Minsang YOO, et al.: Workplace mistreatment and health problems 427 J Occup Health 2015; 57: 427–437 Journal of Occupational Health

Gender and educational level modify the relationship between workplace mistreatment and health problems: a comparison between South Korea and EU countries Minsang Yoo1, Saerom Lee2 and Mo-Yeol Kang1,2

1Department of Preventive Medicine, College of Medicine, Seoul National University, Republic of Korea and 2Occupational Safety and Health Research Institute, Republic of Korea

Abstract: Gender and educational level modify the place did not significantly affect EU workers. Conclu- relationship between workplace mistreatment and sions: Workplace mistreatment is significantly associ- health problems: a comparison between South ated with physical and mental health problems, espe- Korea and EU countries: Minsang YOO, et al. Depart- cially among workers with higher educational levels and ment of Preventive Medicine, College of Medicine, females who work alone in Korea. Seoul National University, Republic of Korea— (J Occup Health 2015; 57: 427–437) Objectives: This study investigated effects of workers’ cultural and personal characteristics on the relationship Key words: Cross-cultural, Educational levels, Gender between workplace mistreatment and health problems difference, Job stress, Workplace mistreatment in both South Korea and EU Countries. Methods: Data were obtained from nationally representative interview Over the past two decades, workplace mistreat- surveys: the third Korean Working Conditions Survey ment has been recognized as a serious problem within (KWCS) in 2011 (50,032 participants) and fifth European the working environment, and this relatively new Working Conditions Survey (EWCS) in 2010 (41,302 research topic has now been studied extensively in participants). The Pressure-State-Response model was several European countries1−4). Even so, there is still adapted to explore differences in the relationship limited information about workplace mistreatment in between mistreatment and health problems according to country, and logistic regression analysis was used Asia, although in recent years, Asian researchers have after stratification of moderating factors. Workplace increasingly focused on workplace mistreatment as an 5, 6) mistreatment, such as discrimination, violence, harass- important social stressor in work contexts . ment, and self-reported health problems, were assessed The scope and terminology of workplace mistreat- by gender and educational level. Results: Among ment varies across the discipline, incorporating KWCS participants, there were 4,321 victims (14.70%) workplace discrimination, violence, and bullying. of workplace mistreatment; among EWCS participants, Workplace discrimination refers to actions of insti- there were 5,927 victims (17.89%). There was a signif- tutions and/or individuals within them that lead to icant positive association between workplace mistreat- unfair terms and conditions that systematically impair ment and self-reported health problems. A stronger the ability of group members to work7). Violence can association was found among workers with higher be defined as any form of aggressive behavior aimed educational levels in Korea (2- to 4-fold higher odds for mental and physical health problems), but there was no toward the goal of harming another human being, and significant difference by education level in workers of it can be both psychological and physical in nature. EU Countries. Female Koreans who worked alone had In the extreme form, violence may involve the use of a higher risk of health problems related to workplace physical force to harm or damage persons or prop- mistreatment than other gender compositions in the erty or to interfere with personal rights or freedom2). workplace (the OR for psychological symptoms reached Another form of aggressive behavior at work is bully- 6.631). In contrast, the gender composition of the work- ing. Workplace bullying includes both physical and nonphysical actions that offend or socially exclude a Received Nov 29, 2014; Accepted May 20, 2015 worker or group of workers; some researchers specify Published online in J-STAGE Jun 25, 2015 that this definition must include an intent to harm Correspondence to: M.-Y., Kang, Occupational Safety and Health either the individual or the organization8). For the Research Institute 400, Jongga-ro, Jung-gu, Ulsan, South Korea, purpose of this paper, we defined workplace mistreat- [681-230] (e-mail: [email protected]) ment to include workplace-related discrimination, 428 J Occup Health, Vol. 57, 2015 violence, abuse, or harassment. stronger or weaker effects on health among workers Cultural differences in construing the meaning of with different cultural and personal backgrounds. specific behaviors make it harder to determine which Given the negative personal and organizational behaviors are abusive9). Less attention has been paid effects of workplace mistreatment, every effort should to national differences in mistreatment, though some be made to understand the mistreatment process and researchers have pointed out the problems of using to prevent the consequent health outcomes. The aim questionnaires about bullying developed in other of this study was to investigate whether workers’ national contexts and approached workplace bullying cultural and personal characteristics affect the relation- cross-culturally6, 9). This context is important because ship between workplace mistreatment and health prob- the prevalence of workplace mistreatment varies lems. Gender and educational level were examined in according to workers’ perceptions10) and their national both Korea and EU countries in order to understand culture4). the health effects of workplace mistreatment and iden- Differences also