Cycle of Violence in Schools: Longitudinal Reciprocal Relationship Between Student’s Aggression and Teacher’s Use of Corporal Punishment

Abstract

The purpose of this study is to examine the longitudinal reciprocal relationship between student’s aggression and teacher’s use of corporal punishment. An autoregressive cross-lagged model was analyzed with the data drawn from 4,051 Korean secondary students (male = 2,084, female = 1,967), in Gyeonggi Education Panel Study for three waves (seventh-ninth grades). Results revealed that student’s aggression provoke teacher’s use of corporal punishment and also teacher’s use of corporal punishment provokes student’s aggression. It is important in that it suggests the cycle of violence with the reciprocal relationship between student’s aggression and teacher’s use of corporal punishment, rather than positing the unidirectional effects. Practically, teachers should keep in mind that corporal punishments, which are at least partially attributable to student’s aggression, actually worsen the problem and lead to a cycle of violence in schools. Accordingly, they should instead respond with alternative disciplinary strategies or direct interventions dealing with the causes of aggression.

Keywords

corporal punishment of teacher Student aggression autoregressive cross-lagged model School violence

School violence is getting much interest from literature all over the world. Especially, aggression is one of the most important variables that has received a great deal of attention from literature. Aggression is associated with various kinds of maladjustments, such as rejection from peers (Choi, Park, & Shin, 2016), delinquency, and substance abuse (Lynne-Landsman, Graber, Nichols, & Botvin, 2011). Also, adolescent aggression could lead to not only later adult aggression but also other issues such as criminal behavior (Farrington, 1991; Huesmann, Eron, Lefkowitz, & Walder, 1984). Thus, concrete understanding of student’s aggression during adolescence is important in that it could suggest the directions for the early prevention of maladjustments. Based on this knowledge, the practicioners in educational fields could arrange appropriate interventions and avoid inappropriate ones for the promotion of social adjustments.

But, as often seen in educational practice, teachers are unable to intervene into student’s aggression appropriately. Rather, some studies found that teachers tend to aggressively punish or discipline students in response to aggression of their students (Lewis, 2001; Lewis, Romi, Qui, & Katz, 2005; Nesdale & Pickering, 2006). What is worse is the fact that most of the students already perceive the teacher’s tendency of aggressive responses (Lewis, 2001). As teachers are the role models in schools, their aggressive responses could spread the violence into schools. A typical form of those kinds of the aggressive response by teachers is a corporal punishment. In other words, student’s aggression could provoke teacher’s corporal punishment and this corporal punishment could also provoke further aggression of students, leading to cycle of violence in schools. But, little prior research has examined the reciprocal relationship between them. If the reciprocal relationship exists, it would suggest special interventions for either teachers or students to break out from the cycle of violence in schools.

Theoretical Conceptualization of the Relationship Between Student’s Aggression and Teacher’s Use of Corporal Punishment

Before examining the relationship between student’s aggression and teacher’s use of corporal punishment, the conceptualization of the aggression and corporal punishment is needed. First, aggression is an umbrella term, whose definition varies by each scholar, but it is often defined as a harmful behavior or stimulus with the offender’s expectation and intention to damage another person against his or her will (Bushman & Anderson, 2001; Geen, 2001). It usually includes both types of aggressions: verbal and physical (Buss & Perry, 1992). The definition of corporal punishment also varies. Straus (1994) defined corporal punishment as “the use of physical force with the intention of causing a child to experience pain but not injury for the purposes of correction or control of the child’s behavior” (p. 4). As there could be various ways to make children experience pain, it is suggested that the definition of corporal punishment should include not only direct but also other indirect kinds of corporal punishments (Hyman, 1995).

There have been theoretical conflicts about the causal directionality in the relationship between aggression and corporal punishment. In other words, about whether teacher’s use of corporal punishment provoke student’s aggression or whether student’s aggression provoke teacher’s use of corporal punishment, the social learning theory model (Bandura, 1978; Straus, 1991) versus the temperament theory model (Muller, Hunter, & Stollak, 1995) conflicts. The social learning theory model (Bandura, 1978; Straus, 1991) suggests that it is teacher’s use of corporal punishment that provokes student’s aggression. It is because of the social learning process of corporal punishment. When students are exposed to corporal punishment, they would not only directly model the violent behavior of teacher but also learn and get used to positive attitudes toward violence with the belief that violence could be a useful means to solve problems (Bandura, 1978; Owen, 2005; Straus, 1991).

On the contrary, the temperament theory model (Muller et al., 1995; Straus, 1991) suggests that it is student’s aggression that provokes teacher’s use of corporal punishment. It is based on prior literature on the child effects model. Child effects literature consistently reports that a substantial proportion of the environments of the children are actually provoked by the children themselves (e.g., O’Connor, Deater-Deckard, Fulker, Rutter, & Plomin, 1998; Scarr & McCartney, 1983). If the child effects model is applied into here, it is the aggressive temperament of students that provokes the environment of being corporally punished, and teachers are intervening with the aggressive behavior of their student by using corporal punishment for disciplinary reasons. Corporal punishment is well known to be commonly used by teachers as a disciplinary strategy, especially in response to antisocial behaviors of adolescents (Dupper & Dingus, 2008; Mamatey, 2010; Robinson, Funk, Beth, & Bush, 2005; Straus, 1991; Youssef, Attia, & Kamel, 1998).

However, this study suggests that the two conflicting models could be compatible, rather than mutually exclusive. Even if the suggestion that teacher’s use of corporal punishment provoke student’s aggression is true, it does not mean that the opposite direction from student’s aggression to teacher’s use of corporal punishment cannot be true. In other words, student’s aggression and teacher’s use of corporal punishment could be longitudinally reciprocal, both being cause and consequence for each other respectively and simultaneously. This “reciprocal model” is based on the prior literature that consistently report the bidirectional causality between the antisocial behavior of adolescents and other various external factors such as peers, parents, and teachers variables over lifetime (Dodge & Pettit, 2003). Until recently, there has not been a study that directly suggest the reciprocal model and empirically test it, whereas it could easily be inferred from prior literature (Straus, 1991).

