Prevalence and Trends in Domestic Violence in South Korea

Abstract

To examine trends in the prevalence of domestic violence since 1997, 1 year prior to the introduction of legislative countermeasures and accompanying services in South Korea, and to analyze what socio-demographic characteristics of perpetrators contribute to spousal violence and whether there were any changes in risk factors over time. This study used two sets of nationally representative household samples: married or cohabiting couples of 1,540 from the 1999 national survey and 3,269 from the 2010 National Survey of Domestic Violence. Frequency analysis was used to measure the prevalence of intimate partner violence (IPV), and cross-tabulation, correlation, and logistic regression analyses were used to look for socio-demographic risk factors of spousal physical violence and patterns of change over time. The frequency analysis showed that the IPV prevalence dropped by approximately 50%, from 34.1% in 1999 to 16.5% in 2010, though it was still higher than many other countries. The cross-tabulation and logistic regression analyses suggested that men with low socio-demographic characteristics were generally more violent, though this tendency did not apply to women. Instead, younger women seemed to be more violent than older women. Last, different levels of household income were associated with different levels of IPV in 2010, but no linear trend was detected. In this study, IPV prevalence trends and risk factors of two different time periods were discussed to provide implications for tackling the IPV problem. Future countermeasures must build on understanding about men with low socio-demographic status and younger women, who were more violent in marital relationships.

Keywords

intimate partner violence prevalence socio-demographic risk factors South Korea

Intimate partner violence (IPV) arises within social contexts, reflecting family structures, socio-psychological factors, and the social environment, as well as biological traits (Jewkes, 2002; J. Kim, 2007). In the Korean society under the residual influence of a patriarchic Confucian culture, IPV was used to be tolerated as an instrument to maintain family order (J. Kim, Park, & Emery, 2009; Shim & Nelson-Becker, 2009). Specifically, the Confucian culture conferred absolute authority to the patriarch of a family, that is, the male head, and resistance of wives was unacceptable for the sake of family harmony (B. Lee, 2006). Moreover, communalism prioritized communal values over personal ones, which forced wives to be obedient to maintain the family hierarchy. These cultural influences made older women to remain silent and powerless during a family conflict and less likely to seek outside help in the event of IPV (Shim & Nelson-Becker, 2009). With growing concerns about women’s victimization, the first national survey of domestic violence was conducted in 1997. The national survey reported that an annual prevalence rate of physical violence against women was 27.9% higher than any other developing country, such as Egypt (13% in 1995/6), Bangladesh-city (19% in 2001), India (10% in 1998/2000), and Peru (provinces; 25% in 2000; United Nations Department of Economic and Social Affairs, 2010).

Following this alarming report, domestic violence was suddenly considered as one of the most urgent issues in South Korea, requiring national efforts to resolve the problem. As a result, starting with the implementation of the Prevention of Domestic Violence and Victim Protection Act and the Special Act for the Punishment of Domestic Violence on July 1, 1998, multilateral efforts took place throughout the first 10 years in the twenty-first century. During that time period, the laws and policies on domestic violence went through many revisions to strengthen legal protections of victims and to expand preventive measures, such as education (Y. R. Park, 2007). After 15 years of nationwide efforts, it is about time to assess the current domestic violence situation so that we may evaluate the effectiveness of previous efforts and provide a blueprint to counter domestic violence for the decades to come.

Similar experiences were observed in the United States. The first National Family Violence Survey in 1975 awakened the United States to understand the gravity of this situation (Straus & Gelles, 1986). In 1975, the annual rate of physical violence against female partners was 16%, followed by a 27% decrease in 1985, a disappointing figure given the introduction of an arrest policy for misdemeanor IPV (Sherman & Cohn, 1989; Straus & Gelles, 1986; Straus, Gelles, & Smith, 1990). Afterward, the United States began to take a different approach by placing more emphasis on the preventive measures, eventually enacting the Violence Against Women Act in 1994. The Act stipulated diverse preventive measures to reduce violence against women including education and training against violence and record keeping and intensified punishments for perpetrators (C. M. Lee, 2004). In 1995 and 1996, the National Violence Against Women (NVAW) survey showed that the IPV rates were as low as in between 0.5% and 1%, which remained relatively unchanged until the 2000s (Breiding, Black, & Ryan, 2008). Unfortunately, this trend could not be substantiated because of the differences in research methodologies (Tjaden & Thoennes, 2000).

As we can learn from the case of the United States as well as from our own, it is crucial to use a consistent and reliable methodology across surveys of different time periods to make a conclusive analysis on the prevalence of domestic violence. Previous studies in South Korea were subject to similar limitations, that is, a lack of reliable longitudinal studies over different periods and an inconsistent use of instruments across studies. To circumvent these limitations, this article presents the IPV prevalence trend from the first national survey in 1997 undertaken prior to the introduction of domestic violence laws to the two surveys in 1999 and 2000, right after the introduction of the laws to the latest national survey of domestic violence conducted in 2010. All these surveys were administered by the Yonsei University Family Welfare team by using an identical instrument of a modified CTS scale to measure IPV in South Korea.

By performing a comparative analysis on national survey results, we intend to examine whether the IPV prevalence has responded after the collaborative efforts from public and private sectors. And then we try to explore the relationships between socio-demographic characteristics on IPV in which conflicting evidences exist about the relationships.

Legal and Institutional Actions in the Past 15 Years

Only two decades ago, there was tacit consensus that domestic violence was a private matter that the outside world should not meddle in (J. Kim, 1998). However, the spread of a series of severe domestic violence cases via mass media turned the tide of social recognition of domestic violence, and consequently the two aforementioned laws were implemented by laying the legal groundwork for the government to step in family matters (H. J. Lee, 2008). The two legislative acts were the Special Act for the Punishment of Domestic Violence, which stipulated special criminal punishment of offenders, and the Prevention of Domestic Violence and Victim Protection Act, which focused on protection for victims (W. J. Kim, 2012). The former legislative act was introduced to provide more protection for the victims by keeping offenders away from them, and the latter was enacted to protect and support victims of domestic violence by aiming the victims to become self-reliant, which was especially difficult for women in a society with many obstacles in labor markets.

