personality and the characteristics of the strain event (Agnew, 1992, 2006). More recently, Agnew has asserted that conditioning factors may mediate or moderate the causal relationship between strain and negative emotions. The conditioning factors proposed by Agnew are mostly well-known risk or protective factors relevant to crime drawn from other major criminological theories such as self-esteem, social support, individual coping strategies, social bonding, and social learning variables.
Unlike other traditional criminological theory, the scope of GST has been extended to clarify gender differences in the explanation of crime by Broidy and Agnew (1997). The GST proposed three hypotheses to explain how and why women and men differ in crime types and rates in the context of GST. Their first hypothesis was that men and women experience gendered strain. That is, males experience strain more conducive to crime than females do, because males tend to be more exposed to financial strain, concern for extrinsic achievement, and experiences of victimization and interpersonal conflicts with their peers. By contrast, females experience more gender-specific strains, such as procedural justice, network relationship problems, and gender-based discrimination, because they tend to emphasize relationships with others. This can help us understand females’ lower crime rates and lower engagement in self-directed behaviors (Agnew, 2006; Jang, 2007; Sharp, Brewster, & Love, 2005).
Broidy and Agnew (1997) suggested a gendered response to negative effects (i.e., anger and depression) in their second hypothesis. According to the hypothesis, males are more likely to externalize the anger (e.g., blaming others), which may be more likely to lead an individual to crime and deviant behaviors. On the contrary, they suggested that females are more likely to internalize anger (i.e., self-harm and suicidal thoughts). In addition, females’ experience of anger is accompanied by other negative emotions (e.g., depression, fear, guilt, and shame) based on their cultural socialization. These differences, in turn, may reduce females’ tendency toward violent crime (Broidy, 2001; Mirowsky & Ross, 1995).
Finally, Broidy and Agnew’s (1997) third hypothesis posited that males respond to strain with physical/illegitimate coping strategies that are more conducive to crime because of gendered differences in available mediating and/or moderating resources. Those mediating and/or moderating factors include opportunity, individual coping skills and knowledge, self-esteem, social ties and networks, personal dispositions (i.e., impulsiveness), and other well-known criminogenic variables pointed out by major criminological theories. Such factors may condition an individual’s response to negative emotions, which may lead to different behavioral outcomes across genders. That is, females have more cognitive, emotional, and social coping strategies than males do, but females have lower levels of self-esteem and self-efficacy than males do. Such differences between males and females clarify that femaleness may be inversely related to crime (Baron, 2007). Therewith, gendered societal proscriptions against crime for females and their greater exposure to social control may reduce females’ criminal responses to strain. To summarize, Broidy and Agnew propose procedural gender differences in GST. Such gender differences are proposed in the pathways from strains through negative effects to different types of coping strategies, namely, delinquency or other self-destructive problematic behaviors. These gender differences are conditioned by different life experiences, characteristics, and various social and situational factors.
Empirical Research on Gender and GST
To date, numerous studies have assessed that the major tenets of GST have received considerable empirical support. However, prior empirical studies testing GST were mostly limited to examining the causality of major types of stressors and strains on a set of delinquent and criminal behaviors, while neglecting the gendered differential pathway and the hypotheses put forth by Broidy and Agnew (1997).
Early studies on gender issues in the context of GST examined the applicability of delinquency and substance use in both males and females, and mostly used gender as a control factor. However, these studies found no significant gender differences in the strain process (Agnew & Brezina, 1997; Hoffmann & Cerbone, 1999; Hoffmann & Su, 1997). For instance, Hoffmann and Su (1997) examined whether distinctive patterns of the socialization of females and males would lead females to more often internalize reactions to stress (drug use) and males to more often externalize reaction to strain (delinquency). They found no significant differences except for a few short-lived differences in the impact of strain on delinquency and drug use by gender.
Some studies, however, reported mixed support for Broidy and Agnew’s (1997) hypothesis suggesting potential gendered pathways in the process of strain and deviance causality (Eitle, 2002; Hay, 2003; Mazerolle, 1998; Sharp et al., 2005). Mazerolle’s (1998) study testing the effects of strain on different types of delinquency found some evidence that strain leads to different types (i.e., violent and property-related delinquency) of crime across genders.
Eitle (2002) investigated whether unfair treatment based on gender was linked to criminality in females. He found that being denied promotion, not being hired, and being unfairly treated were important predictors of crime and substance abuse among females. In a related study, Hay’s (2003) study examining the effects of family-related strain on deviance found that males reported that their parents demonstrated more physical responses to their rule-violating behaviors than did females. He found that parents’ physical punishment type of strain was related to significant and positive effects on males’ delinquent behavior by males but not by females.
Other studies found evidence relevant to the relationship between types of strains and emotional responses to strain (Baron, 2007; Broidy, 2001; Hay, 2003). Broidy (2001) found that respondents who reported anger were more likely to report criminal outcomes than those who reported other negative emotions. She also found that females and males experience similar levels of anger, but that females respond to anger with more legitimate coping strategies than males. Hay (2003) also demonstrated that males and females experience similar levels of anger with regard to family-related strain, but that females are more likely to report feeling guilty. He also argued that the guilt accompanying female anger might reduce delinquency in females (see Hay, 2003).
More recent studies testing the GST hypotheses used a full range of variables. These variables include more severe and gender-specific strains, negative effects of both legitimate and illegitimate coping strategies, and various types of constraints and dispositions conditioning the strain process. Those studies revealed a link between the two types of negative emotions (mostly anger and depression) and different types of coping strategies, which are conditioning variables to strain. Yet, there are mixed findings as well (Baron, 2007; Jang, 2007; Jang & Johnson, 2003; Jennings, Piquero, Gover, & Pérez, 2009; Kaufman, 2009; Piquero & Sealock, 2004; Yun et al., 2014).
Piquero and Sealock (2004) investigated gender difference in the experience of strain, negative effects, and legitimate and illegitimate coping strategies. Their findings were ambiguous, in that there was no difference in the amount of strain experienced by both genders. Furthermore, they found that females reported higher levels of anger and depression than the hypothesis expected. In terms of negative effects, anger had a direct effect on property-related delinquency by males and interpersonal aggression by both males and females, as the theory predicted. Depression, however, was not related to property-related delinquency by males or to interpersonal aggression by males and females. They also found that males reported higher levels of physical and cognitive coping strategies than females contrary to their predictions. Such findings seem to indicate potential procedural differences between males and females in the negative effects of strains on types of behaviors.