None of the social learning model, neither the temperament model nor the reciprocal model, has enough empirical studies that could directly support them. Particularly, there is not a single study that has directly posited the reciprocal model. Although the relationship between student’s aggression and parent’s use of corporal punishment received much interest in the literature, much less empirical studies were carried out with the teacher’s use of corporal punishment (Gershoff, 2010). However, the corporal punishment of teacher is no less detrimental than that of parents. Especially, adolescence is a period when teachers have great influence on student’s adjustments or maladjustment, because they interact with each other frequently everyday in schools (Chang, 2003). Inferring from the empirical studies that examined the relationship between student’s aggression and parent’s use of corporal punishment, which found student’s aggression lead to parent’s use of corporal punishment and vice versa (Gershoff, 2002, 2010), the reciprocal model is expected to be supported.

Effects of Teacher’s Use of Corporal Punishments on Student’s Aggression

Teacher’s use of corporal punishment could easily be regarded as a possible predictor of student’s aggression. This kind of reasoning might be inferred from previous research that found positive relationship between parent’s use of corporal punishments and their children’s antisocial behavior (e.g., Gershoff, 2002, 2010). Even though there have been many theoretical arguments that teacher’s use of corporal punishments would lead to aggressions of students (e.g., Straus, 1991), contrary to expectations, there is a dearth of longitudinal empirical study that directly tested this suggestion. Thus, we could only infer the possible effects based on the prior research which found that parent’s use of punishment leads to student’s aggression (Gershoff, 2002, 2010) both cross-sectionally (e.g., Aucoin, Frick, & Bodin, 2006; Hecker, Hermenau, Isele, & Elbert, 2014) and longitudinally (e.g., Mulvaney & Mebert, 2007; Weaver, Borkowski, & Whitman, 2008), even though some meta-analyses report that their effect sizes are null or too small to interpret meaningfully (Paolucci & Violato, 2004). Confirming the effects of teacher’s corporal punishment on student’s aggression is important in that it would empirically suggest the direction of which disciplinary strategy is to be avoided.

Effects of Student’s Aggression on Teacher’s Use of Corporal Punishments

On the contrary, student’s aggression could also lead to teacher’s use of corporal punishments. However, there is a dearth of longitudinal empirical study that examines it. There are some cross-sectional empirical studies which found that student’s aggression predicts teacher’s use of corporal punishments (e.g., Lewis et al., 2005; Youssef et al., 1998). However, to the best of our knowledge, there has not been a single empirical study that investigated the effects of using longitudinal data. We could only infer the possible effects based on prior research which found that children’s aggression longitudinally leads to the parent’s use of corporal punishment (e.g., Jaffee et al., 2004). There are some other studies with which we could indirectly infer the possible effects of student’s aggression on teacher’s use of corporal punishment. They found that teachers tend to aggressively punish or discipline students in response to their aggressions (Lewis, 2001; Nesdale & Pickering, 2006). The exploration of the causes that make teachers use corporal punishments is important. Not only would it expand the knowledge of how teachers actually react in response to student’s aggression but also could empirically suggest the direction to reduce the use of corporal punishments in the educational practice.

Methodological Limitations of Prior Literature

Even if the theoretical backgrounds of each of the social learning model, temperament model, and reciprocal model are concrete, there is a dearth of study that has directly tested each model. Even a few of existent prior studies that examined the relation between student’s aggression and teacher’s use of corporal punishment have some methodological limitations. Most of them used analytical methods that were not strong enough to infer the causal directionality between student’s aggression and teacher’s use of corporal punishment (i.e., cross-sectional analysis). According to Straus (1991), the models can only be tested with longitudinal and experimental data. However, because of ethical problems, experiments are limited. Randomly assigning students to the corporal punishments group or control group is not possible. In the same manner, manipulating students to be aggressive is not possible. Accordingly, Paolucci and Violato (2004), who have done meta-analysis on the effects of corporal punishments, have suggested that more sophisticated analysis such as structural equation modeling (SEM) is required for the investigation of any potential causal direction between use of corporal punishment and aggression. But few studies have been conducted with SEM and longitudinal data.

In particular, there has not been an empirical study conducted with autoregressive cross-lagged modeling (Curran & Bollen, 2001). It is a recent analytic strategy with longitudinal data with more than three waves. With this analytical strategy, we could relatively strongly infer the causal relationship between two (or more) variables. It is because we could statistically control for prior effects of student’s aggression and teacher’s use of corporal punishment in prediction to subsequent wave student’s aggression and teacher’s use of corporal punishment. Also, we could statistically confirm the assumptions that (a) each variable is measured with the same manner consistently across time (i.e., three waves or more), (b) the conceptualization of latent variables is the same across time, and (c) effects of prior variables on subsequent variables, either autoregressive paths and cross-lagged paths, were the same across all the three times (Curran & Bollen, 2001). Thus, this study tried to use autoregressive cross-lagged modeling to examine the relation between student’s aggression and teacher’s use of corporal punishment, for the relatively stronger inference of the causal relationship between those variables.

The Present Study

Taken together, student’s aggression could lead to teacher’s use of corporal punishments, which then lead to student’s aggression again and vice versa. But few of prior literature investigated this reciprocal relationship with longitudinal empirical data. For the investigation of the causal relationship between student’s aggression and teacher’s use of corporal punishment, an experimental study that proposes a causal direction is required. However, because of the ethical problems, experiments are not possible. Accordingly, we tried to analyze the autoregressive cross-lagged model (Curran & Bollen, 2001), which is possibly the best method for the inference of the causal relationship between two (or more) variables.