exist in terms of the source of tify groups at risk. personal characteristics, such as gender and educa- tional level. It is important to identify risk groups Subjects and Methods of workplace mistreatment in the context of future We used data from the third Korean Working prevention. However, the results from empirical stud- Condition Survey (KWCS) in 2011 and fifth European ies on sociodemographic factors and mistreatment are Working Condition Survey (EWCS) in 2010. These inconsistent. For example, among personal charac- were nationally representative interview surveys and teristics, gender has been the most widely studied. included questions relating to workers’ socioeconomic Some studies have reported more significant effects of data, workplace environment, and social and occupa- mistreatment for women11, 12), while Hoel and Cooper tional health. Informed consent was obtained from found no significant gender differences in their all participants in both surveys, and detailed informa- UK-wide survey of bullying13). tion about these surveys is available at the following There has been substantial research on the health websites: http://www.eurofound.europa.eu/working/ outcomes of workplace mistreatment. For example, surveys/ for the EWCS and http://www.kosha.or.kr/ several cross-sectional studies have found evidence jsp/kwcs/ for the KWCS. The fifth EWCS comprises of correlations between exposure to mistreatment and data from 35 European countries with 43,816 partici- psychological, psychosomatic, and physical symp- pants. The third KWCS, which was based on the toms7, 14). Common symptoms reported by victims in EWCS, was conducted by the Korea Occupational different European countries include musculoskeletal Safety and Health Agency (KOSHA) in 2011. This complaints, anxiety, irritability, and depression15−18). survey recruited 50,032 participants by using strati- Two recent Norwegian longitudinal studies found fied, multistage, random sampling of household units. that exposure to bullying behavior in the workplace For our analysis, we excluded self-employed partici- predicts psychological distress two years later19, 20). In pants (N=8,629 in the EWCS; 20,321 in the KWCS), addition, social isolation, social maladjustment, low those aged under 20 (N=722 in the EWCS; 317 in self-esteem, sleep problems, concentration difficul- the KWCS), and participants who had missing data ties, chronic fatigue, and burnout seem to be common for some variables (N=1,329 in the EWCS); the final symptoms in victims of workplace mistreatment3). study sample included 33,136 participants for the The international cost of outcomes related to work- EWCS and 29,404 participants for the KWCS. place mistreatment is estimated to be between $17 and $36 billion annually21), and an Australian study Variables revealed that total national annual employer costs for Mistreatment lost productivity due to depression were estimated at We defined the “Mistreatment” group as individuals AUD$8 billion per annum22). to whom at least one of the following criteria applied: To date, evidence has supported the hypothesis (1) discrimination, that is, being subjected to discrimi- that workplace mistreatment has negative effects on nation at work because of age, ethnicity, sex, nation- employees’ health. However, most of these studies ality, or religion in the past 12 months; (2) violence, have been based on Western samples. Although work- that is, being subjected to verbal abuse, unwanted place mistreatment shares similar characteristics, other sexual attention, threats and humiliating behavior, or related research suggests that responses to workplace physical violence during the course of work in the mistreatment may be different in Eastern countries23). last month; and (3) bullying, that is, being subjected Therefore, comparison between Western and Eastern to bullying/harassment or sexual harassment during countries has an important meaning when attempting the course of work in the past 12 months. to understand why workplace mistreatment will have Minsang YOO, et al.: Workplace mistreatment and health problems 429

General characteristics of the study subjects Mistreatment score and health outcome score The EWCS and KWCS surveys include ques- The Pressure-State-Response (PSR) model, which tions about a wide array of characteristics; we chose was pioneered by the Organisation for Economic to use age, sex, education level, household income Cooperation and Development (OECD), is a concep- level, employment status, gender composition in tual model of causality. It assumes that human the workplace, and job classification. Education activities exert pressures on the environment, thereby level was classified as “middle school or less” (for affecting environmental quality and the quantity of European countries, this is defined as “lower second- natural resources (“state”); society responds to these ary or second stage of general education or less”), changes through environmental, general economic, and “high school” (European countries: “upper second- sectorial policies and through changes in awareness ary education”), and “college or more” (European and behavior (“societal response”)25). We adapted the countries: “post-secondary non-tertiary education or PSR model, defining mistreatment scores as “pressure”, more”). The gender composition in participants’ calculated by the sum of experienced mistreatments workplace was ascertained by asking, “At your place in the workplace, including discrimination, violence, of work, are workers with the same job title as you and bullying. We then transformed these scores using …?” with a choice of four answers: “mostly men”, a scale of 0 to 10. That is, an individual with no “mostly women”, “more or less equal numbers of experience of mistreatment would have