Jennings et al. (2009) replicated Piquero and Sealock’s (2004) study with a sample of Mexican American youths, finding mixed support for the hypotheses. All strains predicted criminal and delinquent behaviors for males and females. Females, however, reported higher levels of both anger and depression compared with males. Furthermore, anger had significant effects on interpersonal violence and property-related delinquency, but depression predicted only property-related delinquency for both genders. In addition, anger did not completely mediate the effects of strain on deviant acts.
Jang’s (2007) study explicitly tested Broidy and Agnew’s hypotheses with a sample of African Americans. He found supporting evidence that African American women experienced greater levels of gender-related strains such as interpersonal and gender-role strains in the family. African American women were also less likely to adopt illegitimate coping behaviors than their counterparts did because of women’s greater experiences with depression and anxiety.
Kaufman (2009) focused on severe strains relevant to males and females, confirming the theory’s proposition that severe strains (e.g., suicidal behavior by friends and family and violent criminal victimization) strongly affect both types of negative emotions (symptoms of depression and a bad temper) and delinquency outcomes for males and females (e.g., suicidal thoughts, violence, and running away). She also found that the relationship between severe strains and negative effects and the strain process leading to deviant coping strategies differed by gender.
GST, Gender, and South Korean Youths
GST is particularly appropriate to contemporary South Korean society in general and South Korean youths in particular. Today, strains or stresses are noticeably high in South Korea. Several symptoms of high-stress society have been observed, including (a) the rapid increase of deaths related to stress diseases, (b) the highest smoking and alcohol use worldwide, (c) an increasing divorce rate, (d) the highest suicide rate worldwide, and (e) a rapid increase in the number of murder victims (H. Park, 1997). High strain or stress in contemporary South Korean society is structural and deep rooted. For the last three to four decades, South Korea has culturally transformed from a traditional and Confucian society into a modern and Westernized society alongside rapid economic development and democratization. However, certain elements and constituents were transformed not uniformly but at varying degrees and at varying rates. Thus, today’s South Korea has an ambiguous culture, demonstrating some Westernization alongside remnants of the older Confucianism. This collective and structural cognitive dissonance between Western and Confucian culture is a primary cause of high levels of strains or stressors in the country (Ham, 2003).
South Korean youths suffer from this cultural duality. Western and traditional cultural elements coexist and are ambiguously intertwined (Sin & Choi, 2002). Generational cultural tensions are especially evident. Many Westernized youths are confronted by traditional and Confucian older generations and parents (Chong, 2009). As South Korean youths are tightly bonded to family, they are influenced by their parents. Youths passively resist the values and discipline of the older generations through silence and derision, but at the same time, they formally obey their parents, because the family as an institution still strongly controls young people’s daily routines. On the contrary, both online and off, South Korean youths are strongly influenced by the more Westernized values and cultures of their peers. As such, their values and cultures are more individualistic, self-expressive, and nonauthoritative, which conflicts with their parents’ more collectivist, restrained, and authoritative values and cultures. This situation causes daily tensions and strains among youths due to their relatively powerless social status within the family (Y. K. Park, Park, & Kim, 2006).
The recent increase in juvenile delinquency among South Korean youths is associated with their experience of such strains. S. Lee (2003) examined the propositions of GST using a sample of 800 male high school students. The study found support for GST, showing that the boys’ experiences of strain and stress generated by interaction with parents, school, and friends were positively associated with delinquency. Lee also found that situational anger may mediate the effects of some strains and stressors on boys’ delinquency. Furthermore, the study determined that conditioning factors such as limited attachment to parents and association with delinquent peers moderated outcome behaviors among boys.
Unlike their Western counterparts, South Korean youths demonstrate a different type of deviance and other problematic behaviors, because they are caught between deviance facilitating peer association and deviance preventing conventional bonds with parents and teachers. This conflict leads to high levels of self-destructive behavior such as suicide. By contrast, strong conventional bonds contain the level of deviance among South Korean youths; thus, juvenile delinquency in South Korea is still comparatively moderate (Civitas, 2012; Wingfield-Hayes, 2015; Woo, 2014). However, this process from strains to deviance or suicide conditioned by social factors has not yet been empirically examined and systematically explained.
Meanwhile, there is a gender difference in strain processes leading to deviance or other problematic behaviors. Despite decades of gender equalization in South Korean society, differential treatment based on gender remains. This could be another example of incomplete Westernization in the country. Girls are subject to tighter social controls by parents compared with boys. Social and cultural expectations for girls are different from those for boys, and informal sanctions for deviance are harsher for girls than for boys. This leads to differential strains and strain-induced deviance for girls (Seo & Kim, 2014).
However, only a handful of studies test for gender differences in the strain process, leading to deviance among South Korean youths (e.g., Lee, 2003; Moon & Morash, 2005; Morash & Moon, 2007). Morash and Moon (2007) examined the proposed GST and gender hypotheses in a sample of high school students from one city in South Korea, finding some support for GST. Their study indicated that two stressors (abuse by teachers and exam-related stress) have differential impacts on delinquency across genders. Additional empirical studies on gender issues in the strain process need to be conducted on South Korean youths.
Several important aspects should be considered regarding GST and gender in South Korea, which are missing from the limited studies conducted thus far. Existing studies relied on cross-sectional and nonrandom samples, and were generally not representative of South Korean youths. These analyses were also limited, as they did not include various legitimate coping variables and negative effect measures (e.g., anger, fear, guilt), which are theoretically important to mediating variables in the GST hypotheses (Agnew, 2001, 2006). In addition, the operationalization of some variables is not congruent with the theory (e.g., peer hassle; see Morash & Moon, 2007). Such matters need to be empirically examined to fully understand strain-induced gendered pathways to deviance and other problematic behaviors among South Korean youths.
In addition, it appears that the findings of existing Korean studies do not fully examine the propositions by Broidy and Agnew (1997). These studies found that certain types of strains lead to different types of negative emotions by gender, and that there are qualitative gendered differences in coping mechanisms. However, the studies did not focus on whether conditioning factors, such as parental bonding and peer associations, have a differential effect by gender in the process between strain, negative effects, and delinquency or other problematic behaviors (Jang, 2007; Jennings et al., 2009; Morash & Moon, 2007). This is important for South Korean youths because they are equally affected by both parental bonding and peer associations, unlike their Western counterparts who are significantly affected mostly by peers. This issue needs to be empirically tested to be associated with different types of behavioral outcomes (Kim et al., 2010).