Based on the prior literature, it is predicted that there would be longitudinal reciprocal relationship between student’s aggression and teacher’s use of corporal punishment. Examining the longitudinal reciprocal relationship would suggest the directions of interventions for either teachers or students to break out from the cycle of violence. To be more detailed, by confirming the effects of teacher’s corporal punishment on student’s aggression, it would empirically suggest the direction of which disciplinary strategy is to be avoided. Also, by confirming the effects of student’s aggression on teacher’s corporal punishment, it would not only expand the knowledge of reasons why corporal punishment is still used by teachers in education practice (Dupper & Dingus, 2008; Youssef et al., 1998) despite the popular perceptions that corporal punishment is socially undesirable, but also could empirically suggest the direction for reducing the use of corporal punishments in educational practice.

Method

Samples

The data were drawn from 4,051 Korean secondary students (male = 2,084, female = 1,967), in Gyeonggi Education Panel Study (GEPS) for three waves (seventh-ninth grades). In Korean secondary schools, the age of students in seventh grade is about 12 to 13 years old. They first participated in this GEPS in 2012, in each of their schools when they were in seventh grade. The Gyeonggi Institute of Education implemented this panel study with the 63 schools of Gyoenggi Province of South Korea (secondary panel) to follow-up cognitive, affective, and social developments of those students of Gyeonggi Province schools. It was sampled with cluster stratification sampling. The 31 cities of Gyeonggi Province were used as strata and random sampling was used for each stratum in a manner proportional to the population, so that the data could represent the 31 cities proportionately. The panel study was conducted under the approval and cooperation of Gyeounggi Provincial Offices of Education. Before the panel study survey, all students and their parents submitted consent forms of GEPS through newsletter from each school which contained explicit agreements. The survey did not have any content that included any kind of ethical issues. This study got official approval from the Gyeonggi Institute of Education. About 93.4% of students and parents responded for all three waves. To deal with the problem of missing data from those who missed at least one time point, we used full information maximum likelihood (FIML) methods (Arbuckle, 1996; Schafer & Olsen, 1998) in AMOS 20.0. The reason why we used the data from this age group is as follows. First, aggression stands out most in this period (Pellegrini, 2001). Second, teacher influence could be salient in this period because they spend more time than children, and interact with teachers more frequently everyday in schools (Chang, 2003). Third, corporal punishment is known to be mostly salient in this period, with about 70% of students experiencing it (Youssef et al., 1998). Fourth, there is a perception among Korean teachers that use of corporal punishment is necessary for secondary school students (B. Brown, 2009).

Measure of Variables

Student’s aggression

To measure student’s aggressions, participants answered the question of “How many times have you done as follows during the last year? Please select the most appropriate response.” Three items of “Beat others up,” “Swore in the face of a peer,” “Swore or verbally abused someone in online chat rooms or message boards” were followed with a 6-point Likert-type scale (1 = never, 2 = 1 or 2 times per year, 3 = 1 or 2 times per semester, 4 = 1 or 2 times per month, 5 = 1 or 2 times per week, 6 = almost every day). All the items were measured for 3 years each. The Cronbach’s alphas were .60, .56, and .57 for each year in order. The three items constructed the latent variable of student’s aggression for each wave. The selection procedure of items was as follows. First, based on the understanding of literature on aggression, we selected the above three items, those which could be regarded as measuring aggression, within the original 14 items classified in the category of “delinquency” in the GEPS panel data (e.g., smoke, drink alcohol). Second, to empirically test the construct validity of those selected items, explanatory factor analysis was done with the original 14 items of the delinquency category. The explanatory factor analysis resulted in four factors. Among those factors, the selected three items constructed a factor that could be regarded as measuring aggression. This explanatory factor analysis took the role as an ancillary indicator for the final selection among preliminary items that were selected in the first step. Third, with correlations among items and Cronbach’s alpha, the reliability of the finally selected items was confirmed.

Teacher’s use of corporal punishments

To measure teacher’s corporal punishments, participants answered the question of “How many times have you experienced the following from your teacher during the last year? Please select the most appropriate response.” Two items of “Direct corporal punishments (punishments such as hitting with implements, hands, or feet)” and “Indirect corporal punishments (punishments that cause pain without direct physical contact)” were followed with a 6-point Likert-type scale (1 = never, 2 = 1 or 2 times per year, 3 = 1 or 2 times per semester, 4 = 1 or 2 times per month, 5 = 1 or 2 times per week, 6 = almost everyday). All the items were measured for 3 years each. The Cronbach’s alphas were .77, .74, and .74 for each year in order. The two items constructed the latent variable of teacher’s use of corporal punishment. The selection procedure of items was as follows. First, based on the understanding of literature on corporal punishments, we selected the above two items those which could be regarded as measuring aggression within the original five items classified in the category of “human rights situation” in the GEPS panel data (e.g., restrictions on hair or dress style, confiscation of belongings not permitted). Second, with correlations among items and the Cronbach’s alpha, the reliability of the finally selected items was confirmed.

