Anger as a predictor of suicidal ideation in middle-school students in Korea: gender difference in threshold point.

Lee, Jongeun ; Choi, Heeseung ; Kim, Mi Ja 等

INTRODUCTION

Suicide is an overarching social concern that affects all age groups. In recent years, there has been growing concern regarding the increasing rates of suicidal tendencies among adolescents. The National Youth Risk Behavior Survey (YRBS) in the United States showed that 8.4% of high school students attempted suicide one or more times in the previous 12 months (Centers for Disease Control and Prevention, 2007). In Korea, 11.6% of middle and high school students reported attempting suicide, which is a higher rate than that for their U.S. counterparts (H.S. Kim & Kim, 2008). Suicide was one of the two most prevalent causes of death in Korean adolescents, and the first leading cause of death among girls in 2006 (Korea National Statistical Office, 2007).

Persons with suicidal tendencies are categorized into three groups: those with suicidal ideation, those who attempted suicide, and those who have completed suicide (Linehan, 1986). In the YRBS survey, 16.9% of participants (21.8% of girls and 12.0% of boys) reported seriously considering suicide in the past 12 months (Centers for Disease Control and Prevention, 2007). The prevalence of suicidal ideation was higher than the rate of suicide attempt (Centers for Disease Control and Prevention, 2007), and suicidal ideation may be the earliest recognition o-f suicidal tendencies (Keane, Dick, Bechtold, & Manson, 1996). This paper focuses on suicidal ideation, which may lend itself to being an indicator of when to initiate actions most likely to prevent an actual suicide attempt.

Risk factors for suicidal tendencies among adolescents have been reported by numerous studies. These include the relationships between various personality and temperamental factors and adolescents' susceptibility to suicidal tendencies such as anxiety, depression, anger, aggression, impulsivity, and helplessness (Conner, Meldrum, Wieczorek, Duberstein, & Welte, 2004; Fennig et al., 2005; Horesh, Orbach, Gothelf, Efrati, & Apter, 2003; Park, Schepp, Jang, & Koo, 2006). Social, familial, and environmental factors were also studied to determine the extent of their influence on suicide risk in adolescents (Heo, 2007; Park et al., 2006). A lack of social support is a significant predictor of suicidal ideation, whereas strong family support protects adolescents from suicidal ideation (Heo, 2007). Park et al. found that parental divorce and parental alcohol abuse are significant predictors of suicidal ideation, particularly for boys. Sexuality, body image, and social challenges may result in further stress to the adolescent that could lead to a state of depression, which is the most common cause of suicide (Giddens, 2007). Experience at school is another significant factor determining the vulnerability to suicide. Dealing with school-related stress including building and maintaining close peer relationships, completing day-to-day school work, getting good grades, and being accepted in prestigious colleges places adolescents at increased risk for suicidal ideation (Giddens, 2007).

Among those identified risk factors, anger has long been considered crucial in the development of suicidal tendencies (Goldney, Winefield, Saebel, Winefield, & Tiggeman, 1997). If a person is not allowed to or is unable to express anger outwardly, anger turned inwards may result in depression, guilt, shame, anxiety or lethargy (Tavris, 1989). Adolescents encounter many anger-producing situations, such as disappointment, pain, and frustration. A major problem for adolescents is lack of skills and resources required to manage and express anger in acceptable ways (Jones, Peacock, & Christopher, 1992). Adolescents may be unable to express anger to persons in authority positions, such as mothers, fathers or teachers. This is particularly true in Korean culture, which considers suppressing anger as a good deed or a virtue. This can be seen in Chon's report (1999) that "loved ones" as the target of anger was quite low in Korea (2%) compared to the U.S. (33%). Similarly, anger directed at "an acquaintance" was 5% in Korea, while it was 26% in the U.S. Hence, Korean adolescents display a passive pattern of anger, such as using curse words in their routine conversation with friends. Unresolved anger toward the aforementioned negative experiences intensifies gradually and thus makes adolescents vulnerable to suicidal tendencies (Goldney et al., 1997). These findings indicate that anger plays an important role in suicidal tendencies.

