Office harassment : A negative influence on quality of work life

This paper addresses the effect of office harassment on the quality of work life of employees while job stress and organizational commitment were tested as mediators. Results suggested that organizational commitment mediated the relation between office harassment and quality of work life while job stress did not mediate the relationship. Furthermore, analysis confirmed that there existed an inverse relationship between office harassment and quality of work life. In the future, researchers could conduct the study across sectors as well as those sectors where women work under traditional roles settings.


INTRODUCTION
Workplace harassment is one of the most regular happenings in any workplace across the globe and the practice of harassment exists in every organization regardless of nature, size of that organization.The term harassment may be defined as any unwelcomed/ unwanted action/advance toward any employee that would hamper his/her work performance, furthermore, harassment can be termed as an oral or physical action which emits hostility for an individual due to his/her demography or ethnicity or that of people who are related to him/her and the intention is to create hostility and interfering with an individual's work performance and it can also affect negatively, a person's opportunities for the employment.
Harassment comes in many shapes; it can be in form of sexual harassment, gender harassment, generalized workplace harassment, racial harassment and ethnic harassment to name a few.Sexual harassment is a type that constitutes undesired sex based behavior existent as condition for employment or it may give birth to an environment that is hostile for the harassed (Fitzgerald, 1996); as far as the gender harassment is concerned, it is a non-sexual experience based on gender and it can include remarks such as females being incompetent, etc. (Buchanan and Fitzgerald, 2008); the term generalized workplace harassment means negative office interaction that affects the individual's job by affecting the terms and conditions and/or job decisions and such actions do not fall within the boundaries of law; ethnic harassment means a behavior that is hostile as well as one that separates an individual from the rest and the sole stimuli for this action or behavior is that person's ethnicity (Schneider et al., 2000) and racial harassment is the kind where discriminatory and hostile behavior is due to the racial differences among the individuals (Harrick and Sullivan, 1995).For the purpose of this study, gender harassment (GH) and generalized workplace harassment (GWH) have been chosen and because they are studied strictly in an office workplace context, they would be addressed collectively as workplace harassment in this study to address their collective effects on the quality of work life of an employee; quality of work life (QWL) can be explained as a set of elements such as job security, reward systems, training and advancement opportunities and participation in decision making, which in combined form, add positively towards the physical, as well as mental well-being of an employee and also adds towards the work that an employee performs in an organization, quality of work life can also be defined as being a multidimensional construct that includes concepts like job security, reward system, training and opportunities for career advancement and being able to participate in decision making (Dargahi and Yazdi, 2007).
In this study "job stress" and "organizational commitment" would be mediating the effect of harassment on QWL, job stress is stress as a state of disequilibrium in the system of variables, relating people to their environment that results in a change in people's normal levels of well-being (Hart et al., 1993), and organizational commitment is defined as containing in it, an employee's overall positive attitude for his/her organization (Porter et al., 1974).
The target industry for this study is the banking sector of Pakistan; the reason for choosing the banking sector is that it is one of the most stable and developed institutions of the country.