Statistical Analysis

To examine the longitudinal reciprocal relationship between student’s aggression and teacher’s use of corporal punishments, we analyzed the autoregressive cross-lagged modeling with AMOS 20.0. Before the analysis, we tested whether the observed variables had any problems in the normality assumption. It revealed that under the rule of thumb of skewness below |3| and kurtosis below |20|, there were no problems in the normality assumption (Kline, 2005). To deal with the problem of missing data, we used FIML methods (Arbuckle, 1996; Schafer & Olsen, 1998). Model comparisons were analyzed in the order of base model with no constraints at all (Model 1), test of time-varying measurement invariance (Model 2), test of time-varying paths invariance (Model 3), and test of time-varying covariance of error invariance (Model 4). In these analyses, the last model (Model 4) is the final autoregressive cross-lagged model (Curran & Bollen, 2001). During the whole analysis, correlations among errors of observed variables, which are repeatedly measured with the same items in three waves longitudinally (e.g., correlation between e1 and e4, e4 and e7, e1 and e7s in Figure 1), were assumed (Bollen, 1989; Pitts, West, & Tein, 1996). Also, gender (male = 0, female = 1) was input as a control variable over the whole process. Figure 1 shows the detailed research model. But for the simplicity of the figure of the model, those error terms, and their correlations, effects of gender were not shown in the final results of Figure 2. The details of each model in the comparison analysis are as follows:

Model 1: Base model with no constraints at all.

Model 2: As a model for the test of time-varying measurement invariance, each factor loading coefficients from indicators to latent variables of student’s aggression and teacher’s use of corporal punishments for three waves (seventh, eighth, and ninth grade) were constrained to be equal. In Figure 1, for instance, factor loadings from TCP seventh grade to TCP11, TCP12, and TCP13; TCP eighth grade to TCP21, TCP22, and TCP23; TCP ninth grade to TCP31, TCP32, and TCP33 were constrained to be equal.

Model 3: As a model for the test of time-varying paths invariance, each coefficient of autoregressive paths of latent variables, student’s aggression, and teacher’s use of corporal punishments, and cross paths between those latent variables were constrained to be equal. In Figure 1, for instance, paths from TCP seventh to TCP eighth, and from TCP eighth to TCP ninth were constrained to be equal; paths from TCP seventh to SAG eighth, and TCP eighth to SAG ninth were constrained to be equal.

Model 4: As a model for the test of time-varying covariance of error invariance, each covariance of error between latent variables, student’s aggression, and teacher’s use of corporal punishment were constraint to be equal. In Figure 1, for instance, the covariance between d2 and d5, and d3 and d6 was constrained to be equal.

Figure 1.

Multivariate autoregressive cross-lagged model.

Figure 2.

Longitudinal reciprocal relationship between student’s aggression and teacher’s use of corporal punishments in the final models.

For the evaluation of model goodness of fit, the chi-square test, which is sensitive to sample size, was not used regarding the large sample size in the current data. Alternatively, we used RMSEA (root mean square error of approximation) and CFI (comparative fit index), which are relatively insensitive to sample size and able to consider simplicity. We used the standard for interpreting a good model fit as below .05 for RMSEA and more than .95 for CFI (M. W. Brown & Cudeck, 1993; Hu & Bentler, 1999). For model comparison, under the assumption that all the comparing models are in reasonably good model fit, we used the change amount of CFI below .01 as a standard for adopting an equal model fit hypothesis (Cheung & Rensvold, 2002). This decision was based on the fact that not only chi-square test but also chi-square difference test is sensitive to sample size.

Results

Descriptive Statistics and Correlation Among Latent Variables

Table 1 shows descriptive statistics and correlations among latent variables in the measurement model of the base model. First, the means of student’s aggression of each 3 year showed 1.73 to 1.90 out of 6 points. The standard deviation of student’s aggression of each 3 year showed 1.00 to 1.05. The means of teacher’s use of corporal punishment of each 3 year showed 1.44 to 2.42 out of 6 point. The standard deviations of teacher’s use of corporal punishment of each 3 year showed 0.92 to 1.49. Next, correlations among student’s aggressions over three waves (i.e., stabilities) were .56 ~ .68. Also, correlation among teacher’s use of corporal punishments over three waves were .26 ~ .45. The correlations between student’s aggression and teacher’s use of corporal punishments were .19 ~ .36. All the correlations among latent variables were statistically significant below the value of p < .001.

Table 1.

Descriptive Statistics and Correlation Among Latent Variables.

	1	2	3	4	5	6
1. Student’s aggression, seventh grade	—
2. Student’s aggression, eighth grade	.56	—
3. Student’s aggression, ninth grade	.57	.68	—
4. Teacher’s corporal punishments, seventh grade	.28	.21	.19	—
5. Teacher’s corporal punishments, eighth grade	.29	.36	.35	.36	—
6. Teacher’s corporal punishments, ninth grade	.23	.30	.35	.26	.45	—
M	1.90	1.76	1.73	2.42	1.81	1.44
SD	1.05	1.00	1.00	1.49	1.15	0.92

Note. Correlations were analyzed among latent variables in the measurement model of base model in SEM. All the correlations were statistically significant under the value of p < .001. SEM = structural equation modeling.

Test of Model Comparison for the Autoregressive Cross-Lagged Model Between Student’s Aggressions and Teacher’s Use of Corporal Punishments

Table 2 shows the model comparison process for the test of goodness of model fits for autoregressive cross-lagged model between student’s aggressions and teacher’s use of corporal punishments. First, before the model comparisons, all the comparing models were in reasonably good model fit under the standard of RMSEA < .05, CFI > .95. Thus, we analyzed the model comparison for the next step. The results showed that overall pairs of model comparisons were below the standard of ∆CFI < .01, meaning that all the models could be regarded as statistically equal. Accordingly, the last model, Model 4, was adopted for the final model, which is the final autoregressive cross-lagged model between student’s aggressions and teacher’s use of corporal punishments (χ² = 237.162, df = 46, RMSEA = .032, CFI = .983). Figure 2 shows the final model with standardized coefficients.

Table 2.

Test of Model Comparison for the Autoregressive Cross-Lagged Model.