Gender also seems to play an important role in the relationship between anger and suicidal tendencies for adolescents, as reported by several investigators (Ang & Ooi, 2004; Lerner, 1988; Park et al., 2006; Sanford & Donovan, 1985; Swahn & Bossarte, 2007). Some theorists (Lerner, 1988; Sanford & Donovan, 1985) suggest that boys and girls may experience and express anger in different ways, making the study of gender difference in anger an important area for investigation. There are theoretical reasons to expect gender difference in expression of anger. Gilligan's (1993) theory of moral development includes the assertion that people use both justice and care in interpersonal relationships, but men are more likely to use the justice principle and women the care principle (Gilligan & Attanucci, 1988). Girls are more likely to feel guilty after aggressive acts, and boys expect less parental disapproval for their aggression (Perry, Perry, & Weiss, 1989). In spite of the known gender differences in anger expression (Yarcheski, Mahon, & Yarcheski, 2002), no study was available that explored gender differences in the relationship between anger and suicidal ideation.

Considering the fact that other risk factors for suicidal tendencies, such as depression, lack of self-esteem, alcohol abuse, and disruptive behavior, are known to be gender skewed (Kelly, Lynch, Donovan, & Clark, 2001; Wichstrom & Rossow, 2002), it is worth examining whether gender differences exist in the relationship between anger and suicidal ideation. Furthermore, most of the available studies on the relationship between suicide and anger focus on linear relationships, which limits a grasp of the whole picture of the relationships and makes it difficult to identify high-risk groups. Hence, understanding gender differences in this complex relationship would help in the development of gender-specific anger management or suicide preventive programs for adolescents. Therefore, the present study aims to assess the gender differences in: (1) the level of suicidal ideation and anger; (2) predictors of suicidal ideation; and (3) the varying association between suicidal ideation and anger.

METHOD

The target population for this cross-sectional, correlational study was adolescents aged 13 to 15 years who attended middle schools in D City, South Korea. This age group was selected because suicidal ideation rates peak during mid-adolescence (Rueter & Kwon, 2005).

Sampling and Data Collection Procedures

Two sampling strategies were used to identify prospective participants for this study: (1) a proportional stratified random sampling strategy and (2) a convenience sampling strategy. Using a proportional stratified random sampling strategy, 11 middle schools were selected in D City, South Korea. Next, the convenience sample of adolescents in grades 1-3 (equivalent to 6th to 8th grades in the U.S.) was obtained from the selected schools. The data were collected in September 2007.

Researchers and trained assistants contacted school nurses of the selected schools to discuss the study's purpose, significance, and procedure. All nurses who were contacted agreed to assist in recruiting students. First, the nurses presented this proposal to school boards, which consisted of a principal, vice principal, and parents' representatives. After obtaining permission from school boards, the nurses provided a list of classes that agreed to participate in the study and their timetables for recess. The investigators visited the students during recess in the morning and explained the purpose and significance of the study. They assured students that their names and other identifying data would be kept confidential and informed them of their right to withdraw from the study at any time. The investigators then obtained assents from those students who agreed to participate in the study. This is an accepted procedure in Korea. Of 480 students 382 (79.6%) agreed to participate. Each participant filled out a questionnaire that inquired about demographic characteristics, anger, and suicidal ideation. Among the 382 students, 349 returned the questionnaire (91.4%). Data from 258 students who completed the survey were used for data analysis.

Demographic characteristics. Information was obtained on age, gender, school life satisfaction, academic achievement, economic status, and health status. Self-rated academic performance, school life satisfaction, and economic status were recorded into three categories: high, middle, and low. Health status was categorized into three levels: well, usual, and weak.

Anger. Anger was measured with the Korean version of the Multidimensional Anger Inventory (MAI) developed by Siegel (1986) and translated by Lee (1990). The 36-item scale was designed to measure different dimensions of anger, such as frequency, duration, and magnitude, as well as situations that cause anger and the mode of expression. In this study, only the total anger score was used for analysis; other aspects of anger will be discussed elsewhere. Responses ranged from 1 (completely undescriptive of your anger)to 5 (completely descriptive of your anger). The MAI has shown good reliability (Cronbach's [alpha] = .87) in a previous study (Horesh et al., 2003) and in the present study (Cronbach's [alpha] = .88).

Suicidal ideation. Suicidal ideation was measured using the Korean version of the Suicidal Ideation Questionnaire (SIQ), which was developed by Reynolds (1988) and translated by Shin (1992). This 30-item scale assesses frequency of suicidal ideation in the past month. Examples of items include "I wished I were dead" and "I thought that no one cared if I lived or died." Items on the SIQ are scored on a 7-point Likert-type scale, ranging from "almost every day" to "never." A higher score indicates a stronger ideation of suicide. The SIQ has demonstrated excellent reliability (Cronbach's [alpha] = .96) in a previous study (Ang & Ooi, 2004) and in the present study (Cronbach's [alpha] = .97). The SIQ was used to identify adolescents experiencing clinical levels of suicidal ideation using the recommended cutoff score of 41 (Reynolds, 1988).

