The Longitudinal Relationships Between Cyber Delinquency,Aggression,and Offline Delinquency: An Autoregressive Cross-Lagged Model

Abstract

Cyber delinquency in adolescence is a particular area of concern for psychologists owing to its association with several mental health issues and its potential link with offline delinquency. This longitudinal study examined the stability of changes and directions of influence in adolescents’ cyber delinquency, aggression, and offline delinquency across a 4-year period. Our sample consisted of 2,280 adolescents who participated in the Korean Children and Youth Panel Survey conducted by the National Youth Policy Institute in Korea, from 2011 to 2015. Using autoregressive cross-lagged modeling, we found that changes in cyber delinquency, aggression, and offline delinquency were stable over time. There were also multiple cross-lagged effects from early to middle adolescence: cyber delinquency influenced both offline delinquency and aggression, and aggression influenced both cyber and offline delinquency. However, from middle to late adolescence, the only significant effect was that of cyber delinquency on offline delinquency.

Keywords

aggression problem behavior media middle school

The current adolescent generation has considerable experience with both the functionality and dysfunction of cyberspace. The functionality of the Internet provides adolescents with a recreational environment in which they can communicate electronically with their peers, establish new relationships with strangers, and acquire further information and knowledge through Internet searches and the use of online learning platforms. However, various types of crimes also occur on the Internet, such as the distribution of harmful or false information, illegal access of online information, destruction of information systems, and invasion of privacy. Because cybercrime is anonymous and faceless, it is expected to increase in the future. For today’s adolescents, who have been using the Internet since childhood, the Internet is both a necessary tool, by which they socialize and gain access to new information, and a potential trigger for their delinquency, by allowing them to easily engage in illegal activities (Dombrowski, Gischlar, & Durst, 2007). Cyber delinquency is a particular area of concern for psychologists because it can lead to offline delinquency and is associated with several mental health issues in adolescents.

An understanding of the concept of delinquency is important for understanding cyber delinquency. Delinquency is a comprehensive term used to describe crimes committed by adolescents. Criminologists use “delinquency,” rather than “crime,” to describe illegal activity in adolescents (J. S. Lee & Ahn, 2005). In a broad sense, cyber delinquency is defined as all criminal acts and delinquency committed by adolescents in and around cyberspace. The scope of cyber delinquency is defined differently in different cultures, but previous studies have identified the most common cyber delinquency as hacking, virus infringement, the use of false residential registration numbers, searches for pornography, the use of obscene or violent language in online chat rooms, and software piracy (Buzzell, Foss, & Middleston, 2006; Han, 2001; Y. S. Hong & Kim, 2011; Jo & Yang, 2001; Rokven, Weijters, Beerthuizen, & van den Laan, 2018; Siegel & Welsh, 2018).

Cyber Delinquency and Aggression

Previous studies have demonstrated that aggression is a contributor to most acts of cyber delinquency (K.-E. Kim & Yoon, 2012; D. Y. Lee, Park, & Lee, 2006; Oh, 2014). Aggression generally refers to intentional behavior intended to inflict harm on others. However, the concept of aggression includes not only observable behavior but also anger, which is an emotional state that can lead to aggressive behavior (Orpinas & Frankowski, 2001). Aggression can be classified into physical aggression and relational aggression (Orpinas & Frankowski, 2001). Physical aggression refers to behavior intended to harm others through physical acts and verbal attacks or insults, whereas relational aggression refers to nonphysical behaviors intended to damage peer relationships and social status (Crick & Grotpeter, 1995). In cyberspace, users can maintain mutual anonymity, which makes it easier for them to express relational aggression through the dissemination of false facts and the use of harsh language (Holtz & Appel, 2011; Peter, Valkenburg, & Schouten, 2005). However, physical aggression is currently receiving more attention from scholars due to the potential link between aggression and the use of violent computer games by children and adolescents. Today, most computer games contain violent content (Adachi & Willoughby, 2011; Deyreh, 2011; Fischer, Kastenmüller, & Greitemeyer, 2010; Zhen, Xie, Zhang, Wang, & Li, 2011). According to Anderson et al. (2008), more than 90% of computer games available to people over 10 years of age consisted of violent content. Prior research (Anderson & Bushman, 2001; Fischer et al., 2010; Giumetti & Markey, 2007) suggested such violent computer games make users more aggressive in everyday life. However, children and youth continue to play violent video games, and evidence indicates that the younger the computer gamers are, the more likely they are to be affected by the aggressiveness of computer games (Anderson et al., 2008; Zhen et al., 2011). Consistent with the current focus on physical aggression in children and adolescents, we examined the linkages between cyber delinquency and physical aggression in this study.

Causes of Aggression Among Adolescents

Although the Internet allows adolescents to engage in delinquency in an anonymous and faceless environment, without exposing themselves, these facts alone do not explain the root causes of their aggressive online behavior. Several psychological theories may help to explain the causes of cyber delinquency. First, the General Strain Theory (Agnew, 1992) posits that delinquency is caused by everyday strain. According to Agnew’s structural framework, when adolescents experience stress, they experience negative emotions such as depression, despair, and anger. In order to resolve these negative feelings, adolescents commit delinquency.

Second, the Problem Behavior Theory (Jessor & Jessor, 1977) emphasizes that personality characteristics such as emotional vulnerability predict aggression and other problem behaviors. According to this theory, the personality system, perceived environmental system, and behavioral system are all interrelated and influence one another. Adolescents who are emotionally vulnerable and cannot control the expression of their aggression direct their aggression outward, thereby allowing deviant and problem behaviors to be expressed (Goodman, Brogan, Lynch, & Fielding, 1993).

