MANAGING OCCUPATIONAL STRESS IN HUMAN SERVICE NONPROFIT ORGANIZATIONS IN MYKOLAIV, UKRAINE

!e present research responds to the lack of research on occupational stress in human service nonpro"t organizations in Ukraine. !e study identi"ed the main stressors in human service nonpro"ts in Mykolaiv, Ukraine. !en the relationship between occupational stress and the fourteen speci"c stressors was tested in 11 human service nonpro"t organizations in Mykolaiv. !e empirical study revealed a positive relationship between occupational stress and eleven stressors. !e strongest relationships were observed between job demands that fall short of employees’ skills, and job demands that exceed employees’ time to meet them. Two other stressors – low salary and clients’ trauma – were revealed to be the strongest determinants of stress, accounting for over half of its variation. !e study also discovered that employees who work with di#erent client groups experience varying stress levels, and those who work with several groups, as opposed to one, experience more stress. As the "rst study in human service nonpro"t organizations in Mykolaiv and entire Ukraine, the present research lays the groundwork for subsequent research.


Introduction
Occupational stress has been proven by numerous researchers to be a cause of adverse e ects on the emotional, mental and physical abilities of workers that decrease their individual productivity and signi cantly undermine the well-being of entire organizations (Palmer & Gyllensten, 2008). Negative e ects of occupational stress prompt thousands of researchers to examine stress and stress factors and devise stress management mechanisms, and thousands of businesses to seek ways to mitigate stress of their employees.
However, although much research has been done on occupational stress in the past 40 years, it has largely ignored human service nonpro t organizations that are rarely considered as workplaces by researchers despite employing a rapidly increasing number of people and constituting environments prone to occupational stress (Kosny, 2011;Schmid, 2004). e scarce existing research on occupational stress does not o er anything close to holistic examination of occupational stress in nonpro t organizations but rather focuses on speci c isolated issues (Feeney & Bozeman, 2009;Hulbert & Morrison, 2006) in very speci c organizations (Demmer, 2002). e present research, therefore, a empts to ll this gap by examining occupational stress and a number of stressors in a variety of human service nonpro t organizations.
e goal of the present research was to identify stressors in human service nonpro t organizations in Mykolaiv, Ukraine, in order to suggest empirically grounded recommendations for occupational stress reduction.

De ning concepts of stress, occupational stress, stressors, and human service nonpro t organizations
Stress. Stress has been de ned in many di erent ways in the literature. e present research uses the concept of stress as de ned by Health and Safety Executive (HSE, 2001) as "the adverse reaction people have to excessive pressures or other types of demand placed on them" (as cited in Palmer & Gyllensten, 2008, p. 240).
Due to a number of de nitions and viewpoints on stress, it is important to state what stress is not in the present research. Two reservations are worthy of noting. First of all, stress here is not regarded a natural and inevitable part of life, but rather one that causes adverse e ects. Stress indeed has been acknowledged by some researchers to be a natural and inevitable part of life (Kearns, 1973, p. 28;Nelson, Li le & Frazier, 2008, p. 55). Yet, for the purpose of the present research it is regarded as an adverse phenomenon that can and should be eliminated from daily life.
Secondly, stress in the present research does not refer to eustress, but rather to distress, or strain. Most researchers distinguish between positive and negative forms of stress, namely eustress and distress (or strain) respectively (Nelson, Li le & Frazier, 2008, p. 55). Some researchers argue that eustress, de ned as a positive response to a stressor that leads to positive psychological states, should be included in the assessment of stress (Nelson & Simmons, 2004as cited in Nelson, Li le & Frazier, 2008. For the purpose of the present research that aims at uncovering factors that cause adverse e ects on employees in human service nonpro t organizations, eustress is not considered part of stress. e adopted de nition of stress focuses on its negative form that had adverse consequences for employees in organizations. Occupational stress. Occupational stress is a term that has not been de ned exactly in the literature (Storey & Billingham, 2010). It is o en used synonymously with the term work-related stress ( Johnson et al., 2005;Storey & Billingham, 2010;Kosny, 2011) and refers to stress experienced at the workplace, particularly in relation to one's occupation. e European Commission (1999, as cited in Palmer & Gyllensten, 2008 de nes work-related stress as the "emotional, cognitive, behavioral and physiological reactions to aversive and noxious aspects of work, work environments and work organizations" (p.240).
However, there are two considerations with respect to two terms, which caused occupational stress to be preferred and chosen as a term in this research. e rst one is that Cox, Gri th and Houdmont (2006) claim that occupational stress refers to "cases where work is the sole cause of the experience of stress and associated symptoms of ill health" as compared to work-related stress that is originally caused by factors other than work and in which work acts as an aggravating factor (p.3). e second consideration is that since aspects of work, work environments and work organizations di er tremendously across occupations -the research on work-related stress shows vast di erences between stress levels and stress-causing factors across various occupations (Kosny, 2011;Johnson et al., 2005). ese ndings, then, make discussions about undi erentiated work-related stress rather limited; they seem to call for a more distinct term that would incorporate occupational di erences in its de nition. Given these considerations, occupational stress was chosen as a term to be used in this research as the most appropriate and re ective of research in the eld. In addition, it is also more suitable to the goal of research that aims at analyzing stress in the particular occupation.
Stressors. Stressors refer to "environmental factors that function as sources of stress" (Cooper et al., 2001, as cited in Palmer & Gyllensten, 2008. Institute of Work, Health and Organization de nes stressors as "those aspects of work design, and the organization and management of work, and their social and organizational contexts, which have the potential for causing psychological or physical harm" (Cox, Gri ths & Houdmont, 2006).
Human service nonpro t organizations. Human service nonpro t organizations (Schmid, 2004) are characterized by two distinctive features. First, they are non-forpro t organizations, meaning that their establishment and operation are not in uenced by motives of making pro t for the bene t of investors, but rather for improvement of their services (Small Business Administration, N/A). Second, human service nonpro ts are organizations that provide social services to vulnerable groups of people whose needs are not addressed by government or other organizations. Such groups include the unemployed, the homeless, children, elderly, low-income people, people with alcohol and/or drug additions, patients in terminal care, people a icted with severe diseases (e.g., cancer, AIDS, tuberculosis). Human service nonpro t organizations' services include "individual and family services (social counseling, welfare), job training (training, work experience, vocational or rehabilitative courses for the unemployed, underemployed, and physically challenged), day care and residential care (children, elderly), drug counseling, emergency food distribution assistance" and also legal advocacy (Cnaan, 2002, as cited in Schmid, 2004).