Statistical Analysis

Descriptive statistics were used to explain the characteristics of the participants. Cronbach's alpha coefficients were calculated for each scale. To address the specific aims, we used descriptive statistics, multiple linear, and spline regression analysis. For specific aim 1, to compare the levels of suicidal ideation and anger, descriptive statistics and an unpaired t test were used. For specific aim 2, to determine significant predictors of suicidal ideation, multiple linear regression analysis and a specification test were used. For specific aim 3, to identify the varying association between suicidal ideation and anger, Davies's test (Davies, 1987) was used. Data were analyzed using the SPSS/WIN 12.0, STATA 10.0 and R-segmented programs.

RESULTS

Demographic Characteristics of the Sample

A total of 349 students aged 13 to 15 years participated in the study, but only 258 were used for data analysis because 91 (26.1%) had incomplete data. Of the adolescents who completed the survey, 160 (62%) were boys and 98 (38%) were girls. Participants' demographic data are presented in Table 1.

Gender Differences in the Level of Suicidal Ideation and Anger

As seen in Table 2, the level of suicidal ideation differed significantly between boys and girls. The mean score for girls was significantly higher than that of boys (p < .05). When the adolescents were dichotomized into high- and low-risk groups using a standardized cutoff point (SIQ = 41; Reynolds, 1988), 22.5% of girls were in the high-risk group, versus only 11.4% of boys. Girls also reported significantly higher anger scores than did boys (p = .05).

Gender Differences in the Predictors of Suicidal Ideation

While anger was a significant predictor for suicidal ideation only in boys, both school life satisfaction and anger were significant predictors of suicidal ideation in girls. Coefficients for anger among boys and girls were 0.5987 and 0.7711, respectively. The boys' model estimation did not show any diagnostic errors (both tests were not significant), while the girls' model estimation showed specification error (F value = 3.71, p value = 0.0145, df (3, 90). Table 3). The findings indicate a nonlinear relationship between suicidal ideation and anger among girls.

Varying Association between Suicidal Ideation and Anger

To handle the specification error due to linear relations between predictors and suicidal ideation in the girls' model, the linear spline model (threshold regression) was adopted. For the threshold model, the existence and significance of the threshold points for boys' and girls' groups were determined with the Davies test (Davies, 1987). Whereas no significant threshold point was identified for the boys, a significant (p = 0.0059) threshold point at 117.67 was identified for girls, as shown in Figure 1. At the threshold anger point of 117.67, 12.5% of the girls (n = 13) belonged to the high-risk group.

DISCUSSION

Using the descriptive correlational design, this study examined the gender differences in: (1) the level of suicidal ideation and anger; (2) predictors for suicidal ideation; and (3) the varying association between suicidal ideation and anger. It was found that girls' scores were significantly higher on suicidal ideation than were boys. When the SIQ cutoff score of 41 (Reynolds, 1988) was used to judge the severity of suicidal ideation, 22.4% of the girls and 11.4% of the boys were identified as being at risk of suicide. This is similar to the finding for adolescents in the U.S., where 21.8% of girls and 12.0% of boys were reported to have serious suicidal ideation (Centers for Disease Control and Prevention, 2007). Regarding the attempted suicide rate, high school girls were more than twice as likely as boys to attempt to kill themselves (Sperekas, 2000).

With respect to the gender differences in anger, contradicting results have been reported (Jones et al., 1992; K. S. Kim, 2003: Yarcheski et al., 2002). Yarcheski et al's (2002) findings indicated that boys and girls did not differ in anger, whereas Jones et al.'s (1992) and K. S. Kim's studies (2003) indicated that girls had significantly higher anger scores than did boys. These latter results are consistent with our study. A possible explanation for the higher level of anger among girls is that girls may experience anger when they perceive latent discrimination against women in the Korean culture and social systems (Chae, 2001).

Anger is an emotion with interpersonal consequences, both to oneself and others, and its subjective aftereffects are overwhelmingly negative (Averill, 1982). In the present study, a strong positive correlation between anger and suicidal ideation was observed in both boys and girls. Our findings support previous reports that showed a link between anger and suicidal ideation (Goldney et al., 1997; Kotler, Iancu, Efroni, & Amir, 2001; Stein, Apter, Ratzoni, Har-Even, & Avidan, 1998). A high level of anger is an important concurrent marker for adolescents who may be at risk for engaging in suicidal behaviors.

