Delinquency Among Turkish Adolescents: Testing Akers’ Social Structure and Social Learning Theory

Abstract

The purposes of this study are to examine how social learning processes and social structure correlate with delinquency among Turkish adolescents and to articulate to what extent Akers’s social structure and social learning (SSSL) theory explains delinquency in Turkey, which is a different cultural context from Western countries in terms of family structure, level of collectivism or individualism, religion, belief systems, and norms. This study contributes to the existing body of knowledge by providing the first study testing Akers’s theory in the Turkish context. Analyses, relying on a structural equation modeling (SEM) framework, showed that the social learning process accounted for a substantial amount of variation in explaining adolescent delinquency. We also found that social learning process somewhat mediates social structural effects on delinquency.

Keywords

social structure social learning theory delinquency adolescents Turkey

Introduction

In all societies, some individuals tend to behave in a nonnormative way and engage in illegal or antisocial activities compared with others. Also, the frequency of these activities is higher among adolescents than it is in other age groups (Gottfredson & Hirschi, 1990). This diverse phenomenon—adolescent delinquency—is a common problem for most societies all over the world, although most deviant behavior–related theories that are used to explain this phenomenon have been produced in Western countries. Therefore, there has been a great desire among sociologists, criminologists, and policy makers to conduct cross-cultural studies to more ably understand the different aspects of adolescent delinquency and develop solutions for preventing this social problem (Junger-Tas et al., 1994). In addition, in efforts to test the explanatory power of any deviant behavior–related theory, a study of cultural differences can provide a good opportunity to examine whether or not a specific theory can be applied universally, or whether it is more prone to explain a particular society.

As in other societies, delinquency is a social problem in Turkish society and has recently been on the rise (Ozdemir et al., 2013). A dramatic rise has also been observed in the total number of crimes committed by Turkish adolescents, which rose from 62,441 to 115,429 over the last 10 years (TurkStat, 2015). In addition, in comparison with Western countries, the cultural differences in Turkey, with its mixture of Islamic and secular norms, deserves specific attention because such research may reveal how sociological and behavioral dissimilarities affect deviant trajectories among adolescents. In this vein, along with their colleagues from different countries, Turkish criminologists and sociologists have started to test Western crime theories over the last 10 years, in efforts to examine to what extent these dominant theories explain delinquency in Turkey (Ozbay & Ozcan, 2006a, 2006b; Ozdemir et al., 2013; Unal & Cukur, 2011; Yuksek & Solakoglu, 2016).

Among Western crime theories, social learning theory has been one of the most influential approaches in the fields of criminology, sociology, and the behavioral sciences in terms of explaining the motivations behind delinquency. Since 1966, when Burgess and Akers (1966) published the first version of social learning theory, many researchers have benefited from it in their attempts to explain delinquency among adolescents. As a result of the feedbacks from these studies, the theory was enhanced and renamed social structure and social learning (SSSL) theory by Akers in 1998. This new theoretical framework covered a broader context and started to serve as a general theory for explaining crimes in the realm of criminology and sociology.

Although Akers’s SSSL theory, as a general theory of crime aiming at revealing the reasons behind criminal behavior by examining both the micro- and macro-level factors, has been widely used in sociological research, there is still a need for further empirical studies using the SSSL theory as the explanatory power of it is still criticized by some scholars (Pratt et al., 2010). In this vein, this study attempts to contribute to the sociological literature by testing Akers’s SSSL theory in a different context compared with the previous studies conducted in the Western world. In this respect, Turkish society, as the host of a different and mostly unknown culture compared with Western societies, provides a notable venue for testing Western crime theories, as the social domains such as family and community might play a different role in individuals’ daily lives in comparison with their role in Western societies (Karakus et al., 2010). While Akers and Jensen (2006) argue that social learning theory is not directly related to culture and social bonds, as a dominantly Muslim-populated and culturally diverse society, Turkish society might affect how individuals interact with each other and act differently from those residing in Western countries (Wasti, 1998). In this regard, this study, conducted in a dominantly Muslim country for the first time, attempts to reveal the degree to which cultural differences and similarities can explain youth delinquency using Akers’s SSSL theory.

Conceptual Background and Literature Review

The Turkish Context

Before addressing whether Akers’s SSSL theory is applicable to Turkish adolescents, first, we would like to discuss Turkey to provide the readers with a better context with which to understand the purposes of this study and familiarize them with the social structure and epichoric culture of the country that this study focuses on.

Turkey is a country located as a bridge between the Middle Eastern countries and Western countries. Due to its location and its democratic governance despite its Islamic bonds with the Middle East, its sociocultural structure has been under the influence of both Islamic and secular values and norms. Although the majority of people in Turkey are Muslim as in other Middle Eastern countries, its social fabric is substantially different in comparison. Its economic (e.g., ties with World Bank and International Monetary Fund), political (e.g., being a member of United Nations, North Atlantic Treaty Organization [NATO], Organisation for Economic Co-operation and Development [OECD]), and cultural connections with the Western World and communal and religious ties with the Middle Eastern countries set Turkey in a unique and interesting position. Turkey, in this context, can be said to be both a Western and Middle Eastern country (Yuksek & Solakoglu, 2016).

The strong traditions of family and community in Turkey resulting from the role played by the religion of Islam manifest themselves with the presence of strong individual bonds to prosocial others, such as parents, friends, and relatives. This leads, differently from the Western countries, to a more collectivist rather than individualistic culture and closer supervision on adolescents’ behaviors through families, friends, and society in general (Ozbay & Ozcan, 2006b). First, parents exercise social control over adolescents until they become financially independent and get married. Second, neighborliness is strong in the Turkish culture; therefore, neighbors generally know each other well. This leads neighbors to exercise efficient social control over adolescents even if they are not their children. Neighbors do not distinguish adolescents residing in their neighborhood from their own children. Finally, schools also exercise social control through teachers and school administrators. In this respect, the collectivist nature of the Turkish culture normalizes the implementation of the above-reported forms of social control over adolescents, and this contributes to the restriction of adolescent individualism and freedom and to the prevention of adolescent delinquency (Yuksek & Solakoglu, 2016). Although this discussion about Turkey and its collectivist culture seem to be more relevant to a study of social control theory than social learning theory, we argue that the implementation of social control over adolescents through families, neighbors, and schools might prevent the formation of contexts where criminal behavior is learned, moderate the relationship between definitions favorable to law violation and deviant behaviors, and protect adolescents against exposure to deviant models.

As in Western societies, adolescent delinquency has also been one of the most important social problems in Turkey. Accordingly, researchers have addressed adolescent delinquency in the Turkish context and tested Western origin theories to explain delinquency, such as the relationship between adolescent delinquency and social class (Ozbay, 2006), self-control theory and violence among Turkish adolescents (Ozbay & Koksoy, 2009), Hirschi’s social bonding theory (Ozbay & Ozcan, 2006b; Yuksek & Solakoglu, 2016), classic strain theory and delinquency in the context of gender (Ozbay & Ozcan, 2006a), the role of school on delinquency (Unal & Cukur, 2011), and the role of family and self-control on delinquency (Ozdemir et al., 2013). Although some of the above-mentioned studies also addressed components of the SSSL theory, such as peer delinquency/differential association in Ozbay (2006) and Ozbay and Koksoy (2009), there is no single study that attempts to fully capture Akers’s SSSL theory in the Turkish context. In this regard, the current study provides us with the opportunity to expand the scope of the SSSL theory to a different context than the Western societies, to Turkey.