LITERATURE REVIEW
Harassment, an unwelcomed behavior exists in almost all types of industries, the effects of harassment are so profound that large number of researchers have studied this phenomena and the most studied type is that of sexual harassment which is mainly faced by the females in an organization, but for this study, two other types of harassment have been selected which are GH; gender harassment is the most common form of the harassment (Parker and Griffin, 2002), still, the prime target of gender harassment are women, but the nature of harassment is such that the harassing behavior is not necessarily sexualized (Piotowski, 1998), and GWH is the type that is repeatedly aggressive, as well as hostile verbal and nonverbal actions, and this type does not include aggressive or hostile behavior that is physical in its nature (Salin, 2009) and the target of this type of harassment can be both males and females, and the results of this harassment include unreasonable interference with an employee's work performance, affecting an employee's job prospects (U.S. Equal Employment Opportunity Commission, 1993).
For this study, the independent variable is assumed and termed as office harassment and it has two elements that are 1) GWH and 2) GH, both of these harassment types will collectively be addressed as office harassment for the purpose of this study.
For this study, GWH is the first element of the independent variable.Generalized workplace harassment is a form of psychological harassment and it is known by many different names including "generalized workplace abuse" (Richman et al., 1999), "mobbing" (Heinz, 1990) and generalized workplace harassment is also known as "victimization" (Aquino et al., 1999).Scholars have studied generalized workplace harassment under different contexts, for example, Ehrenreich (1999) tried to understand the sex based harassment; besides making a better understanding of the phenomena researchers like Stale and Bjorn (1997) have studied the workplace harassment in the context of victimization of male workers, revealing the ill effects of harassment in males; while others have studied GHW in the context of observing and evaluating the prevalence, perception and its Nazir et al. 10277 degree; and the impact of generalized workplace harassment (Sandvik and Tracy, 2006), has also been studied in the context of its effect on drinking outcomes (Richman et al., 2001); it has been studied that bullying increases risk of protracted duress stress disorder also known as PDSD (Scott and Stardling, 2001).Now, the bullying and generalized workplace harassment are the same, so one could observe the words often used interchangeably when the action is repetitive, continual, abiding and with all of this there is a power imbalance (Sandvik and Tracy, 2006) and others have studied GWH under the context of physical well-being deterioration (Kivimaki et al., 2005).In this study, the GH would be studied, affecting an employee's QWL.No matter the name, this type of harassment includes verbal aggression, humiliation and interruption, and isolation from important work activities (Lim and Cortina, 2005); moreover, generalized workplace harassment can also be looked at as debauching workplace interactions that do not absolutely involve gender (Richman et al., 1999) and it is sub-lethal, non-physical violence, it is deliberate, repeated, reasonably severe interpersonal hostility (Namie, 2003); further elaborating the previous statement, the generalized workplace harassment is that in which any employee at any stage of his/her career, regardless of their ethnicity, demography and gender, can be subjected to this form of harassment (Saunders et al., 2007).Different scholars have addressed GWH in different contexts and for the purpose of this study, the GWH is being studied under the context of its effect on QWL, and for this study, the GWH is defined as a kind of harassment that includes verbal aggression, disrespect, seclusion at workplace, a behavior that is threatening in nature and in addition all of these behaviors are repetitive in nature.For this study, gender harassment is the second element of the independent variable.In a country like Pakistan, women have started working along men, side by side in the organizations, and although women are being welcomed in the organizations across the country, the traditional organizational settings are still male dominated in their setup and as a result, women tend to face resistance in getting accepted as equals in the organizations; the behaviors that the working women most frequently face is the behavior of gender harassment; gender harassment is simply an inequity in gender (Herring, 1999), it can also be defined as the raw verbal or physical or both behaviors that represent themselves hostile and offensive, furthermore, gender harassment is a demeanor that is sexist in nature (Fitzgerald et al., 1997).The gender harassment does not fall under the jurisdiction of law because it is a behavior based on hostility that is without any definitive sexual motive (Lim and Cortina, 2005) and it only triggers an epithetical experience without the cost of evoked sexual coaction aimed towards member of a gender, customarily, women (Lim and Cortina, 2005).Many scholars have studied gender harassment affecting an employee at workplace, for instance, work by Raver and Nishii (2010) and Antecol and Cobb-Clark (2001) have studied gender harassment in the workplace context and observed their effects on the factors affecting an employee; some researchers have studied gender harassment in the context of real versus imagined harassment and they have documented the results of their experiments (Woodzicka and LaFrance, 2001); researchers have also studied gender harassment as a tool for one gender (men) to protest against the opposite gender (women) (Miller, 1997), while some have studied gender harassment's negative consequences on distress and overperformance demand (Parker and Griffin, 2002), others have studied on how to combat gender harassment in cyber space (Citron, 2009); some researchers have studied the effects and how employees cope with the harassment based on gender (Thorpe, 1990) and some researchers have tried to contribute in emerging concept in law regarding Discrimination (Ravitch, 1995).For the purpose of this study, the concept of gender harassment is being studied under the context of QWL, and gender harassment itself for the purpose of this study means the kind of harassment in which the harassed suffers from inequity, uncivil and hostile behavior.