Another identified predictor of suicidal ideation was satisfaction with school life. The present study found that girls were less likely to be satisfied with school life; a lower level of satisfaction with school life was found to be a significant predictor of suicidal ideation in girls.

[FIGURE 1 OMITTED]

There have not been similar studies on the association between satisfaction with school life and suicidal ideation, although Park et al.'s (2006) findings indicated that the level of life satisfaction for girls was slightly lower than for boys, and life satisfaction is a significant predictor of suicidal ideation. This lower level of school satisfaction among girls may be related to the fact that girls in traditional Korean culture have less power and fewer resources than do boys (Park et al., 2006). In order to identify factors which determine the degree of school life satisfaction among girls, further studies are required.

The multiple linear spline regression analysis in the present study, showed that anger had a threshold point for suicidal ideation among girls. This indicated that the risk of suicidal ideation increases considerably beginning at the threshold point. About 12.5% of girls in the present study fell above the threshold point, which was regarded as high risk. The findings suggest that girls are at increased risk for experiencing both suicidal ideation and anger.

It is important to identify anger-provoking situations for girls and teach them how to manage and express their anger in certain situations. Asian culture discourages the expression of anger through direct and aggressive acts in public. A person who expresses anger is regarded as one who has lost control, and therefore lost face (Tavris, 1989). Thus, a significant challenge for most Korean adolescents is to learn how to manage and express anger in culturally acceptable ways (Jones et al., 1992).

These findings support the utility of gender-specific suicide assessment, prevention, and intervention strategies. Preventive interventions have been classified as universal, selective, or indicated depending on the level of risk (Gordon, 1983). Universal preventive interventions include school- or community-based programs designed for all regardless of level of risk. In Korea, it is not feasible for teachers to counsel individual adolescents in anger management due to lack of resources and time. In fact, these limitations do not allow school nurses to place a high priority on suicide prevention. Hence, school-based programs are needed which prepare nurses and teachers to become involved in anger management even when resources are limited.

Selective preventive interventions have been developed for a subgroup of people who exhibit higher level of risk for suicide. Such interventions are known to be cost-effective and efficient (Jones et al., 1992). The findings of the present study have demonstrated the need for developing selective preventive interventions for girls who exhibit high levels of anger. Such interventions should reflect the unique features of anger expression among girls and teach them how to express anger in acceptable ways. These would focus on problem-solving and communication skills. Considering that girls reported low levels of school satisfaction, which is a significant risk factor for suicidal ideation, it is important to provide appropriate school-based services for girls.

The present has several limitations. The first is the self-report bias. Several difficulties are inherent in the assessment of suicidal ideation. Although the 30-item SIQ has adequate reliability and validity, the range of boys' anger scores was wide. This finding may indicate that the anger measurement could have been underreported by boys, who may not have taken the survey as seriously as one would hope. The second is that all participants were from one city in South Korea. Thus recruiting adolescents from different areas, including rural regions on a random basis could increase the generalizability of the findings.

Despite these limitations, results of this study have significant clinical implications. Adolescents experiencing high levels of anger are more likely than adolescents with lower levels of anger to report suicidal ideation. Interestingly, the girls who were above the threshold point of anger score manifested a steep increase in suicidal ideation. Further, this trend has been consistent with the recommended cutoff score for the suicidal risk group. Based on these findings, it is recommend that suicide prevention programs focusing on anger management be offered to adolescents who exhibit early signs of anger problems and suicidal tendencies.

REFERENCES

Ang, R. P., & Ooi, Y. P. (2004). Impact of gender and parents' marital status on adolescents' suicidal ideation. The International Journal of Social Psychiatry, 50, 351-360.

Averill, J. R. (Ed.). (1982). Anger and aggression: An essay on emotion. New York: Springer-Verlag.

Centers for Disease Control and Prevention. (2007). YRBSS: Youth Risk Behavior Surveillance System. Youth Online: Comprehensive Results. Retrieved March 6, 2007 from http://apps.nccd.cdc.gov/yrbss/

Chae, Y. K. (2001). Verification on the model of expression modes and moderating effect of adolescents' anger. Unpublished doctoral dissertation, Chonnam National University, Gwangju, South Korea.

Chon, K. K. (1999). A preliminary study on anger. Journal of Rehabilitation Psychology, 6, 173-190.