Social Learning Theory

The social learning theory is considered as a behavioristic reformulation of Sutherland’s (1947) differential association theory. The differential association theory proposes that in modern societies, people are variously exposed to cultural messages—messages, including both criminal and anti-criminal patterns, techniques, motivations, and definitional stances toward the legal norms—favoring or opposing criminal behavior through interpersonal interactions. Through these interactions with definitions favorable to criminal behavior rather than unfavorable ones, people learn criminal behavior. Also, depending on the intimacy (priority), frequency, longevity (duration), and intensity of these interactions, they have greater effects on behavior (Sutherland, 1947). Akers’s (2009) social learning theory built upon Sutherland’s theory by integrating processes of differential association and definitions from Sutherland’s theory with differential reinforcement and other principles of behavioral acquisition, continuation, and cessation from behavioral learning theory. In this regard, the aim of the social learning theory is not to compete with Sutherland’s arguments but to provide a broader theory to more fully explain criminal behavior.

Social learning theory mainly consists of four core dimensions, which are differential association, definitions/norms, differential reinforcement, and imitation. In this study, we will address the relationship between the three dimensions of social learning theory (differential association, definitions/norms, and differential reinforcement) and delinquency and then generate our hypotheses regarding this relationship.

Differential association

Differential association refers to direct and indirect interactions with an individual or a group, such as friends and family. These interactions provide the social contexts in which all the mechanisms of social learning operate. Too often, social learning theory is mistakenly taken as only a theory of peer influence. However, as Akers and Jensen (2006) indicate, “there is no question that the mutual influences of peer behavior, associations, values, behavior, and reinforcement are central to social learning theory’s explanation of acquisition, maintenance, and change in deviant behavior” (p. 51). Therefore, it is safe to say that the kernel of studies on delinquency lies in the association between peers and the propensity of engaging in delinquency (Haynie, 2001). This relationship is not only a central predictor for studies on delinquency but also the heart of differential association and social learning theory. Moreover, sufficient evidence is available to indicate that there is a strong and positive relationship between peer delinquency and an individual’s own delinquency. Peer delinquency has been used as one of the most important predictors to understand the underlying reasons of delinquency regardless of the names of the theoretical frameworks. In addition, cross-national studies also found that differential association is not only an important predictor for explaining delinquency among adolescents but also plays a major role in understanding crimes in different cultures (Miller et al., 2008; Tittle et al., 2012). Over and above all this, studies conducted in Turkey have demonstrated that peer delinquency follows the same pattern and increases an individual’s own delinquency among Turkish adolescents (Ozbay, 2006; Ozbay & Koksoy, 2009).

Definitions/norms

Definitions/norms basically refer to one’s realization, justification, and neutralization of criminal activities in a positive or a negative way (Akers & Jensen, 2006). It can be asserted that the justification of adolescent delinquency is a major step before engaging in deviant acts (Sykes & Matza, 1957). However, norms against deviant behaviors can be prohibitive for engaging in delinquency, and these norms can originate from a society, religion, family, or a similar social institution in which adolescents live (Akers, 2009). Many empirical studies have provided tangible evidence supporting the assertion that norms favorable to delinquency increase the likelihood of engaging in criminal acts (Lee et al., 2004; Massey & Krohn, 1986; Minor, 1980; Winfree & Griffiths, 1983). In addition, cross-national studies conducted in the United States have produced similar results (Meneses & Akers, 2011; Tittle et al., 2012). Considering the research conducted in Turkey, it can be seen that there are little to no studies that have centered on social learning theory, and studies focusing on beliefs and norms have been examined mostly with theoretical approaches, such as social control theory and containment theory (Ozbay & Ozcan, 2006b, 2008; Yuksek & Solakoglu, 2016). In this regard, although Turkish society provides a different cultural structure than Western societies in terms of religion, customs, and norms, we argue that Turkish and Western societies resemble each other as most of the deviant behaviors were undesirable in both Turkish society and Western societies.

Differential reinforcement

Differential reinforcement refers to a reward or a punishment after the criminal act or deviant behavior (Akers & Jensen, 2006). Differential reinforcement can be constituted by a social or a nonsocial fact, such as emotion, sense, and pleasure that prevents adolescents either from engaging in delinquency or from encouraging them to behave in a delinquent way (Akers, 2009). Reinforcement might involve a cost, a reward, or a negative reaction/punishment to delinquency by others. Considering socialization as a social learning process in which individuals learn from parents, peers, and others the things to do and say in specific contexts, the process of initiating, continuing, or modifying conforming and deviant behavior in the peer group and family context can also be considered as a social learning process in which interactions between peers and between parents and their children expose the children to normative values, behavioral models, and direct reinforcement (Akers, 2009). In this regard, the control efforts of peers and parents through positive reactions to conforming behavior and negative reactions to deviant behavior (through monitoring and sanctioning) might curtail delinquency and support conformity of the youth to prosocial attitudes and differential reinforcement of conformity over deviance (Simons et al., 2004). Similar to other components of social learning theory, the literature reports many studies, showing that cost, reward, and reaction to delinquent behaviors by others have a significant impact on delinquency (Sellers et al., 2003; Winfree & Griffiths, 1983). In addition, studies conducted in many countries have provided empirical evidence for the link between differential reinforcement and delinquency (Hwang & Akers, 2006; Yun & Kim, 2015). Although this is the case for many countries, there is limited evidence as to whether cost of deviant behaviors and reactions by family or others are associated with adolescent delinquency in Turkey.

Following the above discussion, Akers’s (2009, 2012) social learning theory suggests that the process of social learning should explain a substantial amount of within-individual variation in delinquent behaviors. In line with this reasoning, in this study we expect the following:

Hypothesis 1: The social learning process, with its correlates of differential association, definitions/norms, and differential reinforcement, should have a direct significant effect on adolescent delinquency and account for a substantial amount of variation in explaining adolescent delinquency.