Along with the legal and institutional efforts, the effective operation of the supporting system hinged on how the diverse services can be delivered to victims. For the purpose, domestic violence counseling centers began to be established since 1998 under the Prevention of Domestic Violence and Victim Protection Act. The primary roles of the counseling centers were to receive reports of domestic violence incidences, to offer counseling services, and to transfer clients who need immediate protection to relevant institutions or shelters. In addition to services for victims, some centers were assigned to provide counseling and therapeutic services to domestic violence offenders for the prevention of violence from recurring, which were found very effective from numerous studies. For instance, J. Kim, Lee, and Lee (2007) found that 85.3% of program participants ceased to use physical violence immediately after the program, and the rate remained at 71.4% when 375 wives of batterers were asked after 1 to 3 years after the programs. Moreover, in their experimental design studies, diverse programs, such as dialogue method education program (Choi, Yoo, Hong, & Chung, 1999), group therapy programs (E. J. Lee, 2004; S. W. Lee, 2003), and counseling programs (D. H. Park, 2006) led to significant reductions in recidivism.

Once nonexistent, the number of counseling centers began to rise dramatically in the first several years, as the government encouraged the opening of the centers (M. J. Lee, Byun, Hwang, & Lee, 2008). In 1998, only 26 centers were opened, which increased to 82 and 120 in the subsequent 2 years. The number peaked in 2006, reaching 343, and then decreased and stabilized at about 230 centers as complementary institutions such as healthy family supporting centers and multi-ethnic supporting centers share some of the responsibilities (see Table 1; Korea Institute for Health and Social Affairs, 2002; M. J. Lee et al., 2008; Ministry of Gender Equality and Family, 2004-2012). These centers have been major components of the private and governmental efforts to deter domestic violence.

Table 1.

The Number of Domestic Violence Counseling Centers From 1998 to 2012 (Unit: %).

Year	Number	Year	Number	Year	Number
1998	26	2003	136	2008	277
1999	82	2004	136	2009	275
2000	120	2005	268	2010	248
2001	142	2006	343	2011	255
2002	151	2007	302	2012	231

Note. The number of centers was not separately provided for 2003 and 2004 by the Ministry of Gender Equality and Family.

Socio-Demographic Characteristics and IPV

We presume that men with low socio-demographic status are more physically violent in South Korea because men with low socio-demographic characteristics have motivation to control their wives with force especially in the cultural context where men are evaluated by their socio-demographic ability. It is undeniable that more women are now participating in the labor market and egalitarian views are accepted by more people, but a gender norm such that men are asked to be breadwinners and women to be homemakers remains strong (S. Kim & Kim, 2007). In their research about the association of relative income levels between husband and wives with the division of housework hours, S. Kim and Kim (2007) found that women with higher income spent more time in housework to compensate the reversed division of labor, which was in line with the theory of gender compensation. Under this traditional division of labor between men and women, men with low socio-demographic characteristics are more likely to use violence because of inferiority complex. According to J. Kim and Lee (1997) in which 423 people receiving public assistance were studied, inferiority complex, especially in terms of income, work position, and family background, was significantly related with husband-to-wife physical violence. In short, men with low socio-demographic characteristics are more likely to have inferior complex under the cultural influences imposing traditional division of labor and thus to have motivation to control wives with violence while female counterpart are not in the same situation.

Unlike other socio-demographic characteristics, age is expected to be related differently to IPV prevalence between males and females. Compared with younger women with more egalitarian gendered view, older women influenced by the Confucian gender role attitudes are less violent during family conflicts (B. Lee, 2006; Shim & Nelson-Becker, 2009). However, the direction of the effect is not clear for men because older men may use more violence to maintain the patriarchy while younger men are also likely to be situationally violent at the early stage of marriage. This gendered tendency was empirically supported by J. Kim (2007).

Hypotheses

Based on the discussion on the public and private efforts to overcome the IPV problems and the influences of socio-demographic characteristics in predicting IPV over gender, we make the following hypotheses.

Hypotheses 1: The prevalence rates of physical violence by male and female intimate partners have decreased over the last 15 years.

Hypotheses 2: Men with low socio-demographic characteristics perpetrate more physical violence against their female partners, while this pattern does not apply to physical violence by female partners.

Hypotheses 3: The younger women are the more likely they are to use physical violence against their male partners while this relationship does not apply to physical violence by men.

Method

Participants

The main data for this study were collected from the 2010 National Survey of Domestic Violence, which was conducted from August 12, 2010, to October 31, 2010. The survey was supported by the Ministry of Gender Equality and Family, South Korea, and was approved by the Institutional Review Board (IRB) at Yonsei University. For the survey, 3,800 representative households were selected from 200 districts across the country based on the South Korean Population and Housing Census data. In each household, a married or cohabiting adult aged over 19 with the earliest birth date was interviewed from August 12, 2010, to October 31, 2010, by researchers, who were doctoral and master’s students in social work. All the interviewers completed a series of training activities, including face-to-face interview skill sessions and role-playing. Moreover, each team was composed of one male and one female interviewer, and at least one of them was a doctoral student. Among the total sample, we utilize 3,269 married or cohabiting households.

In addition to the 2010 national survey, we used data from three different national surveys: The first national survey conducted in 1997, one year before the implementation of the domestic violence laws in 1998, and two other national surveys conducted in 1999 and 2000. For the 1997 survey, a total of 1,523 married men and women aged over 20 were randomly selected from the phone book and interviewed on the phone from June 1, 1997, to July 31, 1997. This study was supported by the National Research Foundation of Korea. In the 1999 national survey, 1,540 married people were selected via multistage cluster sampling. They were interviewed face-to-face during home visits from August 14, 1999, to September 3, 1999. The 2000 national survey was administered by phone between October 26, 1999, and March 31, 2000, and 1,500 female respondents over 20 years old were interviewed. The questionnaires were given only to females because the objective of the study was to investigate the effects of domestic violence on women’s health. Out of the 1,500 respondents, we selected 1,185 married or cohabiting women. The last two surveys were supported by the Ministry of Justice of South Korea and the Ministry of Health and Welfare of South Korea, respectively. Overall, all the four surveys were based on representative national households, and the 2000 survey had only female respondents. Since the 1997 and 2000 surveys used phone-interviews while the 1999 and 2010 surveys used face-to-face interviews, we analyze data from the 1999 and 2010 surveys to examine the hypotheses of this research.