a mistreat- men and women”, and “nobody else has the same job ment score of 0, whereas an individual experiencing title”. Jobs were classified into 10 groups accord- discrimination, violence, and bullying would score 10. ing to the International Standard Classification of In the same way, the health outcome score (considered Occupation (ISCO) code: “legislators, senior officials the “state” aspect of the PSR model) was defined by and managers”, “professionals”, “technicians and asso- the sum of the 12 self-reported health problems. The ciated professionals”, “clerks”, “service workers and health outcome score was also transformed onto a shop and market sales workers”, “skilled agricultural scale of 0 to 10. and fishery workers”, “craft and related trades work- ers”, “plant and machine operators and assemblers”, Statistical analysis “elementary occupations”, and “armed forces”. We analyzed the EWCS and KWCS data sepa- rately to compare European and Korean participants. Self-reported health problems General characteristics are presented as means and Self-reported health problems were assessed by the standard deviations for continuous variables and question, “Over the last 12 months, have you suffered numbers and prevalences (%) for categorical variables. from any health problems?” The KWCS and EWCS The mistreatment score and health outcome score included 13 health-related questions relating to hearing means were calculated and plotted by nation, in order problems, skin problems, backache, muscular pains to look at national differences. To evaluate the asso- in the shoulders/neck/upper limbs, muscular pains in ciations between experience of mistreatment and each lower limbs, headache/eyestrain, stomach ache, respi- of the self-reported health problems, we performed ratory difficulties, cardiovascular disease, injuries, logistic regression after adjusting for age, sex, and depression or anxiety, overall fatigue, and insom- job classification; for the European data, we further nia/general sleep difficulties. We excluded health adjusted for nation as well as these three variables. problems caused by injury and grouped together the To compare health effects of workplace mistreatment remaining 12 health problems into “somatising tenden- between EU countries and Korea, effect modification cy”, “musculoskeletal pain”, “psychological symp- was investigated by assessing the interaction term toms”, and “overall fatigue”. The “somatising tenden- involving two data sets. We conducted additional cy” category comprised symptoms related to hearing analyses to evaluate gender and educational levels as problems, skin problems, stomach ache, respiratory modifying factors that might affect the relationship difficulties, and headache/eyestrain24). “Musculoskeletal between workplace mistreatment and health outcomes. pain” grouped together muscular pains in the shoul- The stratified analysis was conducted using the vari- ders/neck/upper limbs and lower limbs and backache, ables of gender composition in the workplace, educa- and the “psychological symptoms” category referred tional level, and estimated interaction terms. The to participants who experienced either depression or statistical analysis was conducted using the SAS soft- anxiety or insomnia/general sleep difficulties. The ware (Version 9.22, SAS Institute, Cary, NC, USA). “overall fatigue” category was defined from only one health-related question, that referring to overall fatigue Results itself. Table 1 shows participants’ descriptive character- istics. There were 16,300 males and 16,836 females 430 J Occup Health, Vol. 57, 2015

Table 1. Descriptive characteristics of the study population (N and %) European Working Condition Survey (EWCS) Korean Working Condition Survey (KWCS) Total sample Mistreatment Prevalencec Total sample Mistreatment Prevalencec N % N % % N % N % % Sex Male 16,300 49.19 2,646 44.64 16.23 17,178 58.42 2,417 55.94 14.07 Female 16,836 50.81 3,281 55.36 19.49 12,226 41.58 1,904 44.06 15.57 Education level Middle school 7,692 23.21 1,328 22.41 17.26 3,263 11.1 529 12.24 16.21 High school 13,483 40.69 2,281 38.48 16.92 11,394 38.75 1,740 40.27 15.27 University 11,961 36.1 2,318 39.11 19.38 14,747 50.15 2,052 47.49 13.91 Incomea Lowest 3,852 16.29 529 12.11 13.73 3,558 12.1 503 11.65 14.14 Low-middle 8,721 36.88 1,477 33.81 16.94 11,745 39.95 1,948 45.11 16.59 High-middle 8,274 34.99 1,820 41.66 22.00 8,053 27.39 1,184 27.42 14.70 Highest 2,800 11.84 543 12.43 19.39 6,041 20.55 683 15.82 11.31 Employment statusb Fixed 26,027 95.07 4,660 94.24 17.90 23,182 78.84 3,250 75.21 14.02 Temporary 1,270 4.64 267 5.4 21.02 4,353 14.8 678 15.69 15.58 Day laborer 80 0.29 18 0.36 22.50 1,869 6.36 393 9.1 21.03 Workplace gender composition Mostly men 12,168 36.72 1,994 33.64 16.39 12,176 41.41 1,792 41.47 14.72 Mostly women 11,331 34.2 2,307 38.92 20.36 8,258 28.08 1,342 31.06 16.25 More or less equal num- 6,706 20.24 1,214 20.48 18.10 7,565 25.73 1,020 23.61 13.48 bers of men and women Nobody else has the 2,931 8.85 412 6.95 14.06 1,405 4.78 167 3.86 11.89 same job title Job (ISCO code) Legislators, senior offi- 1,782 5.38 313 5.28 17.56 568 1.93 62 1.43 10.92 cials, and managers Professionals 5,227 15.77 962 16.23 18.40 5,031 17.11 664 15.37 13.20 Technicians and associ- 5,225 15.77 963 16.25 18.43 6,455 21.95 888 20.55 13.76 ated professionals Clerks 3,968 11.97 657 11.08 16.56 2,959 10.06 466 10.78 15.75 Service workers and 6,013 18.15 1,372 23.15 22.82 3,669 12.48 532 12.31 14.50 shop and market sales workers Skilled agricultural and 302 0.91 33 0.56 10.93 158 0.54 14 0.32 8.86 fishery workers Craft and related trades 3,741 11.29 513 8.66 13.71 2,874 9.77 394 9.12 13.71 workers Plant and machine oper- 3,022 9.12 511 8.62 16.91 3,194 10.86 513 11.87 16.06 ators and assemblers Elementary occupations 