Third, Anderson and Bushman (2001) proposed the General Aggression Model, in which aspects of the environment, such as engaging in computer games, activate cognitive structures that trigger aggressive thinking and eventually increase the frequency of attack behaviors. According to this theory, aggression is a personality trait or tendency that is established relatively early in life and is genetically affected (Lahey, Waldman, & McBumett, 1997). It is also relatively stable over time, with aggression in infancy predicting later aggression and delinquency among adolescents (Connor, 2002; Rappaport & Thomas, 2004). Due to the ability to use aggression to predict future maladjustment, many scholars are interested in studying aggression (Loeber & Stouthamer-Loeber, 1987).

Fourth, Gottfredson’s Theory of Circumscription and Compromise explains the relationship between aggression and computer use. According to this theory, a person with a given personality prefers experiences that fit his or her inclinations, and information from the experiences serves to strengthen his or her preferences (Friedman & Schustack, 2009; Swann & Read, 1981). Gottfredson (2002) proposed a cyclical model, in which personality traits influence the specific experiences that individuals pursue, experiences that in turn reinforce their personality traits. Gottfredson (2002) argued that one member’s unique experience, rather than a shared experience, has a greater impact on the development of personal characteristics, and the cyclical relationship between traits and experiences makes individual traits more pronounced with age. Applying Gottfredson’s theory to aggression and cyber delinquency suggests that highly aggressive children and adolescents seek to use computers that contain aggressive activity, which in turn increases their aggressiveness.

Despite these theories, the relationship between aggression and cyber delinquency is complicated, and it is unclear whether aggression leads to, is exacerbated by, or results from cyber delinquency. Some studies have found that cyber delinquency was a consequence of psychopathological symptoms (Davis, 2001; C. S. Ko, 2012). However, in other studies, cyber delinquency preceded significant psychiatric, functional, psychosocial, and health-related issues (Kuss, Griffiths, Karila, & Billieus, 2014; Weinstein & Lejoyeux, 2010). Additional studies have suggested a causal link between cyber delinquency and mental health issues, with cyber delinquency inducing or exacerbating symptoms of psychopathology including aggression, depression, social anxiety, distress, and self-injurious behavior such as substance use (Dong, Lu, Ahou, & Zhao, 2011; C. H. Ko, Yen, Yen, Chen, & Chen, 2012; Rashid, Muhammad, Muhammad, Altaf, & Mehvish, 2014; Sun et al., 2012).

In examining the causal relationship between cyber delinquency and aggression, it is necessary to consider the influence of time. As many scholars have noted (e.g., Durkin & Barber, 2002; Kutner & Olson, 2008; Olson, 2004; Savage & Yancey, 2008; Sherry, 2006; Unsworth, Devilly, & Ward, 2007), computer use does not immediately lead to increased aggression in adolescents. Recent studies have examined the effects of computer use on aggression within specific time intervals (Möller & Krahé, 2009; Shibuya, Sakamoto, Ihori, & Yukawa, 2008; Williams & Skoric, 2005). However, these studies examined very short time intervals using a longitudinal path model for which the data were limited to two time points.

Although the evidence regarding the relationship between cyber delinquency and aggression is contradictory, the research base offers much clearer evidence regarding the relationship between aggression and offline delinquency. Numerous longitudinal studies have demonstrated that aggression predicts offline delinquency in adolescents (Jessor, Van Den Bos, Vanderryn, Costa, & Turbin, 1995). These studies identified aggression as having a developmental pattern; it is developed early in childhood and is linked to delinquency, drug use, and aggression in adolescence (Broidy et al., 2003; Brook, Whiteman, Finch, & Cohen, 1996; Farrell, Kung, White, & Valois, 2000). As these studies suggest, children and youth must be followed for multiple years to assess the time-cumulative effects of cyber delinquency. To exclude other confounding factors, researchers should employ panel surveys that collect data on participants at relatively frequent intervals.

The Current Study

Based on the studies suggesting that aggression is observed relatively early in life (Lahey et al., 1997) and aggression in early adolescence predicts aggression and delinquency in late adolescence (Connor, 2002; Rappaport & Thomas, 2004), this study examined the relationships between cyber delinquency, aggression, and offline delinquency using longitudinal data collected from adolescents over a 4-year period. Utilizing autoregressive cross-lagged modeling (ACLM), the purpose of this study was to examine the longitudinal effects of cyber delinquency, aggression, and offline delinquency in early (14 years), middle (16 years), and late (18 years) adolescence.

Method

Participants

The study analyzed longitudinal data from the Korean Children and Youth Panel Survey (KCYPS) conducted by the National Youth Policy Institute in Korea. The KCYPS is a national longitudinal panel study that collects data regarding school and family life, cultural activities, use of mass media, and problematic behavior in Korean adolescents. The sample was chosen using stratified multistage cluster sampling based on the 2009 school statistics published by the Ministry of Education in Korea. The KCYPS was initiated in 2010, and data are collected between September and December of each year.

The data for this study were collected for 3 years, from 2011 to 2015. The KCYPS measured cyber delinquency and offline delinquency annually, except for 2010, and aggression in 2011, 2012, 2013, and 2015. Therefore, this study analyzed data from the second (2011, Time 1), fourth (2013, Time 2), and sixth (2015, Time 3) years of the KCYPS, as cyber delinquency, aggression, and offline delinquency were all measured for these 3 years.