Measures

In analyzing IPV, we concentrate on physical violence by men and women. It was intended to maintain our focus to look at how IPV interventions and legal punishments in South Korea were associated with the changes of their key target problem, that is, physical violence.

Intimate partner violence (IPV)—physical violence

To measure physical violence between intimate partners, all the surveys (1997, 1999, 2000, and 2010) utilized physical violence questions from the Revised Conflict Tactics Scale (CTS2; Straus, Hamby, Boney-McCoy, & Sugarman, 1996). CTS2 used 7-point Likert-type scale, but we dichotomized them to see whether each type of physical violence was present or not. The questions included the annual prevalence of mild physical violence (“Threw something at me,” “Pushed, grabbed, or shoved me,” “Slapped me”) and severe incidences (“Hit or tried to hit me with something,” “Beat me up,” “Chocked me,” “Threatened me with a knife or gun or used them”). However, “Kick, bit, or punched partner” was dropped in the 2010 national survey because this question is hardly distinguishable from “Beat me up” in the Korean context and thus may have caused confusion when the participants filled out the questionnaire. The Cronbach’s alpha of the items on IPV perpetration and victimization were .792 and .824 in 1999 and .768 and .743 in 2010, respectively, which fell within the satisfactory range and which are very close to each other.

Analysis Strategy

All statistical analyses were carried out using the SPSS software package (Version 21.0; SPSS Inc., Chicago, Illinois). As a preliminary analysis, missing data were checked and 64 out of 3,269 subjects from the 2010 sample were dropped by the complete case analysis, leaving 3,205 for final analysis. Imputation approach was not taken because the missing values were basic socio-demographic and IPV variables. First, the Cronbach’s alpha of the IPV measures was calculated to test the reliability. Second, the annual IPV prevalence over time and socio-demographic characteristics of 1999 and 2010 samples were presented by frequency analysis. The frequency analysis on IPV prevalence was used to test Hypothesis 1. Third, chi-square analyses were conducted to examine the group differences in the IPV severity by socio-demographic variables, followed by logistic regression analyses to identify what socio-demographic variables were significant predictors of IPV for 1999 and 2010. These methods were used to test Hypotheses 2 and 3.

Results

Trends in the Prevalence of Intimate Physical Violence

As shown in Table 2, the prevalence of IPV remained relatively unchanged immediately before and after the implementation of the domestic violence laws. In 1997 and 1999, over 30% of the individuals reported the occurrence of physical violence between intimate partners. However, the rate plummeted almost by 50% in 11 years to 16.5% from 34.1% in 1999. This pattern of changes was also observed in both IPV by men and women. The rates of prevalence for physical violence by men and women remained approximately at 30% and 16% to 18%, respectively, until 2000, and dropped to 13.5% and 8.8% in 2010. Similar to the overall physical violence trend, the rates of mild and severe physical violence had no notable changes between 1997 and 2000, but there were huge drops in 2010 for both men and women (see Table 3). Overall, the decreases in the prevalence of physical violence support Hypothesis 1.

Table 2.

Trend in Annual IPV Prevalence: Physical Violence by Perpetrator (Unit: %).

Perpetrator Category	1997 (n = 1,523)	1999 (n = 1,540)	2000 (n = 1,185)	2010 (n = 3,205)
Total	31.4	34.1	—	16.5
Violence by men	27.9	29.5	29.4	13.5
Violence by women	15.8	17.7	—	8.8

Note. IPV = intimate partner violence.

Table 3.

Trend in Annual IPV Prevalence: Physical Violence by Category (Unit: %).

Violence Category	1997 (n = 1,523)		1999 (n = 1,540)		2000 (n = 1,185)		2010 (n = 3,205)
Violence Category	PVM	PVF	PVM	PVF	PVM	PVF	PVM	PVF
Mild violence	27.5	15.5	28.7	16.8	28.9	—	12.8	8.7
Severe violence	7.9	2.8	8.4	3.5	6.9	—	2.8	1.2

Note. IPV = intimate partner violence. “PVM” and “PVF” refer to physical violence by male partners and physical violence by female partners, respectively.

Demographic Information of the 1999 and 2010 Survey Respondents

First, we provide demographic information about the respondents of two different time periods. The frequency analysis showed that the respondents from 1999 were younger in general than those from 2010, although those in the 30 to 39 age group constituted the largest proportion for each year. As for education, 35.1% and 34.6% of the respondents in the respective years reported having at least 2-year vocational college degrees, while 64.9% and 65.5% did not have a high school diploma. Only 39.2% in 1999 and 46.2% in 2010 were employed, while the rest were either unemployed or not in the labor force. With regard to household monthly income level, nearly 44% of the respondents from 1999 reported income between US$1,000 and US$1,999, and the households in 2010 were uniformly distributed over income categories with a slight skew toward higher levels. The detailed demographic information is presented in Table 4.

Table 4.

Demographic Characteristics.

		N (%)
Variable	Response Category	1999 (N = 1,540)	2010 (N = 3,205)
Age	−29	179 (11.6)	99 (3.2)
	30-39	640 (41.6)	776 (24.8)
	40-49	458 (29.7)	765 (24.4)
	50-59	192 (12.5)	657 (21.0)
	60+	71 (4.6)	832 (25.5)
Education	Middle school or less (0-9)	248 (16.2)	883 (28.4)
	High school (10-12)	744 (48.7)	1,153 (37.1)
	Vocational college (13-14)	126 (8.3)	354 (11.4)
	College or higher (13-18)	409 (26.8)	722 (23.2)
Employment	Employed	603 (39.2)	1,393 (46.2)
Employment	Unemployed	937 (60.8)	1,622 (53.8)
Monthly income	Under US$1,000	171 (11.4)	750 (23.4)
	US$1,000-US$1,999	661 (44.2)	593 (18.5)
	US$2,000-US$2,999	412 (27.5)	799 (24.9)
	US$3,000-US$3,999	141 (9.4)	490 (15.3)
	Over US$4,000	111 (7.4)	573 (17.9)