eories of occupational stress
ere has been much controversy and debate regarding the process in which stressors cause stress, which consequently gave rise to several theories, the most prominent of which are interactional and transactional (Palmer & Gyllensten, 2008, p.240;Storey & Billingham, 2010, p.660;Cox, Gri ths & Houdmont, 2006).
Interactional theories. Interactional theory includes the famous French and Caplan's Person-Environment Fit model and Karasek's Job Demands and Control model (Cox, Gri ths & Houdmont, 2006). ese classic and still widely used models have nevertheless been severely criticized and signi cantly revised. Person-Environment Fit model sees stress as a result of a mis t between the person and the work environment (whether between supplies and values, or demands and abilities) that leads to psychological, physiological and behavioral strains (Edwards & Harrison, 1993). is model has been criticized for its theoretical, methodological and empirical insu ciencies (Edwards & Cooper, 1990;Edwards & Harrison, 1993).
e Job Demand and Control Model focuses on the interaction between objective demands of the work environment and the amount of control allo ed to the worker. e original model has been generally found too limited in its focus on only two concepts, so it had to be supplemented by a variety of factors that a ect the original dichotomous relationship: social support (Hall, 1988as cited in Goh, Sawang & Oei, 2010, work experience (Dollard & Wine eld, 1998), workplace learning (Panari et al., 2010), and culture (Gyorkos, 2012). e theory was also criticized for ambiguous conceptualization and operationalization of the decision latitude construct, the nature of relationship between demand and control (Cox, Gri ths & Houdmont, 2006). Transactional theory. Although the two interactional models are still applied and used in current research, they have been generally supplemented and even supplanted by the transactional theory of stress, developed by R. Lazarus and recognized to be more advanced than interactional theories (Cox, Gri ths & Houdmont, 2006, p.4). e strength of the transactional theory is that it views stress as a dynamic relationship between the person and the work environment, which accommodates subjective experience of an individual in a way that interactional theories do not (Cox, Gri ths & Houdmont, 2006, p.4). In transactional theory an individual's stress level is determined by his/her subjective appraisal of the environment, namely, the demand placed on him/her, and the availability of coping resources to meet that demand (Lazarus & Launier, 1978as cited in Guinn, Vinvent & Dugas, 2009Lazarus, 1990, as cited in Cox, Gri ths & Houdmont, 2006). e cause of stress does not exist exclusively in the individual or the environment, but rather arises in the transaction between the two (Cooper et al., 2001, as cited in Palmer & Gyllensten, 2008. e theory, therefore, acknowledges and accounts for the variation of responses to the same conditions as expressed by di erent individuals. Transactional theory is considered to be a more advanced theory, because it o ers a more comprehensive model of the stress process. According to Lazarus andFolkman (1984, as cited in Goh, Sawang &Oel, 2010), the model is linear and consists of four main stages and three paths (Figure 1). Its central tenet is that a potentially stressful event will trigger the primary appraisal process in which an individual assesses the degree to which this threat relates to his/her well-being (p.12). If an event is perceived as threatening or challenging, the secondary appraisal process starts, whereby the individual assess his/her coping resources to manage the threat. If coping resources are available, the person will implement them. Coping strategies refer to a "complex, organized sequence of behaviors that include cognitive appraisal, action impulses, and pa erned somatic reactions" (Goh, Sawang & Oel, 2010, p.13). e e ectiveness of one's coping processes, then, determines the eventual psycho-physiological reactions of the potentially stressful event.   Figure 2). First, the new model added a stress outcome stage (Stress Symptoms 1) between secondary appraisal and coping strategies, which is more re ective of the actual process of an individual experiencing stressful events. Stressful psycho-physiological arousal (Stress Symptoms 1) then in uences the initiation of coping strategies to manage the stressor, which leads to a new level of psycho-physiological stress experience (Stress Symptoms 2), based on Stress Symptoms 1. Second, the new model recognized that the process of experiencing stress starts when one's cognitive appraisals are activated by a stressful encounter in the Primary Appraisal stage, which led to the addition of another path. irdly, the new model recognized a path from Stress Symptoms 1 to psycho-physiological experience a er coping (Stress Symptoms 2). is is explained by the fact that psycho-physiological FIGURE 2: Revised 6-path Transactional Model (adapted from Goe, Sawang & Oei, 2010) responses are continuous experiences in the process of stress and coping, so they are expected to be continued a er coping. is revised transaction model demonstrates the dynamic nature of psychophysiological stress experience and provides a sound theoretical model for the empirical assessment of occupational stress.