The dependent variable for this study is QWL; originally, the term of quality of work life was used for the first time at international labor relations conference in 1972 (Hian and Einstein, 1990).QWL is very important for any organization as it is liable for attracting new workforce and retention of the existing workforce in any organization (Saraji and Dargahi, 2006).As far as the meaning of QWL is concerned, it is an umbrella term that includes many concepts (Krueger et al., 2002), quality of work life is a concept that aims at bringing the human aspect to the mechanized work environment of today, the aim is to ensure mental, as well as physical well-being of an employee and this concept can further be enhanced by inclusion of the notion of welfare that is obtained by the employee during the course of employment into the concept of quality of work life.QWL can also be defined as the consideration for the exigency and longing of an employee with regards to the working conditions, remuneration, and chances of professional development, work-family role balance, safety and social interactions at workplace and social relativity of employee's own work (Walton, 1973;Sirgy et al., 2001;Tzafrir and Gur, 2007).Some researchers have termed QWL as management philosophy which improves the dignity of an employee and their well-being and it paves the way for an organizational culture change (Ivancevich and Matteson, 1990).Some scholars define QWL as a combination of strategies, procedures and ambiance related to a workplace that altogether, enhance and sustain the employee satisfaction by aiming at improving work conditions for the employees of the organizations and as a result of that, improved QWL improves the effectiveness of the organizations for the employers (Saraji and Dargahi, 2006) and for some, QWL is favorable conditions and ambiances at a workplace which sustains and enhances employee satisfaction through rewards, professional growth chances and job security (Lau and May, 1998).Different scholars have been studying QWL under different contexts broadening the general understanding of QWL and how it affects and reacts in different situation, for example, scholars have observed how a company's perceived QWL affects its market and financial performance (Lau and May, 1998).Some researchers have studied the QWL in the context of retention of employees (Gilford et al., 2002) while some have studied the assessment of quality of the people's working life (Kirby and Harter, 2001).Quality of work life for this study is with regards to the gender and generalized workplace harassment and how these two types of harassment affect the QWL of an employee.For this study, the QWL is defined as the concept that aims at providing employee job satisfaction through safe healthy working conditions (Anderson et al., 1996), social equity and collaboration at workplace, in other words, QWL is studied through a psychosocial aspect.In this study, QWL is mainly evaluated through the context of job satisfaction; job satisfaction is further taken as a multidimensional psychosocial construct, and from a multidimensional point of view job satisfaction is considered as a global attitude, that employees experiment when they face during their work, which results from an employee's own perception regarding the work, organizational characteristics and individual need differences and individual values (Churchill Jr et al., 1974).
In this study, job stress and organizational commitment would mediate the relationship between independent and dependent variable.Stress occurs when the perceived pressure has exceeded one's perceived ability to cope such pressure (Palmer et al., 2003).Some researchers have defined stress as a relation between the "perception" of what is demanded from an individual and to that individual's "perception" of his/her capability to fulfill such demand, and an imbalance would exceed stress threshold, amounting to stress (Clancy and McVicar, 2002).The term job stress is used to distinguish stress from other sources from the stress generated from specific job of an employee.Job stress is defined as combined effect increased demand from job and low control over such job (Bosma et al., 1998).Furthermore, job stress is an emotional experience that contains the elements of fear, alarm, worry, vexation, anger, sorrow (Motowidlo et al., 1986), and job stress occurs in the attendance or the vacancy of more enduring departure from usual working (Parker and DeCotis, 1983).Scholars have used the job stress as mediating variables under different conditions; researchers have conducted a study observing the mediating effects of job stress in the relationship between safety attitudes and accident rates (Siu et al., 2004); furthermore, scholars have also studied optimism and job burnout in college students and they in this context have used job stress as a mediator (Chang et al., 2000).Another research studied job stress as mediator in the context of self-efficacy and how it may protect an individual from job strain experiences resulting in low chances for increased burnout (Schwarzer and Hallum, 2008).In this study, the job stress is taken as a mediator to mediate the relation between office harassment and quality of work life.
Organizational commitment is the second mediator in this study.Organizational commitment can be characterized by a willingness of an employee to expend much effort on behalf of the organization (Mowday et al., 1979); researchers suggest that the organizational commitment is a multidimensional concept that includes three components that are: normative, affective and continuance commitment (Meyer and Allen, 1991) while another study defined organizational commitment as having three main components which are, (a) firm belief in organization's goals and acceptance of such goals, (b) consent in exertion of extensive effort on the behalf of the organization and (c) an absolute desire to retain organizational identity (Porter et al., 1974); number of researches suggest that organizational commitment is connected to behaviors which bolster organizational effectiveness (Mayer and Schoorman, 1992;O'Reily III and Chatman, 1986) and this effectiveness in part affects the quality of work life because effectiveness would come from members of organization behaving in such a way that it is supportive of organizational goals which would only happen if their QWL is positive.For this study, organizational commitment is an employee's willingness to retain organizational membership and desire to associate him/her-self with the organization and its overall goals.