Differences in the Prevalence of IPV Across Socio-Demographic Groups

The results of the cross-tabulation analysis of the demographic characteristics of men and their physical violence perpetration are presented in Table 5. No statistically significant association was observed between age and physical violence by men for all violence severity categories in 1999 and 2010. However, education showed a strong association with physical violence by men, for both levels of violence severity in 1999 and 2010 (p < .001 and p < .002 for mild violence and severe violence, respectively, in 1999, and p < .007 and p < .001 in 2010). In 1999, at least 20% of each education group perpetrated mild physical violence annually, but men of middle school or lower education levels were the most violent group with a probability of 42.6%, followed by men with high school degrees. In 2010, the most violent group was composed of men with high school degrees (15.3%), and the rates fell with more education. The probability of IPV perpetration in each group in 2010 was generally lower than their 1999 counterparts. An analogous pattern was observed in the case of severe violence; men with lower levels of education were physically more violent for both years (13.5% and 4.4% for men with middle school or less education in 1999 and 2010, respectively), and the overall percentages were lower in 2010. Regarding occupation, it was significantly associated with mild violence in 1999 (p < .002) and with severe violence in 2010 (p < .015). In 1999, blue-collar male partners in services, sales, agriculture, forestry, and fishery, technician, mechanics, and manual labors were more violent by the mild violence standards (32.1%), followed by “others” (28.4%), who were mostly unemployed; white-collar workers (23.2%) in managerial, professional or quasi-professional, office work, and military were the least violent. In 2010, 3.6% of the men in the “other” category perpetrated severe physical violence, whereas 2.9% and 1.3% of blue- and white-collar men did so against their female partners. Last, household income was found to have a significant association with mild and severe physical violence only in 2010. Men in the “US$1,000 to US$1,999” group showed the highest perpetration of mild physical violence, and those in the “less than US$1,000” group perpetrated severe physical violence the most, and the percentages fell as household income increased.

Table 5.

Cross-Tabulation Between Demographic Characteristics and IPV by Male Partners (Unit: %).

	1999 (N = 1,540)				2010 (N = 2,959)
	Mild Violence		Severe Violence		Mild Violence		Severe Violence
	Yes	No	Yes	No	Yes	No	Yes	No
Age
Under 30	33.9	66.1	9.9	90.1	7.7	92.3	0.0	100.0
30-39	26.2	73.8	7.1	92.9	11.0	89.0	1.6	98.4
40-49	30.9	69.1	10.4	89.6	14.0	86.0	2.8	97.2
50-59	30.6	69.4	7.9	92.1	15.9	84.1	3.1	96.9
Over 60	21.9	78.1	4.8	95.2	11.9	88.1	2.3	97.7
χ² (sig)	7.206 (.125)		6.166 (.187)		8.218 (.084)		4.464 (.347)
ϕ	.068		.063		.057		.042
Education
Middle school	42.6	57.4	13.5	86.5	13.6	86.4	4.4	95.6
High school	30.5	69.5	9.7	90.3	15.3	84.7	3.2	96.8
2-year college	23.6	76.4	4.6	95.4	12.6	87.4	2.5	97.5
4-year college or more	23.0	77.0	5.9	94.1	9.9	90.1	0.7	99.3
χ²(sig)	26.629 (.000)		14.497 (.002)		12.103 (.007)		21.345 (.000)
ϕ	.140		.098		.069		.092
Occupation
White collar	23.2	76.8	6.8	93.2	11.3	88.7	1.3	98.7
Blue collar	32.1	67.9	8.8	91.2	14.2	85.8	2.9	97.1
Others	28.4	71.6	12.2	87.8	14.2	85.8	3.6	96.4
χ² (sig)	12.462 (.002)		4.777 (.092)		4.033 (.133)		8.383 (.015)
ϕ	.091		.092		.042		.061
Household income
−US$1,000	36.8	63.2	7.6	92.4	9.9	90.1	4.2	95.8
1,000-2,000	30.3	69.7	8.1	91.9	16.2	83.8	3.6	96.4
2,000-3,000	26.0	74.0	11.2	88.8	13.8	86.2	2.8	97.2
3,000-4,000	26.2	73.8	6.4	93.6	11.8	88.2	1.9	98.1
4,000 +	25.2	74.8	6.3	93.7	11.9	88.1	1.1	98.9
χ² (sig)	8.722 (.068)		5.712 (.222)		11.782 (.019)		13.277 (.010)
ϕ	.076		.062		.063		.067

Note. Age and household income were discretized to examine the group differences of IPV prevalence. IPV = intimate partner violence.

Unlike violence by men, physical violence by women did not have a significant association with any of the primary demographic variables except age and household income for certain violence severity and year combinations. For all age groups in 1999 and for those in the “under 30” group in 2010, at least 10% used mild physical violence, and women under 30 were the most violent, reaching 22.9% in 1999 and 18.8% in 2010. Moreover, women were more severely violent as they were younger in 1999 (p < .05), yet this relationship did not exist in 2010. With regard to household income, there was a significant association only with mild physical violence in 2010 such that women with household income between US$1,000 and US$1,999 were the most violent (10.9%; see Table 6).

Table 6.

Cross-Tabulation Between Demographic Characteristics and IPV by Female Partners (Unit: %).