	χ²	df	RMSEA	CFI	∆CFI
Model 1	121.453	37	.024	.993
Model 2	208.371	41	.032	.985	−.008
Model 3	230.992	45	.032	.984	−.001
Model 4 (Final model)	237.162	46	.032	.983	−.001

Note. Model 1: base model with no constraints at all; Model 2: model for the test of time-varying measurement invariance; Model 3: model for the test of time-varying paths invariance; Model 4: model for the test of time-varying covariance of error invariance. RMSEA = root mean square error of approximation; CFI = comparative fit index.

In the final model, the standardized effects of gender on student’s aggression of seventh, eighth, and ninth grades were −.56, −.28, and −.46, revealing more aggression for male. Also, standardized effects of gender on teacher’s use of corporal punishment of seventh, eighth, and ninth grade were −.26, −.34, and −.23, revealing more corporal punishment for male. Standardized autoregressive coefficients (i.e., stabilities) among student’s aggression were .55 for each, with relatively high stabilities. Standardized autoregressive coefficients among teacher’s use of corporal punishments were .36 and .35 for each. Standardized cross-lagged coefficients in the effects of student’s aggression on the teacher’s use of corporal punishments were .13 and .15 for each. On the other direction, standardized cross-lagged coefficients in the effects of teacher’s use of corporal punishments on the student’s aggression were .10 and .08 for each. All the standardized cross-lagged coefficients were statistically significant below the value of p < .001. Thus, the longitudinal reciprocal relationship that student’s aggression leads to teacher’s use of corporal punishments, and vice versa, teacher’s use of corporal punishments lead to student’s aggression, is confirmed with the autoregressive cross-lagged model.

Discussion

This study examined the longitudinal reciprocal relationship between student’s aggression and teacher’s use of corporal punishment with the analysis of the autoregressive cross-lagged model. In other words, we tried to test if student’s aggression provokes teacher’s use of corporal punishment, and in the other direction, if teacher’s use of corporal punishment also provokes student’s aggression. It is meaningful in that we empirically verified the longitudinal reciprocal relationship between student’s aggression and teacher’s use of corporal punishment, not just for a unidirectional relationship. Concrete discussions according to the main results are as follows.

First, the results revealed that student’s aggression leads to teacher’s use of corporal punishment s. In other words, we can say that students, who are exhibiting aggressive behavior, actually provoke their teacher’s use of corporal punishments. This could be the reason why corporal punishments are still used by teachers even though it is legally banned worldwide, including Korea. It is because students themselves are attributable, at least in part, to being punished. It is consistent with the literature on the child effects, which has theoretically and empirically suggested that substantial proportion of the environments of the children are actually provoked from the children themselves (e.g., O’Connor et al., 1998; Scarr & McCathy, 1983). However, it does not mean that use of corporal punishment could be justifiable because it is at least partially attributable to students’ aggressive behavior themselves. One of the most important roles of a teacher is to be a role model for social guidance. Thus, their violent reactions to violent behavior of students would rather worsen the problem. Future research should find the moderating variables that could buffer the paths from student’s aggressions to teacher’s use of corporal punishments. One of possible moderating variables that could affect the paths from student’s aggressions to teacher’s use of corporal punishments could be teacher’s implicit attitudes toward corporal punishments that strong punishments are effective in reducing student’s aggression (Kilimci, 2009; Mamatey, 2010; Robinson et al., 2005; Straus, 1991). By confirming the effects of student’s aggression on teacher’s use of corporal punishment, this study expands the knowledge of reasons why corporal punishment is still often used by teachers in educational practice (Dupper & Dingus, 2008; Youssef et al., 1998) despite the common perceptions that corporal punishment is socially undesirable.

The results might be attributable to and only specifically applied to Korean culture, even though we did not assume any cultural differences. Teacher’s use of corporal punishment has been legally prohibited in Korea since 2011, and there is a common perception that corporal punishment is socially undesirable. Even so, our descriptive data showed that many Korean secondary students still report that they have been corporally punished in 2012, 2013, and 2014, when they were seventh, eighth, and ninth grade, respectively. It might be vestige of Confucian culture in which corporal punishments were regarded as the rod of love and were in common practice. Even, in Korea, corporal punishment has been regarded as the duty of teacher to discipline and guide students who are exhibiting misbehavior, a right of the teacher to protect their authority and for efficient class management (B. Brown, 2009; Cho, 2001; Shin & Koh, 2005). There is an idiom in Korea that goes “The king, the teacher and the father are one,” which indicates the strong authority of teachers. These vestiges of Confucian culture of Korea might be the reason why student’s aggression provokes teacher’s use of corporal punishment in our data. Teachers might have reacted with corporal punishments in response to student’s aggression for the purpose of discipline and guidance, or to efficiently manage the class and demonstrate their authority. However, it should be cautioned that there is no empirical evidence to support it.

Second, the results revealed that teacher’s use of corporal punishment also leads to student’s aggression. It means that even if teachers use corporal punishment as a disciplinary strategy in response to their student’s aggression, this punishment would actually worsen the aggressive problem of students. It is the expected result based on the prior research that found the positive effects of parent’s use of corporal punishments on their children’s aggression (Gershoff, 2002, 2010) However, only little of prior research empirically tested this model with longitudinal data. Gershoff (2010) argued that the dearth of empirical research might have contributed to maintaining the implicit myth among teachers that strong punishments are effective in reducing student’s aggression. Thus, this study would serve as empirical evidence that, contrary to many teachers’ implicit myths, it is not an effective form of disciplinary strategy for dealing with student’s aggression but actually worsens it. Based on the result, teacher education practice should provide anticorporal punishment programs. Teaching the ineffectiveness of corporal punishments to teachers has been found to be effective in some studies (e.g., Robinson et al., 2005).