Social Structure Model

In addition to the core dimensions of social learning theory discussed above, Akers (2009) proposed his general theory of crime, the SSSL theory, to expand and elaborate upon his original theory. In developing an SSSL model of criminal behavior, Akers (2009) benefited from the structural variation that stems from macro-level predictors of offending (e.g., age, gender, social class, race) while simultaneously accounting for individual-level, or group-level, variation in crime causation as explained by the principles of social learning theory (Morris & Higgins, 2010). According to Akers (2009), social structure produces variations of delinquency by affecting the process by which individuals learn to refrain from or commit acts that comprise delinquency. Akers (2009) specifies four dimensions of social structure that provide the context that the social learning concepts exist and the learning process takes place: (a) differential social organization, referring to the ecological, community, or geographical differences of an individual; (b) differential location in the social structure, referring to the sociodemographic characteristics that place individuals and groups in their specific categories, such as class, age, gender, race, and ethnicity; (c) theoretically defined structural variables, referring to criminogenic conditions of societies, such as anomie, social disorganization, and group conflict; and (d) differential social location in groups, referring to membership of the individual in primary, secondary, and reference groups. The foundation of SSSL is that individual differences in (a) social organization, (b) one’s location in the social structure, (c) the impact from structural theories of crime, and (d) one’s position in a social group will explain variation in rates of crime and deviance and these differences are mediated through the principal components of social learning theory (differential association, definitions/norms, differential reinforcement, and imitation; Morris & Higgins, 2010). In other words, empirical models incorporating dimensions of SSSL with delinquency as the dependent variable should show (a) significant direct effects of structural variables on the social learning process; (b) nonsignificant or at least substantially reduced direct effects of social structure variables on the dependent variable; (c) significant direct effects of the social learning process on the dependent variable (Baron & Kenny, 1986). In this section, due to data restrictions, we will address two social structure dimensions, which provide a basis for our hypotheses. These two dimensions are differential location in the social structure and differential social location in groups.

Differential location in the social structure (age, gender, and social class)

Differential location in the social structure covers the sociodemographic variables, such as gender, age, class, or race/ethnicity, which are generally used in crime-related studies as control and background variables (Akers, 2012). Age is a well-known predictor in studies related to social learning theory and other theoretical approaches developed in criminology (Akers & Lee, 1999). There is a general agreement in the literature that there is a positive relationship between age and delinquency during adolescence, and it becomes a negative relationship in later adulthood (Lee et al., 2004). As Thornberry (1987) argues, attachment to parents has a stronger influence on children during adolescence where parents play a key role in controlling the behavior of youth. As children age, their activities gradually move from home to school and peer networks. Therefore, the overall strength of parental influence on adolescents’ behavior (the constraining effects of the family and parents on delinquency) gets weaker, and their exposure to contexts where they can learn criminal and deviant behaviors increase. Research conducted in Turkey has provided mixed results on age indicating both positive and negative relationships with delinquent activities. For example, Ozbay and Ozcan (2006a) found that although age has a positive effect on school delinquency, it has a negative effect on public disturbance. In contrast to Ozbay and Ozcan (2006a), Unal and Cukur (2011) found no relationship between age and delinquency. In line with the SSSL theory regarding age as an indicator of location in the social structure and the limited evidence as to whether age is associated with delinquency based on the previous literature, we expect the following:

Hypothesis 2a: There will be a significant relationship between age and delinquency: Older children will be more likely than younger children to be engaged in delinquency.

Also, in line with the SSSL model, the effect of age on delinquency should be mediated by the social learning process. Therefore, we hypothesize the following:

Hypothesis 2b: The direct effect of age on delinquency will be mediated through the social learning process. That is, age will have a significant direct effect on the social learning construct and a nonsignificant effect on delinquency.

Gender has always been one of the central characteristics examined in the context of criminal research. Although gender as a variable exists in most of the criminological studies as a control variable, there have been many studies centered on gender to understand the motivations behind delinquency, that is, whether males have a stronger tendency than females to engage in delinquent behaviors (Bartusch & Matsueda, 1996; Hagan et al., 1985; Piquero et al., 2005). For example, Canter (1982) argues that females spend more time at home than do males and are more concerned with family relationships. Their lesser involvement in delinquency stems from closer parental supervision and monitoring. Akers (2009) thinks similarly and argues that females have a lower level of tendency to engage in delinquency compared with males as the behaviors of females follow a more confirmative way of behaving in comparison with males, due to their higher level of exposure to social control by their families, compared with males. In this sense, males’ greater involvement in delinquency is due to the combination of their higher exposure to risk and lower exposure to protective factors (Fagan et al., 2007). Thus, males’ higher exposure to an excess of definitions favorable to law violation over definitions unfavorable to law violation may increase their likelihood to engage in delinquency more than girls. In addition, studies conducted in the Turkish context (Ozbay & Ozcan, 2006a, 2008; Yuksek & Solakoglu, 2016) indicate the significant influence of gender on the tendency of delinquency. Therefore, we hypothesize the following:

Hypothesis 3a: There will be a significant relationship between gender and delinquency: Boys will be more likely than girls to be engaged in delinquency.

Also, similar to age, we expect the following:

Hypothesis 3b: The direct effect of gender on delinquency will be mediated through the social learning process. That is, gender will have a significant direct effect on the social learning construct and a nonsignificant effect on delinquency.

Social class is another characteristic that has been considered as an important factor in explaining crime and delinquency. In the SSSL, social class is argued to influence delinquency to the extent that it affects the social learning process (Lee et al., 2004). Social class might affect social learning by way of class-related interpersonal stress, as in lower class households, where there may be more stress due to financial hardships and more negative family relationships that may lead children to criminal behavior (McCord, 1991), or social capital, as in upper-middle and upper class households, where children have more extensive social networks to attain role models that encourage them to succeed through legitimate activities (Krivo & Peterson, 1996; Lee et al., 2004). Despite these arguments, social class might also be positively associated with delinquency, as in the case of alcohol consumption (Akers, 1992). However, in the SSSL, the mediation of structural effects by social learning variables is not dependent on the direction of those effects. Therefore, we expect the following:

Hypothesis 4a: There will be a significant relationship between social class and delinquency.

Also, we expect the following:

Hypothesis 4b: The direct effect of social class on delinquency will be mediated through the social learning process. That is, social class will have a significant direct effect on the social learning construct and a nonsignificant effect on delinquency.

Differential social location in groups (family structure)

Differential social location in groups refers to individuals’ membership in primary, secondary, or reference groups, such as the family, peer groups, or colleagues (Akers, 2012). Family is the primary social institution in which associations are differentiated (Akers & Jensen, 2006). Moreover, family is the main actor or agency for providing social control for adolescents (Gottfredson & Hirschi, 1990). In this sense, family structure, namely, whether or not adolescents live with their both parents or live in the context of other arrangements, can have an impact on social control, social learning, integration into society, and the regulation of all who have the capacity to encourage/discourage adolescents from engaging in delinquency. There has been sufficient support suggesting that living with two parents has a negative effect on delinquency (Akers & Jensen, 2006; Lee et al., 2004). Despite the previous studies that associate adolescents living in single-parent and nonparent households compared with two-parent households with higher likelihoods of engaging in delinquency, based on the collectivist Turkish culture, we argue that the implementation of strong social control over adolescents of single-parent and nonparent households protect them against the learning environments of definitions favorable to law violation and exposure to deviant models. Therefore, we expect to have a nonsignificant relationship between family structure and delinquency. In other words, children of two-parent households and those of single-parent and nonparent households will not differ in their likeliness in engaging in delinquency.