	1999 (N = 1,540)				2010 (N = 2,951)
	Mild Violence		Severe Violence		Mild Violence		Severe Violence
	Yes	No	Yes	No	Yes	No	Yes	No
Age
Under 30	22.9	77.1	6.1	93.9	18.8	81.2	1.7	98.3
30-39	17.2	82.8	4.3	95.7	9.8	90.2	1.0	99.0
40-49	16.4	83.6	2.1	97.9	8.5	91.5	1.1	98.9
50-59	10.1	89.9	1.3	98.7	8.6	91.4	1.3	98.7
Over 60	11.1	88.9	0.0	100.0	5.3	94.7	0.7	99.3
χ² (sig)	12.104 (.017)		12.322 (.015)		24.136 (.000)		1.481 (.830)
ϕ	.089		.090		.093		.023
Education
Middle School	17.0	83.0	2.7	97.3	7.1	92.9	1.0	99.0
High School	17.9	82.1	3.9	96.1	9.1	90.9	1.6	98.4
2-year college	14.7	85.3	5.5	94.5	9.3	90.7	0.6	99.4
4-year college or more	15.5	84.5	2.9	97.1	8.4	91.6	0.7	99.3
χ² (sig)	1.372 (.712)		2.583 (.460)		2.680 (.444)		4.199 (.241)
ϕ	.030		.041		.031		.039
Occupation
White collar	16.5	83.5	3.3	96.7	7.9	92.1	0.8	99.2
Blue collar	18.7	81.3	3.4	96.6	9.7	90.3	1.8	98.2
Others	16.2	83.8	3.7	96.3	8.5	91.5	0.7	99.3
χ² (sig)	1.285 (.526)		0.090 (.956)		1.094 (.579)		5.236 (.073)
ϕ	.029		.008		.022		.047
Household income
−US$1,000	19.9	80.1	2.3	97.7	7.0	93.0	1.8	98.2
1,000-2,000	18.0	82.0	4.1	95.9	10.9	89.1	1.6	98.4
2,000-3,000	15.0	85.0	3.2	96.8	10.2	89.8	1.2	98.8
3,000-4,000	17.0	83.0	4.3	95.7	7.6	92.4	0.6	99.4
4,000 +	13.5	86.5	3.6	96.4	7.0	93.0	0.5	99.5
χ² (sig)	3.550 (.470)		1.673 (.796)		10.145 (.038)		6.043 (.196)
ϕ	.049		.033		.059		.045

Note. Age and household income were discretized to examine the group differences of IPV prevalence.

Correlations Among Variables

The results of the correlation analysis between the demographic characteristics of IPV offenders and physical violence are shown in Table 7. For violence by men, education and household income were strongly associated with IPV in 1999 (r = −.148, p < .01 for education; r = −.054, p < .05 for household income), while only education had a significant association with IPV by women (r = −.064, p < .01). For IPV by women, age was the only factor which was significantly correlated with IPV in 1999 and 2010 (r = −.101, p < .01 in 1999; r = −.068, p < .01 in 2010).

Table 7.

Correlations Among Variables.

Characteristics of Male Partners and Violence
	1999 (N = 1,540)					2010 (N = 2,959)
	1	2	3	4	5	1	2	3	4	5
1. Age	1					1
2. Education	−.163**	1				−.499**	1
3. Employment	−.174**	.056*	1			−.420**	.239**	1
4. Income	.001	.141**	−.009	1		−.384**	.496**	.347**	1
5. PVM	−.024	−.148**	.019	−.054*	1	.017	−.064**	.006	−.007	1
Characteristics of Female Partners and Violence
	1999 (N = 1,540)					2010 (N = 2,951)
	1	2	3	4	5	1	2	3	4	5
1. Age	1					1
2. Education	−.270**	1				−.692**	1
3. Employment	.026	−.021	1			−.170**	.149**	1
4. Income	−.013	.123**	−.038	1		−.474**	.556**	.161**	1
5. PVF	−.101**	−.030	.015	−.030	1	−.068**	.018	.011	−.014	1

Note. Education was 1 = middle school or less, 2 = high school, 3 = 2 year college, 4 = 4 year college or more; the reference group for employment is the unemployed; acronyms are used to save space, and the meanings are as follows: PVM = physical violence by male partners; PVW = physical violence by female partners. Both PVM and PVW are dummy variables taking either 0 or 1.

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

Prediction of IPV Using Individual Socio-Demographic Characteristics

To evaluate whether socio-demographic factors—age, education, employment status, and household income—related to the perpetration of spousal physical violence, logistic regressions were used, and all models were significant in predicting IPV occurrences (see Table 8). First, age was negatively associated with physical violence against women in 1999 (B = −.014, odds ratio = .986), though this relationship disappeared in 2010. Second, men with less education perpetrated more mild physical violence than men with higher education in each year. Last, employment and household income were not statistically significant predictors of IPV by men.

Table 8.

Logistic Regression Results.

Variables	Physical Violence by Male Partners						Physical Violence by Female Partners
	1999 (N = 1,540)			2010 (N = 2,959)			1999 (N = 1,540)			2010 (N = 2,951)
	B	SE	Odds Ratio	B	SE	Odds Ratio	B	SE	Odds Ratio	B	SE	Odds Ratio
Demographic variables
Age	−.014*	.006	0.986	−.006	.006	0.995	−.039***	.009	0.962	−.032***	.008	0.968
Education	−.281***	.047	0.755	−.173**	.055	0.841	−.144*	.063	0.866	−.074	.074	0.928
Employment	.117	.233	1.124	.135	.161	1.144	.103	.139	1.108	−.002	.053	0.998
Income	−.001	.000	0.999	−.002	.046	0.998	.000	.000	1.000	−.093	.053	0.912
Model statistics
χ²	45.074***			12.528*			23.314***			21.879***
df	4			4			4			4

Note. Demographic variables represent the characteristics of relevant perpetrators of each model; education was 1 = middle school or less, 2 = high school, 3 = 2-year college, 4 = 4-year college or more; the reference group for employment is the unemployed.

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

However, two models to predict IPV by women produced slightly dissimilar results. First, younger women were found to be more violent than the older counterparts in both 1999 and 2010 (B = −.039, p < .001; B = −.032, p < .001). Second, education was an important predictor of IPV by women in 1999 (B = −.144, p < .05), though this was not the case in 2010. Last, employment and household income were not useful predictors. Overall, Hypothesis 3 is partly supported, in that only education is a good predictor of violence by men among many socio-demographic characteristics, whereas violence by women was largely irrelevant with the socio-demographic characteristics except education in 1999.

Discussion

In this study, the prevalence of IPV over the past 15 years in South Korea were compared to determine whether the situation improved after the launch of legal and institutional efforts in the late 1990s. Despite a noticeable drop in physical violence between spouses, the latest national survey showed there is still room for improvement. To provide implications for IPV interventions, socio-demographic characteristics as risk factors were examined.