THEORETICAL FRAMEWORK
The study for this research work would be a non-intervention study as no intervention would be made and only data would be gathered from the researchable situation.The study includes an element of both qualitative and quantitative study.

Research model
In this model, office harassment is the independent variable which comprises further of two elements (1) gender harassment and (2) generalized workplace harassment; job stress and organizational commitment are two mediators in the relationship between independent and dependent variables that have been taken into account; in the model, the employee's quality of work life is the dependent variable which is affected by the office harassment.

Hypotheses
From the proposed model, for the purpose of this study, hypotheses have been drawn.It is hypothesized that office harassment has a negative effect on the quality of work life of an employee and hence, it is hypothesized that as the office harassment increase, the employee's quality of work life would decrease.Also, it is hypothesized that job stress mediates the relation between office harassment and an employee's quality of work life and so, the second hypothesis is formed, which states that job stress acts as a mediator for this model and carries the effect of office harassment to the employee's quality of work life.
Lastly, organizational commitment in the proposed model mediates the relation between office harassment and an employee's quality of work life and this forms the hypothesis which states that organizational commitment acts as a mediator and carries the effect of office harassment to the employee's quality of work life.

Sample size
For this study, the target sample was the banking sector of Pakistan, and females were the main respondents, being women employees.The type of sampling was convenience sampling due to the budget and time constraints.To collect the data, a questionnaire was used which consisted of questions regarding gender harassment, generalized workplace harassment and the level of quality of work life of the bank employees.The questionnaire was administered in banks that are top players in the banking industry and represents the industry of the country.Questions related to job stress and organizational commitment was also included in the questionnaire to analyze the mediation between these two variables provided.Of the questionnaires circulated, overall response rate was 38%.
The software used for this study was SPSS; this particular software was used due to its easy availability and also due to its ease of use and data importability and exportability to other MS office packages.SOBEL test was run to find out mediation between independent and dependent variable; further conditions set by Baron and Kenny (1986) were also considered and checked.