3,670 11.08 571 9.63 15.56 4,421 15.04 781 18.07 17.67 Armed forces 186 0.56 32 0.54 17.20 75 0.26 7 0.16 9.33 For total sample 33,136 100 5,927 100 17.89 29,404 100 4,321 100 14.70 aThere was no information on income variables for 9,496 EWCS participants (1,561 participants affiliated with the mistreatment group). bThere was no information on income variables for 5,759 EWCS participants (982 participants affiliated with the mistreat- ment group). cPrevalence of experiencing workplace mistreatment in past 12 months. Minsang YOO, et al.: Workplace mistreatment and health problems 431 in the EWCS, among which 16.23% of males and The mean mistreatment scores and health outcome 19.49% of females experienced some type of mistreat- scores (from the PSR model) differed by nation ment. In the KWCS, there were 17,178 males and (Supplementary Table 1). Figure 1 shows the associa- 12,226 females, and the prevalence of mistreatment tion between mistreatment score and health outcome was 14.07% for males and 15.57% for females. The score according to nation. The majority of coun- mean age was similar for both survey samples. In tries in Europe reported more frequent mistreatment the EWCS, the mean age was 41.41 (SD=11.52); compared with Korea, and the health outcome score for participants reporting mistreatment, it was 40.66 for Korea was the lowest of all the countries. Among (SD=11.40). In the KWCS, the mean age was 41.75 European countries, Finland, Latvia, and Estonia (SD=11.76), while that for participants reporting showed the highest mistreatment and health outcomes, mistreatment was 41.47 (SD=12.11). The Korean and Kosovo and Ireland had the lowest mistreatment results show a trend for more educated participants and self-reported health problems, respectively. to experience less mistreatment, but the trend was in The frequencies and associations between mistreat- the opposite direction for EU countries. Prevalence ment and self-reported health problems for all partici- of mistreatment according to employment status and pants and both of the respective surveys are shown in gender composition in the workplace was similar in Table 2. There was a difference in the frequencies both surveys, but the prevalence of mistreatment was of health outcomes between the EWCS and KWCS. greater overall for EWCS participants. Day laborers About half of the EWCS participants reported soma- and individuals who worked mostly with women had tizing tendency, musculoskeletal pain, and overall a higher prevalence of mistreatment. Regarding job fatigue, whereas the frequencies of these health classification, mistreatment was higher in legislators, outcomes in the KWCS were from 19.84 to 38.66%: senior officials, and managers; professionals; and tech- the differences in frequencies between the EWCS nicians and associated professionals in Europe than in and KWCS were about 20%. All health outcomes Korea. were significantly associated with mistreatment in

Fig. 1. Mean mistreatment and health outcome scores according to country. 432 J Occup Health, Vol. 57, 2015

Table 2. Mistreatment frequency and multiple logistic regression analysisa results for the association between mistreatment and self-reported symptoms (OR=odds ratio; 95% CI=95% confidence intervals) European Working Korean Working Total Self-reported Condition Survey (EWCS)b Condition Survey (KWCS) p- symptoms interaction N (%) OR (95% CI) N (%) OR (95% CI) N (%) OR (95% CI) Somatizing 22,746 2.307 16,912 2.126 5,834 2.722 <0.0001 tendency (36.37) (2.209−2.409) (51.04) (2.003−2.256) (19.84) (2.536−2.921) Musculoskeletal 31,631 2.165 20,263 2.004 11,368 2.268 0.005 pain (50.58) (2.069−2.265) (61.15) (1.880−2.135) (38.66) (2.120−2.426) Psychological 9,398 2.640 8,485 2.643 913 3.400 0.0012 symptoms (15.03) (2.509−2.778) (25.61) (2.490−2.806) (3.11) (2.955−3.912) 20,727 1.952 13,733 1.925 6,994 1.895 Overall fatigue 0.5483 (33.14) (1.869−2.039) (41.44) (1.818−2.038) (23.79) (1.767−2.032) aThe logistic regression model was adjusted for age, sex, and job classification. bThe EWCS data was adjusted for nations. both surveys, but how closely workplace mistreatment overall fatigue=2.060). On the other hand, men who was associated with health outcomes was somewhat worked with mostly women showed a greater risk of different in the two surveys. Psychological symp- somatizing tendency due to workplace mistreatment toms had the strongest association with mistreatment (OR=3.633). In female KWCS participants, workers compared with other variables, and this association who worked alone (“nobody else has the same job was stronger for KWCS participants (OR=3.400, 95% title”) showed a strong association between mistreat- CI=2.955−3.912) than EWCS participants (OR=2.643, ment and health outcomes, especially psychological 95% CI=2.490−2.806). Somatizing tendency and symptoms (OR=6.631). The OR of all females musculoskeletal pain also showed a stronger associa- for musculoskeletal pain was 2.300, while that of tion with mistreatment for Korean participants than females who worked alone was 3.111. For these two for European participants, and all health outcomes groups of females, the ORs of psychological symp- except overall fatigue showed statistically significant toms were 3.642 and 6.631, respectively. On the interaction terms between in the two surveys. That other hand, females who worked with mostly men is, significant health outcomes were observed among showed a more increased risk of somatizing tendency participants who experienced mistreatment in their and psychological symptoms related to workplace workplace in both surveys; however, the risk eleva- mistreatment (OR for somatizing tendency=3.633, OR tions were greater in the KWCS than in the EWCS. for psychological symptoms=6.772). Table 3 shows the factors affecting