First, the dramatic fall in IPV prevalence can be considered as the most successful achievements anywhere in the world. Unlike the United States, where the prevalence remained largely unchanged for more than a decade following governmental interventions in the 1970s, South Korea was able to reduce the rate by half within a decade. During this period, the enforcement of the Prevention of Domestic Violence and Victim Protection Act and Special Act for the Punishment of Domestic Violence, an expansion of supporting services, such as shelters, counseling, education, and training, as well as the reinforcement of legal interventions were expected to contribute to the improvement over time (M. J. Lee et al., 2008). Above all, domestic violence counseling centers have been a main factor in the systematic overhaul to fend off critical family problems. Besides immediate protection and supporting services for victims, the centers play a pivotal role in preventing recidivism by providing batterer intervention programs. Consequently, the dramatic decrease of IPV in the past 15 years established a good precedent for the counties experiencing the same social problem. From the example of South Korea, the balance between legal punishment and the provision of correctional and therapeutic programs as well as protective measures for the victims are important elements of the improvement. Also, accessible domestic violence counseling centers have played important roles in increasing awareness of problematic nature of domestic violence and in providing protective services for victims and intervention programs for batterers. Although these multilateral efforts well paid off, the prevalence was still higher than in most other countries. To maintain the positive changes, it is important to review and improve existing approaches to the IPV problem, reflecting our new understanding of current patterns and risk factors for IPV.

In addition to the influences of IPV intervention policy and supporting services, other factors might have affected the prevalence of IPV, including demographic changes in South Korea. Due to the fastest rates of aging, 2010 Census showed that the proportion of people over 65 years old increased from 7.3% to 11.3% in 10 years. Along with aging population, people are getting married late in their lives, so the average age of married and cohabiting couples has risen. Considered previous studies that showed older people are less violent, a higher percentage of older age groups in 2010 might have resulted in a lower IPV rate. However, this is not strongly substantiated because of the following two reasons. First, all age groups have shown decreases in physical violence in 2010 compared to their counterpart age groups in 1999. That is, changes in the composition of age alone cannot explain the uniform decreases of IPV across age. Second, logistic regression models suggest that age is not associated with husband-to-wife violence in 2010 unlike 1999. Although it is not clear when this discrepancy has occurred, age is not a factor any more in predicting IPV by men. In terms of education, people with less education, that is, middle school or less, became more numerous in 2010. This change is believed to be brought out by the relative increase of older generations, not by the changes in accessibility toward education. For instance, no significant difference was observed in the 4-year college entrance rates of general high school students (84.5% in 1999 and 81.5% in 2010). It is true that the college entrance rates of vocational high school students mostly into 2-year vocational college have dramatically increased from 38.5% to 71.1%. But the impact on the overall education level is relatively minor because vocational high school students are just one fifth of total students (Ministry of Education, 2013). Last, a rise in the employment rate could have partially affected by an increase in the labor-force participation of women (from 47% in 1999 to 49% in 2010), Yet the lower employment rate in 1999 is more of a temporary surge in the unemployment rate to 6.2% from the usual 3% to 4% due to the foreign exchange crisis in the late 1990s. Given the diverse social changes measured by age, education, and employment, future research are suggested to study the influences of aging population and more vocational college education on the IPV prevalence.

Second, low socio-demographic characteristics, such as low levels of education and household income, were generally important risk factors for physical violence by men, though this was not true for violence by women. This gender difference in the association between IPV and socio-demographic characteristics can be interpreted by sex role strain paradigm proposed by Pleck (1981), especially self-role discrepancy theory as the underlying theory. According to the paradigm, those who violate socialized sex roles suffer from negative psychological consequences (Pleck, 1981). In a Confucian culture, where the division of labor remains strong, men are asked to assume the role of breadwinners in their households. For men with low socio-demographic status, they are likely to believe they are not meeting up to the roles and to have psychological difficulties, such as depression and inferiority complex (J. Kim & Lee, 1997; Song & Lee, 2012). For instance, J. Kim and Lee (1997) on their study of low-income men showed that education was the key characteristics associated with inferior complex, and inferior complex was found to be a significant risk factor of physical violence by male partners (S. Kim & Kim, 2007; Lim, 2009). However, the value of women is not judged by socio-demographic characteristics in the cultural context with division of labor. For this reason, women have little motivation to use physical violence against their partners even if they have low socio-demographic status. Provided the gendered difference in the effects of socio-demographic characteristics on IPV, different approaches to the family problem are required. Based on the analysis results and the previous studies, an empirical investigation about the negative psychological consequences from gender roles and low socio-demographic characteristics is an important first step to enhance our intervention efforts to alleviate the higher IPV problem of men with low socio-demographic characteristics.

Unlike the general relationships between socio-demographic characteristics and IPV, for certain characteristics, that is, age for male partner violence and education for female partner violence, their significance in predicting IPV disappeared over time. For age on male partner violence, the insignificance resulted from the larger fall in violent behaviors among younger males in 2010. Organized education on domestic violence along with the delayed first marriage led to a larger drop in IPV of younger male, while the most violent young male group in 1999 formed the middle or older groups in 2010. The age differences in the magnitude of reductions offset the previous age gap in the IPV prevalence. For education on female partner violence in 1999, women with higher educational attainment had more opportunity to aware of their rights and respond actively to protect themselves from violence. However, education, especially of domestic violence, is offered at all educational institutions and governmental offices, so more women have a better access to the information, which may help them understand their rights and take proper responses against domestic violence.

Third, regardless of the marginality of violence by women with respect to prevalence and severity, it should not be neglected in family violence intervention considerations. In the United States, the 1975 and 1985 national surveys reported the somewhat puzzling phenomenon that wives were as violent as husbands, and the prevalence increased during that decade (Straus & Gelles, 1986). Many researchers attempted to understand this phenomenon, concluding that women used violence for retaliation or self-defense against male partners (Downs, Rindels, & Atkinson, 2007; Hamberger, 1997). Similarly, the influence of a patriarchal culture made women in Asia less likely engage in violent acts against their male partners (J. Kim, 2007). This cultural trait was also found in the 2010 national survey, in which only 3.1% households had violence by women only, less than half the percentage of the households with violence by men only. Similarly, the percentage of households where only women perpetrated severe physical violence (0.4%) was far less than the rate of households where only men perpetrated severe physical violence (2%). However, the seriousness of violence by women may lead to more severe violence by husbands (Straus, 2005; Straus & Gelles, 1986). In addition to this potential danger for women, family violence witnessed by children poses another problem. A large body of previous literature has indicated that witnessing inter-parental violence adversely influenced many developmental aspects of children, including their behavioral, cognitive, and emotional functioning, along with their long-term developmental outcomes (Adamson & Thompson, 1998; Covey, Menard, & Franzese, 2013; Edleson, 1999; Margolin & Gordis, 2004; Margolin & Vickerman, 2011; Silvern et al., 1995). To address physical violence problems by women, close attention must be paid to households with younger women and lower incomes. In particular, early interventions are necessary to ensure that victimized women have no choices but to stand up against offenders on their own, endangering themselves and their children.