Survey participants were selected only if they had used the Internet during each of the 3 years for which data were analyzed. The sample consisted of 2,280 adolescents (1,128 girls and 1,152 boys). The participants were in the eighth grade (14 years) at Time 1, tenth grade (16 years) at Time 2, and 12th grade (18 years) at Time 3. Their average annual household income was US$39,669.07 (SD = US$2,481.00) at Time 1, US$41,055.95 (SD = US$2,687.020) at Time 2, and US$42,340.77 (SD = US$2,564.172) at Time 3.

Measures

Cyber delinquency

In this study, cyber delinquency was defined as general crime and delinquent acts committed by adolescents, mainly in cyberspace. Delinquency was measured using the following six items: posting false information on a chat-room message board, software piracy, using another person’s identity or residential registration number without permission, misrepresenting one’s sex or age in a chat-room, hacking another person’s computer or website, and swearing or using violent language on chat-room message boards. Participants were asked to indicate whether they had engaged in each of these six behaviors in the last 6 months (1 = yes, 0 = no). The total score for these six items represented the level of cyber delinquency.

Aggression

Prior research has shown that physical aggression in cyberspace is more likely to be linked to aggression in everyday life (Anderson & Bushman, 2001; Fischer et al., 2010; Giumetti & Markey, 2007). Therefore, this study limited its focus to physical aggression and participants were asked to respond to the following five items based on their behavior during the last 6 months; “I hit other people if I get very angry,” “If people hit me, I hit them back,” “I fight more frequently than other people do,” “If I get angry, I sometimes feel compelled to throw things,” and “Sometimes, I cannot help hitting others.” Responses were rated on a 5-point scale ranging from very unlikely (1) to very likely (5). The five items were considered somewhat reliable as a scale, with a Cronbach’s alpha that exceeded .70 for each year of data collection: for Time 1, α = .73, for Time 2, α = .72, and for Time 3, α = .75.

Offline delinquency

Offline delinquency was defined as illegal or immoral behavior in adolescents. Adolescent delinquency is generally divided into the following types: status, learning, rebellious, violent, and sexual (Grotpeter & Crick, 1996). In this study, offline delinquency was measured using status, violent, and sexual delinquency. Behavior pertaining to status delinquency included smoking, drinking, and running away from home. Behavior pertaining to violent delinquency included hitting others violently, group fighting, buying things using other people’s money, stealing other people’s money or possessions, teasing or mocking others severely, threatening others, and bullying. Behavior pertaining to sexual delinquency included engaging in sexual intercourse and sexual violence or harassment. Participants were asked to indicate whether they had engaged in each of these 14 behaviors in the last 6 months (1 = yes, 0 = no), and the total score represented the level of offline delinquency.

Statistical Analysis

ACLM

ACLM was performed to assess the longitudinal relationships between cyber delinquency, aggression, and offline delinquency. In ACLM, the scores at time t − 1 (an earlier time point) were used to explain the scores at time t (a later time point); therefore, parameter values at Time 2 depended on those observed at Time 1 (Bast & Reitsma, 1997; Curran & Bollen, 2001). There were two types of longitudinal relationships in the model: the autoregressive and cross-lagged relationships. The parameter values for autoregressive relationships were determined by regressing a measure at a later time point (t), onto the same measure at an earlier time point (t − 1). The parameter values for cross-lagged relationships were determined by regressing one construct at a later time point (t) onto another construct at an earlier time point (t − 1). It is important to note that the cross-lagged relationships controlled for the autoregressive relationships, that is, the influences of the constructs at the earlier time point (t − 1) on values for the same construct at a later time point (t) (S. Hong, Yoo, You, & Wu, 2010). As a result, the outcome construct at a later time point (t) was predicted by both the same construct and the other construct at an earlier time point (t − 1). We performed ACLM via structural equation modeling (SEM) using AMOS (Version 24.0).

Assessment of invariance across time

To evaluate autoregressive and cross-lagged effects, we explored the assumptions of configural invariance and invariance of error covariance. We compared the model fit indices of 11 nested models, which were organized by adding invariance constraints hierarchically, as shown in Figure 1.

Figure 1.

Autoregressive cross-lagged model.

The baseline model did not include invariance constraints. In Figure 1, configural invariance and invariance of error covariance across time are represented using letters next to the paths. Paths A, B, and C represented the configural invariance of the autoregressive coefficients. Paths D, E, F, and G represented the configural invariance of the cross-lagged coefficients. For example, we set the autoregressive coefficients on cyber delinquency_T1 → cyber delinquency_T2 and cyber delinquency_T2 → cyber delinquency_T3 equal to A and restricted the cross-lagged coefficients on cyber delinquency_T1 → aggression_T2 and cyber delinquency_T2 → aggression_T3 equal to D. Once configural invariance was achieved, the identity of the autoregressive and cross-lagged effects over time was revealed (J. H. Kim, Kim, & Hong, 2009). We developed Models 1, 2, and 3 by entering A, B, and C, in order, to examine the configural invariance of the autoregressive effects. We then developed Models 4, 5, 6, and 7 by entering D, E, F, and G to test the configural invariance of the cross-lagged effects between the variables over time. Paths H, I, and J were the correlation paths (i.e., d1 ↔ d4; d2 ↔ d5, d3 ↔ d6, d4 ↔ d7, d5 ↔ d8, d6 ↔ d9, d1 ↔ d7, d2 ↔ d8; d3 ↔ d9) between the structural errors (d1, d2, d3, . . ., d9) across time and represented the invariance of error covariance. The invariance of error covariance was established to exclude causal influence over time when the reciprocal relationships between cyber delinquency, aggression, and offline delinquency were examined (J. H. Kim et al., 2009). We developed Models 8, 9, and 10 by adding H, I, and J to test the invariance of error covariance. We then compared the model fit indices of all 11 models sequentially to choose the final model. The χ² value is sensitive to sample size, inflates Type 1 error, and applies strict standards for rejecting the null hypothesis (Bollen, 1989; Jöoreskog & Sörbom, 1996). Therefore, we used three indices to compare model fit (S. Hong, 2000; J. H. Kim et al., 2009): the comparative fit index (CFI), normed fit index (NFI), and root mean square error of approximation (RMSEA). CFI and NFI values of more than .90 demonstrate an adequate fit (Bentler, 1990; Tucker & Lewis, 1973). In addition, RMSEA values of less than .05 demonstrate acceptable and adequate fit, respectively (Browne & Cudeck, 1993). To satisfy the assumptions of configural and error covariance invariance, the model with more invariance constraints should be improved with better or equal fit indices (Cheung & Rensvold, 2002).