Stressors
Various researches compile di erent lists of factors that cause occupational stress. ey range from ve ( Johnson et al., 2005), six (Storey & Billingham, 2010), seven (Schuler, 1982) to 24 factors (Williams & Cooper, 1998). Due to the limited space, the most widely cited factors were chosen for the present research, and grouped into ve categories: job demands, job rewards, roles in the organization, relationships at work, and work with clients.
Job demands. is stressor arises out of the person-environment t theory proposed by French and Caplan and is characterized by a mismatch between job demands and person's abilities (Edwards, 1996, p.296). Abilities include the personal resources that an individual can draw upon in order to meet job demands. ey include skills and knowledge that grow with use, and time and energy that diminish with use. Demands refer to the quantitative (quantity) and qualitative (level of di culty) requirements placed on the person. Demands can be both objective and subjective, but stress arises when perceived/subjective demands deviate from the person's abilities to meet them. Mismatch between demands and abilities results in work overload or underload that have been widely acknowledged to function as sources of stress ( Judith & Storey, 2010;Schuler, 1982;Sadri & Marcoulides, 1997).
Work overload. Work overload causes stress when demands excess abilities in either quantity or quality (hence qualitative and quantitative overload in Schuler, 1982). Work overload has been acknowledged to be an important stressor in the human service nonpro t organizational environment (Kosny, 2011, Zhuk et al., 2009). According to Kosny (2011), high workload is caused by the processes of deinstitutionalization and welfare restructuring, whereby government transferred its responsibility of delivering social services to human service nonpro t organizations, and failed to provide them with any support (Sidelnyk, 2010). Downsizing or outright lack of enough employees in human service nonpro t organizations o en causes employees to work more and to take on additional responsibilities outside of their level of expertise. Feeney and Bozeman (2009) indirectly support this claim in their study of work hours in public versus nonpro t sectors in the U.S., where they discovered that workers in the nonpro t sector spend more time at work (p.472). Zhuk et al. (2009) in their research on Ukrainian counter-tra cking nonpro t organizations support this point -NPOs in Ukraine o en encourage employees to take on additional responsibilities of fundraising, human resource or institution management.
Work underload. Work underload causes stress when abilities exceed job demands in two cases, either depletion or interference (Edwards, 1996, p.298). Stress in depletion arises when insu cient demands cause unused skill to atrophy, so that future demands cannot be met, which leads to work overload. us, poor working schedules may cause employees to experience either work overload or underload, hence being in a constant state of stress. Interference causes stress when developing and maintaining specialized skills beyond those needed to meet job demands and prevents the person from learning other required skills.
Bureaucratic work. Overwhelming bureaucratic work has been cited as a signi cant stressor in the human service nonpro t organizations (Kosny, 2011;Demmer, 2002). Kosny cites the research of Baines et al. (2002) that concludes that stringent reporting and documentation requirements that accompany funding create additional work for employees. As a result, they are most likely to do it in their own time in addition to their usual responsibilities (p.476). It is this bureaucratic work that causes the most severe stress, according to Demmer's research in AIDS organizations (2002).
Job rewards. is stressor arises out of the E ort-Reward Imbalance model proposed by Siegrist that holds that stress develops from an imbalance between e ort expended and rewards received (Cox, Gri ths & Houdmont, 2006, p.3). E orts include work time, skills, and emotions that an employee invests in order to meet job demands. Rewards distributed by the employer consist of wages, job security and career opportunities ( Jonge et al., 2008).
Low salary. Ongori and Agolla (2008) suggest that one of the major factors that causes stress for employees is low or inadequate salary. According to their study, 73 per cent of the respondents report low salary to be one of the stressors that they experience on the job (2008). Kosny (2011) supports this nding by her assessment of the environment of human service nonpro t organizations, which not only does not focus on nancial rewards due to its non-pro t nature, but also does not possess enough resource to distribute as rewards. National governments o en transfer to the nonpro ts the delivery of services to marginalized populations that are incapable of paying for services received. Sindelnyk (2008) agrees that human service nonpro t organizations su er from irregular and insu cient funding. is insu ciency means that organizations cannot employ the needed workforce and thus they burden their few employees with additional, "background" work that is necessary to the smooth functioning of the organization (Kosny, 2011, Zhuk et al., 2009. is work, however, is neither recognized nor rewarded by funding bodies, so employees must carry out tasks that surpass their formal job description without any rewards (Kosny, 2011, Zhuk et al., 2009. Job insecurity. Job insecurity is another important factor ( Johnson et al., 2005) that is, however, characterized by some controversy. Lack of su cient and stable nancial resources forces human service nonpro t organizations to lay-o their sta , which creates the feelings of anxiety and job insecurity among employees (Kosny, 2011). Although job insecurity and threat of turnover are emphasized by Kosny (2011) as being very important determinants of stress, Demmer's research (2002) in AIDS social organizations revealed that workers rank job insecurity as having the lowest severity of causing stress. Indeed, Kosny (2011) admits that many workers in human service nonpro t organizations are paid very low wages or are not paid at all, which in addition to the organizations' chronic lack of sta creates hardly any reason for the feelings of job insecurity.
Career development. is factor has been named by several researchers (Schuler, 1982;Johnson et al., 2005) to be an important stress factor. Stress from uncertainty about promotion arises when people express concerns about their future and expect that their work will be rewarded, which leads them to seek opportunities for advancement at work and waiting to be promoted (Sadri & Marcoulides, 1997;Ongori & Agolla, 2008). However, organizational structures in small human service nonpro t organizations are usually at, which means that there are hardly any hierarchical layers, hence opportunities, for promotion.
Roles in the organization. With respect to roles in the organization, most stress arises from role con ict and role ambiguity (Schuler, 1982;Johnson et al., 2005;Storey & Billingham, 2010).
Role ambiguity. Role ambiguity refers to employee's uncertainty about his/her duties and responsibilities, expectations and evaluation criteria (Storey & Billingham, 2010). Gilboa et al. (2008) further de ne it as the relative unpredictability of the outcome of an individual's behavior, which includes the lack of input from the environment to guide the individual's behavior (p.230-231). ey argue that it is a rather strong stressor because of the very few coping processes that could possible counteract its negative e ects. Role ambiguity is structurally determined, which means that when it is high, an individual faces di culty in pursuing job assignments because of his inability to modify them. Employee's lack of knowledge of what is expected of him/her hampers any purposeful e ort to a ain job objectives. us, role ambiguity is the most detrimental stressor to job performance (Gilboa et al., 2008, p.250).
Role con ict. Role con ict arises when an individual is confronted with two or more conflicting or opposing role expectations and demands from others (Kahn et al., 1964 as cited in Schmidt et al., 2012, p.1). is psychological con ict leads to the inability of the person to ful ll expectations regarding every role. Although most o en considered to be a stress factor ( Jonson et al., 2005), Teh, Yong & Lin (2012) cite research that considers role con ict a mediator variable between workaholism and job demands, burnout and well-being (p.3). is means that role con ict may actually be negatively related to stress.
Work Relationships. Work relationships have been considered an important part of organizational climate and a common source of occupational stress (Schuler, 1982;Johnson et al., 2005;Storey & Billingham, 2010). Stress may arise from relationships with supervisors, co-workers and the society external to the organization.
Dysfunctional relationships with supervisors. According to Schuler (1982), stress from relationships with supervisors arises from the various job rules and constant pressure to do more (p.10). Constraining relationships with supervisors deny the individual's ful llment of the need to have control over the job, and also the need for recognition and acceptance as a free and competent individual. Incorrect management style and excessive authority centralization (Bucurean & Costin, 2011) and discrimination and favoritism (Sadri & Marcoulides, 1997) make workers feel pressured at the workplace.
Relationships with co-workers. Although Kosny (2011) contends that environments of nonpro t organizations are usually caring and supportive (p.470), con icts with co-workers can be common sources of stress in the workplace (Sadri & Marcoulides, 1997). Because organizational environments of human service nonpro t organizations are emotionally complex and fragile, not only do con icts (negative phenomena) act as stressors, but also lack of support (absence of positive phenomena) is shown to be a source of stress. Lack of support from co-workers can be detrimental to the workers' well-being. Demmer's research (2002), for instance, identi ed lack of support as the most important stressor.
Absence of institutional mechanisms that would enable care-givers to talk about their experiences exacerbates the e ect of other stressors. Denial and suppression make the impact of other stressors much more acute, as evident in the research done in the intensive care unit in a public hospital in Sydney, Australia (Sorensen & Iedema, 2009).
e negative e ect of the lack of the institutional process for sharing emotions, fears and concerns was also expressed in Zhuk et al. 's study (2009).
Negative social reactions. Lack of support and appreciation of human service nonpro t organizations' employees from the larger society can also cause stress. Nurses who participated in the research of Gossman and Silverstein (1993) reported negative reactions of their families and relatives toward their work with people infected with HIV/AIDS as a strong factor of stress.
Work with clients. e peculiar feature of the work in human service nonpro t organizations is that a large share of its employees' work involves work with clients who have high emotional and physical needs. Kosny (2011) stressed that human service workers experience high levels of stress as a result of "working with clients who have high needs" (p. 474). According to her research, working with marginalized populations is characterized by three major stressors: clients' need for emotional support, clients' trauma and clients' violence.
Clients' need for emotional support. e rst factor that has crucial importance on employees who work with clients with AIDS or those in palliative care is work of supporting entire families through illness and bereavement of their members (Demmer, 2002;Kosny, 2011;Hulbert & Morrison, 2006;Sorensen & Iedema, 2009). In order to support them, social workers engage in emotional labor, which is defined as the e ort involved when employees "regulate their emotional display in an a empt to meet organizationally-based expectations specific to their roles" (Brotheridge & Lee, as cited in Sorensen & Iedema, 2009). In palliative care, emotional labor is stressful for those caregivers who have not developed a positive a itude to death and experience discomfort in their interactions with dying people (Sorensen & Iedema, 2009, p.9). Emotional labor is further exacerbated if feelings and struggles associated with illness and death are denied; which rst leads to anxiety and then to stress (Obholzer, 2005, as cited in Sorensen & Iedema, 2009. Clients' trauma. Another stressor is clients' trauma that is su ered by social workers vicariously. ose who work with clients that have experienced various types of trauma must deal with witnessing either directly or indirectly their clients' su erings. Caregivers thus o en identify themselves with and "adopt" their clients (Kosny, 2011;Grossman & Silverstein, 1993). Care-givers who work with patients su ering from AIDS and patients in palliative care must also deal with their patients' death (Grossman & Silverstein, 1993, Hulbert & Morrison, 2006. e severity of clients' death factor among AIDS care-givers seems to diminish in its function as a source of stress. e study on a itudes of employees in AIDS service organization in New York, conducted by Demmer (2002), revealed that active antiretroviral therapy that transformed caregivers' work from terminal care toward chronic disease management, made the factor of patients' death less severe. As death moved to the background, caring for patients with chronic illnesses emerged to the forefront as a severe factor, as argued by Grossman & Silverstein (1993), because it requires more time and an on-going e ort.
Clients' violence. e third factor that was revealed by research to be a stressor for employees who work with clients is the experience of violence as in icted by their clients (Kosny, 2011). Violence can take forms of verbal abuse, threat of physical violence and sexual harassment (McDonald & Sirotich, 2005). Kosny (2011) suggests that Job demands -Demands exceeding abilities (work overload) -Demands falling short of abilities (work underload) -Excessive amount of bureaucratic work Job rewards -Low salary -Job insecurity -Lack of career development opportunities Roles in the organization -Role ambiguity -Role con ict Work relationships -Dysfunctional relationships with supervisor -Con icts with co-workers -Negative social reactions Work with clients -Clients' need for emotional support -Clients' trauma -Clients' violence Occupational Sress violence is most commonly in icted by clients who have substance use or mental health problems. Stress from experiencing violence by social workers is further exacerbated by their a achment to clients and normalization of violence. If workers feel responsible for the well-being of their clients they may risk their health instead of pressing criminal charges (Kosny, 2011).