self-reported Education level also modifies the relationship health problems among those who experienced any between workplace mistreatment and health problems type of mistreatment. Overall, workplace mistreat- in Korea in contrast with EU countries. Individuals ment was most closely associated with psychological with a higher level of education were more likely symptoms, but the effect size of workplace mistreat- to report health problems when they experienced ment was different according to gender composition mistreatment, especially psychological symptoms and educational level. In the EWCS, there were no (OR=4.016). Furthermore, all health outcomes in significant effect differences across gender composi- Korea showed a significant interaction between tion in the workplace among females; however, male mistreatment and educational level; this was not the subjects who worked with mostly men were more case for any health outcome in European countries. likely to report musculoskeletal pain (OR=2.244), and those who performed unique jobs (“nobody else has Discussion the same job title”) suffered more from psychological Our aim in this study was to investigate effects problems (OR=3.984) when they experienced mistreat- of workers’ cultural and personal characteristics on ment at work. There was little difference in the ORs the relationship between workplace mistreatment and of health problems according to education level. health problems, exploring cultural areas, gender and The Korean data show different patterns. Male educational levels as modifying variables in these KWCS participants who worked mostly with men relationships. Generally speaking, the prevalence reported more frequent psychological symptoms and of workplace mistreatment and health problems was overall fatigue when they experienced mistreatment higher in EU countries than in Korea. However, (OR for psychological symptoms=3.862, OR for the association between workplace mistreatment and Minsang YOO, et al.: Workplace mistreatment and health problems 433

Table 3. Odd ratios of mistreatment for self-reported health problems according to workplace gender composition and educational levels (EWCS=European Working Condition Survey; KWCS=Korean Working Condition Survey)

EWCS KWCS Somatizing Musculoskeletal Psychological Overall Somatizing Musculoskeletal Psychological Overall tendency pain symptoms fatigue tendency pain symptoms fatigue Sex Male 2.180 2.003 2.956 2.336 2.705 2.289 3.316 1.86 Mostly men 2.224 2.244 2.943 2.394 2.713 2.319 3.862 2.060 Mostly women 2.575 1.633 2.858 2.857 3.633 2.129 2.941 1.607 More or less equal num- 2.029 1.758 2.909 2.059 2.457 2.249 2.189 1.503 bers of men and women Nobody else has the 1.877 1.704 3.984 2.406 2.306 2.493 2.579* 1.354* same job title p-interactiona 0.4636* 0.0309 0.5226* 0.5389* 0.4992* 0.8954* 0.0265 0.0016 Female 2.161 2.020 2.530 2.270 2.839 2.300 3.642 1.965 Mostly men 1.756 1.903 2.351 1.919 3.179 2.509 6.772 2.352 Mostly women 2.298 2.238 2.582 2.327 2.903 2.089 3.317 1.997 More or less equal num- 2.151 1.755 2.618 2.225 2.698 2.576 3.286 1.643 bers of men and women Nobody else has the 2.098 1.848 2.452 2.720 2.097 3.111 6.631 2.497 same job title p-interactiona 0.5285* 0.1012* 0.6269* 0.3282* 0.1112* 0.259* 0.6659* 0.4092* Education level Middle school 2.225 1.863 2.695 2.284 1.543 1.989 2.355 1.504 High school 2.274 2.295 2.870 2.349 3.356 1.937 3.218 1.781 University 2.024 1.851 2.558 2.241 2.738 2.716 4.016 2.15 p-interactionb 0.1839* 0.5807* 0.7967* 0.5521* 0.015 <0.0001 0.021 0.0003 aInteraction between mistreatment and workplace gender dominancy. bInteraction between mistreatment and educational level. *Not significant. health problems was stronger in Korea than in EU mistreatment varies according to the workers’ national countries, and the pattern of these associations varied culture, comparison of two samples is possible and depending on gender, gender composition in the worthwhile. The antecedents and consequences of workplace, and educational level. The results of our workplace mistreatment have received much atten- study support the hypothesis that certain kinds of indi- tion in psychology and occupational medicine1, 4, 7), but vidual characteristics and cultural factors may exacer- cross-cultural examinations of workplace mistreatment bate the damaging effects of workplace mistreatment, have been rare6, 9). Some studies have investigated especially in Korea. differences in the way that employees conceptualize The survey data indicated that 17.89% of workers workplace mistreatment27). In our study, all health in EU countries and 14.7% of workers in Korea expe- outcomes were significantly associated with mistreat- rienced some kind of mistreatment, such as workplace ment in both surveys, but how closely related work- violence, bullying, and harassment, over the last year. place mistreatment and health outcome were differed Previous studies have revealed that approximately somewhat between the EU countries and Korea. 