Conner, K. R., Meldrum, S., Wieczorek, W. F., Duberstein, P. R., & Welte, J. W. (2004). The association of irritability and impulsivity with suicidal ideation among 15- to 20-year-old males. Suicide & Life-Threatening Behavior, 34, 363-373.

Davies, R. B. (1987). Hypothesis testing when a nuisance parameter is present only under the alternative. Biometrika, 74, 33-43.

Fennig, S., Geva, K., Zalsman, G., Weizman, A., Fennig, S., & Apter, A. (2005). Effect of gender on suicide attempters versus nonattempters in an adolescent inpatient unit. Comprehensive Psychiatry, 46, 90-97.,

Giddens, S. (Ed.). (2007). Suicide: Coping in a changing world. New York: Rosen Publishing.

Gilligan, C. (Ed.). (1993). In a different voice: Psychological theory and women's development. Cambridge, MA: Harvard University Press.

Gilligan, C., & Attanucci, J. (1988). Two moral orientations: Gender differences and similarities. Merrill-Palmer Quarterly, 34, 223-237.

Goldney, R., Winefield, A., Saebel, J., Winefield, H., & Tiggeman, M. (1997). Anger, suicidal ideation, and attempted suicide: A prospective study. Comprehensive Psychiatry, 38, 264-268.

Gordon, R. S. Jr. (1983). An operational classification of disease prevention. Public Health Reports, 98, 107-109.

Heo, J. S. (2007). The effect of a school type and psychological environment on high school students' suicide ideation. Unpublished master's thesis. Catholic University of Daegu, Daegu, South Korea.

Horesh, N., Orbach, I., Gothelf, D., Efrati, M., & Apter, A. (2003). Comparison of the suicidal behavior of adolescent inpatients with borderline personality disorder and major depression. The Journal of Nervous and Mental Disease, 191, 582-558.

Jones, M. B., Peacock, M. K., & Christopher, J. (1992). Self-reported anger in black high school adolescents. Journal of Adolescent Health, 13, 461-465.

Keane, E. M., Dick, R. W., Bechtold, D. W., & Manson, S. M. (1996). Predictive and concurrent validity of the Suicidal Ideation Questionnaire among American Indian adolescents. Journal of Abnormal Child Psychology, 24, 735-747.

Kelly, T. M., Lynch, K. G., Donovan, J. E., & Clark, D. B. (2001). Alcohol use disorders and risk factor interactions for adolescent suicidal ideation and attempts. Suicide & Life-Threatening Behavior, 31, 181-193.

Kim, H. S., & Kim, H. S. (2008). Risk factors for suicide attempts among Korean adolescents. Child Psychiatry and Human Development, 39, 221-235.

Kim, K. S. (2003). The relationship among adolescents' anger, anger coping strategies, and school adjustment. The Korean Journal of Educational Psychology, 17, 393-410.

Korea National Statistical Office. (2007). Cause of death in Korean population at year 2006. Korea Government, Seoul, Korea.

Kotler, M., Iancu, I., Efroni, R., & Amir, M. (2001). Anger, impulsivity, social support, and suicide risk in patients with posttraumatic stress disorder. The Journal of Nervous and Mental Disease, 189, 162-167.

Lee, B. K. (1990). Effect of biofeedback training on the reduction of blood pressure in essential hypertensives. Unpublished doctoral dissertation, University of Seoul, Seoul, South Korea.

Lerner, H. G. (Ed.). (1988). Women in therapy. Northvale, NJ: Aronson. Linehan, M. M. (1986). Suicidal people. One population or two? Annals of the New York Academy of Sciences, 487, 16-33.

Park, H. S., Schepp, K. G., Jang, E. H., & Koo, H. Y. (2006). Predictors of suicidal ideation among high school students by gender in South Korea. The Journal of School Health, 76, 181-188.

Perry, D., G., Perry, L. C., & Weiss, R. J. (1989). Sex differences in the consequences that children anticipate for aggression. Developmental Psychology, 25, 312-319.

Reynolds, W. M. (1988). Suicide Ideation Questionnaire: Professional Manual. Odessa, FL: Psychological Assessment Resources.

Rueter, M. A., & Kwon, H. K. (2005). Developmental trends in adolescent suicidal ideation. Journal of Research on Adolescence, 15, 205-222.

Sanford, L. T., & Donovan, M. E. (Eds.). (1985). Women and self-esteem. London: Penguin.