Conclusion
A discussion of two major theoretical approaches revealed superiority of transactional model that views stress as a dynamic relationship between the person and the work environment, and that locates stressors in the transaction between an individual and the environment. Based on the limited research in nonpro t organizations, research among social workers, and research in a variety of other workplace environments, 14 stressors were identi ed and grouped into ve categories (Figure 3). e gure below summarizes ve stressors' categories with relevant factors in each of them and illustrates their hypothesized relationship with the occupational stress.

Methodology
In order to answer the main research question of the study "How to manage occupational stress in human service nonpro t organizations in Mykolaiv, Ukraine?" the following hypotheses were raised.
Hypothesis 1: ere is a positive relationship between employees' stress and job demands: (a) job demands that exceed their abilities in quantity and quality (work overload); (b) job demands that fall short of their abilities in quantity and quality (work underload); and (3) excessive amount of bureaucratic work.
is hypothesis is largely derived from the person-environment t theory by French and Caplan, and it intends to test whether the mismatch between job demands and employees' abilities is positively related to stress. Abilities refer to the personal resources that person can draw upon in order to meet job demands (skills and knowledge, time and energy); demands refer to the quantitative (quantity of work) and qualitative (level of di culty) requirements placed on the person. Mismatch between abilities and demands can occur when demands exceed abilities (work overload) or when they fall short of abilities (work underload). Both instances have been widely acknowledged to cause stress by multiple researchers ( Judith & Storey, 2010;Schuler, 1982;Sadri & Marcoulides, 1997;Kosny, 2011, Zhuk et al., 2009. Bureaucratic work is a speci c job demand that has been widely cited by researchers to cause stress when it exceeds employees' abilities (Kosny, 2011;Demmer, 2002).
Hypothesis 2: ere is a positive relationship between employees' stress and their job rewards: (a) low salary; (b) job insecurity; (c) lack of career development opportunities.
is hypothesis is generally derived from Siegrist's E ort-Reward Imbalance model that suggests that stress arises out of the imbalance between e orts expended and rewards received (Cox, Gri ths & Houdmont, 2006, p.3). E orts include work time, skills, and emotions that an employee invests in the work in order to meet job demands. Rewards distributed by the employer consist of wages, job security and career opportunities ( Jonge et al., 2008).
is hypothesis, then, intends to test the relationship between stress and the imbalance in the job rewards. Low salary has been suggested as a major source of stress by Ongori and Agolla (2008), in whose study 73% of respondents reported experiencing stress from it. Job insecurity was suggested as a signi cant stressor by Johnson et al. (2005), whereas Demmer's research (2002) revealed it to be of minor importance. Lack of career development opportunities was derived as a factor from the research by Sadri and Marcoulides (1997) and Ongori and Agolla (2008). Although the researchers suggested that stress arises out of uncertainty about promotion, this factor was modi ed to the lack of career development opportunities, which is common to nonpro t organizations (as explained in the Situation Analysis section of the present document) and which in large determines uncertainty about promotion. e lack of such opportunities is, therefore, hypothesized to be positively related to stress.
Hypothesis 3: ere is a positive relationship between employees' stress and their roles: (a) ambiguity of their roles; and (b) the role con ict.
is hypothesis is derived from the work of Schuler (1982), Johnson et al. (2005) and Storey & Billingham (2010), who showed that role con ict and role ambiguity are the two most important stressors that relate to employees' roles in their organizations. Role ambiguity refers to the employee's uncertainty about his/her duties and responsibilities, expectations and evaluation criteria, and was shown by Gilboa et al. (2008) to be a strong stressor. Role con ict arises when an individual is confronted with two or more conflicting or opposing role expectations and demands. is factor was shown to be positively related to stress by some researchers (Storey & Billingham, 2010), and negatively related by others (Teh, Yong & Lin, 2012).
Hypothesis 4: ere is a positive relationship between employees' stress and their work relationships: (a) dysfunctional relationships with their supervisor; (b) con icts with co-workers; and (c) negative social reactions.
Work relationships have been regarded an important part of organizational climate and a common source of occupational stress by multiple researchers (Schuler, 1982;Johnson et al., 2005;Storey & Billingham, 2010). Dysfunctional relationships with supervisor were found to be a source of stress in the research of Bucurean and Costin (2011) and Sadri and Marcoulides (1997), who explained this relationship by ine ective management, supervisors' excessive authority, favoritism and discrimination. Con icts with co-workers were also illumined to be positively related to stress by Sadri and Marcoulides (1997) and Demmer (2002). Lastly, research of Gossman and Silverstein (1993) revealed negative perceptions of workers in human service organizations by the larger society to be a strong factor of stress.
Hypothesis 5: ere is a positive relationship between employees' stress and their work with clients: (a) clients' need for emotional support; (b) clients' trauma; and (c) clients' violence.
is hypothesis is derived from Kosny (2011), who contended that human service workers experience high levels of stress due to their work with clients who have high needs. Clients' need for emotional support was revealed to be a stress factor particularly among employees who work with clients with AIDS and with patients in palliative care, because in order to support them, human service workers o en engage in emotional labor (Demmer, 2002;Kosny, 2011;Hulbert & Morrison, 2006;Sorensen & Iedema, 2009). Clients' trauma was shown to be an especially signi cant stressor for human service workers who take care of patients with AIDS and those in palliative care, as they o en vicariously su er their clients' pain and death (Grossman & Silverstein, 1993, Hulbert & Morrison, 2006. Lastly, violence of clients was found to be a source of stress for those who work with clients with substance abuse and mental health problems (Kosny, 2011).
Hypothesis 6: Bureaucratic work is the most severe stressor for the employees of human service non-pro t organizations.
is hypothesis is derived from the research by Demmer (2002) in organizations that deal with people who live with AIDS. He found that bureaucratic work was the stressor highest in severity (p.235).
Participants e empirical study was conducted in the entire population of 11 active human service nonpro t organizations in Mykolaiv, Ukraine. Human services refers to a variety of delivery systems such as social welfare service, education, mental health services and other forms of healthcare. e entire workforce of these organizations, comprising 156 participants, was included in the empirical study. us, the present research did not utilize sampling strategies, and was conducted as a census. However, due to the spread of the workforce of these organizations across the region, not all employees could be accessed for the participation in the study. Consequently, a total of 96 employees took part in the study, which represents 62% of the population.