Results

Descriptive Statistics and Correlations

The descriptive statistics for cyber delinquency, aggression, and offline delinquency are shown in Table 1. Mean differences between time points were statistically significant. As the age of the participants increased, cyber delinquency and aggression decreased, but offline delinquency increased. In addition, the correlations between cyber delinquency, aggression, and offline delinquency are shown in Table 2.

Table 1.

Descriptive Statistics for the Main Variables.

	Time	$\bar{X}$	SD	Min	Max	Range	F
Cyber delinquency						0-6	364.735***
	1	0.84	1.12	0	6
	2	0.27	0.64	0	6
	3	0.16	0.47	0	5
Aggression						1-5	105.032***
	1	2.13	0.57	0	10
	2	1.97	0.52	0	7
	3	1.93	0.53	0	9
Offline delinquency						0-14	42.473***
	1	0.33	0.91	1	4.00
	2	0.34	0.81	1	3.83
	3	0.50	0.88	1	4.00

***

p < .001.

Table 2.

Correlations Among the Main Variables.

Variable	1	2	3	4	5	6	7	8
1. Cyber delinquency_T1
2. Cyber delinquency_T2	.263***
3. Cyber delinquency_T3	.235***	.359***
4. Aggression_T1	.177***	.094***	.021
5. Aggression_T2	.108***	.070***	.023	.379***
6. Aggression_T3	.044	.024	.048	.307***	.465***
7. Offline delinquency_T1	.262***	.109***	.095***	.168***	.136***	.083***
8. Offline delinquency_T2	.159***	.170***	.116***	.115***	.151***	.075***	.351***
9. Offline delinquency_T3	.158***	.132***	.213***	.080***	.103***	.152***	.316***	.429***

Note. A suffix of “_T1” indicates Time 1, “_T2” indicates Time 2, and “_T3” indicates Time 3.

***

p < .001.

Nested Model Comparisons

The results of comparing the 11 nested models are shown in Table 3. All models estimated the longitudinal relationships between cyber delinquency, aggression, and offline delinquency without potentially confounding associations over time. Relative to the other models, the fit indices were the highest for the baseline model, which was selected as the final model for the study.

Table 3.

Fit Indices for the Nested Model Comparisons.

Model	χ²	df	NFI	CFI	RMSEA (95% CI)
Baseline	199.598	13	.922	.926	.078 [.069, .088]
Model 1	233.866	14	.909	.913	.082 [.073, .091]
Model 2	256.956	15	.900	.904	.083 [.074, .092]
Model 3	290.579	16	.887	.891	.085 [.077, .094]
Model 4	292.339	17	.886	.892	.083 [.075, .092]
Model 5	293.972	18	.885	.891	.081 [.073, .089]
Model 6	296.017	19	.885	.890	.079 [.071, .087]
Model 7	296.286	20	.885	.890	.077 [.069, .085]
Model 8	352.691	22	.863	.869	.080 [.073, .087]
Model 9	360.046	24	.860	.867	.077 [.070, .084]
Model 10	433.157	26	.831	.838	.082 [.075, .088]

Note. NFI = normed fit index; CFI = confirmatory fit index; RMSEA = root mean square error of approximation; CI = confidence interval.

Structural Regression Analysis

The estimated structural regression coefficients for the final model are shown in Table 4. Cyber delinquency at Times 1 and 2 positively influenced cyber delinquency at Times 2 and 3, respectively, indicating that the autoregressive effect of cyber delinquency was significant across the 4 years examined. In addition, aggression at Times 1 and 2 positively influenced aggression at Times 2 and 3, and offline delinquency at Times 1 and 2 positively influenced offline delinquency at Times 2 and 3; therefore, the autoregressive effects of aggression and offline delinquency were significant over the 4 years examined.

Table 4.

Structural Regression Analysis Results for the Final Model.

Parameter	Estimate	SE	CR	p
Cyber delinquency_T1 → Cyber delinquency_T2	.146	.012	11.887	.000
Cyber delinquency_T2 → Cyber delinquency_T3	.263	.016	16.782	.000
Aggression_T1 → Aggression_T2	.341	.019	18.064	.000
Aggression_T2 → Aggression_T3	.473	.020	23.715	.000
Offline delinquency_T1 → Offline delinquency_T2	.277	.019	14.950	.000
Offline delinquency_T2 → Offline delinquency_T3	.445	.022	20.6351	.000
Cyber delinquency_T1 → Aggression_T2	.024	.010	2.422	.015
Cyber delinquency_T2 → Aggression_T3	−.006	.017	−0.388	.698
Cyber delinquency_T1 → Offline delinquency_T2	.050	.016	3.213	.001
Cyber delinquency _T2 → Offline delinquency_T3	.078	.028	2.756	.006
Aggression_T1 → Cyber delinquency_T2	.058	.024	2.433	.015
Aggression_T1 → Offline delinquency_T2	.074	.030	2.502	.012
Aggression_T2 → Cyber delinquency_T3	−.001	.019	−0.063	.949
Aggression_T2 → Offline delinquency_T3	.065	.034	1.907	.056

Note. CR = critical ratio.