4−10% of the working population in Europe has been Three categories of health outcome showed a stronger exposed to workplace mistreatment3). Although the association with mistreatment among Korean partici- prevalence of workplace mistreatment is different by pants than among European participants. Evidence method and sampling technique, it is obvious that suggests that cultural beliefs and values may mutually workplace mistreatment is reaching epidemic levels, influence how individuals perceive and respond to and the World Health Organization has declared this aggressive acts12). For example, Salin (2003) argues phenomenon a significant public health threat26). One that workplace mistreatment is perceived as a “normal” of the advantages in our study is that we used the and acceptable way of getting things done in some same questionnaires to survey two different samples, cultures, while in others it is viewed as unacceptable. because the KWCS was based on the EWCS. Northern European countries, the UK, and Australia, Therefore, even if the perception of workplace which are characterized by negative attitudes towards 434 J Occup Health, Vol. 57, 2015 signs of abuse of power, low power distance, more threat, anger, and depression or demoralization, which feminine values, and individualism, tend to have a in turn evoke sustained autonomic arousal. The lower threshold for categorizing inadequate behav- Whitehall II study findings showed that ERI predicts iors as mistreatment and may be more ready to use the incidence of self-reported coronary heart disease the right to complain about it9). The results of our (CHD)32), psychiatric disorder33), and poor health func- PSR model showed that in European countries, both tioning34) during follow-up. mistreatment and response scores were higher than in A complementary concept, CATS, proposes that a Korea (Fig. 1). Drawing on power distance research, stress response occurs when an individual perceives we supposed that while workplace mistreatment is a discrepancy between what should be and what is— universally perceived as a negative situation, its effects between the set value and the real value. The core will differ from low power-distance cultures (EU assumption of the CATS model is that sustained or countries) to high power-distance cultures (Eastern chronic arousal produced by stress, such as worry or Asian countries). This suggests that mistreatment may rumination, may produce physiological activation and be tolerated to some extent by employees in cultures lead to health impairment. Eventually, an adverse that accept a hierarchically based power disparity. stress reaction develops, which is characterized by The main factor linking exposure to workplace physiological arousal (e.g. increased heart rate or mistreatment and health outcomes is stress: Workplace blood pressure). Thus, the increase in arousal as a mistreatment can be considered a major job stress response to stressful situations such as mistreatment factor. Once workplace mistreatment is perceived as at work may lead to the development of health prob- a traumatic event, cognitive processes produce chang- lems35). es in the victim’s thoughts and beliefs28). According Workplace mistreatment may contribute to health to the social cognition view, individuals hold certain disparities by having differential effects on minor assumptions or schemas regarding the world and the groups compared with dominant groups. For exam- self, which are stable and generally positive29). A ple, one study examined the relationship between the traumatic event presents information that is incom- incidence of workplace bullying and everyday experi- patible with pre-existing schemas regarding personal ences of members of ethnic and racial minorities in invulnerability, generosity of the world, meaning, self- the American workplace and found that, even though worth, and relations with others30). Eventually, an experiences of workplace mistreatment were signifi- adverse emotional reaction develops. cantly associated with negative emotional reactions There is now substantial evidence linking work for all targets, African-American and Hispanic/Latino characteristics, including workplace mistreatment, to workers who were targets of general bullying were the prevalence of work-related health problems2, 4). more likely to respond actively than white work- Given the complexity and variability of stressful ers36). A prospective cohort study of working women experience at work, it is essential to identify its harm- in the United States indicated that women exposed ful components with the help of theoretical models. to low social support at work are at significantly The aim of work stress models is to understand how higher risk of poor functional status and a decline in stress influences the pathogenesis of disease in work mental health over time37). The results of the current life. Two such models in this area have received study also highlighted a difference in the perception special attention: the effort–reward imbalance (ERI) of workplace mistreatment and health complaints model and cognitive activation theory of stress (CATS) between minor and dominant groups (Supplementary model. Table 2). Participants who worked with different Siegrist’s ERI model proposes that health risks gender co-workers showed a relatively higher risk derive from the mismatch between high efforts at of health symptoms due to workplace mistreatment, work and low rewards received31). This model focuses particularly female workers in Korea (Table 3). on the reciprocity of exchange in occupational life, We found an overall stronger relationship between where high-cost/low-gain conditions are considered workplace mistreatment and health problems for particularly stressful. Work effort is part of a social female workers than for male workers in Korea contract, which operates with three types of reward: (Table 3). Various studies have suggested the impor- money, esteem, and career opportunities, includ- tance of gender in the mistreatment process11, 12), with ing job security. Among these rewards, workplace results indicating that the effects of mistreatment tend mistreatment could be interpreted as a low level of to be greater in women. Gender differences in work- social approval, inducing lack of esteem. Lack of place mistreatment can by explained in part by the esteem and approval, despite high effort, may produce tendency of women to experience traumatic events emotional distress. The imbalance between high that differ fundamentally from those experienced by efforts and low rewards elicits recurrent