Research Tool
Self-administered questionnaire was chosen as the data collection method most suitable for the purposes of present research both conceptually and practically. e questionnaire contained 48 questions in total, of which 41 are closed-ended and 7 are open-ended. Four types of closed-ended questions are used in the questionnaire: list questions, category questions, rating questions and scale questions (using 7-point and 5-point Likert-style rating scales). e questionnaire was developed by the researcher with the use of elements from the Generic Job Stress Questionnaire developed by the National Institute for Occupational Safety and Health (N/A). Table  1 provides an operationalization table that speci es the questions and measurement scales that correspond to each variable. e questionnaire, however, is not limited to the questions that collect data for the testing of hypotheses, and contains additional questions regarding the organizational environment, nature of employees' work and demographics. ese additional questions enabled the collection of valuable data that allowed some important research ndings to be made.

Empirical Findings
Analysis of the data is based on 92 questionnaires out of possible 156, which constitutes 59.0% response rate. e initial response rate was higher at 62%, or 96 questionnaires, but four of the initially obtained responses had to be excluded from the analysis for containing a signi cant amount of missing data. Table 2 presents the demographic characteristics of the participants in the empirical study. us, almost three quarters of the respondents (71.7%) are women, and almost 80% of them are under the age of 45. Over 80% have higher education, and among them 53% have their degrees in the elds of psychology, social work or pedagogy. us, the participants of the study are mostly young and highly educated women. e majority of participants (44.6%) are social workers or psychologists who work directly with clients whom the organizations serve. e second largest group is composed of mid-level managers (17.4%) that include project and area managers, specialists and coordinators. Among the respondents, 12% were in the top management, of which 8% (7 respondents) were chairpersons in the organizations. Accountants, documentators, lawyers, medical workers and maintenance sta together comprise 26.2%. Among all participants, over 40% have worked in their organizations over 5 years, and they are mostly represented by top and mid-level managers and accountants. e other 60% that have worked less than 5 years mostly include social workers. Social workers also represent a predominant majority in the 33% of employees who worked in their organizations for less than 2 years. Lastly, in these human service organizations social workers are not the only ones who work directly with clients. While they comprise 44.6% of all respondents, almost 74% of all the employees surveyed work directly with clients.

The organizational environment
In order to o er appropriate recommendations for stress management, the questionnaire a empted to assess the environment of the eleven organizations and the nature of their employees' work. us, this section presents a descriptive overview of each of the ve categories of stressors. Job demands. Data collected through the questionnaires revealed that less than half of respondents (47.83%) work a standard 36-40-hour workweek. A signi cant number of respondents work a larger number of hours: over 35% of respondents work more than 40 hours a week, of whom a third work 46-55 hours. Such long hours seem to be partially explained by the fact that 87% of respondents do additional work that is not part of their formal job description. Whereas 60% of them spend just 1-5 hours a week on this work, another 40% spend 6-21 hours. Despite working longer hours, however, over half (54.35%) of respondents indicated at least sometimes lacking time and energy to complete the required work assignments. For the majority of respondents, such high work load is constant, i.e., it does not vary in its amount -only 23% of respondents indicated having schedules with uneven workloads. Bureaucratic work is a job demand that was revealed to be of great signi cance, as 65% of respondents reported having to do too much of it.
Job rewards. Analysis of the collected data revealed that the average salary of respondents is around 1,860 UAH (Ukrainian hryvnas), which is 30% above the subsistence wage set by the state (Ministry of Finance of Ukraine, 2013), but which in reality does not cover the minimal needs. Furthermore, the salaries of 11% of respondents fell even below the state-determined subsistence wage. e salaries of only 31.5% of respondents were higher than 2,401 UAH -a salary that can cover the subsistence needs. us, 82% of respondents indicated that their salary was inadequate, and on average, wished a 1,734 UAH increase. e majority of respondents (93%) indicated their concern with job instability. e mean level of instability on the scale from 1 (stable) to 7 (very unstable) was 4.18, which is slightly above the middle point. No statistically signi cant correlations were observed between the job instability and the number of years worked in the organization (p=0.492, α=0.01), which signi es that job instability is a common perception of all respondents regardless of the length of their employment in their organizations.
Almost two thirds of the employees experienced career development during their time in their organizations. However, 78% of respondents are dissatis ed with existing career development opportunities. No statistically signi cant relationships were observed between satisfaction with career development and the number of years worked in the organization, which means that the problem is not that respondents did not yet have enough time to advance their careers.
Roles in the organization. e majority of employees reported having a clear understanding of their duties and evaluation criteria. However, 29% reported at least sometimes lacking the understanding of their responsibilities and expectations. Over a third of employees also reported experiencing role con ict, when people demand from them the ful llment of incompatible tasks. No statistically signi cant relationships were observed between the length of employment in the organization and role ambiguity or role con ict, which means that respondents encounter di culties with their roles regardless of how long they have worked in their organizations.
Work relationships. Analysis of the responses to the questionnaires revealed work relationships in the organizations to be, in general, positively assessed by the respondents. Almost 84% of them reported being satis ed with the relationships with their supervisors. Almost a third of respondents at least sometimes experience con icts with coworkers, but 80% of respondents a rm that they can rely on emotional support from their colleagues. A slightly worse situation is reported with regard to relationships with the society, as 40% of respondents indicated experiencing negative reactions of the society towards their work in the non-pro t sector with marginalized populations. However, when asked to rate how much they worry about those negative reactions on a scale from 0 (no worry) to 7 (worry very much), respondents tended to choose the lower rating, so that the average rating was 2.65.
Work with clients. Although the mission of social service nonpro t organizations is closely connected to the vulnerable populations they serve, only 74% of respondents indicated that they work directly with their clients. Of them, 88% work with several groups of clients, ranging from 2 to 11, with the average of 4. Figure 4 below shows the share of respondents that work with each of the 11 groups of clients.
On average, respondents spend 13-16 hours a week working directly with their clients. e load of work with clients, however, has a rather signi cant range: almost half of respondents work 1-12 hours a week, a third work 13-28 hours, and almost a h work 29-37 hours. Among those respondents who work directly with clients, 56% reported the need to provide emotional support to their clients, 69% testi ed that they adopt and su er through the trauma of their clients, and 22% indicated that they have at least sometimes experienced verbal or physical violence. Table 3 provides the means and standard deviations for a dependent variable of stress and for 14 independent variables. Since the variables were measured along di erent Likert-scales (scale boundaries are presented by Min. and Max. values in the table), for the purposes of comparison, the table presents their means as percentages of the maximum value on each scale. Table 3 shows that the mean stress level is moderately high at the level of 4.2 on the 0-7 Likert-scale, where 0 signi ed no stress and 7maximum stress level. us, the mean stress level for all respondents equals 60% of the maximum. e means of all 14 independent variables exceed 34% of the maximum, which signi es that all stressors are experienced by the respondents. e dispersion of the stress variable can be seen in Figure 5 above (mean=4.2, standard deviation=1.652). e means for a dependent variable "stress" and for 14 independent variables were calculated for each of the 11 organizations in the study. Although the di erences among organizations were observed, the ANOVA test revealed that none of the di erences were statistically signi cant (F (10.81)=1.750; p =0.084, α=0.05). erefore, the data analysis was performed cumulatively for all organizations together.

General results
Independent-Samples T-tests were performed to test the signi cance of the di erence of means between male and female respondents, and between respondents who do and do not work with clients. T-tests revealed no signi cant di erences in the mean stress levels among genders (t(90)=1.158, p=0.119, α=0.05) or among those who do and do not work with clients (t=0.321, p=0.750, α=0.05). erefore, the data analysis was performed for the sample as a whole, without further division into gender and work with clients. Data for respondents who work with clients were analyzed separately only for the tests that assessed the impact of three stress factors that pertain to those respondents only who work with clients -clients' need for emotional support, clients' trauma and clients' violence.