Data and Measures

Data

The data used in this study come from a survey conducted by the European Cities Against Drug (ECAD)¹ organization. ECAD is a nonprofit organization aimed at promoting a drug-free Europe. For this purpose, in 2005, the organization launched a comparative research project under the name of “Youth in Europe”² in corporation with member cities to develop a solution and policy in response to drug problems among adolescents. The project covers ECAD member cities, and in the context of this project, Youth in Europe Survey of Istanbul/Bagcilar, the data used in this study, was conducted in 2007 as part of this comparative project. The survey was administrated to 2,445 high school students using a random sampling selection technique.³

Bagcilar, where the survey was conducted, is a district of Turkey’s most populated and cosmopolitan city, Istanbul. Although Bagcilar is not considered a city and is part of Istanbul, it is highly populated and denser than most of the Turkish cities. According to TurkStat (2015), Bagcilar’s population in 2007 was 719,267 and most of its residents belong to mainly lower middle and lower classes.

The survey questions, originated from the Youth in Europe surveys, were prepared by the Icelandic Center for Social Research and Analysis (ICSRA).⁴ The survey was designed to measure social problems among youth, covering delinquent behaviors, school attachment, health status, academic achievement, smoking and alcohol usage, leisure activities, emotional features of adolescence, and suicidal attitudes.

Dependent Variable

Six items were used to capture the likelihood to engage in delinquency. Respondents were asked how often (if ever) they had done any of the following during the last 12 months: punched somebody, knocked somebody over, kicked somebody, hit/slapped somebody, held somebody by their neck, and threatened somebody with violence. Response categories range from never to 18 times or more. To create the delinquency scale, each item was coded as “0” if the respondent was never involved in delinquency in the last year and “1” if the respondent was involved in delinquency once or more times.⁵ Higher scores on the scale indicated a greater likelihood of engaging in delinquency. The internal consistency of the six items was large (α = .93).

Independent Variables

Differential association was measured using eight items. Following Krohn et al. (1985) and Morris and Higgins (2010), the respondents were asked, to the best of their knowledge, how many of their friends have done the following during the last 12 months?—stolen something worth more than three movie tickets, broken into a building or a car to steal, damaged or vandalized things that did not belong to you, searched out for fights, picked fights, smoked cigarettes, smoked hash or marijuana, and drank alcohol. The response options are none (0), a few (1), some (2), most (3), and all (4). Higher scores on the scale indicated more association with delinquent peers.⁶ Principal components factor analysis using varimax rotation, shown in Table 1, supported the use of a single factor (α = .85).

Table 1.

Principal Components Factor Analysis Results for the Measures of Social Learning Theory.

Social Learning Theory Measures	Factor loadings	M	SD
Indicators of differential association
Stolen something	0.583	0.099	0.468
Broken into a building	0.589	0.079	0.448
Damaged things	0.634	0.137	0.531
Searched out for fights	0.773	0.636	1.072
Picked fights	0.792	0.831	1.144
Smoked cigarettes	0.748	0.947	1.151
Smoked marijuana	0.746	0.258	0.748
Drank alcohol	0.777	0.641	1.016
Eigenvalue	4.035
Variance explained	50.40%
α	0.850
Indicators of definitions/norms
Situation 1	0.628	2.722	1.080
Situation 2	0.704	2.572	1.039
Situation 3	0.722	2.186	1.049
Situation 4	0.606	2.381	1.109
Situation 5	0.641	1.344	0.779
Situation 6	0.640	1.282	0.714
Situation 7	0.588	1.229	0.682
Eigenvalue	2.948
Variance explained	52.10%
α	0.765
Indicators of cost of differential reinforcement
Drinking alcohol	0.890	4.316	1.082
Smoking cigarettes	0.883	4.268	1.060
Smoking cannabis	0.854	4.377	1.106
Rebelling against rules	0.595	3.809	1.297
Stealing from shops	0.738	4.471	1.061
Eigenvalue	3.204
Variance explained	64.10%
α	0.846
Indicators of parental reaction to delinquency
Smoking cigarettes	0.836	3.863	0.481
Drinking alcohol	0.855	3.876	0.464
Smoking cannabis	0.840	3.934	0.328
Eigenvalue	2.136
Variance explained	71.20%
α	0.783

An important issue concerning differential association that needs to be considered is that using respondents’ perception of their friends’ delinquency as a measure of differential association and then using that to explain respondents’ self-report of their own delinquency has been criticized as producing an empirical tautology (Lee et al., 2004). However, both cross-sectional and longitudinal research have shown that respondents’ report of friends’ delinquency is not simple a reflection of one’s own delinquency. Rather, these two measures point to distinct phenomenon, and independently of one’s report of friends’ delinquency, self-reported delinquency is strongly associated with peer delinquency (Akers, 2009).

Definitions/norms was measured using seven items (situations) referring to respondents’ reports whether or not they justify engaging in delinquent behaviors. These seven items are “sometimes there are situations that justify people being beaten up or hit,” “when someone treats me badly I think it is okay to beat up him or her or hit him or her,” “sometimes you need to hit or punch people to protect your honor in your peer group,” “one who does not respond to a personal attack is considered a coward in my group of friends,” “sometimes it is necessary to smoke cigarettes in order not to be left out of the peer group,” “sometimes it is necessary to drink alcohol in order not to be left out of the peer group,” and “sometimes it is necessary to smoke cannabis in order not to be left out of the peer group.” Answer possibilities are strongly agree (4), agree somewhat (3), disagree somewhat (2), and strongly disagree (1). Higher scores on the scale indicated more favorable attitudes toward delinquency. Principal components factor analysis using varimax rotation (see Table 1) supported a singular construct (α = .76).

Cost of differential reinforcement was measured using five items. Respondents were asked to what extent their selected delinquent actions were considered as a reason for losing respect from their friends: drinking alcohol, smoking cigarettes, smoking cannabis, rebelling against the rules of adults, and stealing from shops. Response categories included increases respect a lot (1), increases respect somewhat (2), has no effect (3), decreases respect somewhat (4), and decreases respect a lot (5). Higher scores on the scale indicated higher cost, namely, lower levels of respect. Principal components factor analysis using varimax rotation (see Table 1) supported a singular construct (α = .84).

Parental reaction to delinquency was constructed using three items. Respondents were asked how their parents would react if they would engage in delinquent behaviors, such as smoking cigarettes, drinking alcohol, and smoking cannabis. Response possibilities are totally against (4), very much against (3), rather much against (2), and they would not care (1). Higher scores on the scale indicated higher levels of reaction by parents to delinquency. Principal components factor analysis using varimax rotation (see Table 1) supported a singular construct (α = .78).

Our research includes two of the four dimensions of social structure specified in the SSSL theory. Age, gender, and social class were employed to represent differential location in the social structure, whereas family structure was used to represent differential social location in groups. One’s location in the social structure indicated by age, gender, social class, and family structure is expected to affect one’s chances of learning deviant and criminal behavior. Because, as Lee et al. (2004) state, these locations “structure one’s exposure to models, associations, reinforcements, attitudes, and other aspects of the learning process” (p. 18). In this context, age is a numerical variable ranging from 12 to 16 years. Gender was recoded as male (1) and female (0). As the respondents in the sample are adolescents who were still in school and not employed full time, we used an equal weighed combination of mother’s education, father’s education, mother’s occupation, father’s occupation, and family’s financial status to measure social class (Lee et al., 2004). An adolescent’s mother’s and father’s education consist of five categories, ranging from “primary school or less” to “college graduate.” An adolescent’s mother’s and father’s occupation consist of three categories: mother (father) works at home, works part-time outside the home, and works full-time outside the home. An adolescent’s families’ financial status is created using four items referring to the financial status of the parents: (a) my parents are poorly off financially, (b) my parents cannot afford to have a car, (c) my parents hardly have enough money to pay for necessities, and (d) my parents do not have enough money to pay for the extracurricular activities. Response categories range from almost never (5) to almost always (1). Mean scores of these four items were calculated to form the scale (α = .79). Higher scores meant stronger financial status. Finally, a dummy variable was created for family structure (1 = living with both parents and 0 = living in other arrangements). Descriptive statistics regarding the dependent and independent variables are displayed in Table 2.