The cross-lagged effects of cyber delinquency, aggression, and offline delinquency over time were as follows. Although cyber delinquency at Time 1 was significantly associated with aggression at Time 2 and showed a positive relationship, cyber delinquency at Time 2 was not significantly associated with aggression at Time 3. In addition, cyber delinquency at Times 1 and 2 was significantly associated with offline delinquency at Times 2 and 3 and showed static relationships. Aggression at Time 1 was significantly associated with both cyber delinquency and offline delinquency at Time 2 and showed static relationships, but aggression at Time 2 was not significantly associated with cyber delinquency or offline delinquency at Time 3.

Discussion

This study examined the relationships between cyber delinquency, aggression, and offline delinquency with longitudinal data collected from adolescents over a 4-year period. Cyber delinquency is a particular area of concern for psychologists because it can lead to offline delinquency and is associated with several mental health issues in adolescents. During the period that eighth-grade students progressed to the 12th grade, the overall levels of cyber delinquency and aggression decreased, while the level of offline delinquency increased.

We found that cyber delinquency, aggression, and offline delinquency at Time t − 1 exerted a positive influence on cyber delinquency, aggression, and offline delinquency at Time t, suggesting that the inclination toward cyber delinquency, aggression, and offline delinquency persisted throughout the transition from early to late adolescence. This finding is consistent with those from previous studies. For example, J. H. Kim and Jung (2009) found that juvenile delinquency is consistent and stable over time and impacts delinquency in adolescence. Also, Thornberry, Lizotte, Krohn, Farnworth, and Jang (1991) reported that past delinquency, more strongly than other social variables, influences current delinquency. Because aggression that begins early in life is linked to adolescents’ aggression as well as delinquency and drug use, previous aggression becomes an important factor for predicting later delinquency (Broidy et al., 2003; Brook et al., 1996; Farrell et al., 2000).

Although our cross-lagged estimates were small in magnitude, there were nevertheless multiple positive influences between aggression, cyber delinquency, and offline delinquency from early to middle adolescence. Cyber delinquency exerted a positive influence on both aggression and offline delinquency, and aggression positively influenced cyber delinquency and offline delinquency during this period. Although cyber delinquency and aggression exerted cross-lagged effects on one another in middle adolescence, the strength of the association between them declined toward late adolescence. As the level of cyber delinquency was decreasing during middle and late adolescence, the level of offline delinquency was increasing, regardless of the level of aggression. Together, these results indicate that aggression had a mediating effect on the relationship between cyber and offline delinquency during early adolescence, but the strength of this effect had decreased by late adolescence. In understanding this result, it is essential to note that our measure of aggression only captured physical aggression and was based on the results of previous studies suggesting that physical aggression is more directly related to aggression in everyday life than is relational aggression (Anderson & Bushman, 2001; Fischer et al., 2010; Giumetti & Markey, 2007).

Although cyber delinquency and aggression during early adolescence exerted a positive influence on offline delinquency, aggression in middle adolescence did not exert an effect on cyber delinquency or offline delinquency during late adolescence. Similarly, in a study by D. K. Kim, Jeon, and Lee (2008), aggression exerted an influence on cyber delinquency in early adolescence, but cyber delinquency was not associated with aggression during late adolescence. Accordingly, the inconsistency between these findings and those of previous studies that suggested a link between cyber delinquency and aggression in late adolescence (K.-E. Kim & Yoon, 2012; D. Y. Lee et al., 2006; Oh, 2014) could be a result of how aggression was measured. The current longitudinal analysis suggested that the relationship between cyber delinquency and aggression may have changed from early to late adolescence, which could also explain the disparity in the results of current and past studies.

According to Gottfredson’s (2002) cyclical model, personality characteristics influence environmental experiences, and the experiences serve to reinforce the characteristics of the individual. Although the overall levels of cyber delinquency, aggression, and offline delinquency varied over time, their autoregressive influences were stable over the 4-year study. In particular, the autoregressive effect for cyber delinquency suggests that the provision of education regarding cyber delinquency during early adolescence could significantly impact the level of cyber delinquency during late adolescence. This study also found that delinquency, aggression, and offline delinquency influenced each other during the period from early to middle adolescence, suggesting that education regarding cyber delinquency could have a more substantial impact if it is provided to adolescents as early as possible to influence the experiences they seek out on the Internet.

There were several limitations of this study. First, the study did not measure the variables precisely, because it selected and analyzed data from the KCYPS, a project with far-reaching goals beyond this study. Second, this study analyzed Korean adolescents, and to generalize the results, future studies of youth from other cultural backgrounds are needed. Third, in this study, the subjectivity of the responses could not be ruled out due to the use of a self-report questionnaire. Therefore, it is necessary to employ additional methods for measuring aggression and delinquency in future studies. Fourth, although our final model fit the data well, the cross-lagged estimates were small. Therefore, future research is needed to verify these cross-lagged relationships in other samples. Fifth, in this study, the descriptive statistics for the variables indicates very low levels of engagement in cyber delinquency, offline delinquency, and aggression. Therefore, future research is needed to investigate the extent to which these results are replicated in samples (e.g., youth involved with the criminal justice system) with higher levels of delinquency.