The Current Study
This study aims to examine gender differences in deviant behaviors and suicidal thoughts among South Korean youths in the theoretical framework of GST. The study adds to the literature regarding GST and gender in several ways: First, this study uses a nationally representative longitudinal random sample collected from the entire South Korean nation. This could contribute to the GST tradition by extending GST to a different cultural context from Western settings, such as South Korea, where the suicide rate is high and deviance is moderate from the global perspective. Second, this study contains a wide range of measures relevant to and important for testing females’ criminality in the GST context in analysis models. These variables include various types of individual strain and stress measures, negative emotions (anger, depression), and various conditioning variables suggested by previous studies and major criminological theories. Thus, this study is more suitable for examining gendered procedural differences in GST. Third, this study includes independent variables rarely used in previous GST studies, such as appearance stress, attitude toward socialized gender roles (feminine and masculine), and suicidal thoughts as a gendered, nondeviant behavioral outcome. These measures are particularly interesting in the study of South Korean youths. Appearance stress is severe among South Korean youths, especially girls, and is related to the prospering plastic surgery industry in South Korea (J. Lee, 2012). Socialized gender roles also differentially affect the extent of parental controls on boys and girls (Seo & Kim, 2014). In addition, suicidal thoughts are an important measure when examining South Korean society, where the suicide rate, as mentioned, is substantially high. Finally, this study compares deviance to suicidal thoughts. The former is other-hurting behavior, whereas the latter is self-hurting. Strains or stressors could result in various coping strategies; however, pathways toward different types of coping strategies should be more closely examined. This study serves that purpose.
The current study examines the following hypotheses, considering Broidy and Agnew’s (1997) major propositions and other related literature reviews within the contexts of gender and South Korean culture:
Method
Data
This study used nationally representative data collected from the second and third waves of the National Youth Policy Institution (NYPI). Data collection was conducted as part of the Korea Youth Panel Survey (KYPS) projects (2003-2008) developed by the NYPI in South Korea to construct a longitudinal database providing resources for academic research by supplying a wide range of information about South Korean youths. Using a stratified, multistage cluster sampling method, the KYPS project followed a nationally representative panel sample of junior high school students aged 14 years for 6 years. Data were collected in three stages: First, trained interviewers visited randomly selected schools to conduct face-to-face interviews with participants. During interviews, they filled out questionnaires asking about various aspects of participants’ educational experiences, social relationship experiences, and so on.
Second, participants were asked to take part in a self-administered survey to identify their deviance and problematic behaviors (e.g., theft, engaging in violence, smoking, drinking, suicidal thoughts). Third, participants’ parents also took part in telephone interviews, providing information such as family monthly income, parental education levels, and family structure. All participants were assured of the confidentiality of the survey before they took part (K. S. Lee & Baek, 2007). The response rate of the data is remarkable. In the Wave 1 data collection, a total of 3,449 youths and their parents took part in the first survey, resulting in a response rate of 93.2% in 2003. In 2004, a total of 3,188 of those who participated in the first wave survey took part in the Wave 2 data collection, resulting in a response rate of 92.4%. Finally, in 2005, a total of 3,125 participants who took part in the Wave 2 data collection took part in the Wave 3 data collection, with a 90.6% follow-up rate from youths and their parents.
Measures
Dependent variables
This study used two types of coping behaviors as the dependent variables: deviant behaviors and suicidal thoughts (Wave 3). Deviant behaviors represent other-directed behavior, whereas suicidal thoughts are a self-destructive behavior. Dependent variables were selected based on the theoretical proposition that the types and significance of stress and strain measures and conditioning factors may predict different types of delinquency that differ according to gender. In theory, it is predicted that, for example, females are more likely to respond to strain and stress with self-destructive behaviors (e.g., purging behaviors, suicidal thoughts, and eating disorders), whereas males are more likely to respond to strains and stress with other-directed and expressive deviant behaviors and aggressions (e.g., violent fighting) (Jennings et al., 2009; Kaufman, 2009; Sharp, Terling-Watt, Atkins, Giliam, & Sanders, 2001).
The deviance variable was measured by asking participants if they had been involved in any of the following other-directed and expressive deviant behaviors during the past year: (a) truancy, (b) running away from home, (c) violent fighting, (d) joining gang activities, (e) money extortion, (f) theft, (g) verbal abuse, (h) aggressive behavior, and (i) bullying classmates. Self-destructive deviance, such as substance use, was not included in this measure. Participants’ responses on a binary scale for each behavior (never = 0, yes = 1) were summed as an index and recorded as a dichotomous variable (0 = never, 1 = one or more).
Suicidal thoughts were measured through participants’ responses to the statement that “I sometimes want to kill myself for no reason” on a Likert-type scale ranging from “strongly disagree” (0) to “strongly agree” (4). Presumably compared with deviance, suicidal thoughts as a coping strategy are particularly relevant to females, although this has rarely been used in previous studies (Kaufman, 2009). Therefore, using gender-specific outcome behavior, suicidal thoughts may be useful in examining the gendered nature of the strain and problematic behavior process in the GST context (Kaufman, 2009).
Independent variables
In this study, strain/stress measures were presented as noxious stimuli. These measures were developed based on theory and previous studies from data collected in Wave 2. Agnew (2006) recommended using various types of individual measures for strain measures rather than using a composite strain scale. He argued that using a composite strain scale would cause major analytical problems, resulting in a failure to find supporting evidence on the GST propositions regarding gender and crime in previous studies (Baron, 2007). Therefore, this study utilized various types of individual measures of strain and stress identified as key strain measures in previous studies, including subjective and objective strain measures (Agnew, 2001, 2006). Subsequently, this study used seven strain measures from the KYPS survey. The survey asked participants to respond to a series of questions involving personal problems they might experience in their lives, namely, (a) health problems, (b) perceived psychological/mental problems, (c) family financial problems, (d) experience of parental verbal abuse, (e) parental physical punishment, and (f) being suspended from school. Each single item was measured on a 5-point Likert-type scale (1 = strongly disagree, 5 = strongly agree) and used as an individual strain measure in the analyses.