Table 2.

Correlation Matrix and Descriptive Statistics for Study Variables.

No.	Measures	1	2	3	4	5	6	7	8	9	10	11	12	13	14
1	Delinquent Act 1	1.000
2	Delinquent Act 2	0.776***	1.000
3	Delinquent Act 3	0.746***	0.811***	1.000
4	Delinquent Act 4	0.712***	0.728***	0.786***	1.000
5	Delinquent Act 5	0.691***	0.697***	0.715***	0.711***	1.000
6	Delinquent Act 6	0.550***	0.586***	0.580***	0.569***	0.644***	1.000
7	Differential association	0.354***	0.356***	0.330***	0.355***	0.342***	0.339***	1.000
8	Definitions/norms	0.338***	0.324***	0.320***	0.313***	0.339***	0.323***	0.295***	1.000
9	Cost of differential reinforcement	−0.170***	−0.175***	−0.188***	−0.171***	−0.195***	−0.214***	−0.279***	−0.401***	1.000
10	Parental reaction	−0.187***	−0.193***	−0.181***	−0.184***	−0.192***	−0.205***	−0.307***	−0.129***	0.174***	1.000
11	Gender	0.309***	0.254***	0.256***	0.199***	0.293***	0.203***	0.084***	0.304***	−0.145***	−0.090***	1.000
12	Age	0.093***	0.088***	0.080***	0.059**	0.083***	0.081***	0.100***	0.112***	−0.108***	−0.102***	0.192***	1.000
13	Family structure	−0.004	−0.030	−0.006	−0.007	−0.020	−0.003	−0.030	0.010	0.022	0.004	−0.067**	0.018	1.000
14	Social class	0.018	0.019	0.023	0.026	0.017	0.034	0.004	0.021	−0.023	−0.019	−0.074**	−0.123**	0.146***	1.000
	M	0.244	0.242	0.271	0.319	0.233	0.170	3.630	13.719	21.243	11.674	0.497	13.362	0.921	2.593
	SD	0.429	0.428	0.444	0.466	0.423	0.376	4.875	4.234	4.426	1.078	0.500	0.686	0.269	0.604

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

Analyses

Method of Research

Testing the hypotheses proposed in this study within the context of the SSSL theory requires mediational analysis. Although mediational analysis can be performed by both multiple regression and structural equation modeling (SEM), in general, SEM is considered the preferred method for several reasons: (a) SEM can control for measurement error; (b) it allows for the exploration of the direct and indirect relationships between variables and constructs; (c) it provides information on the degree of the fit of the entire model; and (d) it is more robust and flexible than multiple regression. For example, it allows the researchers to include multiple outcome variables, multiple predictor variables, and multiple mediating variables (as in our case) into the model (Baron & Kenny, 1986; Morris & Higgins, 2010). For the reasons above, we used an SEM modeling framework for the analyses in this study.

Results

Table 2 presents the bivariate correlations for all the variables used in the analyses. Table 2 shows that the observed measures for delinquency (the delinquent acts) are all associated (r = .550 to .811). The table also shows that the measures of the social learning theory are all associated (r = −.401 to .356). Gender has correlations with delinquent acts (r = .199 to .309) and all social learning measures (r = −.145 to .304). Age is correlated with delinquent acts (r = .059 to .093) and all social learning measures (r = −.108 to .112). However, family structure and social class have no independent effect on the correlates of social learning theory and delinquency. The level of intercorrelations among the variables shown in Table 2 indicates little cause for concern about multicollinearity. Also, despite being components of the social learning process, it can be seen that there are no potential problems of multicollinearity among the social learning measures.

Table 3 presents the results of the measurement model, which includes confirmatory factor analysis (CFA) results for the latent variables representing social learning theory and delinquency. The CFA allows us to test how well the observed measures indicate the latent measures. In the present study, six delinquency measures represent a latent measure of delinquency, and differential association, definitions/norms, cost of differential reinforcement, and parental reaction to delinquency represent a latent measure of social learning theory.

Table 3.

Measurement Model for Delinquency and SLT Measures.

Measures	Standardized estimates
Delinquency
Delinquent Act 1	0.846***
Delinquent Act 2	0.884***
Delinquent Act 3	0.897***
Delinquent Act 4	0.851***
Delinquent Act 5	0.816***
Delinquent Act 6	0.678***
SLT
Differential association	0.607***
Definitions/norms	0.586***
Cost of differential reinforcement	−0.494***
Parental reaction to delinquency	−0.366***
Correlation between SLT and delinquency	0.607***
Fit statistics
χ² test of model fit	478.923***
Confirmatory fit index	.960
Root mean square error of approximation	.080
Standardized root mean of the residual	.040

Note. SLT = social learning theory.

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

As can be seen in Table 3, the fit between the hypothesized measurement models and the models generated by the data was good. The chi-square was statistically significant (χ² = 478.92, p < .001), and the confirmatory fit index (CFI) and standardized root mean of the residual (SRMR) are all suggestive of good (in terms of root mean square error of approximation [RMSEA] acceptable) model fit (Hu & Bentler, 1999; Kline, 2004). All the factor loadings for delinquency and all the factor loadings for social learning theory (except for parental reaction to delinquency) were above .50 and statistically significant. Despite its low factor loading, we retained the measure of parental reaction to delinquency due to its theoretical relevance. The correlation between the latent measures was large (.607), suggesting that social learning and delinquency are significantly related.

The mediational analysis was conducted with two models by applying an SEM framework, as shown in Table 4. The first model (referred to as the base model) included four exogenous predictor variables (gender, age, social class, and family structure) and endogenous delinquency construct. In the second SEM (referred to as the SSSL model), the social learning construct was added as a potential mediator of the structural factors. In both models, the fit statistics were used to determine how well each model fit the data. The chi-square test statistics were 322.92 and 648.48, respectively, and both were statistically significant. The other fit statistics, CFI (.97 and .95), RMSEA (.07 and .06), and SRMR (.02 and .03) were within their proper ranges, suggesting that the models fit the data. Figure 1 presents the general SSSL tested here.

Table 4.

Structural Equation Model Results: Base Model Versus SSSL Model.