This study is limited for the following reasons. First, certain differences in methodology, that is, data collection and research subjects, existed across the surveys, which made it difficult to compare all the surveys in a one-on-one basis. For instance, IPV prevalence between the 2000 survey and other surveys cannot be directly compared because the 2000 survey studied female respondents only. Because other surveys contain the responses from male respondents, who may want to underreport their problematic behaviors, IPV prevalence may have been underestimated. However, this discrepancy poses little threat to our analysis because the 1999 and 2010 surveys are compared, which contain both male and female respondents and interviewed representative households door-to-door. Moreover, a modification was made to the domestic violence instrument in the 2010 national survey. This poses problems in a direct comparison between this survey and the previous ones. But the sub-question, “Kick bit, or punched partner,” had to be dropped because the linguistic resemblance with “Beat me up” in the Korean context to prevent confusion for respondents. Second, the surveys in this study were cross-sectional data, so it is difficult to pinpoint the legal and institutional interventions as the sole factor causing the dramatic cut in domestic violence. Despite our presumption that the legal and service interventions played an important role in reducing IPV, further research is required to prove the causality of the relationships. Third, the last limitation is the unavailability of the response rates of the surveys. Despite the lack of the response rates, the surveys were approved by Statistics Korea (governmental branch administering national statistics) as national data by conforming to the following survey procedures. Within the randomly selected 200 districts across the nation, targeted households were revisited up to 3 times when no eligible survey respondents were present, and then another household within the same district was chosen. By these procedures, it is likely that response rates of both surveys are high, but the limitation still remains.

Despite the several limitations, there are strengths in this study as well. This study was based on four nationally representative surveys carried out by an identical research team, which provided the most reliable and consistent data given there is no longitudinal studies available. The changes and predictors of the IPV at different time periods could be used as a blueprint in designing future intervention strategies. Specifically, the effects of socio-demographic characteristics on IPV by men and the gendered differences in the relationships may be used as resources to improve our understanding of the nature of the family problem. Also, the public and private interventions would serve as a good model for the countries that have similar patriarchal cultural backgrounds and serious IPV problems.

In addition to these strengths, future studies may contribute to the literature in several ways. Although physical violence is the interest in this study, an investigation on the other forms of violence such as emotional abuse is also requested. This would provide a more comprehensive understanding about the IPV problem. Moreover, we approached to interpret the more violent behaviors of men with lower education by drawing on the Pleck’s gender role theory, but it was not probed here whether the theoretical framework could be similarly applied to female groups of different socio-demographic characteristics. Thus, future studies are suggested to look at how gendered roles play in the IPV perpetration by women across socio-demographic groups and to explain the phenomenon in the Confucian countries with patriarchal cultural background. Finally, an examination of other risk factors such as emotional, cognitive, and behavioral aspects of individuals, families, or communities would enhance our understanding on the nature of IPV.

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) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This study was supported by the National Research Foundation of Korea, the Ministry of Law, the Ministry of Health and Welfare, and the Ministry of Gender Equality and Family in South Korea.

Author Biographies

Jae Yop Kim, PhD, is a professor in the School of Social Welfare at Yonsei University, South Korea. He received his PhD in social service administration from University of Chicago. He served as the principal investigator for national surveys in 1997, 1999, 2000, and 2010, supported by different branches of the Korean government. Also, he developed a family therapy, “TSL family therapy.” His areas of research interest include conjugal violence, the relationship between women’s experience with intimate partner violence, mental health outcomes, and neurobiological perspectives for social work practice.

Sehun Oh, MA, is a PhD student in the School of Social Welfare at Yonsei University. He received his MA in economics from University of Rochester and MSW from Yonsei University. His research interests include poverty and inequality, family dynamics, and social policy.

Seok In Nam, PhD, is an associate professor in the School of Social Welfare at Yonsei University. He received his PhD from Yonsei University and was a postdoctoral fellow in the Edward R. Roybal Institute on Aging, School of Social work, University of Southern California. His research interests include the adjustment of elderly people in retirement, neurobiological perspectives for social work practice, medical social work in health and mental health, and intimate partner violence among the aging.

References

Adamson

J. L.

Thompson

R. A.

(1998). Coping with interparental verbal conflict by children exposed to spouse abuse and children from nonviolent homes. Journal of Family Violence, 13, 213-232. doi:10.1023/A:1022896804777

Breiding

M. J.

Black

M. C.

Ryan

G. W.

(2008). Prevalence and risk factors of intimate partner violence in eighteen US states/territories, 2005. American Journal of Preventive Medicine, 34, 112-118. doi:10.1016/j.amepre.2007.10.001

Choi

K. R.

Yoo

E. H.

Hong

S. J.

Chung

(1999). Development and evaluation of program for coping with spouse abuse. Family and Environment Research, 132, 159-173.

Covey

H. C.

Menard

Franzese

R. J.

(2013). Effects of adolescent physical abuse, exposure to neighborhood violence, and witnessing parental violence on adult socioeconomic status. Child Maltreatment, 18, 85-97. doi:10.1177/1077559513477914

Downs

W. R.

Rindels

Atkinson

(2007). Women’s use of physical and nonphysical self-defense strategies during incidents of partner violence. Violence Against Women, 13, 28-45. doi:10.1177/1077801206294807

Edleson

J. L.

(1999). Children’s witnessing of adult domestic violence. Journal of Interpersonal Violence, 14, 839-870. doi:10.1177/088626099014008004

Hamberger

L. K.

(1997). Female offenders in domestic violence: A look at actions in their context. Journal of Aggression, Maltreatment & Trauma, 1, 117-129. doi:10.1300/J146v01n01_07

Jewkes

(2002). Intimate partner violence: Causes and prevention. The Lancet, 359(9315), 1423-1429. doi:10.1016/S0140-6736(02)08357-5

Kim

(1998). Spousal abuse and social class in Korean National Family Violence Survey. Korean Journal of Social Welfare, 35, 133-155.