One of the possible mechanisms from teacher’s corporal punishments to student’s aggressions could be through inflicting the belief that sometimes could be a useful mean to solve a problem (Bandura, 1978; Owen, 2005; Straus, 1991). Although those kinds of theoretical explanations have been suggested by some researchers (e.g., Owen, 2005; Straus, 1991), they have never have been empirically tested. Future research would empirically test those mechanisms with longitudinal data. Some say that the problem is not what kind of disciplinary strategy itself, but what matters is how to use it and in which context, that affects the effectiveness of disciplines (e.g., Baumrind, 1995). It is true that the same corporal punishments could sometimes lead to student’s aggression, but sometimes not. Even sometimes, it could have positive effects in reducing aggression just as the implicit perceptions of teachers. Or some teacher’s use of corporal punishment are relatively more acceptable than that of others because there are some variations in the relation. But what is important is that our data suggest, despite the variations, the negative effects of teacher’s use of corporal punishments offset the null effects or positive effects. In other words, it is more likely that teacher’s use of corporal punishment could cause student’s aggression in general. However, it is managed and from, regardless of the situation or person, corporal punishment from teachers generally worsens the aggression of students.

Although the longitudinal reciprocal relationship between student’s aggression and teacher’s use of corporal punishments were confirmed, the paths from student’s aggressions to teacher’s use of corporal punishments were 1.5 to 2 times bigger than that of the opposite paths. It suggests that the temperament theory model is more competitive in explaining the relationship between student’s aggressions and teacher’s use of corporal punishments in comparison with the social learning theory model. However, it should be interpreted with caution. It is because even if the path from one side is bigger than the other one, both of the paths were statistically significant. Also, the results in this study alone are not enough to conclude or generalize the relative power of the effects. Furthermore, research is needed to clearly identify the relative power of the effects.

Limitation of the Study

This study is not free from the critiques of Paolucci and Violato (2004) that the effect size of (parental) corporal punishment against aggression was too small to interpret meaningfully. Even though the overall model fit of the data was very reasonable, the standardized coefficients between teacher’s use of corporal punishments and student’s aggressions were quite small. But the criteria of effect size are dependent on each research field. When we consider the high stability of aggression (Badaly, Kelly, Schwartz, & Dabney-Lieras, 2013; Huesmann et al., 1984; Sentse, Kretschmer, & Salmivalli, 2015), the small effects of environmental variables, which are relatively changeable, could be academically and practically meaningful. Especially, teacher variables are known to be quite changeable through trainings (Alvarez, 2007). According to prior literature, even the effect sizes of personal characteristic variables on aggression were similar with this study (e.g., Kerig & Stellwagen, 2010). Also, our model included autoregressive paths, which controlled for the previous year corporal punishments and aggressions. These autoregressive paths could offset the total effects, decreasing the variances. Second, we used only two to three items for each latent variable, which is a relatively small number of items. Thus, the latent variables of student’s aggression and teacher’s use of corporal punishments in this study could have different constructs from other studies, which could impede the strict comparisons. It is a limitation of longitudinal panel data. But, according to literature from clinical psychology, the validity of two to four items is well known to be reasonable (e.g., Löwe et al., 2010). Third, the Cronbach’s alphas of student’s aggression were a bit low (.56-.60). Thus, the construct validity of the aggression in this study seems weak. But, it is well known that low Cronbach’s alphas such as below .50 do not severely impede the validity (Schmitt, 1996). Fourth, because we used only self-report methods in measuring latent variables, it is not free from the limitations of the shared method variances problem. Fifth, although we interpreted the effects between corporal punishment and aggression as effects based on causality, we cannot say that it is superior to the causality of experimental data. However, we can say that it is practically the best way available for inferring causality, because of associated ethical problems surrounding experiments with corporal punishment and aggressive behavior.

Conclusion and Implications of the Study

Overall, the results revealed that the relationship between student’s aggression and teacher’s use of corporal punishment was longitudinally reciprocal, both being cause and consequence for each other simultaneously and respectively. It means that our reciprocal model is supported, combining the social learning theory model and the temperament theory model (Bandura, 1978; Muller et al., 1995; Straus, 1991). It is meaningful in that we synthesized the conventional competing models on the relationship between corporal punishments and aggression, which posited the unidirectional effects, one being cause and the other being consequence. Our results could be interpreted as a cycle of violence between student’s aggression and teacher’s use of corporal punishment, in that violence (student’s aggression) leads to another incidence of violence (teacher’s use of corporal punishment s) and this violence again leads to further violence (additional student’s aggression). Although this cycle of violence model could be inferred from the prior literature that supports each model respectively (e.g., Lewis et al., 2005; Straus, 1991), there has not been a study that empirically tested the reciprocal relationship. Thus, future research should rather not assume the unidirectional effects between aggression and corporal punishment, based on our results of the reciprocal relationship. Also, we suggest the importance of exploring the negative teacher variables as well as positive teacher variables that could affect student’s aggression, such as corporal punishment, for more multilateral approach (Lewis et al., 2005). This is because it could suggest the direction of what is to be avoided. Based on this understanding, teachers could arrange appropriate interventions and discipline strategies, while avoiding inappropriate ones. Furthermore, research on other negative teacher variables is needed in the future.

Methodologically, this study expanded the prior literature with the analysis of autoregressive cross-lagged modeling with longitudinal data. Prior research was mainly conducted with cross-sectional data or two time points at best when longitudinal (Gershoff, 2010; Muller et al., 1995). Thus, the results of these were not strong enough to infer the causal directionality between student’s aggression and teacher’s use of corporal punishment. For the strong inferences of causality, experimental data are needed. However, because of the ethical problems, experiments with corporal punishment and aggression are not possible. In these backgrounds, autoregressive cross-lagged modeling with longitudinal data is practically the best way available for inferring the causality (Curran & Bollen, 2001; Paolucci & Violato, 2004). Until now, no empirical study was conducted with autoregressive cross-lagged model (Curran & Bollen, 2001). As Paolucci and Violato (2004) suggested, future research would need more sophisticated analysis with longitudinal data, for the investigation of any potential causal direction between the use of corporal punishment and aggression.