Hypothesis 1:
ere is a positive relationship between employees' stress and job demands: (a) job demands that exceed their abilities in quantity and quality (work overload); (b) job demands that fall short of their abilities in quantity and quality (work underload); and (3) excessive amount of bureaucratic work.
is hypothesis was tested using correlation analysis, in particular Pearson correlation (Table 4 contains the correlations output). Correlation analysis yielded FIGURE 5. Percentage of Respondents that Chose Each of the Levels of Stress Stress Percent positive statistically signi cant Pearson coe cients for job demands exceeding (r=0.404, p=0.000, α=0.01) and falling short of abilities (r=0.440, p=0.000, α=0.01) in quality and quantity cumulatively. e coe cient for the excessive amount of bureaucratic work was also positive and statistically signi cant, but weaker (r=0.267, p=0.005, α=0.01). us, there is a positive relationship between employees' stress and job demands, so Hypothesis 1 is supported for all three variables.
A more detailed correlation analysis produced the following results. With regard to job demands that exceed employees' abilities (work overload), correlation analysis showed a stronger relationship between stress and the excess in quantity than in quality (r=0.440, p=0.000, α=0.01 for quantity and r=0.254, p=0.007, α=0.01 for quality). By contrast, with regard to job demands that fall short of employees' abilities, the stronger relationship with stress was observed for demands that fall short in quality rather than in quantity (r=0.388, p=0.000, α=0.01 for quality; r=0.201, p=0.027, α=0.05 for quantity). erefore, the stress of employees in nonpro ts in the study is correlated with quantitative work overload (having to do too much work that exceeds their time and energy) and from qualitative work underload (having to do too mundane and unchallenging tasks).
Correlation analysis between job demands revealed the following relationships to be signi cant at α=0.05: -Demands exceeding abilities in quantity and demands falling short of abilities in quality (r=0.526, p=0.000); -Demands exceeding abilities in quality and demands exceeding abilities in quantity (r=0.493, p=0.000); -Demands falling short of abilities in quality and demands falling short of abilities in quantity (r=0.390, p=0.000); -Demands falling short of abilities in quality and work in addition to formal job responsibilities (r=0.354, p=0.000); -Demands exceeding abilities in quality and demands falling short of abilities in quality (r=0.262, p=0.007); -Demands exceeding abilities in quantity and work in addition to formal job responsibilities (r=0.253, p=0.007); -Demands exceeding abilities in quantity and the excessive amount of bureaucratic work (r=0.216, p=0.019).
Hypothesis 2: ere is a positive relationship between employees' stress and their job rewards: (a) low salary; (b) job insecurity; and (c) lack of career development opportunities.
is hypothesis was tested using correlation analysis, in particular Pearson correlation (Table 4 contains the correlations output). Correlation analysis showed a positive and statistically signi cant relationship between stress and low salary (r=0.214, p=0.02, α=0.05). e perceived low salary was measured by the discrepancy (arithmetic di erence) between actual salary and that which respondents consider to correspond be er to the e orts they expend on their work. Correlation analysis revealed a statistically insigni cant relationship between the stress and job instability (p=0.075, α=0.05) and between stress and lack of career development opportunities (p=0.270, α=0.05). erefore, Hypothesis 2 is partially supported: it is supported with regard to low salary, but rejected with regard to job insecurity and lack of career development opportunities.
Hypothesis 3: ere is a positive relationship between employees' stress and their roles: (a) ambiguity of their roles; and (b) the role con ict. is hypothesis was tested using correlation analysis, in particular Pearson correlation (see Table 4 for the correlation output). Correlation analysis revealed a positive statistically signi cant relationship for the relationships between stress and role ambiguity (r=0.249, p=0.008, α=0.01), and stress and role con ict (r=0.305, p=0.002, α=0.01). us, hypothesis 3 is supported for all three independent variables.
Hypothesis 4: ere is a positive relationship between employees' stress and their work relationships: (a) dysfunctional relationships with their supervisor; (b) con icts with co-workers; and (c) negative social reactions.
is hypothesis was tested using correlation analysis, in particular Pearson correlation (see Table 4 for the correlations output). Correlation analysis revealed a positive and statistically signi cant relationship between stress and con icts with co-workers (r=0.259 and p=0.006, α=0.01), and between stress and negative social reactions (r=0.241 and p=0.010, α=0.01). e relationship between stress and dysfunctional relationships with supervisor, however, was shown to be insigni cant (p=0.163, α=0.01). us, hypothesis 4 is only partially supported. It is supported for the variables of con icts with co-workers and negative social reactions, but rejected for the dysfunctional relationships with the supervisor.
Hypothesis 5: ere is a positive relationship between employees' stress and their work with clients: (a) clients' need for emotional support; (b) clients' trauma; and (c) clients' violence.
Hypothesis 6: Bureaucratic work is the most severe stressor for the employees of the human service nonpro t organizations.
Multiple linear regression analysis was used to test this hypothesis. e analysis excluded respondents who do not work with clients, as they do not experience three out of 14 stressors -clients' need for emotional support, clients' trauma and clients' violence. erefore, multiple regression analysis was performed with the data of 68 respondents that work directly with clients.
is means that of every unit of change in stress, 0.228 is explained by the low salary, and 0.322 is explained by clients' trauma. Since "clients' trauma" has the highest ß coe cient, it impacts stress level more signi cantly, and, therefore, can be considered the most severe stressor. Bureaucratic work is thus revealed to be statistically insigni cant (p=0.315, α=0.05). erefore, Hypothesis 6 is rejected.

Work with clients
Since work with clients constitutes a signi cant part of the work of 68 respondents, several statistical tests were conducted in order to assess the impact of this work on stress in a greater detail. ese tests revealed rather interesting results. e number of groups of clients that respondents work with ranges from 1 to 11, and the majority of respondents (88.24%) work with more than one group (Figure 6 provides details on the number of groups and the corresponding share of respondents who work with them).

FIGURE 6. Percentage of Employees that Work with Each Number of Client Groups
Number of groups of clients Percent e correlation analysis was performed in order to test the relationship between the number of groups and respondents' stress. e analysis revealed this relationship to be positive and statistically signi cant (Pearson r=0.260, p=0.016, α=0.05). So the more groups of clients respondents work with, the more stress they experience.
Data analysis revealed that the mean stress level of respondents who do and do not work with certain groups of clients di ers. In order to assess the signi cance of the di erence of means, a series of Independent Samples T-tests was conducted. ese tests revealed a statistically reliable di erence with regard to two groups of clients. e rst one is people living with HIV/ AIDS: there is a signi cant di erence between the mean stress level of employees who work with people living with HIV/AIDS (mean=4.41, s=1.46) and those who do not (mean=3.42, s=1.88), t(66)=2.030, p=0.46, α=0.05. e second group is women in sex business: there is a signi cant di erence between the mean stress level of employees who work with women in sex business (mean=4.88, s=1.53) and those who do not (mean=3.83, s=1.48), t(66)=1.051, p=0.007, α=0.05. T-tests did not reveal statistically reliable di erences with regard to other 9 groups, which signi es that there is no statistically signi cant di erence among those who work and do not work with each of the remaining 9 groups of clients. Correlation analysis was performed in order to test the relationship between stress level of respondents who work with clients and other variables. e analysis revealed a negative statistically signi cant relationship between stress and emotional support from co-workers (Pearson r=-0.255, p=0.018, α=0.05). is means that the more emotional support respondents receive from their colleagues, the lower is their stress level. Correlation analysis revealed no statistically signi cant relationship between stress and the number of years worked in the organization (p=0.159, α=0.05), and between stress and the amount of hours a week spent on work with clients (p=0.086, α=0.05). is means that respondents' stress does not vary with the length of their employment or with the length of time spent working with clients.