Measures					Base model	SSSL model
Measures					Coefficient (SE)	Coefficient (SE)
Direct effects
Gender	→	Delinquency			0.215 (0.016)***	0.091 (0.017)***
Age	→	Delinquency			0.024 (0.012)*	−0.020 (0.011)
Family structure	→	Delinquency			−0.005 (0.029)	0.004 (0.028)
Social class	→	Delinquency			0.032 (0.013)*	0.007 (0.012)
Social learning	→	Delinquency			—	0.074 (0.004)***
Gender	→	Social learning			—	1.657 (0.160)***
Age	→	Social learning			—	0.601 (0.117)***
Family structure	→	Social learning			—	−0.144 (0.286)
Social class	→	Social learning			—	0.338 (0.128)**
Total indirect effects
Gender	→	Social learning	→	Delinquency	—	0.123 (0.013)***
Age	→	Social learning	→	Delinquency	—	0.044 (0.008)***
Family structure	→	Social learning	→	Delinquency	—	−0.010 (0.021)
Social class	→	Social learning	→	Delinquency	—	0.025 (0.009)**
Fit statistics
Confirmatory fit index					.970	.949
Root mean square error of approximation					.071	.066
Standardized root mean of the residual					.019	.036
χ² test of model fit					322.925***	648.483***
R²
Delinquency					0.094	0.378
Social learning theory						0.127

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

Figure 1.

Theoretical model for social structure, social learning, and adolescent delinquency.

The base model presents the direct effects of gender, age, social class, and family structure on adolescent delinquency without taking into account the effect of the social learning process. It can be seen that delinquency is significantly associated with all structural variables, except for family structure. That is, male (0.215), older (0.024), and upper-class (0.032) adolescents are more likely than their female, younger, and lower-class counterparts to be engaged in delinquency. These findings provide support for Hypotheses 2a (age), 3a (gender), and 4a (social class). As expected, we also find a nonsignificant direct effect of family structure on delinquency. The base model explained 9.4% of the variance in the delinquency measure.

The SSSL model in Table 4 shows that the social learning construct has a statistically significant direct effect on adolescent delinquency. Thus, this finding provides support for Hypothesis 1 that social learning theory accounts for a considerable amount of individual variation in delinquency.

The SSSL model also provides support for the mediating effects of the social learning component proposed in Akers’s SSSL theory. The introduction of the social learning construct into the analysis substantially reduced the effects of the structural factors on delinquency. Indeed, the effects of age and social class were rendered nonsignificant after the social learning construct is included into the analysis. On the contrary, although a substantial amount of gender effect on delinquency is mediated, it should be noted that gender, while mediated to some extent, retains significant and substantial effects on delinquency (0.091), unmediated by the social learning construct. According to Akers (2009), even modest reductions in the direct effects accord with the SSSL theory. Therefore, it can be claimed that these results provide modest support for Hypotheses 2b (age), 3b (gender), and 4b (social class).

Finally, as proposed in the SSSL theory, we should expect that the structural variables that have significant effects on delinquency to have significant effects on the social learning construct and that structural variables that have nonsignificant direct effects on delinquency to have nonsignificant direct effects on the social learning construct. That is what the results in the SSSL model in Table 4 reveal: All but family structure has significant direct effects on both the delinquency and the social learning constructs. The SSSL model provided a considerable improvement over the base model and explained 37.8% of the variance in the delinquency measure. The SSSL model also explained 12.7% of the variation in the social learning construct used here.

Discussions and Conclusion

The purposes of this study were (a) to examine how social structural effects and social learning process correlate with delinquency among Turkish adolescents and (b) to articulate to what extent Akers’s SSSL theory explains delinquency in Turkey, which is a different cultural context from Western countries in terms of family structure, level of collectivism or individualism, religion, belief systems, and norms. Accordingly, this study provides two main contributions to the literature. First, this study fills a gap in the criminological literature by showing that Akers’s SSSL theory is applicable to Turkish society, and thus strengthens the theory’s generalizability through covering all kinds of societies in explaining the underlying reasons of delinquency among adolescents. Second, as the Turkish literature on delinquency is not mature in comparison with Western literature, this study serves to expand the knowledge on delinquency, which has been developed insufficiently so far in Turkey.

The SEM models incorporating the predictor, mediating, and outcome variables permitted testing Hypotheses 2a to 4a that social structural factor has significant (and in the case of family structure nonsignificant) direct effects on delinquency, testing Hypothesis 1 that the social learning process has a significant direct effect on delinquency, and testing Hypotheses 2b to 4b that whatever effects social structural factors have on delinquency, they would be substantially mediated by the social learning process. These hypotheses were supported and the data fit theoretical expectations.

The findings of the study indicate that social learning process accounts for substantial portions of the variations in adolescent delinquency and mediates a substantial amount of the effects of the social structural factors on adolescent delinquency. Our analyses show that gender retained significant net effects on delinquency, which suggests that social learning may not mediate as much as it moderates the gender ratio (Baron & Kenny, 1986; Lee et al., 2004). The magnitude of the net effects was 57% weaker than the direct effects of gender on delinquency. Although this finding supported the theoretical expectation that social learning substantially mediates the relationship between gender and adolescent delinquency, it also raises questions about the ability of the social learning construct to mediate the gender effect on delinquency. In the case of social class and age, however, substantial portions of the direct effects were mediated. Regarding Akers’s (2009) argument that empirical support for the SSSL does not require that the social learning process accounts for all of the variation in the social structure factors, we could say that the results in this study lend modest support to the SSSL theory.

The results suggested that family structure was not associated with delinquency and therefore was not mediated by the social learning process. This suggests that living with arrangements other than with two parents might not explain the likeliness in engaging in delinquency in Turkey, possibly due to its more collectivist culture. Instead of a nuclear family including only parents, the context of family in Turkey covers relatives, such as uncles and aunts, grandfathers, and grandmothers who help single mothers or single fathers to increase social control over adolescents and provide more care. Thus, the disadvantages of single parenthood or broken families might easily be removed by the presence of strong traditions of family in Turkey. Also, Akers (2009) indicates that “differences in the societal or group rates of criminal behavior are a function of the extent to which cultural traditions, norms, and social control systems provide socialization, learning environments, and immediate situations conducive to conformity or deviance” (p. 323). Therefore, it can be speculated that the Turkish lifestyle including not only families but also neighbors and schools implementing strong social control over adolescents might provide the more immediate contexts that discourage the criminal behavior and encourage the development of self-control. In addition, regarding family structure in our sample, 92% of the adolescents live in two-parent households, which clearly indicates the need for additional analyses of samples including respondents with more diverse experiences. Future studies with more adequate measures of family structure are clearly needed.

In addition to testing the applicability of the SSSL theory in the Turkish context, another, relatively of secondary importance, purpose of this study was to address whether the SSSL theory would produce different results in the Turkish context compared with its tests in Western countries. Although Turkey has a different and mostly unknown culture to Western societies with a more collectivist nature normalizing the implementation of strong social control over adolescents, we found that Turkish society is more like Western societies in encouraging and discouraging adolescent delinquency. Although we found similar results in terms of social learning process, gender, age, and social class with the previous studies conducted in Western countries (Akers & Lee, 1999; Lee et al., 2004; Piquero et al., 2005), our results also showed the idiosyncratic nature of family structure among Turkish adolescents. This finding would be one of the contributions of this study to the literature, with some caveats, specific to the Turkish context.