feelings of males38). Additionally, there is strong evidence of a Minsang YOO, et al.: Workplace mistreatment and health problems 435 sex-related hemispheric lateralization of amygdala questionnaires that lacked highly structured questions function in memory of a traumatic experience, espe- or clear definitions. Forth, this study considered the cially in the processing of strong emotional memo- country as representing the culture, but there are many ries39). Similarly, another study revealed that post- other cultural determinants in addition to country that traumatic symptoms were more prevalent among should be considered, such as language, proximity women (49%) than men (35.3%) when exposed to and topography, religion, economic development, tech- workplace mistreatment11). On the basis of these find- nological development, political boundaries, industry ings, it has been suggested that gender differences in type, and climate. Although we adapted the PSR the cognitive appraisal of traumatic stressors are an model to compare overall workplace mistreatment important issue in explaining female–male differences, and health outcome according to nation, considering as well as gender-specific acute psychobiological reac- nations, which are heterogeneous in the EU, as repre- tions to trauma40). senting cultures would be inadequate. It is also noteworthy that educated workers in Notwithstanding these limitations, our sample Korea have a greater risk of health problems related included a very large number of employees from the to workplace mistreatment, but this is somewhat general working population, allowing us to study a difficult to interpret. When more educated Korean nonselective population, as well as men and women people experience mistreatment or harassment in the separately, which has been shown to be crucial. workplace, they more frequently they tend to report Moreover, as we compared representative data from health problems. This effect of educational level is Korea and EU countries, surveyed using almost iden- not apparent for European participants. There may tical methodologies and survey questionnaires, it was be cultural reasons for this difference between Korea possible to conduct in-depth comparison analyses. and European countries. One possible explanation In conclusion, workplace mistreatment is closely is that subjective social class identification is closely associated with physical and mental health prob- connected with educational level in Korea. There lems, especially among workers with higher educa- is a widespread belief among Koreans that higher tional levels and females who work alone in Korea. social status can be obtained as a result of educa- Awareness of these gender and educational differences tional achievement41). Therefore, higher educational is an important step in targeting preventive measures level is equated with higher social class. Moreover, and improving their efficiency. Special attention the general trend of change has been increasing job should be paid to these potentially vulnerable groups insecurity for white-collar workers in Korea and an in designing strategies to reduce health complaints due intensification of educational pursuits42), which is an to workplace mistreatment. expression of high effort. If someone with a higher educational level experiences mistreatment at work, References they are more likely to regard that event as distressing 1) Hodgins M, MacCurtain S, Mannix-McNamara and unacceptable than a worker with a lower educa- P. Workplace bullying and incivility: a systematic tional level. Workplace mistreatment being regarded review of interventions. International Journal of as low reward, despite high effort to gain high educa- Workplace Health Management 2014; 7: 54−72. tional level, may produce emotional distress. The 2) Hills D, Joyce C. A review of research on the prev- imbalance between high efforts and low rewards alence, antecedents, consequences and prevention of workplace aggression in clinical medical practice. would elicit psychological stress, which in turn would Aggression and Violent Behavior 2013; 18: 554−69. evoke somatic symptoms. 3) Einarsen S, Hoel H, Zapf D, Cooper C. Bullying The current study has several methodological short- and harassment in the workplace: developments in comings. First, the cross-sectional design of this theory, research, and practice. Crc Press: 2010. study did not allow us to assume a causal relation- 4) Moayed FA, Daraiseh N, Shell R, Salem S. ship between individual characteristics and the health Workplace bullying: a systematic review of effects of workplace mistreatment. There is a prob- risk factors and outcomes. Theoretical Issues in ability that health problems lead to mistreatment. Ergonomics Science 2006; 7: 311−27. Hence, the potential for overestimation or underesti- 5) Kim S-S, Williams DR. Perceived discrimination mation of risk remains. Second, our study was based and self-rated health in South Korea: a nationally on self-reported health problems, with no clinical representative survey. PloS One 2012; 7: e30501. 6) Lim JJC. A Cross-cultural Comparison and diagnosis. However, we believe that self-reported Examination of Workplace Bullying in Singapore assessment of work stress is not a source of major and the United States. University of Houston; 2011. bias in our study. Third, the different types of work- 7) Okechukwu CA, Souza K, Davis KD, de Castro place mistreatment, such as discrimination, harass- AB. Discrimination, harassment, abuse, and bullying ment, and violence, were assessed by a simple set of in the workplace: contribution of workplace injustice 436 J Occup Health, Vol. 57, 2015