Correlations among independent variables
Correlation analysis between 11 independent variables revealed multiple statistically signi cant correlations. For the clarity of presentation, they have been grouped in accordance with strength into strong, moderate and weak. Relationships that are signi cant at 0.01 level were market with two asterisks (**), and those signi cant at 0.05 level were market with one asterisk (*).

Discussion
e present study on occupational stress was the rst one conducted in human service nonpro t organizations in Mykolaiv, Ukraine. It was also the rst study conducted in Ukraine that included organizations, as well as employees, that serve various groups of clients (11 di erent groups) as opposed to only one (e.g., research of Zhuk et al. (2009) involved only organizations against human tra cking). Such peculiarities of the study invested it with signi cant relevance and allowed it to make rather interesting ndings.
Operating within the transactional model of stress, which views stress as a dynamic relationship between the person and the work environment and emphasizes subjective experience of stress by each individual, the study focused on respondents' assessment of the aspects of organizational environments that make them feel stressed. Despite the di erences among organizations (size of the workforce, scale of work, categories of clients served), the study discovered no signi cant di erences in the stress experience among respondents belonging to them. is nding allowed data analysis to be performed cumulatively for the entire sample and yielded to the results a high degree of generalisability. e rst major nding was that the mean stress level of respondents was revealed to be rather high (60% of the maximum), which signi es that stress is a real problem in the environments of human service nonpro ts in Mykolaiv, which corroborates the ndings of earlier research that deemed nonpro t environments (Kosny, 2011) and human service work ( Johnson et al., 2005) to be highly stressful. All 14 hypothesized stressors were found to be experienced by respondents of the 11 organizations (their means expressed as percentages of their maximum level were over 34% for all 14 variables). However, not all of them were found to be related to stress. Out of 14, the following eleven factors were discovered to be positively related to stress: demands exceeding abilities, demands falling short of abilities, excessive amount of bureaucratic work, low salary, role ambiguity, role con ict, con icts with co-workers, negative social reactions, clients' need for emotional support, clients' trauma and clients' violence. However, the other three were revealed to be statistically insigni cant: job insecurity, lack of career development opportunities, and dysfunctional relationships with supervisor.
Out of the eleven factors that were shown to be positively related to stress, two were revealed to be in a strong positive relationships with stress (0.4 < Pearson r < 0.49): job demands exceeding abilities and job demands falling short of abilities. is nding corroborates French and Caplan's theory of the relationship between stress and a mismatch between job demands and person's abilities, and the conclusions of multiple researchers that work overload and underload ( Judith & Storey, 2010;Schuler, 1982;Sadri & Marcoulides, 1997) are positively related to stress. With regard to demands exceeding abilities, the strength of the relationship lies primarily in the quantity of demands: employees experience stress when the amount of work they need to complete exceeds their time and energy. With regard to demands falling short of abilities, the strength of the relationship lies in the quality of demands: employees experience stress when their work assignments fall far below their knowledge and skills, and are therefore perceived to be too unchallenging.
Some interesting ndings emerged from the correlations between various aspects of job demands. A strong relationship was discovered between demands exceeding abilities in quantity and job demands falling short of abilities in quality. is means that assignments that fall below employees' skills and knowledge (are boring and unchallenging) also tend to be overwhelming in quantity. Correlation analysis revealed that such assignments are correlated with bureaucratic work and work that employees perform over and beyond their formal responsibilities in their organizations. us, it is these two areas of responsibilities that by being boring and overwhelming in amount are strongly related to stress. Other correlations revealed a moderately strong relationship between the quantity and quality of demands: those employees who lack time and energy to complete all their assignments also lack necessary skills; and on the contrary -employees whose skills go far beyond their assignments o en also would like to have more work to do. ree other stressors were revealed to have a positive moderately strong relationship to stress: clients' trauma, clients' need for emotional support and role con ict. Positive relationship between clients' trauma and stress supports the conclusion of Kosny (2011) and Grossman and Silverstein (1993) that social workers who work with clients that have been traumatized adopt their su erings and experience stress. Death of clients was not revealed to signi cantly impact the stress level, as no statistically signi cant di erence emerged between employees who do and do not work with terminally ill patients in palliative care, which supports the conclusions of Grossman and Silverstein (1993) that death of patients may not be more severe than long-term care. e discovered positive relationship between stress and clients' need for emotional support, which results from the emotional labor in which human service workers o en engage in order to support their clients, corroborated earlier ndings by Demmer (2002), Hulbert and Morrison (2006), Sorensen and Iedema (2009).
Discovered positive relationship between stress and role con ict supported the dominant view in research (e.g., Johnson et al., 2005). ese results, however, ran contrary to Teh, Yong and Lin (2012) who considered role con ict to be a mediator variable that prevents employees from workaholism and burnout, and therefore is negatively related to stress. Further, correlation analysis revealed role con ict to be positively related to demands that exceed abilities in quantity (strong relationship), demands that fall short of abilities in quality (moderately strong relationship), and the excessive amount of bureaucratic work (weak relationship). is means that those employees that experience con icting demands have to do a large amount of bureaucratic work that exhausts their time and energy leaving none for their other job assignments.
Lastly, six out of eleven stressors were discovered to have a positive, but weak, relationship to stress: excessive amount of bureaucratic work, clients' violence, con icts with co-workers, role ambiguity, low salary and negative social reactions. Positive relationships of the excessive amount of bureaucratic work and clients' violence to stress have been identi ed by previous research (Kosny, 2011;Demmer, 2002). Just as in other organizational environments (Sadri & Marcoulides, 1997), in human service organizations, con icts with co-workers are positively related to stress. eir e ect, however, is exacerbated in that con icts with co-workers trigger a decrease in emotional support, which was shown to negatively correlate with stress. Con icts with co-workers were also discovered to be positively related to role con icts, which means that employees who experience con icting demands put on them by their colleagues also experience con icts with them.
Positive relationship between role ambiguity and stress supports the nding of Gilboa et al. (2008). Role ambiguity was also revealed to relate positively to role con ict and con icts with co-workers, which suggests that lack of clarity about roles is accompanied by the perception of con icting responsibilities and con icts with coworkers. e positive relationship between stress and discrepancy in the actual and desirable salary falls in line with the Siegrist's E ort-Reward Imbalance model that holds that stress develops from the imbalance between e orts expended and rewards received, and coincides with ndings of Ongori and Agolla (2008) that showed low salary to be positively related to stress. Lastly, the study's ndings regarding negative social reactions that positively relate to stress coincide with the ndings of Gossman and Silverstein (1993). Negative social reactions were also revealed to be positively related to demands exceeding abilities and demands falling short of abilities, role ambiguity and role con ict. e possible explanation for such relationship may lie in the nature of negative reactions. Employees of human service nonpro ts in Mykolaiv su er predominantly from social prejudices about money laundering by nonpro ts. us, con icting or unclear responsibilities, and inadequate job assignments may make employees feel uncon dent about their role and work in their organizations. e three factors not related to stress in a signi cant manner included career development opportunities, job insecurity and dysfunctional relationships with supervisor. Lack of relationship between job insecurity and stress runs counter to Kosny (2011) and Johnson et al. (2005) who considered it to be an important factor, but aligns with Demmers' (2002) research ndings, in which job insecurity was the least important stressor. Lack of relationship between stress and lack of career development opportunities falls out of line with previous research, where desire and inability to be promoted caused stress (Sadri & Marcoulides, 1997;Ongori & Agolla, 2008). is nding can be partially explained by the fact that lack of career development opportunities is inherent in the work of nonpro ts and is therefore assumed to be part of the normal work environment. Furthermore, it may not be bothersome because given the low salary, a great amount of work demands, job insecurity and lack of career development, workers do not seem to choose work in nonpro ts for the sake of bene ts external to the work, and therefore, may not be bothered by the lack thereof. Lastly, regarding dysfunctional relationships with supervisor, this factor does not seem to be experienced by employees in Mykolaiv nonpro ts, as 84% of respondents reported to be satis ed with their relationships with supervisors.
Although no signi cant relationship was discovered between stress and the three factors above, they were revealed to be positively and signi cantly related to each other and to demands falling short of abilities. is suggests that employees whose skills and knowledge outweigh their job demands feel insecure about their jobs desire more opportunities for career advancement and have con icts with their supervisors. e regression model with 14 factors showed a signi cant goodness of t (R Square=0.663, Adj. R Square=0.438) and a high level of reliability. Among all 14 factors, the model revealed two to most signi cantly account for the change in the stress variable: low salary and clients' trauma together explain more than half the change in stress. Among the two, clients' trauma had more weight, and was, therefore, found to be the most severe stressor for the employees in human service nonpro ts in Mykolaiv.
Analysis of the groups of clients with whom Mykolaiv nonpro ts work revealed that the more groups a human service worker contacts, the more stress he/she experiences. Furthermore, the analysis of means with regard to each particular group revealed that people living with HIV/ AIDS and women in sex business are two groups that signi cantly a ect the stress level of employees who work with them. is nding -that a number of groups of clients has bearing on stress, and that various groups of clients have varying impact on stress -seems to be a new nding that should be explored in a greater depth to discover what exactly makes a particular group more stressful than another, and several of groups more stressful than one.
is study has two major limitations. First, it used a quantitative research method of a questionnaire, which limited the study's ability to explain the reasons and dynamics that underlie its particular ndings. e study can be e ectively complemented by a qualitative method of an in-depth interview, for instance. However, it must be noted that, given the peculiarity of the nonpro t environment and the purposes of the present study, the questionnaire was the most suitable method. Since no research had been previously conducted in human service nonpro ts in Mykolaiv -hence, no understanding of their organizational environment, and the problem of stress in particular, was available -a questionnaire was the best method to get the accurate picture of the stress experience in these organizations through the collection of a large amount of data from a sizable number of employees. By the means of the questionnaire, then, this study laid the groundwork for further research. It explored the experience of stress in Mykolaiv nonpro ts and o ered a wide array of peculiar ndings, each of which can be used by further research for deeper investigation.
e second limitation of this study is that it was cross-sectional, i.e., it assessed employees' stress at a single point in time. is circumstance could make the results of the study too speci c to the period of time when it was conducted, and less generalizable over time. However, it is important to note that the study took precaution against this, and the questions that employees answered in the questionnaire asked them about their experience in general rather than at a particular point in time (e.g., most of the questions were Likert-style frequency scales that asked respondents to state how o en they experience such-and-such factor). Longitudinal studies with a range of several months to half a year would nevertheless be useful in assessing the changes over time.
Since the present study was the rst one conducted in human service nonpro ts in Mykolaiv, and in its exploratory nature it laid the groundwork for subsequent research, the following areas of further inquiry can be suggested. First, ndings of the present research can be assessed in greater depth and detail, with regard to which three particular areas of further inquiry can be identi ed. First, correlations among independent variables (e.g., job demands, dysfunctional relationships with supervisor, role con ict) that the present research revealed could be assessed in greater depth and tested for causal relationships. Second, the di erence in the stress level between employees who work with a particular group of clients, or several of them, should be explored for underlying reasons. It would be particularly valuable to discover what makes a particular group more stressful that another. ird, the lack of signi cant relationship between stress and job instability and career development, which are considered to be important stress factors in environments other than of nonpro t organizations, should be further explored in order to understand why they do not a ect stress of employees in nonpro ts.
Second, subsequent research can concentrate on exploring at least three themes and frameworks that the present research assumed. First, since the need for stress management programs was revealed to be of vital importance, subsequent research might concentrate on the e ectiveness of speci c stress management techniques that may be employed by individuals and organizations. Second, the present research implied the Transaction model of stress and focused on factors that bring it about, so subsequent research could valuably explore each of the identi ed stressors along the Transaction model in order to locate a speci c location of the problem of each stressor -in primary appraisal, secondary appraisal or coping strategies. ird, subsequent research could replicate the present study in human service nonpro ts in other cities in Ukraine in order to see whether generalizations can be made across the country.
Lastly, with regard to research methods, further research should consider using qualitative methods (e.g., in-depth interviews) in order to assess the reasons and dynamics that underlie the ndings of the present study. Longitudinal studied can also be conducted in order to assess whether stress experience of employees changes over time, and also in order to test speci c stress management techniques that constitute an important area of further inquiry.