Sixth, data regarding cyber delinquency, aggression, and offline delinquency were collected every 2 years, which is a somewhat long period for measuring changes in adolescent development. Future studies should measure the effects of adolescent maturation more precisely by using shorter time intervals.

Footnotes

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

ORCID iD

Su-Jung Nam

Author Biography

Su-Jung Nam, PhD, is a professor of the Department of Home Economics Education, College of Education, Jeonju University, Republic of Korea. She has undertaken research in the areas of Communication, digital divide, literacy and consumer well-being, and safety behavior. Her research primarily focuses on the consumer issues in digital divide and digital literacy.

References

Adachi

P. J. C.

Willoughby

(2011). The effect of violent video games on aggression: Is it more than just the violence? Aggression and Violent Behavior, 16, 55-62. https://doi.org/10.1016/j.avb.2010.12.002

Agnew

(1992). Foundation for a general strain theory of crime and delinquency. Criminology, 30, 47-88. https://doi.org/10.1111/j.1745-9125.1992.tb01093.x

Anderson

C. A.

Bushman

B. J.

(2001). Effects of violent video games on aggressive behavior, aggressive cognition, aggressive affect, physiological arousal, and prosocial behavior: A meta-analytic review of the scientific literature. Psychological Science, 12, 353-359.

Anderson

C. A.

Sakamoto

Gentile

D. A.

Ihori

Shibuya

Yukawa

… Kobayashi

(2008). Longitudinal effects of violent video games on aggression in Japan and the United States. Pediatrics, 78, 772-790. https://doi.org/10.1542/peds.2008-1425

Bast

Reitsma

(1997). Matthew effects in reading: A comparison of latent growth curve models and simplex models with structured means. Multivariate Behavioral Research, 32, 135-167. https://doi.org/10.1207/s15327906mbr3202_3

Bentler

P. M.

(1990). Comparative fit indices in structural models. Psychological Bulletin, 107, 238-246. https://doi.org/10.1037/0033-2909.107.2.238

Bollen

K. A.

(1989). Structural equations with latent variables. Hoboken, NJ: John Wiley. https://doi.org/10.1002/9781118619179

Broidy

L. M.

Nagin

D. S.

Tremblay

R. E.

Bates

J. E.

Brame

Dodge

K. A.

… Vitaro

(2003). Developmental trajectories of childhood disruptive behaviors and adolescent delinquency: A six-site, cross-national study. Developmental Psychology, 39, 222-245. https://doi.org/10.1037/0012-1649.39.2.222

Brook

J. S.

Whiteman

Finch

Cohen

(1996). Young adult drug use and delinquency: Childhood antecedents and adolescent mediators. Journal of the American Academy of Child and Adolescent Psychiatry, 35, 1584-1592. https://doi.org/10.1097/00004583-199612000-00009

10.

Browne

M. W.

Cudeck

(1993). Alternative ways of assessing model fit. In Bolen

K. A.

Long

J. S.

(Eds.), Testing structural equation models (pp. 136-162). Thousand Oaks, CA: SAGE.

11.

Buzzell

Foss

Middleston

(2006). Explaining use of online pornography: A test of self-control theory and opportunities for deviance. Journal of Criminal Justice and Popular Culture, 13, 96-116.

12.

Cheung

G. W.

Rensvold

R. B.

(2002). Evaluating goodness-of-fit indexes for testing measurement invariance. Structural Equation Modeling, 9, 233-255. https://doi.org/10.1207/S15328007SEM0902_5

13.

Connor

D. F.

(2002). Aggression and antisocial behavior in children and adolescents: Research and treatment. New York: Guilford Press.

14.

Crick

N. R.

Grotpeter

J. K.

(1995). Relational aggression, gender, and social-psychological adjustment. Child Development, 66, 710-722. https://doi.org/10.1111/j.1467-8624.1995.tb00900.x

15.

Curran

P. J.

Bollen

K. A.

(2001). The best of both worlds: Combining autoregressive and latent curve models. In Collins

L. M.

Sayer

A. G.

(Eds.), Decade of behavior: New methods for the analysis of change (pp. 107-135). Washington, DC: American Psychological Association. https://doi.org/10.1037/10409-004

16.

Davis

R. A.

(2001). A cognitive-behavioral model of pathological Internet use. Computers in Human Behavior, 17, 187-195.

17.

Deyreh

(2011). Psychological pathology of computer and video games among elementary students. Procedia—Social and Behavioral Psychology, 15, 3095-3097. https://doi.org/10.1016/j.sbspro.2011.04.251

18.

Dombrowski

S. C.

Gischlar

K. L.

Durst

(2007). Safeguarding young people from cyber pornography and cyber sexual predation: A major dilemma of the Internet. Child Abuse Review, 16, 153-170. https://doi.org/10.1002/car.939

19.

Dong

Ahou

Zhao

(2011). Precursor or sequela: Pathological disorders in people with internet addiction disorder. PLoS ONE, 6(2), e14703. https://doi.org/10.1371.journal.pone.0014703

20.

Durkin

Barber

(2002). Not so doomed: Computer game play and positive adolescent development. Applied Development Psychology, 23, 273-292. https://doi.org/10.1016/S0193-3973(02)00124-7

21.