10.

Kim

(2007). Domestic violence of South Korea. Seoul, South Korea: Hakjisa.

11.

Kim

Lee

(2007). An analysis of the effectiveness of domestic violence offenders’ correction treatment programs. Seoul, South Korea: Ministry of Gender Equality and Family.

12.

Kim

Lee

(1997). The effects of inferiority complex and gender role attitudes on husband-to-wife physical violence and a study on therapeutic models: Based on low income families. Korean Journal of Social Welfare, 33, 71-93.

13.

Kim

Park

Emery

C. R.

(2009). The incidence and impact of family violence on mental health among South Korean women: Results of a national survey. Journal of Family Violence, 24, 193-202. doi:10.1007/s10896-008-9220-5

14.

Kim

(2007). Housework and economic dependency among dual-earner couples in Korea: Economic exchange or gender compensation. Korean Journal of Sociology, 41(2), 147-174.

15.

Kim

W. J.

(2012). The legal study on prevention of domestic violence about marriage immigrant women and the protection. Chosun Law Journal, 19(1), 447-474.

16.

Korea Institute for Health and Social Affairs. (2002). Measures to strengthen the function and role of domestic violence and sexual violence counseling centers or protection facilities. Seoul: Ministry of Gender Equality and Family.

17.

Lee

(2006). Caring-self and women’s self-esteem: A feminist’s reflection on pastoral care and religious education of Korean-American women. Pastoral Psychology, 54, 337-353.

18.

Lee

C. M.

(2004). A study on the impacts and effectiveness of the Violence Against Women Act (VAWA) in the United States. Korean Journal of Criminology, 16(1), 277-304.

19.

Lee

H. J.

(2008). Analysis of domestic violence law and practice over the last 10 years. Korean Criminology Review, 75, 127-169.

20.

Lee

M. J.

Byun

W. S.

Hwang

J. I.

Lee

S. Y.

(2008). Improving service for the victims of domestic violence and sexual assault improving service for the victims of domestic violence and sexual assault. Seoul, South Korea: Korean Women’s Development Institute.

21.

Lee

S. W.

(2003). A study for development of batterer treatment program of domestic violence and its effectiveness (Doctoral thesis). Yonsei University, Seoul, South Korea.

22.

Lim

G. W.

(2009). A study on the risk factors of violence by analyzing the characteristics of domestic violence offenders. The Correction Welfare Studies, 15, 105-119.

23.

Margolin

Gordis

E. B.

(2004). Children’s exposure to violence in the family and community. Current Directions in Psychological Science, 13, 152-155. doi:10.1111/j.0963-7214.2004.00296.x

24.

Margolin

Vickerman

K. A.

(2011). Posttraumatic stress in children and adolescents exposed to family violence: I. Overview and issues. Couple and Family Psychology: Research and Practice, 1(S), 63-73. doi:10.1037/2160-4096.1.S.63

25.

Ministry of Education. (2013). 2013 Education statistical year book. Seoul: Ministry of Education.

26.

Ministry of Gender Equality and Family. (2004-2012). Guidelines for women’s right promotion project in 2004 to 2012. Seoul: Ministry of Gender Equality and Family.

27.

Park

D. H.

(2006). A study on evaluation of the batterer treatment programs for domestic violence—Focusing on the prosecution’s suspension of indictment policy on condition of obligatory counsel (Doctoral theses). Yonsei University, Seoul, South Korea.

28.

Park

Y. R.

(2007). Changing the needs of the victims of domestic violence and the paradigm shift of victim protection policy from a feminist perspective. Journal of Korean Women’s Studies, 23(3), 189-214.

29.

Pleck

(1981). The myth of masculinity. Cambridge, MA: The MIT Press.

30.

Sherman

L. W.

Cohn

E. G.

(1989). The impact of research on legal policy: The Minneapolis domestic violence experiment. Law & Society Review, 23, 117-144.

31.

Shim

W. S.

Nelson-Becker

(2009). Korean older intimate partner violence survivors in North America: Cultural considerations and practice recommendations. Journal of Women & Aging, 21, 213-228. doi:10.1080/08952840903054773

32.

Silvern

Karyl

Waelde

Hodges

Starek

Heidt

Min

(1995). Retrospective reports of parental partner abuse: Relationships to depression, trauma symptoms and self-esteem among college students. Journal of Family Violence, 10, 177-202. doi:10.1007/BF02110599

33.

Song

Lee

(2012). Gender role attitude and depressive symptoms: Comparisons across gender and age groups. Korean Journal of Population Studies, 35(3), 87-116.

34.

Straus

M. A.

(2005). Women’s violence toward men is a serious social problem. Current Controversies on Family Violence, 2, 55-77.

35.

Straus

M. A.

Gelles

R. J.

(1986). Societal change and change in family violence from 1975 to 1985 as revealed by two national surveys. Journal of Marriage and Family, 48(3), 465-479.

36.

Straus

M. A.

Gelles

R. J.

Smith

(1990). Physical violence in American families: Risk factors and adaptations to violence in 8,145 families. New Brunswick, NJ: Transaction.

37.

Straus

M. A.

Hamby

S. L.

Boney-McCoy

Sugarman

D. B.

(1996). The Revised Conflict Tactics Scales (CTS2) development and preliminary psychometric data. Journal of Family Issues, 17, 283-316. doi:10.1177/01925139-6017003001

38.

Tjaden

Thoennes

(2000). Prevalence and consequences of male-to-female and female-to-male intimate partner violence as measured by the National Violence Against Women Survey. Violence Against Women, 6, 142-161.

39.

United Nations Department of Economic and Social Affairs. (2010). The world’s women: Trends and statistics. New York, NY: United Nations.