Shin, M. S. (1992). An empirical study of the mechanism of suicide: Validation of the scale for escape from the self. Unpublished doctoral dissertation, University of Yonsei, Seoul, South Korea.

Siegel, J. M. (1986). The multidimensional anger inventory. Journal of Personality and Social Psychology, 51, 191-200.

Sperekas, N. B. (Ed.). (2000). Suicide Wise: Taking steps against teen suicide. Berkeley Heights, NJ: Enslow Publishers.

Stein, D., Apter, A., Ratzoni, G., Har-Even, D., & Avidan, G. (1998). Association between multiple suicide attempts and negative affects in adolescents. Journal of the American Academy of Child and Adolescent Psychiatry, 37, 488-494.

Swahn, M. H., & Bossarte, R. M. (2007). Gender, early alcohol use, and suicide ideation and attempts: Findings from the 2005 youth risk behavior survey. Journal of Adolescent Health, 41, 175-181.

Tavris, C. (Ed.). (1989). Anger: The misunderstood emotion (Rev. ed.). New York: Touchstone.

Wichstrom, L., & Rossow, I. (2002). Explaining the gender difference in self-reported suicide attempts: A nationally representative study of Norwegian adolescents. Suicide & Life-Threatening Behavior, 32, 101-116.

Yarcheski, A., Mahon, N. E., & Yarcheski, T. J. (2002). Anger in early adolescent boys and girls with health manifestations. Nursing Research, 51, 229-236.

Jongeun Lee, Mi Ja Kim, and Chang Gi Park, College of Nursing, University of Illinois at Chicago; Dong-soo Shin, Division of Nursing, Hallym University College of Medicine, South Korea.

Requests for reprints should be sent to Heeseung Choi, College of Nursing, University of Illinois at Chicago, 845 S. Damen Ave. Chicago, IL 60612. E-mail: hchoi20@uic.edu

Table 1. Demographic Characteristics of Subjects

Factors Total Males Females

 n % n % n %
School life satisfaction
High 69 26.7 52 32.5 17 17.3
Medium 158 61.3 91 56.9 67 68.4
Low 31 12.0 17 10.6 14 14.3
Academic achievement
High 72 28.3 42 26.4 30 31.3
Medium 138 54.1 85 53.5 53 55.2
Low 45 17.6 32 20.1 13 13.5
Economic status
High 31 12.0 24 15.0 7 7.1
Medium 206 79.9 123 76.9 83 84.7
Low 21 8.1 13 8.1 8 8.2
Health status
Well 165 64.0 115 71.8 50 51.0
Usual 85 32.9 42 26.3 43 43.9
Weak 8 3.1 3 1.9 5 5.1
Total 258 100.0 160 62.0 98 38.0

Note: Because of missing data, the total number of responses varies.

Table 2
Mean Difference of Suicidal Ideation and Anger by Gender

 Males Females

Variables (N = 160) (N = 98) t (P)

 Mean (SD) Mean (SD)

Suicidal ideation 20.09 (27.35) 31.82 (35.48) -2.98 (0.0032)
Anger 87.53 (20.01) 92.86 (21.68) -1.97 (0.0500)

Table 3. Variances in Suicidal Ideation Explained by Demographic
Variables and Anger for Boys and Girls

 Multiple linear regression

 Female

School life satisfaction -18.8278 **

Economic status -5.5716

Health status -1.4361

Anger .7711 **

Anger (2nd a slope after
117.67 for female and
95.4 for male
Constant -12.7105 ***

Adjusted [R.sup.2] 0.3851

F-test F(4,93) = 16.18

(p value) (p [less than or equal to] .0001)

Specification test F(3, 90) = 3.71
 (p = .0145)

 Multiple linear regression

 Male

School life satisfaction -1.1919

Economic status 2.7223

Health status 3.2549

Anger .59872 **

Anger (2nd a slope after
117.67 for female and
95.4 for male

Constant -39.3441 **

Adjusted [R.sup.2] 0.1663

F-test F(4,155) = 8.93

(p value) (p [less than or equal to] .0001)

Specification test F(3, 152) = 0.41
 (p = .7489)

 Multiple linear-spline regression
 Female

School life satisfaction -17.6906 **

Economic status -6.2811

Health status -1.947

Anger .4131 **

Anger (2nd a slope after 1.8661 **
117.67 for female and
95.4 for male
Constant 17.2940 *

Adjusted [R.sup.2] 0.4376

F-test F(5, 92) = 16.10

(p value) (p < .0001)

Specification test F(3, 89) = 0.96
 (p = .4163)