Farrell

A. D.

Kung

E. M.

White

K. S.

Valois

R. F.

(2000). The structure of self-reported aggression, drug use, and delinquent behaviors during early adolescence. Journal of Clinical Child Psychology, 29, 282-292. https://doi.org/10.1207/S15374424jccp2902_13

22.

Fischer

Kastenmüller

Greitemeyer

(2010). Media violence and the self: The impact of personalized gaming characters in aggressive video games on aggressive behavior. Journal of Experimental Social Psychology, 46, 192-195. https://doi.org/10.1016/j.jesp.2009.06.010

23.

Friedman

H. S.

Schustack

M. W.

(2009). Personality: Classic theories and modern research (4th ed.). Boston, MA: Pearson Allyn & Bacon.

24.

Giumetti

G. W.

Markey

P. M.

(2007). Violent video games and anger as predictors of aggression. Journal of Research in Personality, 41, 1234-1243. https://doi.org/10.1016/j.jrp.2007.02.005

25.

Goodman

S. H.

Brogan

Lynch

M. E.

Fielding

(1993). Social and emotional competence in children of depressed mothers. Child Development, 64, 516-513. https://doi.org/10.1111/j.1467-8624.1993.tb02925.x

26.

Gottfredson

L. S.

(2002). Gottfredson’s theory of circumscription, compromise, and self-creation. In Brown

(Ed.), Career choice and development (pp. 85-148). San Francisco, CA: Jossey-Bass. https://doi.org/10.1007/s13398-014-0173-7.2

27.

Grotpeter

J. K.

Crick

N. R.

(1996). Relational aggression, overt aggression, and friendship. Child Development, 67, 2328-2338. https://doi.org/10.1111/j.1467-8624.1996.tb01860.x

28.

Han

J. W.

(2001). An empirical study on the social bond elements of juvenile delinquents in cyber space. Korean Association of Police Science, 3, 357-384.

29.

Holtz

Appel

(2011). Internet use and video gaming predict problem behavior in early adolescence. Journal of Adolescence, 34, 49-58. https://doi.org/10.1016/j.adolescence.2010.02.004

30.

Hong

(2000). The criteria for selecting appropriate fit indices in structural equation modeling and their rationales. The Korean Journal of Clinical Psychology, 19, 161-177.

31.

Hong

Yoo

S. K.

You

C. C.

(2010). The reciprocal relationship between parental involvement and mathematics achievement: Autoregressive cross lagged modeling. The Journal of Experimental Education, 78, 419-439. https://doi.org/10.1080/00220970903292926

32.

Hong

Y. S.

Kim

D. K.

(2011). Structural relation of juvenile stress, self-esteem, self-control and cyber delinquency: Testing a sex difference. Korea Youth Research Association, 18, 127-148. https://doi.org/10.4324/9780203823453

33.

Jessor

S. L.

(1977). Problem behavior and psychosocial development: A longitudinal study of youth. Cambridge, MA: Academic Press.

34.

Jessor

Van Den Bos

Vanderryn

Costa

F. M.

Turbin

M. S.

(1995). Protective factors in adolescent problem behavior: Moderator effects and developmental change. Developmental Psychology, 31, 923-933. https://doi.org/10.1037/0012-1649.31.6.923

35.

N. K.

Yang

D. K.

(2001). The relationships among perceived adolescents’ social supports, internet addiction tendency, and internet-related delinquency. Korean Psychological Association, 8, 111-133.

36.

Jöoreskog

K. G.

Sörbom

(1996). LISREL 8: Users’ reference guide. Chicago, IL: Scientific Software International.

37.

Kim

D. K.

Jeon

J. Y.

Lee

S. J.

(2008). Tests of the mediating effect of stress and aggression in the relations between juvenile leisure constraint and cyber delinquency. Korean Journal of Youth Studies, 15, 157-179.

38.

Kim

J. H.

Jung

H. W.

(2009). Testing the autoregressive cross-lagged effects between parent attachment and delinquency: Multigroup analysis across gender. Korean Criminology Review, 20, 125-148.

39.

Kim

J. H.

Kim

Hong

(2009). Writing articles using structural equation modeling. Seoul, Korea: Communication Books.

40.

Kim

K.-E.

Yoon

(2012). Associations between adolescents’ victimization of violence, tolerance toward violence, and cyber violence offending behavior. Journal of Korean Society of Child Welfare, 39, 213-244.

41.

C. H.

Yen

J. Y.

Yen

C. F.

Chen

C. S.

Chen

C. C.

(2012). The association between internet addiction and psychiatric disorder: A review of literature in adolescents. Journal of Adolescent Health, 44, 598-605. https://doi.org/10.1016/j.jadohealth.2008.11.011

42.

C. S.

(2012). The effects of addictive cell-phone use of adolescence on mental health. Forum for Youth Culture, 30, 8-36.

43.

Kuss

D. J.

Griffiths

M. D.

Karila

Billieus

(2014). Internet addiction: A systematic review of epidemiological research for the last decade. Current Pharmaceutical Design, 20, 4026-4052. https://doi.org/10.2174/13816128113199990617

44.

Kutner

Olson

(2008). Ground theft childhood: The surprising truth about violent video games and what parents can do. New York: Simon & Schuster.

45.

Lahey

B. B.

Waldman

I. D.

McBumett

(1997). The development of antisocial behavior: An integrative causal model. Journal of Child Psychology and Psychiatry, 40, 669-682. https://doi.org/10.1017/S0021963099004047

46.