Five composite stressor measures exerting contemporaneous effects (i.e., at the time of the interview) on participants’ lives were measured. In the KYPS survey, a series of questions asked participants to measure if they were currently experiencing the following stressors: (a) parental stress (a summed index of four items; α = .89), (b) school-related stress (a summed index of four items; α = .82), (c) stress from friends (a summed index of three items; α = .87), (d) appearance stress (a summed index of four items; α = .81), and/or (e) monetary stress (a single item). All items used to measure stressors were recorded on a 5-point Likert-type scale (1 = strongly disagree, 5 = strongly agree).
Victimization was operationalized by asking if the participants had been victims (yes = 1 or no = 0) of any of the following incidents during the past year: (a) threats, (b) extortion of money, (c) beating, (d) sexual abuse or rape, (e) bullying, and/or (f) malicious alienation by peers. In GST, victimization is one of the major strain measures reflecting the criterion “seen as unjust, high in magnitude,” and is predicted to increase the tendency to engage in criminal and deviant behaviors for the victims (Agnew, 1992, 2001, 2006). Responses of participants’ survey questions were summed as a single scale and recorded as binary variables (0 = never, 1 = one or more than once).
Negative emotionality
Measures of anger and depression were used to represent two forms of trait-based negative emotionality. Although the theory originally specified using situational negative emotions rather than trait-based negative emotionality, situational negative effects were not available in the KYPS data. Mazerolle, Burton, Cullen, Evans, and Payne (2000), however, found evidence that trait-based negative emotions are still relevant in testing GST and gender. For example, individuals reporting a higher level of anger are more likely to demonstrate a higher level of situational anger due to repeated exposure to strain and stress (Jennings et al., 2009; Mazerolle et al., 2000). Furthermore, Agnew (2006) found that trait-based anger is more conducive to criminal behaviors than situational-based anger as it relates to chronic strains (Jennings et al., 2009).
In this study, a composite scale of anger was measured by asking whether participants agreed with the following survey questions: “I have a hot temper,” “I feel that I have a volatile personality like an explosive that is about to blow up,” “I sometimes cannot control my temper when I get mad,” and/or “I become very aggressive when I am upset” (1 = strongly disagree, 5 = strongly agree; α = .73). Depression was also a composite scale comprising four items asking whether participants agreed with the following statements: “I always worry about everything,” “I am sometimes very lonely [and/or ‘sad’ and/or ‘fearful’] for no reason” (1 = strongly disagree, 5 = strongly agree; α = .84).
Conditioning factors, coping strategies, and control variables
The current study included various conditioning factors, namely, constraints and dispositions, proposed by major criminological theories. Peer delinquency is a composite measure operationalized by five items reflecting respondents’ delinquent peer associations. Respondents were asked how many of their close friends had, in the past year, ever (a) engaged in truancy, (b) been a bully, (c) been arrested by the police, (d) been in a fight, or (e) stole something. Responses from each item by participants were summed to form a composite index (range = 0-62; α = .90). Parental attachment and parental supervision measures were used as constraints. Parental attachment is a composite scale of five items referring to participants’ perception of a positive relationship with parents (1 = strongly disagree, 5 = strongly agree; α = .85). The parental supervision variable was comprised of two items tapping each participant’s perception of parental monitoring on their acts (1 = strongly disagree, 5 = strongly agree; α = .86; see the appendix).
This study also includes two types of legitimate coping strategy measures: cognitive and physical. Agnew (1992, 2006) suggested that the impact of strain on an individual’s participation in deviant and other self-destructive problematic behavior might be ameliorated by coping skills. However, legitimate coping skill strategies have, until now, been largely ignored in the extant literature. Only a few recent studies have capitalized on legitimate coping strategies in examining GST (i.e., self-esteem, self-efficacy)—see Jang (2007), Kaufman (2009), and Jennings et al. (2009). First, the data relied upon here included requisite information on two types of cognitive coping strategies: self-esteem and negative self-image. Self-esteem was measured through three distinct items asking if participants perceived themselves as “having a good personality,” “being an able person,” and/or “being a valuable person” (1 = strongly disagree, 5 = strongly agree; α = .83). Negative self-image is also a rarely used measure indicating individuals’ perceived self-image regarding their inability to handle problems and the likelihood of self-blame. It was measured using three items asking whether participants consider themselves “worthless,” “a bad person,” and/or “a loser” (1 = strongly disagree, 5 = strongly agree; α = .76). Second, physical coping strategies reflect physical activities that can serve as coping strategies to reduce the adverse effects of strains and stressors (Agnew, 2006). This was measured using an item asking about participants’ involvement in athletic activities over the past year (1 = yes, 0 = no).
Attitudes toward socialized gender roles, feminine and masculine, were measured according to participants’ responses on the survey questions and statements reflecting their attitudes and perceptions toward the gendered socialized roles of women and men. The feminine socialized role measure was an additional measure comprising three survey questions asking if participants believe that women should “have greater value in marriage than social success,” “pay more attention to men’s opinion than their own,” and/or “comply with men” (1 = strongly disagree, 5 = strongly agree; α = .77). The masculine socialized role measure was operationalized by three survey questions asking if participants agreed with the following statements: “The most important thing for men is social success,” “Men should assert their own opinions rather than comply with others,” and “Men need to have power to control others” (1 = strongly disagree, 5 = strongly agree; α = .76). Finally, two control measures were included in analyses: age and family income. Participants’ age is a continuous variable measured at the time of the interview. Family income is a three-item categorical variable consisting of high- (>US$5,000 per month), medium- (a reference category), and low family income (<US$1,500 per month). The survey questions and variables used in this study are provided in the appendix.