Conclusion
e present research set out with a goal to identify stressors and their e ect on employees stress in human service nonpro t organizations in Mykolaiv. e study identi ed and discussed fourteen di erent stressors from ve categories: job demands -demands exceeding abilities, demands falling short of abilities, excessive amount of bureaucratic work; job rewards -low salary, job insecurity, lack of career development opportunities; roles in the organization -role ambiguity, role con ict; work relationships -dysfunctional relationships with supervisor, con icts with coworkers, negative social reactions; and work with clients -clients' need for emotional support, clients' trauma and clients' violence.
Six hypotheses were formulated about the relationships between stress and fourteen stressors. To test these hypotheses, quantitative research with a survey strategy and a self-administered questionnaire as a data collection method were chosen as most appropriate for the empirical study. e study was conducted in the entire population of human service nonpro t organization in Mykolaiv, which is comprised of 11 entities.
e study received 92 out of 156 possible responses, which constitutes 62% response rate.
Data analysis fully supported three hypotheses, partially supported two, and rejected one hypothesis. e following stressors were revealed to be positively related to stress: demands exceeding abilities and demands falling short of abilities revealed strong relationship to stress; clients' need for emotional support, clients' trauma and role con ict revealed moderately strong relationship to stress; excessive amount of bureaucratic work, low salary, role ambiguity, con icts with co-workers, negative social reactions and clients' violence were more weakly related to stress. e other three factors -dysfunctional relationships with supervisor, job insecurity and lack of career development opportunities -revealed no statistically signi cant relationship with stress. e regression model for all 14 independent variables and stress as a dependent variable revealed to have high goodness of t, and showed 2 out of 14 factors to signi cantly account for the change in stress: clients' trauma and low salary together explain over half of the variation in stress variable. e empirical study also revealed a signi cant di erence in mean stress levels among people who do and do not work with two particular groups of clients: people living with HIV/ AIDS and women in sex business.
In the virtue of being the rst study on stress in the human service nonpro t environment in Ukraine, the present research laid the groundwork for subsequent research in the area of stress management in the nonpro ts in Mykolaiv and Ukraine.