Lee

D. Y.

Park

J. Y.

Lee

D. H.

(2006). Validation of relationship model among adolescents dysfunctional psychological traits, internet addiction and cyber-related delinquencies: Meditated by internet addiction. Korean Journal of Counseling, 7, 103-115.

47.

Lee

J. S.

Ahn

Y. Y.

(2005). A study of use of computer by elementary schoolers and cyber-related delinquency. Korean Youth Research Association, 16, 225-254.

48.

Loeber

Stouthamer-Loeber

(1987). Prediction. In Quay

H. C.

(Ed.), Handbook of juvenile delinquency (pp. 370-371). Hoboken, NJ: John Wiley.

49.

Möller

Krahé

(2009). Exposure to violent video games and aggression in German adolescents: A longitudinal analysis. Aggressive Behavior, 35, 75-89. https://doi.org/10.1002/ab.20290

50.

I. S.

(2014). The relations of traditional bullying, cyberbullying and types of aggression in terms of gender. Korean Journal of Counseling, 15, 1871-1885. https://doi.org/10.15703/kjc.15.5.201410.1871

51.

Olson

C. K.

(2004). Media violence research and youth violence data: Why do they conflict? Academic Psychiatry, 28, 144-150. https://doi.org/10.1176/appi.ap.28.2.144

52.

Orpinas

Frankowski

(2001). The aggression scale: A self-report measure of aggressive behavior for young adolescents. Journal of Early Adolescence, 21, 50-67. https://doi.org/10.1177/0272431601021001003

53.

Peter

Valkenburg

P. M.

Schouten

A. P.

(2005). Developing a model of adolescent friendship formation on the Internet. Cyberpsychology and Behavior, 8, 423-430. https://doi.org/10.1089/cpb.2005.8.423

54.

Rappaport

Thomas

(2004). Recent research findings on aggressive and violent in youth: Implications for clinical assessment and intervention. Journal of Adolescent Health, 35, 260-277. https://doi.org/10.1016/j.jadohealth.2003.10.009

55.

Rashid

B. K.

Muhammad

Altaf

Mehvish

(2014). The relationship between internet addiction and anxiety among students of University of Sargodha. International Journal of Humanities and Social Science, 4, 288-293.

56.

Rokven

J. J.

Weijters

Beerthuizen

M. G. C. J.

van den Laan

A. M.

(2018). Juvenile delinquency in the virtual world: Similarities and differences between cyber-enabled, cyber-dependent and offline delinquents in the Netherlands. International Journal of Cyber Criminology, 12, 27-46. https://doi.org/10.5281/zenodo.1467690

57.

Savage

Yancey

(2008). The effects of media violence exposure on criminal aggression: A meta-analysis. Criminal Justice and Behavior, 35, 1123-1136. https://doi.org/10.1177/0093854808316487

58.

Sherry

J. L.

(2006). Violent video games and aggression: Why can’t we find links? In Priess

R. W.

Gayle

B. M.

Burrell

Allen

N., M.

Bryant

(Eds.), Mass media effects research: Advances through meta-analysis (pp. 231-248). Mahwah, NJ: Lawrence Erlbaum.

59.

Shibuya

Sakamoto

Ihori

Yukawa

(2008). The effects of the presence and context of video game violence on children: A longitudinal study in Japan. Simulation and Gaming, 39, 528-539. https://doi.org/10.1177/1046878107306670

60.

Siegel

L. J.

Welsh

B. C.

(2018). Juvenile delinquency: Theory, practice, and law (13th ed.). Boston, MA: Cengage Learning.

61.

Sun

Johnson

C. A.

Palmer

Arpawong

T. E.

Unger

J. B.

Xie

… Sussman

(2012). Concurrent and predictive relationships between compulsive internet use and substance use: Finding from vocational high school students in China and the USA. International Journal of Environmental Research and Public Health, 9, 660-673. https://dx-doi-org.web.bisu.edu.cn/10.3390./ijerph9030660

62.

Swann

W. B.

Read

S. J.

(1981). Acquiring self-knowledge: The search for feedback that fits. Journal of Personality and Social Psychology, 41, 1119-1128. https://doi.org/10.1037//0022-3514.41.6.1119

63.

Thornberry

T. P.

Lizotte

A. J.

Krohn

M. D.

Farnworth

Jang

S. J.

(1991). Testing interactional theory: An examination of reciprocal causal relations among family, school, and delinquency. The Journal of Criminal Law and Criminology (1973-), 82(1), 3-35. https://doi.org/10.2307/1143788

64.

Tucker

L. R.

Lewis

(1973). A reliability coefficient for maximum likelihood factor analysis. Psychometrika, 38, 1-10. https://doi.org/10.1007/BF02291170

65.

Unsworth

Devilly

Ward

(2007). The effect of playing violent video games on adolescents: Should parents be quaking in their boots? Psychology, Crime, and Law, 13, 383-394. https://doi.org/10.1080/10683160601060655

66.

Weinstein

Lejoyeux

(2010). Internet addiction or excessive internet use. The American Journal of Drug and Alcohol Abuse, 36, 277-283. https://doi.org/10.3109/00952990.2010.491880

67.

Williams

Skoric

(2005). Internet fantasy violence: A test of aggression in an online game. Communication Monographs, 72, 217-233. https://doi.org/10.1080/03637750500111781

68.

Zhen

Xie

Zhang

Wang

(2011). Exposure to violent computer games and Chinese adolescents’ physical aggression: The role of beliefs about aggression, hostile expectation, and empathy. Computers in Human Behavior, 27, 1675-1687. https://doi.org/10.1016/j.chb.2011.02.006