Analytic Strategy
To test the proposed hypotheses, gender-specific processes of strains leading toward outcome behaviors were analyzed. First, means-difference tests were conducted for all variables by gender. This test only provided preliminary information and did not intend to identify a significant difference between males and females. Second, two ordinary regression analyses (ordinary least squares [OLS]) were conducted to investigate the effects of all strain/stress measures on two types of negative emotions (anger and depression). Third, one multiple logistic regression analysis and one OLS regression were estimated. Multiple logistic regression analyses were utilized to examine the effects of strain/stress measures on delinquency, given the measurement of dependent variables (binary). For suicidal thoughts, OLS regression analyses were conducted, given the nature of dependent measurement (Likert-type scale). Both logistic and OLS regression analyses consisted of three steps: First, baseline models (Model 1) were developed to examine strain and stress effects on two types of dependent variables: delinquency and suicidal thoughts. Next, two types of negative emotionality, anger and depression, were included, respectively, in the baseline models for assessing the mediation effects of negative emotions on individual dependent variables (Models 2 and 3). Finally, two separate full models were estimated to test GST and gender hypotheses. That is, it examines whether the strain process to deviance was mediated by negative emotions and if conditioning factors had a different impact on relationship by gender (Model 4). Such analytical approaches in tests of GST are favored and used by many previous studies (for similar approaches in tests of GST, see Jennings et al., 2009; Kaufman, 2009).
Results
Descriptive Statistics and Means-Difference Test
Table 1 provides the t-test results for males and females. Although not all variables were significantly different, some mean differences between the dependent and independent variables across genders were evident. Regarding strain/stress measures, males reported higher levels of parental physical punishment, parental stress, and peer stress, whereas females reported higher levels of victimization, health problems, appearance stress, and monetary problems. For negative emotions, females reported greater anger and depression than males. Regarding conditioning factors, males reported greater levels of coping resources such as self-efficacy but demonstrated greater levels of delinquent peer associations. Females reported having a greater level of attachment to parents and parental supervision than males. Males appear to have a more positive attitude toward gendered social roles. Finally, males were significantly more involved in deviant behaviors than females, while there was no meaningful difference in suicidal thoughts across genders.
Descriptive Statistics and Means-Difference Test for Variables.
p < .05. **p < .01. ***p < .001.
Main Effect of Strains and Stressors on Anger and Depression
Table 2 demonstrates the main strain effects of individual strain/stress measures on negative emotion (anger, depression) across genders. Strains/stressors exerted a significant and positive impact on depression, which did not differ by gender. Both males and females experienced depression based on their victimization, perceived psychological/mental problems, appearance stress, and stress from relationships with parents and friends as well as schoolwork. No gendered processes from these strains to depression were found. The Z tests for coefficient equality showed no statistically significant difference for the effect size across genders for all variables regarding significant effects on depression (Z < ±1.64).
Effects of Strain on Negative Emotions by Gender.
p < .05. **p < .01. ***p < .001.
However, strains leading to anger did emerge, and these differed by gender. For males, three strain variables—parental physical punishment, monetary success, and suspension from school—were significant. For females, six strain variables were significant: psychological/mental problems, parental verbal abuse, parental stress, appearance stress, monetary stress, and suspension from school. The results showed some similarities between South Korean youths and their Western counterparts. They also demonstrated cultural features of contemporary Korea. For example, males more often responded to strains of an overt and physical nature (parental physical punishment) with anger, whereas females more often responded to psychological (psychological/mental problems), emotional (parental verbal abuse), and social relationship (appearance stress) strains with anger. This finding was similar to that of Western studies.
Meanwhile, both males and females were angered by monetary stress and suspension from school. This is associated with the Westernization of South Korean society. Westernization emphasizes monetary success and good school performance, which is the primary channel for upward social mobility equally for both males and females. Thus, both genders responded with anger to such strains. The Z tests for coefficient equality showed no gender difference in monetary stress (Z < ±1.64), but gender differences were found concerning suspension from school (Z > ±1.64). This can be attributed to unequal social stigmatization for girls in South Korean society, meaning that girls are more negatively branded by school suspension than are boys. Thus, it could indicate unequal Westernization between boys and girls in South Korea.
Multiple Regression Models: Logistic and OLS Models
After the main effects of strain and stress on anger and depression were examined, a series of gender-specific multiple logistic regression and OLS regression models were conducted to examine the mediation effects of negative emotions and conditioning effects of coping strategies on two dependent variables, as proposed by the GST hypotheses.
The results in Table 3 were derived from a series of gender-specific logistic regression models on delinquency for males and females. In Model 1, monetary stress and being suspended from school were significant for delinquency in males, whereas parental verbal abuse and being suspended from school were significant for females. This finding indicates the gendered nature of strains leading to delinquency. Males are more responsive to material strains (monetary stress), whereas females are more responsive to parental relationships (parental verbal abuse). Suspension from school led to delinquency in both genders. However, there was a gender difference in effect size (Z > ±1.64). Compared with males, females were more strongly affected by suspension from school leading toward delinquency.
Logistic Regression of Delinquency on Measures of GST by Gender.
Note. GST = general strain theory; OR = odds ratio.
p < .1. *p < .05. **p < .01.
Results from Models 2 and 3 suggest that anger had no direct effect on delinquency for both genders. At the same time, the direct effects of one strain variable for males—suspension from school—and two strain measures for females—parental verbal abuse and suspension from school—were sustained. However, the meanings differed according to the strain measures. Anger partially mediates suspension from school for males but moderates parental verbal abuse for females. A gender difference was thus observed. Depression was significantly related only to delinquency for males. No significant effect of depression on delinquency was found for females. For males, the mediation effect emerged, and a direct effect of depression on delinquency was observed. For females, the findings indicated no direct effect of depression on delinquency, and the direct effect of suspension from school was constant. In addition, there was a gender difference in Model 3.
Model 4 presents the logistic regression results of a full set of strain measures, negative emotions, and conditioning and control variables on delinquency for males and females. For the final full model, no strain measures and negative emotions had direct effects on delinquency. Suspension from school was completely mediated by two intervening factors: peer delinquency and negative self-image for males, and peer delinquency and parental supervision for females. Peer delinquency was a significant and positive factor for both males and females. Negative self-image had significant and positive effects on delinquency only for males, while parental supervision was a single significant factor deterring delinquency in females. This finding indicates a gendered cultural context for South Korean youths. Strains and negative emotions do not directly lead to delinquency, because various social factors are also in play. Strains and negative emotions leading to delinquency interacted with delinquent peer influence, conventional social stigmatization, negative self-image, or parental supervision. Without such conditioning factors, strains and negative emotions may not develop into delinquency. Furthermore, strains leading to delinquency differ according to gender. Unlike males, females were more directly controlled by parents within the family. By contrast, males were more responsive to socialized control of their self-image or reputation outside the family.