Practically, this study suggests the directions of interventions for either teachers or students to breakout from the cycle of violence. Teachers should keep in mind that corporal punishments, which are at least partially attributable to student’s aggression, actually worsen the problem and leads to a cycle of violence in schools. Thus, teachers should instead respond with alternative disciplinary strategies or direct interventions when dealing with the causes of aggression. Some alternative disciplinary strategies such as social skills training are well known to be effective (Dupper & Dingus, 2008; Goldstein, 1999). To do this, first of all, teachers need to recognize the fact that they have a tendency to use corporal punishments, in response to student’s aggression and try to avoid using it by themselves. In regard to this, Alvarez (2007) found that previous training predicted teachers’ better disciplinary strategies in response to their student’s aggression, which suggest the possibility of positive changes of teachers. The results are meaningful in that it expands the literature and knowledge about which coping strategies teachers tend to use, in response to their student’s aggression. Even though there has been some research about teacher’s reaction to student’s aggression (e.g., Nesdale & Pickering, 2006), most of them just asked teachers about their general reactions with self-reports at a single time point, which has limitations in inferring the causality on the effects of student’s aggression to teacher’s use of corporal punishment.

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.

Author Biography

Boungho Choi is a researcher and police officer at the same time at the Education department of Korean National University. His research interests focus on youth violence, school bullying and victimization, juvenile delinqeuncy and quantitative research methodology.

References

Alvarez

H. K.

(2007). The impact of teacher preparation on responses to student aggression in the classroom. Teaching and Teacher Education, 23, 1113-1126.

Arbuckle

J. L.

(1996). Full information estimation in the presence of incomplete data. In Marcoulides

G. A.

Schumacker

R. E.

(Eds.), Advanced structural equationmodeling: Issues and techniques (pp. 243–277). Mahwah, NJ: Lawrence Erlbaum.

Aucoin

K. J.

Frick

P. J.

Bodin

S. D.

(2006). Corporal punishment and child adjustment. Journal of Applied Developmental Psychology, 27, 527-541.

Badaly

Kelly

B. M.

Schwartz

Dabney-Lieras

(2013). Longitudinal associations of electronic aggression and victimization with social standing during adolescence. Journal of Youth and Adolescence, 42, 891-904.

Bandura

(1978). Social learning theory of aggression. Journal of Communication, 28, 12-29.

Baumrind

(1995). Child maltreatment and optimal caregiving in social contexts. New York, NY: Garland Publishing.

Bollen

K. A.

(1989). A new incremental fit index for general structural equation models. Sociological Methods & Research, 17, 303-316.

Brown

(2009). Perceptions of student misconduct, perceived respect for teachers, and support for corporal punishment among school teachers in South Korea: An exploratory case study. Educational Research for Policy and Practice, 8, 3-22.

Brown

M. W.

Cudeck

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

K. A.

Long

J. S.

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

10.

Bushman

B. J.

Anderson

C. A.

(2001). Is it time to pull the plug on hostile versus instrumental aggression dichotomy? Psychological Review, 108, 273-279.

11.

Buss

A. H.

Perry

(1992). The aggression questionnaire. Journal of Personality and Social Psychology, 63, 452-459.

12.

Chang

(2003). Variable effects of children’s aggression, social withdrawal, and prosocial leadership as functions of teacher beliefs and behaviors. Child Development, 74, 535-548.

13.

Cheung

G. W.

Rensvold

R. B.

(2002). Evaluating goodness-of-fit Indexes for testing measurement invariance. Structural Equation Modeling, 9, 233-255.

14.

Cho

D. S.

(2001). Corporal punishment revisited. Journal of Elementary Education, 14, 137-158.

15.

Choi

Park

Shin

(2016). The effects of aggression and self-blame style on the longitudinal relationship between perceived peer acceptances: A moderated mediation analysis. Asian Journal of Education, 17, 171-194.

16.

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.), New methods for the analysis of change (pp. 107-135). Washington, DC: American Psychological Association.

17.

Dodge

K. A.

Pettit

G. S.

(2003). A biopsychosocial model of the development of chronic conduct problems in adolescence. Developmental Psychology, 39, 349-371.

18.

Dupper

D. R.

Dingus

A. E. M.

(2008). Corporal punishment in US public schools: A continuing challenge for school social workers. Children & Schools, 30, 243-250.

19.

Farrington

D. P.

(1991). Childhood aggression and adult violence: Early precursors and later-life outcomes. In Pepler

D. J.

Rubin

K. H.

(Eds.), The development and treatment of childhood aggression (pp. 5-29). Hillsdale, NJ: Lawrence Erlbaum.

20.

Geen

R. G.

(2001). Human aggression. Buckingham, UK: Open University Press.

21.

Gershoff

E. T.

(2002). Corporal punishment by parents and associated child behaviors and experiences: A meta-analytic and theoretical review. Psychological Bulletin, 128, 539-579.

22.

Gershoff

E. T.

(2010). More harm than good: A summary of scientific research on the intended and unintended effects of corporal punishment on children. Law and Contemporary Problems, 73, 31-56.

23.

Goldstein

A. P.

(1999). Aggression reduction strategies: Effective and ineffective. School Psychology Quarterly, 14, 40-58.

24.

Hecker

Hermenau

Isele

Elbert

(2014). Corporal punishment and children’s externalizing problems: A cross-sectional study of Tanzanian primary school aged children. Child Abuse & Neglect, 38, 884-892.

25.