to occupational health disparities. Am J Ind Med Stress 2013; 27: 321−38. 2014; 57: 573−86. 23) Kanetsuna T, Smith PK. Pupil insights into bully- 8) Grubb PL, Roberts RK, Grosch JW, Brightwell WS. ing, and coping with bullying: a bi-national study Workplace bullying: what organizations are saying. in Japan and England. Journal of School Violence Empl Rts & Employ Pol’y J 2004; 8: 407−523. 2002; 1: 5−29. 9) Escartín J, Zapf D, Arrieta C, Rodriguez-Carballeira 24) Farioli A, Mattioli S, Quaglieri A, Curti S, Violante A. Workers’ perception of workplace bullying: a FS, Coggon D. Musculoskeletal pain in Europe: cross-cultural study. European Journal of Work and the role of personal, occupational, and social risk Organizational Psychology 2011; 20: 178−205. factors. Scand J Work Environ Health 2014; 40: 10) Escartín J, Rodríguez-Carballeira A, Zapf D, Porrúa C, 36−46. Martin-Pena J. Perceived severity of various bully- 25) OECD Environmental Indicators: Development, ing behaviours at work and the relevance of expo- Measurement and Use. Paris: OECD; 2330. sure to bullying. Work & Stress 2009; 23: 191−205. 26) World Health Organization (2006) World Health 11) Rodríguez-Muñoz A, Moreno-Jiménez B, Sanz Report 2006: Working Together for Health. WHO, Vergel AI, Garrosa Hernández E. Post-traumatic Geneva. [Online]. 2006 [cited 2014 Jul 7]; Available symptoms among victims of workplace bullying: from: URL: http:// www.who.int/whr/2006/en exploring gender differences and shattered assump- 27) Liefooghe AP, Davey KM. 11 Explaining bullying tions. Journal of Applied Social Psychology 2010; at work. Bullying and emotional abuse in the work- 40: 2616−35. place. International Perspectives in Research and 12) Salin D. Ways of explaining workplace bullying: Practice 2003: 219. a review of enabling, motivating and precipitating 28) Janoff-Bulman R. Posttraumatic growth: three structures and processes in the work environment. explanatory models. Psychological Inquiry 2004: Human relations 2003; 56: 1213−32. 30−4. 13) Hoel H, Cooper CL. Destructive conflict and bully- 29) Mruk CJ. Self-esteem research, theory, and practice: ing at work. Manchester (UK): Manchester School Toward a positive psychology of self-esteem. NY: of Management, University of Manchester Institute Springer Publishing Company; 2006. of Science and Technology; 2000. 30) Tedeschi R, Tedeschi RG, Calhoun LG. Trauma and 14) Hansen ÅM, Hogh A, Mikkelsen EG. Physiological transformation: Growing in the aftermath of suffer- Response as Biomarkers of Adverse Effects of the ing. Newbury Park (CA): Sage Publications; 1995. Psychosocial Work Environment 2012; 565−80. 31) Siegrist J. Adverse health effects of high-effort/ 15) Vartia MA. Consequences of workplace bullying low-reward conditions. J Occup Health Psychol with respect to the well-being of its targets and the 1996; 1: 27. observers of bullying. Scand J Work Environ Health 32) Kivimäki M, Virtanen M, Vartia M, Elovainio M, 2001: 63−9. Vahtera J, Keltikangas-Järvinen L. Workplace bully- 16) Hansen ÅM, Hogh A, Persson R, Karlson B, Garde ing and the risk of cardiovascular disease and AH, Ørbæk P. Bullying at work, health outcomes, depression. Occup Environ Med 2003; 60: 779−83. and physiological stress response. J Psychosom Res 33) Ferrie JE, Head J, Shipley MJ, Vahtera J, Marmot 2006; 60: 63−72. MG, Kivimäki M. Injustice at work and incidence 17) Niedhammer I, David S, Degioanni S. Association of psychiatric morbidity: the Whitehall II study. between workplace bullying and depressive Occup Environ Med 2006; 63: 443−50. symptoms in the French working population. J 34) Stansfeld SA, Bosma H, Hemingway H, Marmot Psychosom Res 2006; 61: 251−9. MG. Psychosocial work characteristics and social 18) Vie TL, Glasø L, Einarsen S. How does it feel? support as predictors of SF-36 health functioning: Workplace bullying, emotions and musculoskeletal the Whitehall II study. Psychosomatic Medicine complaints. Scand J Psychol 2012; 53: 165−73. 1998; 60: 247−55. 19) Nielsen MB, Hetland J, Matthiesen SB, Einarsen 35) Rodríguez-Muñoz A, Notelaers G, Moreno-Jiménez S. Longitudinal relationships between workplace B. Workplace bullying and sleep quality: the mediat- bullying and psychological distress. Scand J Work ing role of worry and need for recovery. Behavioral Environ Health 2012; 38: 38−46. Psychology 2011; 19: 453−68. 20) Finne LB, Knardahl S, Lau B. Workplace bully- 36) Fox S, Stallworth LE. Racial/ethnic bullying: ing and mental distress—a prospective study of Exploring links between bullying and racism in the Norwegian employees. Scand J Work Environ US workplace. Journal of Vocational Behavior 2005; Health 2011: 276−87. 66: 438−56. 21) McCarthy P, Mayhew C. Safeguarding the organiza- 37) Cheng Y, Kawachi I, Coakley EH, Schwartz J, tion against violence and bullying: an international Colditz G. Association between psychosocial work perspective. NY: Macmillan; 2004. characteristics and health functioning in American 22) McTernan WP, Dollard MF, LaMontagne AD. women: prospective study. Bmj 2000; 320: 1432−6. Depression in the workplace: an economic cost 38) Crick NR, Grotpeter JK. Relational aggression, analysis of depression-related productivity loss gender, and social-psychological adjustment. Child attributable to job strain and bullying. Work & Development 1995; 66: 710−22. Minsang YOO, et al.: Workplace mistreatment and health problems 437

39) Cahill L, Uncapher M, Kilpatrick L, Alkire MT, Psychological bulletin 2007; 133: 183. Turner J. Sex-related hemispheric lateralization of 41) Hong D-S. Social change and stratification. Social amygdala function in emotionally influenced memo- Indicators Research 2003; 62: 39−50. ry: an FMRI investigation. Learning & Memory 42) Koo H. The changing faces of inequality in South 2004; 11: 261−6. Korea in the age of globalization. Korean Studies 40) Olff M, Langeland W, Draijer N, Gersons BP. 2007; 31: 1−18. Gender differences in posttraumatic stress disorder.