Table 4 provides the results of a series of gender-specific OLS regression analyses for suicidal thoughts. Model 1 reveals that several strain and stress measures were significant and positive for suicidal thoughts for both males and females: parental verbal abuse, stresses from parents and peers, and suspension from school. By contrast, appearance stress only appeared significant for suicidal thoughts by males, while monetary stress and stress related to schoolwork were predicted for females’ suicidal thoughts (Model 1).
OLS Regression of Suicidal Thought on Measures of GST by Gender.
Note. GST = general strain theory.
p < .05. **p < .01. ***p < .001.
Models 2 and 3 indicate that anger and depression generate significant and positive effects on suicidal thoughts for both males and females. In addition, anger and depression partially mediate most significant stressors significant in Model 1. Anger completely mediates the positive and significant direct effects of suspension from school for both males and females, and parental verbal abuse for females only (Model 2). Depression completely mediates the relationship between appearance stress for males and males’ suicidal thoughts, as well as the relationship between females’ monetary stress and school stress (Model 3). However, the impact of health problems on females’ suicidal thoughts became significant when anger was included in Model 2. Family dysfunction and stress related to schoolwork became significant for males when depression was included in Model 3.
Finally, unlike the results in the logistic regression models predicting delinquency, the full model pertaining to suicidal thoughts showed that negative emotions completely and partially mediated nearly all the strains and stressors for both males and females (Model 4). The findings demonstrate that many of the direct impacts of strain and stressors on suicidal thoughts disappeared after the mediation and/or conditioning effects of negative emotionality and/or conditioning variables. Furthermore, mediation and/or conditioning effects weakened the direct and positive influence of parental verbal abuse on males’ suicidal thoughts and weakened the effects of parental stress and monetary stress on females’ suicidal thoughts. Like the mediating effects of negative emotionality, several conditioning factors, such as negative self-identity and feminine and masculine social role identities, were significantly associated with suicidal thoughts for males and females. While negative self-identity and feminine social role identity increased the likelihood of suicidal thoughts for both males and females, masculine social role identity decreased the likelihood of suicidal thoughts for males and females. Furthermore, having more parental attachments appeared to significantly reduce suicidal thoughts for males, while self-efficacy seemed to have negative effects on suicidal thoughts for females. A control measure, namely, a higher level of family income, significantly increased the likelihood of suicidal thoughts for males only. That is, males from high-income families were more likely than males from low-income families to respond to suicidal thoughts if they experienced certain strain and stress measures with anger or depression.
Discussion
The purpose of this study was to examine the applicability of hypotheses drawn from Broidy and Agnew’s (1997) theory to explain gender differences in the strain-deviance process of South Korean youths who can be perceived as typical non-Western participants. As such, this study added to the existing GST literature by exploring the impact of various types of strains and negative emotionality on different types of problematic behaviors—deviant and self-destructive—by using a full set of conditioning variables. Specifically, this study included a dependent variable that, while not a typical maladaptive behavior, is harmful to youths. This dependent variable, namely, suicidal thoughts, is important in the study of youths in South Korea where the rate of such behavior is high. A recent study by Kaufman (2009) was the only one to utilize suicidal thoughts as a dependent variable to explore the influence of gender on the strain-deviance process. In addition, the current study contributed to developing the literature on gender and GST by examining the generalizability of the theory to South Korean youths, which meant expanding theoretical generalizability beyond the conventional Western context.
The results of a series of analyses generated mixed support for the proposed hypotheses (Hypotheses 1-4). First, the results partially supported Hypothesis 1, and some differences according to gender were observed. For depression, no gendered difference was evident. These strain variables affected depression for both genders. For anger, the experience of several strain/stress measures differed somewhat according to gender. This finding was intermingled with the gendered difference, the Westernization of contemporary South Korean society, and the cultural duality of unequal Westernization, all of which differed by gender. Males more often responded to overt and physical strains with anger, whereas females responded to psychological, emotional, and interpersonal relationship-based strains with anger. Here, females were more controlled by parents than males. Thus, the conflict between more traditional parents and Westernized female youths was more intense than that between parents and male youths, because parents’ higher control expectations collided with the independent aspirations of Westernized female youths. On the contrary, the cultural transition to Westernization emphasizes individualized competition, and values material success and social upward mobility through education for males and females equally. Not surprisingly, strains stemming from these factors triggered anger for both genders. The t-test results indicate that females experienced higher levels of health problems, appearance stress, and monetary problems, whereas males experienced higher levels of victimization, parental physical punishment, parental stress, and peer stress.
Hypotheses 2 to 4 intended to examine the mediation proposition of negative emotions on the relationship between strain and deviant behaviors and suicidal thoughts. To examine the mediation process, it needed to test if strains were significantly associated with negative emotions (Hypothesis 2), and then if negative emotions were significantly related to outcome behaviors (Hypothesis 3). Finally, it needed to assess the mediation effects of negative emotions on outcome behaviors by examining if the direct effects of strain were weakened by including negative emotions in the full model (Hypothesis 4).
For Hypothesis 2, the results showed some gender differences but differed from the proposition. Gendered differences were found in negative emotionality leading to delinquency. Unlike Hypothesis 2, depression was more likely to predict delinquency for males. Anger was not associated with delinquency for males and females. Nonetheless, a gender difference was observed for delinquency. With the inclusion of anger and depression, the mediation effects of negative emotionality emerged, while the moderation effects of negative emotionality were suspected. Furthermore, the direct effect of negative emotionality on delinquency appeared only for males. This could indicate the gendered procedural nature of negative emotionality leading to delinquency. However, a strong gendered difference was not found in negative emotionality leading to suicidal thoughts. For both genders, strong direct effects of anger and depression were equally found to differ from Hypothesis 2. However, a weak gendered difference could be argued for negative emotionality leading to suicidal thoughts. For males, both anger and depression demonstrated consistent partial or complete mediation effects. For females, this is more complex. For strains related to parents, anger demonstrated partial or complete mediation effects. By contrast, for strains related to peers and school, depression showed partial or complete mediation effects. This could indicate the gendered nature of procedural details in the process from strains through negative emotionality to suicidal thoughts.