L. T.

Bentler

P. M.

(1999). Cutoff criteria for fit indexes on covariance structure analysis: Conventional criteria versus new alternatives. Structual Equation Modeling, 6, 1-55.

26.

Huesmann

L. R.

Eron

L. D.

Lefkowitz

M. M.

Walder

L. O.

(1984). Stability of aggression over time and generations. Developmental Psychology, 20, 1120-1134.

27.

Hyman

I. A.

(1995). Corporal punishment, psychological maltreatment, violence, and punitiveness in America: Research, advocacy, and public policy. Applied & Preventive Psychology, 4, 113-130.

28.

Jaffee

S. R.

Caspi

Moffitt

T. E.

Polo-Tomas

Price

T. S.

Taylor

(2004). The limits of child effects: Evidence for genetically mediated child effects on corporal punishment but not on physical maltreatment. Developmental Psychology, 40, 1047-1058.

29.

Kerig

P. K.

Stellwagen

K. K.

(2010). Roles of callous-unemotional traits, narcissism, and machiavellianism in childhood aggression. Journal of Psychopathology and Behavioral Assessment, 32, 343-352.

30.

Kilimci

(2009). Teachers’ perceptions on corporal punishment as a method of discipline in elementary schools. The Journal of International Social Research, 2, 242-251.

31.

Kline

R. B.

(2005). Principles and practice of structural equation modeling (2nd ed.). New York, NY: Guilford Press.

32.

Lewis

(2001). Classroom discipline and student responsibility: The students’ view. Teaching and Teacher Education, 17, 307-319.

33.

Lewis

Romi

Qui

Katz

Y. J.

(2005). Teachers’ classroom discipline and student misbehavior in Australia, China and Israel. Teaching and Teacher Education, 21, 729-741.

34.

Löwe

Wahl

Rose

Spitzer

Glaesmer

Wingenfeld

. . . Brähler

(2010). A 4-item measure of depression and anxiety: Validation and standardization of the Patient Health Questionnaire-4 (PHQ-4) in the general population. Journal of Affective Disorders, 122, 86-95.

35.

Lynne-Landsman

S. D.

Graber

J. A.

Nichols

T. R.

Botvin

G. J.

(2011). Trajectories of aggression, delinquency, and substance use across middle school among urban, minority adolescents. Aggressive Behavior, 37, 161-176.

36.

Mamatey

(2010). South Korean EFL teachers perceptions of corporal punishment in school: Cultural vs. educational system factors. Novitas-Royal (Research on Youth and Language), 4, 190-200.

37.

Muller

R. T.

Hunter

J. E.

Stollak

(1995). The intergenerational transmission of corporal punishment: A comparison of social learning and temperament models. Child Abuse & Neglect, 19, 1323-1335.

38.

Mulvaney

M. K.

Mebert

C. J.

(2007). Parental corporal punishment predicts behavior problems in early childhood. Journal of Family Psychology, 21, 389-397.

39.

Nesdale

Pickering

(2006). Teachers’ reactions to children’s aggression. Social Development, 15, 109-127.

40.

O’Connor

T. G.

Deater-Deckard

Fulker

Rutter

Plomin

(1998). Genotype–environment correlations in late childhood and early adolescence: Antisocial behavioral problems and coercive parenting. Developmental Psychology, 34, 970-981.

41.

Owen

S. S.

(2005). The relationship between social capital and corporal punishment in schools: A theoretical inquiry. Youth & Society, 37, 85-112.

42.

Paolucci

E. O.

Violato

(2004). A meta-analysis of the published research on the affective, cognitive, and behavioral effects of corporal punishment. The Journal of Psychology, 138, 197-222.

43.

Pellegrini

A. D.

(2001). A longitudinal study of heterosexual relationships, aggression, and sexual harassment during the transition from primary school through middle school. Journal of Applied Developmental Psychology, 22, 119-133.

44.

Pitts

S. C.

West

S. G.

Tein

J. Y.

(1996). Longitudinal measurement models in evaluation research: Examining stability and change. Evaluation and Program Planning, 19, 333-350.

45.

Robinson

D. H.

Funk

D. C.

Beth

Bush

A. M.

(2005). Changing beliefs about corporal punishment: Increasing knowledge about ineffectiveness to build more consistent moral and informational beliefs. Journal of Behavioral Education, 14, 117-139.

46.

Scarr

McCartney

(1983). How people make their own environments: A theory of genotype→ environment effects. Child Development, 54, 424-435.

47.

Schafer

J. L.

Olsen

M. K.

(1998). Multiple inputation for multivariate missing-data problems: A data analyst’s perspective. Multivariate Behavioral Research, 33, 545571.

48.

Schmitt

(1996). Uses and abuses of coefficient alpha. Psychological Assessment, 8, 350-353.

49.

Sentse

Kretschmer

Salmivalli

(2015). The longitudinal interplay between bullying, victimization, and social status: Age-related and gender differences. Social Development, 24, 659-677.

50.

Shin

Koh

M. S.

(2005). Korean education in cultural context. Essays in Education, 14, 1-10.

51.

Straus

M. A.

(1991). Discipline and deviance: Physical punishment of children and violence and other crime in adulthood. Social Problems, 38, 133-154.

52.

Straus

M. A.

(1994). Beating the devil out of them: Corporal punishment in American families. San Francisco, CA: Jossey-Bass/Lexington.

53.

Weaver

C. M.

Borkowski

J. G.

Whitman

T. L.

(2008). Violence breeds violence: Childhood exposure and adolescent conduct problems. Journal of Community Psychology, 36, 96-112.

54.

Youssef

R. M.

Attia

M. S. E. D.

Kamel

M. I.

(1998). Children experiencing violence II: Prevalence and determinants of corporal punishment in schools. Child Abuse & Neglect, 22, 975-985.