RESULTS AND DISCUSSION
Reliability and validity of data is presented in Table 1.Result of the SOBEL test is presented in Tables 2 to 5. In this study, the independent variable is office harassment, which contains two elements, which are GH and GWH.A research suggests that in the banking sector, 11.6% of the employees got harassed in the last 6 months while 24.6% were harassed in the last 5 years and 39.6% witnessed GWH in the last 5 years; the study suggested that women (11.4%) more than men (9.9%), suffered from GWH, and the most common GWH experience was found to be withholding of information which was 54% occasionally and 13.9% on a weekly or daily basis.This practice was followed by task given to employees that had unreasonable (49.3%) or impossible (7.8%) targets or deadlines (Hoel and Cooper, 2000).GWH has been found to have significant effects on employees, the α levels were as follows: α for verbal aggression was 0.80 for females and males; α for disrespectful behavior was 0.80 for females and 0.81 for males; α for isolation/ exclusion was 0.77 for females and 0.76 for males, and α for physical aggression was 0.39 for females and 0.63 for males (Richman et al., 2001).Study conducted regarding organizational effects of GWH revealed, commitment (β = -0.20,p < 0.01), job satisfaction (β = -0.29,p < 0.01),   turnover intentions (β = -0.27,p < 0.01), anxiety and depression (β = -0.28,p < 0.01), and physical symptoms (β= -0.26, p < 0.01) and the relationship with life satisfaction was also significant (β = -0.18,p < 0.01) which meant that GWH was negatively related to organizational commitment (β = -0.20,p <0.01), job satisfaction (β = -0.29,p < 0.01), turnover intention (β = 0.27, p < 0.01), anxiety and depression (β = 0.28, p < 0.01) and physical health symptoms (β =0.26, p < 0.01), (Raver and Nishii, 2010).GWH also has been found to be positively related to the stress r (398) = 0.341, p < 0.001, and furthermore, GWH was found to be related negatively with the job satisfaction r (401) = 0.589, p < 0.001 (Sandvik and Tracy, 2006).The second element of independent variable was GH.Study revealed that GH negatively affects organizational commitment (β = -0.17,p < 0.01), job satisfaction (β = -0.23,p < 0.01) and turnover intentions (β = -0.28,p < 0.01) and gender harassment was also found negatively related to organizational commitment (β = -0.17,p < 0.01) (Raver and Nishii, 2010).In another study conducted, it was found that due to increase in gender harassment, women employees felt that they needed to over perform and that caused psychological distress; gender harassment was positively related to over-performance demands (β = 0.27, p < 0.01) and over-performance was found in the study to positively related with psychological distress (β = 0.28, p < 0.001), furthermore, women reported more gender harassment t (574) = 2.65, p < 0.01 and overperformance demands than males t (574) = 11.83,p < 0.001 with means 1.93 (1.04) and 1.70 (1.02) for females and males respectively (Parker and Griffin, 2002), suggesting that the majority of the affected are females.
Researchers have found that gender harassment causes symptoms for psychological trauma (Berg, 2006) while other scholars have established that daily encounter of gender harassment leads to anger, anxiety and depression (Swim et al., 2001) that leads to stress which disrupts the overall job satisfaction of an employee, decreasing his/her QWL.Studies have further confirmed that gender harassment causes loss in productivity, decrease in job satisfaction and professional relationship, and increase in turnover intentions and behaviors (Barling et al., 2001;Langhout et al., 2005), and some scholars have established that gender harassment causes, besides low psychological well-being, greater physical health problems as well (Culbertson and Rosenfeld, 1994;Fitzgerald et al., 1997;Richman et al., 2002) and an increase in an employee's job stress (which results from lower psychological well-being) causes decrease in the employee's job satisfaction with the correlation between job stress and job satisfaction of r = -0.463,p < 0.01 and correlation between job stress and skill development at r = -326, p < 0.01 (Imtiaz and Ahmad); further researches by scholars have also reinforced these findings that gender harassment causes elevated symptoms of turnover intentions averaging 0.48 on Cohen's effect size (d), increase in coworker dissatisfaction was at 0.79 and decrease levels of work satisfaction was recorded at 0.58, Cohen's effect size (d) for organizational commitment, emotional and physical health also ranged from 0.22 to 0.79 averaging at 0.48 (Leskinen et al., 2010).Here, the Cohen's effect size (d) suggests that if the values are around 0.30, then the effect is small, if value is around 0.50, the effect is medium and beyond that, the effect is large.From the results of scholars, the first hypothesis is proved that OH decreases the QWL of an employee as it is proven by scholars that both the constituents (GH, GWH) of office harassment cause significant adverse effects on the elements that account for the QWL of an employee and one of the important elements is that of job satisfaction, although it is not QWL, perception of QWL is usually measured by job satisfaction surveys (Krueger et al., 2002); furthermore, job satisfaction is direct positively linked with job performance (standardized path coefficient = 0.49, t = -4.81)and job satisfaction is linked negatively with turnover intention (standardized path coefficient = 0.34, t = -2.37)which was further positively and strongly linked with turnover intention (standardized path coefficient = 0.39, t = -2.70)(Ruyter et al., 2001) and this shows the chain reaction of this element of job satisfaction on many aspects of QWL.The results obtained by applying the SOBEL test in SPSS software revealed fulfilling the conditions set by Baron and Kenny (1986); in the 1 st step, the relationship between independent variable (gender harassment) and dependent variable (quality of work life) was p = 0.000, and in the 2 nd step, the relationship between independent variable (gender harassment) and mediator (job stress) was significant at p = 0.000) while at 3 rd step, the relation between mediator (job stress) and independent variable (gender harassment) was found not significant at p = 0.242 and the relation between independent variable (gender harassment) and dependent variable remained significant at p = 0.019 which suggested that the mediation requirements set by Baron and Kenny were not fulfilled, hence, no mediation.Furthermore, there was a significant initial relationship between the gender harassment and quality of work life ( = -0.62,p = 0.00) that was not affected after controlling for the mediator ( = -0.45,p = 0.01) which indicates that job stress does not mediate the relationship between the GH and QWL.