As for Hypothesis 2, the findings yielded mixed support for Hypothesis 3, although gender differences were largely withheld. Gender differences and the mediation effects of negative emotionality depended on the types of outcome behavior and negative emotionality. For delinquency, depression had some mediation effects for males. However, there were no mediation effects for both anger and depression among females. Meanwhile, for suicidal thoughts, the gender difference and mediation effects of negative emotionality emerged and were interrelated. For males, both anger and depression demonstrated consistent partial or complete mediation effects. For females, this is more complex, and the mediation effects of negative emotionality were associated with sources of strains. For strains related to parents, anger demonstrated partial or complete mediation effects. By contrast, for strains related to peer and school, depression showed partial or complete mediation effects. In summary, some gender differences in the emotional response toward outcome behaviors—and thus Hypothesis 3—were supported. However, the mediation effects of anger and depression were not persistent and uniform but differed by gender and type of outcome behavior, negative emotionality, and sources of strain. Thus, Hypothesis 3 was partially supported for this matter.
Finally, the results revealed partial evidence for Hypothesis 4 and varied according to types of outcome behaviors. Nevertheless, gendered differential pathways toward deviance and suicidal thoughts through negative emotionality and conditioning factors were largely supported. For delinquency, the direct effects of all strain measures disappeared in the full model. Conditioning factors completely mediated strain variables and depression for males. This might explain the overall moderate delinquency rate of South Korea compared with those of other Western countries. In South Korea, delinquency needs to be mediated by social conditioning factors, and no direct effects of strains and negative emotionality emerged. In addition, only two strain variables for males and one strain variable for females were related to delinquency through the mediations of negative emotionality and conditioning factors for males.
Gendered pathways were also found. For males, negative self-image and peer delinquency were significant mediating variables for delinquency. For females, parental supervision and peer delinquency were significant mediation-conditioning variables. These could be gender-specific findings explaining unequal Westernization between the genders in South Korea. Peer influence affected both genders, but parental influence differentially affected females only, which means that traditional parental control was stronger for females to deter delinquency. By contrast, males were more likely to be affected by socialized stigmatization, which occurs mostly at school outside the family.
For suicidal thoughts, more diverse sources of initial and intermediate causes were related. This explains why South Korea experiences high suicide rates compared with Western countries. Direct effects of strains and negative emotionality, both anger and depression, were sustained for both genders. In addition, conditioning effects partially or completely mediated strains and negative emotionality—anger and depression—toward suicidal thoughts. Interestingly, peer influence was not observed for suicidal thoughts compared with delinquency. Rather, self-esteem, as a positive and socially desirable coping skill, was significantly associated with pathways toward suicidal thoughts for both genders. This means that suicidal thoughts are not a socialized product.
Overall, the pathways toward suicidal thoughts were similarly observed for both genders; however, procedural details differed slightly by gender. For males, parental supervision was a preventive factor for suicidal thoughts. Interestingly, young males from high-income families were more likely to have suicidal thoughts. This finding makes sense, because young males from high-income families are more likely to be disconnected from their delinquent peers and delinquent subcultures in South Korea. Their pathway to delinquency is therefore blocked, making suicidal thoughts another available route for them. For females, the direct effects of appearance stress and monetary stress were sustained, unlike for males. This explains the distinctive characteristics of contemporary South Korean society. Appearance is a form of social status among young females who tend to believe that beauty is an important social capital for social success. However, this is not applicable to young males whose social status and success depend more on merits and family background. Furthermore, monetary success was another source of strain leading to suicidal thoughts for females. This can be attributed to blocked opportunities for delinquent behaviors for females. Both genders experience monetary stress, because material well-being is socially emphasized equally for both. However, males responded to this strain through delinquency through the mediating effects of negative emotionality, whereas females responded through suicidal thoughts because their opportunities for delinquent behaviors are generally blocked, as mentioned.
Taken as a whole, the current study found some evidence for gendered pathways to delinquency and weak evidence for suicidal thoughts. Delinquency was an outcome of socially mediated processes from strains to delinquent behavior, and this process of mediation is gendered. Although the influence of peers was evident for both genders, females were more affected by parental factors, whereas males were more affected by socialized stigmatization. The gendered difference was relatively weak for suicidal thoughts. Most strain, negative emotionality, and conditioning factors had direct and mediation effects for both genders. Nevertheless, some gender differences were also found. Compared with males, females more often responded to gender-specific strains, appearance stress, and monetary success through suicidal thoughts. The perseverance of the direct effects of these strains supports these gendered differences. Furthermore, the suicidal thoughts of males, unlike those for females, were more related to family status and parental factors. For males, parents’ control and parental supervision deterred suicidal thoughts. However, for females, parents’ control had no effect on suicidal thoughts but deterred delinquency. Higher family status and a high income facilitated suicidal thoughts. This could be because of blocked delinquent opportunities for males from families with higher status, who were more likely to respond with suicidal thoughts. No such dynamics were found for females.
In summary, the current study supports the hypothesis by Broidy and Agnew (1997) regarding gender and GST by showing some generalizability of the theory to the South Korean context. For strain variables, significant gender differences were evident depending on the type of outcome behavior. Some strain variables related to either deviant behaviors or suicidal thoughts demonstrated differentially gendered direct or indirect effects. Furthermore, the findings of the full multivariate models revealed that the mediation effects of negative emotionality were related to types of outcome behaviors. For delinquency, the mediation effect was limited for males in the case of depression. For suicidal thoughts, the mediation effect was substantial, while the mediation effect of conditioning factors was quite robust. The mediation effect was complete for delinquency and partial or complete for suicidal thoughts. These mixed findings are consistent with previous studies conducted on Western samples and those using South Korean samples (Baron, 2007; Hay, 2003; Jang, 2007; Jennings et al., 2009; Kaufman, 2009; S. Lee, 2003; Morash & Moon, 2007).