Although this research holds several merits, some limitations should be mentioned. First of all, the results were based on students’ perceptions and were assumed to be dominantly related to the school climate. Therefore, they might cause validity concerns (McEvoy & Welker, 2000). A social desirability problem may also exist in the study due to school climate as a self-report system cannot be verified by other resources and crime statistics (Wang et al., 2010). In addition, although the survey was conducted in one of the most crowded provinces of Istanbul, results might not be generalizable to the entire population of Turkish adolescents.

In addition to all that, there are the following limitations resulting from the data set. First, the data we used in this study are cross-sectional data, which limit the causal inferences we can make about the relationship between exposure variables (the social learning construct and social structure variables) and delinquency. Also, due to our cross-sectional data, we could not explicitly determine whether the observed relationship between one’s own delinquency and peer delinquency is due to peer influence or self-selection into delinquent peer groups based on behavioral dispositions. Although Akers and Jensen (2006) propose that the sequence of events, in which deviant associations precede the onset of delinquent behavior, will occur more frequently than the sequence of events, in which the onset of delinquency precedes the making of delinquent associations, our expectation is that future longitudinal studies would allow determination of whether peer influence or self-selection is the primary mechanism responsible for similarities in adolescent delinquent behavior.

Second, the models tested here did not incorporate two dimensions of the social structure identified by Akers (2009), namely, differential social organization and theoretically defined structural variables, due to data restrictions. Moreover, the social learning theory measures used here did not capture imitation, also due to data restrictions. However, the measures capturing differential association, definitions/norms, and differential reinforcement are a proper set of measures for a partial test of Akers’s SSSL theory.

Future studies should address how the social learning processes occur in different ecological environments, such as schools and neighborhoods. A well-designed survey measuring all the four dimensions of the social structure specified in the SSSL theory, all kinds of differential associations, including parents, social networks, relatives, and siblings, and all kinds of imitation processes, such as the media, siblings, peers, and family, can present more valid results. Similarly, questions measuring less serious deviant behaviors, including cheating and violating social rules, such as obeying traffic signs or fastening seat belt can be administered into a new survey to assess the validity of the SSSL theory.

To conclude, although not a complete test of the SSSL theory, this study stood out as a preliminary for testing the theory in Turkey. Even with limited measures and some caveats, we found support for Akers’s SSSL theory. As policy implications, considering the findings of this study, it can be offered that following Akers’s (2009) Iowa study, a role model among athletes or pop stars can be introduced to students as good examples for them to imitate. Also, video clips showing the hazardous effects of substance usage or similar deviant acts on people’s health and social life can increase opportunities for positive imitations. The Big Brother/Big Sister chosen from successful students might be another approach to show and provide role models (Williams & McShane, 2010). Another implication would be to arrange parent conferences regularly to inform them about peer selection processes and parental control applications over their children. As the results of this study provided support for the SSSL theory, similar prevention and treatment programs developed under the light of this theory and conducted in the United States can be replicated in Turkey. For example, Highfields and Essexfield’s prevention programs, which were designed as “Guided Group Interaction Sessions” that aimed at increasing the level of norms favorable to conformity (Nicholson & Higgins, 2017), may be useful for adolescents in Turkey.

Finally, as argued by Yuksek and Solakoglu (2016), school is the primary formal organization that influences adolescents’ life courses. It is the primary context for social interactions, the cultivation of interpersonal skills, the formation of peer groups, self-expression, and the development of self. Adolescents in our sample spend more than 40 hr a week at school. In this regard, whether they like their school or not determines whether they will use this considerable period of time effectively regarding societal rules or not. Considering differential association with peers as the most consistent component of social learning theory predicting delinquency (Akers & Jensen, 2006; Haynie, 2001), schools might provide the most favorable learning environment where adolescents “associate differently with peers who are deviant or tolerant of deviance, learn definitions favorable to delinquent behavior, are exposed to deviant models that reinforce delinquency, then initiate or increase involvement in that behavior” (Akers, 2012, p. 72). Therefore, it is important for parents, teachers, and school administrators to be aware of the fact that strengthening adolescents’ bonds to their school should be an educational priority to prevent delinquency. The reason is that students who like (or strongly attached to) their school are expected to be more active in their curricular and extracurricular activities and be more focused in their educational objectives. In that case, it is expected that they will be more eager to stay away from deviant peers (deviant role models) and thus less exposed to definitions favorable to law violation. Following this argument, we suggest that schools can play an important role in creating safer environments for adolescents by ensuring desirable, attractive, and effective curricular and extracurricular activities, which have the potential for (a) encouraging adolescents to participate and get involved with these activities, (b) contributing adolescents’ feelings of attachment to their school, and (c) providing better social learning processes for them.

Footnotes

Authors’ Note

The authors accept that this article has not been published elsewhere and that it has not been submitted simultaneously for publication elsewhere.

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.

ORCID iD

Durmus A. Yuksek

Notes

References

Akers

R. L.

(1992). Drugs, alcohol, and society: Social structure, process, and policy. Wadsworth.

Akers

R. L.

(2009). Social learning and social structure: A general theory of crime and deviance. Transaction Publishers.

Akers

R. L.

(2012). Criminological theories: Introduction and evaluation. Routledge.

Akers

R. L.

Jensen

G. F.

(2006). The empirical status of social learning theory of crime and deviance: The past, present, and future. In Cullen

F. T.

Wright

Blevins

(Eds.), Taking stock: The status of criminological theory (Vol. 1, pp. 37–76). Transaction Publishers.

Akers

R. L.

Lee

(1999). Age, social learning, and social bonding in adolescent substance use. Deviant Behavior, 20, 1–25.

Baron

R. M.

Kenny

D. A.

(1986). The moderator–mediator variable distinction in social psychological research: Conceptual, strategic, and statistical considerations. Journal of Personality and Social Psychology, 51, 1173–1187.

Bartusch

D. J.

Matsueda

R. L.

(1996). Gender, reflected appraisals, and labeling: A cross-group test of an interactionist theory of delinquency. Social Forces, 75, 145–176.

Bendixen

Endresen

I. M.

Olweus

(2003). Variety and frequency scales of antisocial involvement: Which one is better? Legal and Criminological Psychology, 8, 135–150.

Burgess

R. L.

Akers

R. L.

(1966). A differential association-reinforcement theory of criminal behavior. Social Problems, 14, 128–147.

10.

Canter

R. J.

(1982). Family correlates of male and female delinquency. Criminology, 20, 149–167.

11.

Fagan

A. A.

Van Horn

M. L.

Hawkins

J. D.

Arthur

M. W.

(2007). Gender similarities and differences in the association between risk and protective factors and self-reported serious delinquency. Prevention Science, 8, 115–124.

12.

Gottfredson

M. R.

Hirschi

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

13.

Hagan

Gillis

A. R.

Simpson

(1985). The class structure of gender and delinquency: Toward a power-control theory of common delinquent behavior. American Journal of Sociology, 90, 1151–1178.

14.

Haynie

D. L.

(2001). Delinquent peers revisited: Does network structure matter? American Journal of Sociology, 106, 1013–1057.

15.

L. T.

Bentler

P. M.

(1999). Cutoff criteria for fit indexes in covariance structure analysis: Conventional criteria versus new alternatives. Structural Equation Modeling, 6, 1–55.