SOBEL test indicates p = 0.244 which suggests that job stress is a not a significant mediator between gender harassment and quality of work life.Secondly, fulfilling the conditions set by Baron and Kenny (1986), in the 1 st step, the relationship between independent variable (gender harassment) and dependent variable (quality of work life) was p = 0.001, and in the 2 nd step, the relationship between independent variable (gender harassment) and mediator (job stress) was significant at p = 0.000 while at 3 rd step, the relation between mediator (job stress) and independent variable (gender harassment) was found significant at p = 0.013 and the relation between independent variable (gender harassment) and dependent variable became insignificant at p = 0.249 which suggested that the mediation requirements set by Baron and Kenny were fulfilled.
Initially, the relationship between the generalized workplace harassment and quality of work life is = -0.49,p = 0.00 and after controlling for the mediator, the relationship showed = -0.19,p = 0.24 which indicates that job stress mediates the relationship between the GWH and QWL.SOBEL test indicates p = 0.018, which means that job stress is a significant mediator between generalized workplace harassment and quality of work life.
Thirdly, fulfilling the conditions set by Baron and Kenny (1986), in the 1 st step, the relationship between independent variable (gender harassment) and dependent variable (quality of work life) was p = 0.000, and in the 2 nd step, the relationship between independent variable (gender harassment) and mediator (job stress) was significant at p = 0.000 while at 3 rd step, the relation between mediator (job stress) and independent variable (gender harassment) was found significant at p = 0.000 and the relation between independent variable (gender harassment) and dependent variable became insignificant at p = 0.187 which suggested that the mediation requirements set by Baron and Kenny were fulfilled.
There was a significant initial relationship between the gender harassment and quality of work life ( = -0.62,p = 0.00) that was non-significant after controlling for the mediator ( = -0.16,p = 0.18), which indicates that organizational commitment mediates the relationship between the GH and QWL.SOBEL test indicates p = 0.000 which means that organizational commitment is a significant mediator between gender harassment and quality of work life.
Lastly, fulfilling the conditions set by Baron and Kenny (1986), in the 1 st step, the relationship between independent variable (gender harassment) and dependent variable (quality of work life) was p = 0.001, and in the 2 nd step, the relationship between independent variable (gender harassment) and mediator (job stress) was significant at p = 0.001, while at 3 rd step, the relation between mediator (job stress) and independent variable (gender harassment) was found significant at p = 0.000, and the relation between independent variable (gender harassment) and dependent variable became insignificant at p = 0.202 which suggested that the mediation requirements set by Baron and Kenny were fulfilled which suggested that the mediation requirements set by Baron and Kenny were fulfilled.
There was a significant initial relationship between the generalized workplace harassment and quality of work life ( = -0.49,p = 0.00) that was non-significant after controlling for the mediator ( = -0.13,p = 0.20) which indicates that organizational commitment mediates the relationship between the GWH and QWL.SOBEL test indicates p = 0.001, which means that organizational commitment is a significant mediator between generalized workplace harassment and quality of work life.The first hypothesis was proved as there was negative relation observed between office harassment and quality of work life, in second hypothesis job stress mediated the relation between office harassment and an employee's quality of work life but job stress did not mediate gender harassment which was an element of office harassment, and organizational commitment successfully mediated relation between office harassment and quality of work life.

CONCLUSION AND RECOMMENDATIONS
As the Pakistani culture is conservative, it is the reason that women in banking sector were much hesitant in responding to the questions related to harassment and so, the data gathering was very difficult and also, much resistance was shown on respondent's part where many denied responding to the questionnaire, and as the present study has attempted to shed light on how office harassing behaviors affects the quality of work life of an employee, while job stress and organizational commitment mediate this process, in future, it is recommended that research study regarding harassment should be conducted and the respondents be educated about the issue of harassment so that more confidence is gained from the respondents to get accurate and true response, keeping this in mind, there still exists a greater need to explore the issue of harassment in Pakistani work environment where the problem of harassment exists but people are afraid to talk about it.Another future research path can be to replicate the study in other sectors where women carry out traditional work roles and are in majority.A researcher can further study the effects of studied type of harassment on the quality of work life of female industrial worker of the country.

Table 1 .
Cronbach's alpha for questionnaire items

Table 2 .
SOBEL test for mediation between QWL, GH and JS

Table 3 .
SOBEL test for mediation between QWL, GWH and JS.

Table 4 .
SOBEL test for mediation between QWL, GH and OC.

Table 5 .
SOBEL test for mediation between QWL, GWH and OC.