The current findings reveal the South Korean cultural contexts of unequal Westernization and cultural duality. Cultural conflict exists between more traditional and Confucian parents and more Westernized and liberal youths. This conflict triggers strains for youths but deters juvenile delinquency overall. Thus, for both males and females, many strains and negative emotions do not develop into delinquency. Strains and negative emotions should be mediated by weakened parental controls. However, a high level of strains triggered by such conflicts leads to suicidal thoughts through negative emotionality. This could be attributed to the fact that high-income families, in which parental control is usually high, have greater effects on suicidal thoughts for males. In addition, parental control is unequally enforced on children across genders. Parents more tightly control females, while they are more permissive to males. On the contrary, males are more resistant and confrontational to their parents, whereas females are more compliant and submissive. Thus, males’ delinquency was positively related to peer-related factors, whereas females’ delinquency was negatively related to parental supervision. Socially, Westernization affects males and females differentially in South Korea. Individualized material success and upward mobility facilitated by Westernization are equally emphasized for both genders. However, the means of obtaining such culturally desired goals were gendered. Unlike for males, the opportunity for delinquent behavior to resolve material stress was blocked for females. Material strains thus lead to suicidal thoughts. In South Korea, females are socially evaluated primarily by appearance, and males primarily by merits. Thus, although males experience appearance strains, they have other ways to compensate. However, for females, appearance strain is a critical strain source that cannot be resolved by other means. The findings of the current study imply this social context in contemporary South Korean society.
However, the findings of this study should be interpreted with caution. One primary concern for the current research results is the possibly hidden effect of social learning and social bonding variables. For example, suspension from school, peer stress, and parental verbal abuse appeared as significant strain predictors for each deviant behavior. However, although the current study counted these variables as strain measures, they can also be perceived as social learning and/or social bonding measures; for example, while suspension from school is a stressor for a youth, it may also facilitate social gatherings by providing unsupervised hours and the opportunity for differential association toward delinquency. The limitation of survey items in the current secondary data means that this issue cannot be resolved. It is merely assumed that both strain and social learning processes may occur together.
In addition, other limitations needed to be addressed. This study did not use other types of negative emotions, such as fear, anxiety, and guilt. Because the functions of other types of negative emotions are not clearly identified, it may be important to examine their mediation effects to fully assess gender and GST hypotheses. Furthermore, although this study attempted to utilize various types of strains and stressors, including gender-specific strain measures (i.e., appearance stress), these strain measures are still less than satisfactory because of the limitation of the secondary data. More gender-specific strain measures not included in this study need to be identified and included in the model, and the strain measures included in the model in this study may need further elaboration.
In spite of these limitations and concerns, this study has merits in terms of the development of the gender and GST literature. It uses longitudinal and representative data on South Korean youths; thus, it can examine the generalizability of gender and GST issues beyond the Western context. In addition, the variables used in the current study include various types of strains (including serious strain and gender-specific strain), two important types of negative emotions, legitimate coping strategies (physical and cognitive coping strategies), and various types of conditioning factors drawn from major criminological theories. Therefore, this study offers a full test of gender and GST, and provides temporal ordering and representative samples to secure a greater level of confidence in the generalizability of its results. In addition, this study explains South Korean society, which, consequent to undergoing an insufficient and unequal transition to Westernization, demonstrates cultural duality, unlike Western societies. Despite its limitations, the current study raises important arguments that might facilitate future studies focusing on such transitional societies.
Footnotes
Appendix
Descriptions of Strain/Stress Measures and Conditioning Measures.
| Variable | Scale | Survey Items |
|---|---|---|
| Strain | ||
| Victimization | Composite scale | “Have you ever been victimized by the following crime during the past year?” (yes/no): Seriously beaten, money extortion, sexual abuse or rape, threatened, being bullied, and malicious behavior (six items) |
| Health problems | A single item a | I have a health problem now |
| Psychological/mental problems | A single item a | I have a psychological and mental problem now |
| Family financial problem | A single item a | My family has financial problems now |
| Parental verbal abuse | A single item a | My parents often seriously insult me verbally |
| Parental physical punishment | A single item a | My parents sometimes use psychical punishment seriously |
| Stress | “I am under stress because . . .” | |
| Parental stress | Composite scale a | “My parents give me stress because my bad grade,” “I have conflicts with my parents,” “My parents excessively interfere me,” and “I have trouble in communicating with parents” |
| School stress | Composite scale a | “My bad score or grade at school,” “homework or examinations,” “concerns about getting job or entering college in the future,” and “I am sick and tire of studying” |
| Peer stress | Composite scale a | “My friends ignore or tease me badly,” “I could not win recognition from friends,” and “I feel I am inferior to my friends” |
| Appearance stress | Composite scale a | I am overweight or underweight. I am too tall or too short. I am not a good looking and I cannot wear fancy outfit |
| Monetary stress | A single item a | I don’t have enough money to buy what I want to buy |
| Suspended (school) | A single item a | I have been a subject of disciplinary punishment or warning from school |
| Negative affect | ||
| Anger | Composite scale a | “I feel that I am a very volatile person like explosives to be blown up,” “I have a hot temper,” and “I cannot control my feeling to beat other when I upset” |
| Depression | Composite scale a | “I always worry for everything,” “I am, sometimes, very lonely with no reason,” and “very depressed/sad with no reason” |
| Conditioning factors | ||
| Physical coping | A single item | “Have you joined physical activity during the last year?” (yes/no) |
| Peer delinquency | Composite scale | “How many of your close friends involved in or was . . . ?” (yes/no): Aggressive behaviors at school, arrested by police, truancy, violently beating, extorting money and property and Stealing (7 times) |
| Parental attachment | Composite scale a | “My parents try to spend enough time with me,” “always show their affection for me.” “I talk often about what happened to me outside home,” and “I feel easy to share my thought and feeling with my parents” |
| Parental supervision | Composite scale a | “My parents usually know where I am, while I am away from home,” “with whom I am, while I am away from home” |
| Self-esteem (cognitive coping) | Composite scale a | “I believe I have a good personality,” “I believe I am an able person,” and “I believe that I am a very precious person” |
| Negative self-image (cognitive coping) | Composite scale a | “Sometimes, I think I am worthless,” “a bad person,” and “I am a loser” |
| Feminine identity | Composite scale a | “Women should have greater value in marriage than social successes,” “Women should be amenable than assertive,” and “Women should be submissive and obedience” |
| Masculine identity | Composite scale a | “The most important thing for men is social success,” “Men should carry their own point persistently than follow other’s opinion,” and “Men who can dominate others are real men” |
1 = strongly disagree, 2 = disagree, 3 = neither agree nor disagree, 4 = agree, and 5 = strongly agree.
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) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This project is supported by a government-funded organization, the National Youth Policy Institution (NYPI) in South Korea.