16.

Hwang

Akers

R. L.

(2006). Parental and peer influences on adolescent drug use in Korea. Asian Journal of Criminology, 1, 51–69.

17.

Junger-Tas

Terlouw

G. J.

Klein

M. W.

(1994). Delinquent behavior among young people in the western world: First results of the international self-report delinquency study. Kugler Publications.

18.

Karakus

McGarrell

E. F.

Basibuyuk

(2010). Fear of crime among citizens of Turkey. Journal of Criminal Justice, 38, 174–184.

19.

Kline

R. B.

(2004). Principle and practice of structural equation modeling. Guildford Press.

20.

Krivo

L. J.

Peterson

R. D.

(1996). Extremely disadvantaged neighborhoods and urban crime. Social Forces, 75, 619–650.

21.

Krohn

M. D.

Skinner

W. F.

Massey

J. L.

Akers

R. L.

(1985). Social learning theory and adolescent cigarette smoking: A longitudinal study. Social Problems, 32, 455–473.

22.

Lee

Akers

R. L.

Borg

M. J.

(2004). Social learning and structural factors in adolescent substance use. Western Criminological Review, 5, 17–34.

23.

Massey

J. L.

Krohn

M. D.

(1986). A longitudinal examination of an integrated social process model of deviant behavior. Social Forces, 65, 106–134.

24.

McCord

(1991). Family relationships, juvenile delinquency, and adult criminality. Criminology, 29, 397–418.

25.

McEvoy

Welker

(2000). Antisocial behavior, academic failure, and social climate: A critical review. Journal of Emotional and Behavioral Disorders, 8, 130–140.

26.

Meneses

R. A.

Akers

R. L.

(2011). A comparison of four general theories of crime and deviance: Marijuana use among American and Bolivian university students. International Criminal Justice Review, 21, 333–352.

27.

Miller

H. V.

Jennings

W. G.

Alvarez-Rivera

L. L.

Miller

J. M.

(2008). Explaining substance use among Puerto Rican adolescents: A partial test of social learning theory. Journal of Drug Issues, 38, 261–283.

28.

Minor

(1980). The neutralization of criminal offense. Criminology, 18, 103–120.

29.

Morris

R. G.

Higgins

G. E.

(2010). Criminological theory in the digital age: The case of social learning theory and digital piracy. Journal of Criminal Justice, 38, 470–480.

30.

Nicholson

Higgins

G. E.

(2017). Social structure social learning theory: Preventing crime and violence. In Teasdale

Bradley

M. S.

(Eds.), Preventing crime and violence (pp. 21–31). Springer.

31.

Ozbay

(2006). Class and self-reported juvenile delinquency: Evidence from Turkey. Journal of Criminal Justice, 34, 237–249.

32.

Ozbay

Koksoy

(2009). Is low self-control associated with violence among youths in Turkey? International Journal of Offender Therapy and Comparative Criminology, 53, 145–167.

33.

Ozbay

Ozcan

Y. Z.

(2006a). Classic strain theory and gender: The case of Turkey. International Journal of Offender Therapy and Comparative Criminology, 50, 21–38.

34.

Ozbay

Ozcan

Y. Z.

(2006b). A test of Hirschi’s social bonding theory: Juvenile delinquency in the high schools of Ankara, Turkey. International Journal of Offender Therapy and Comparative Criminology, 50, 711–726.

35.

Ozbay

Ozcan

Y. Z.

(2008). A test of Hirschi’s social bonding theory: A comparison of male and female delinquency. International Journal of Offender Therapy and Comparative Criminology, 52, 134–157.

36.

Ozdemir

Vazsonyi

A. T.

Cok

(2013). Parenting processes and aggression: The role of self-control among Turkish adolescents. Journal of Adolescence, 36, 65–77.

37.

Piquero

N. L.

Gover

A. R.

MacDonald

J. R.

Piquero

A. R.

(2005). The influence of delinquent peers on delinquency: Does gender matter? Youth & Society, 36, 251–275.

38.

Pratt

T. C.

Cullen

F. T.

Sellers

C. S.

Winfree

L. T.

Madensen

T. D.

Daigle

L. E.

. . .Gau

J. M.

(2010). The empirical status of social learning theory: A meta-analysis. Justice Quarterly, 27, 765–802.

39.

Sellers

C. S.

Cochran

J. K.

Winfree

L. T.

(2003). Social learning theory and courtship violence: An empirical test. In Akers

R. L.

Jensen

G. F.

(Eds.), Social learning theory and the explanation of crime (Vol. 1, pp. 109–127). Transaction Publishers.

40.

Simons

R. L.

Simons

L. G.

Wallace

L. E.

(2004). Families, delinquency, and crime: Linking society’s most basic institution to antisocial behavior. Roxbury Publishing.

41.

Sutherland

E. H.

(1947). Principles of criminology. J. B. Lippincott.

42.

Sweeten

(2012). Scaling criminal offending. Journal of Quantitative Criminology, 28, 533–557.

43.

Sykes

G. M.

Matza

(1957). Techniques of neutralization: A theory of delinquency. American Sociological Review, 22, 664–670.

44.

Thornberry

T. P.

(1987). Towards an interactional theory of delinquency. Criminology, 25, 863–892.

45.

Tittle

C. R.

Antonaccio

Botchkovar

(2012). Social learning, reinforcement, and crime: Evidence from three European cities. Social Forces, 90, 863–890.

46.

TurkStat. (2015, December 12). Turkish Statistical Institute. http://www.turkstat.gov.tr/VeriBilgi.do?alt_id=61

47.

Unal

Cukur

C. S.

(2011). Okula Baglilik, Okulda Uygulanan Disiplin Teknikleri ve Okulda Magdurlugun Lise Ogrencilerinin Sapkin Davranislari Uzerine Etkisi [The effects of school bonds, discipline techniques in school and victimization on delinquency of high school students]. Educational Sciences: Theory and Practice, 11, 547–570.

48.

Wang

M. T.

Selman

R. L.

Dishion

T. J.

Stormshak

E. A.

(2010). A Tobit regression analysis of the covariation between middle school students’ perceived school climate and behavioral problems. Journal of Research on Adolescence, 20, 274–286.

49.

Wasti

S. A.

(1998). Cultural barriers in the transferability of Japanese and American human resources practices to developing countries: The Turkish case. International Journal of Human Resource Management, 9, 608–631.

50.

Williams

F. P.

McShane

M. D.

(2010). Criminology theories. Prentice Hall.

51.

Winfree

L. T.

Griffiths

C. T.

(1983). Social learning and adolescent marijuana use: A trend of study of deviant behavior in a rural middle school. Rural Sociology, 48, 219–239.

52.

Yuksek

D. A.

Solakoglu

(2016). The relative influence of parental attachment, peer attachment, school attachment, and school alienation on delinquency among high school students in Turkey. Deviant Behavior, 37, 723–747.

53.

Yun

Kim

(2015). Illicit drug use among South Korean offenders: Assessing the generality of social learning theory. International Journal of Offender Therapy and Comparative Criminology, 59, 1166–1187.