CYNICISM AS MEDIATING VARIABLE BETWEEN LEADERSHIP SUPPORT AND EMOTIONAL BURNOUT: ADMINISTRATIVE SUPPORT STAFF IN TURKISH UNIVERSITIES

Abstract Studies in management and organization scholarship prefer to focus on personnel who constitute the core in institutions having bilateral employee status, academicians in the case of our research. Yet, organizational issues pertaining to administrative employees, as support personnel, seem understudied in the literature. Thus, in this study, we focus on issues experienced by administrative employees in Turkish universities. We have employed leader-member exchange theory as theoretical background, and conducted a fi eld research through a questionnaire form to test the model. Organizational cynicism, emotional burnout and leadership support constitute the main dimensions of the research theme. We argue that leadership support affects emotional burnout through organizational cynicism. Analysis shows that administrative support personnel in universities suffer from emotional burnout, mainly sourcing from organizational cynicism on which leadership support is highly infl uential. Thus superiors in bilateral institutions are expected to support administrative employees to eliminate cynicism, otherwise they are in danger of emotional burnout.


Introduction
Organizational studies diff er in a range from for-profi t organizations to public organizations, including non-profi t organizations.Besides type-based diff erentiations, organizations in the public sector can be classifi ed based on a number of factors such as personnel structures, autonomy levels or institutionalized aims.Yet, organizational issues, or problems pertaining to personnel, are not rare in all kinds of organizations, if not very common.Defi ning these issues is of crucial importance, organizational sett ings have enormous impact on human resources.These eff ects might vary from temporary absence to permanently leave in simple terms if management fails to defi ne issues, or to develop solutions when necessary.Thus, in this study we focus on organizational issues pertaining to administrative employees in universities.Having bilateral employee status, universities are comprised of academicians as the core personnel and administrative personnel having the back-up and support status in addition to students.While there are many studies focusing on academicians (Watt s and Robertson, 2011; Toker, 2011) or on students (Wei, Wang and Macdonald, 2015), there is litt le scholarly att ention devoted to administrative personnel who perform many duties.Thus, by using leader-member exchange theory, we focus on a number of organizational att ributes of administrative personnel to present and evaluate their particular att itudes and expectations on leadership support, cynicism and emotional burnout.
The organizational eff orts in public institutions have increased the work load with a raise on expectations on the management side.This increase in expectations from public employees should be balanced with some organizational precautions that would eliminate negative outcomes.We argue that emotional burnout and organizational cynicism, among others, have negative impacts over administrative personnel if management fails to eliminate them.Thus, this paper takes for granted that leadership support would be of great importance for administrative personnel to carry out the increasing rate of administrative burden and work load.If less or no leadership support is perceived by administrative employees, then emotional burnout due to cynicism might be expected on the employees' side.If emotional burnout is the case, the following problems might occur: temporary or permanent absenteeism, work leave, high turnover rate among some issues; and all of them could deteriorate the work environment.For this problem, studies indicate toward direct or indirect eff ects of leadership support on employee burnout process (Greco, Laschinger and Wong, 2006).
Diff erentiated from core personnel with regard to working conditions, fi nancial gains or employee status, we argue that management has to provide administrative personnel with leadership support for bett er achievement of tasks.In this context, administrative personnel particularly might be in need of motivational empowerment, if not accompanied with additional gains like social, fi nancial or institutional benefi ts.Leadership support including participation in decision making might help the administrative personnel to feel themselves more valued.Studies show that employees with high self-esteem and equipped with organizational citizenship aff ect the organi-zational climate in positive ways (Randhawa and Kaur, 2015;Mayer, Fraccastoro and McNary, 2007).
We organized the paper in four sections.First, we present the conceptual and theoretical background of the study.We explain the methodology in section 2. In section 3 we focus on the fi ndings of the research, and in section 4 we discuss the fi ndings by comparing with other studies in literature.We draw conclusions in the fi nal section.

Conceptual and theoretical background
Employee burnout 'is a unique type of stress syndrome characterized by emotional exhaustion, depersonalization and reduced personal accomplishment' (Schulz, Greenley and Brown, 1995, p. 333;Maslach, 2003, p. 2).It is an aff ective phenomenon with various contributors as organizational context, organization and management processes, work environment, individual employee characteristics, work satisfaction, client severity (Schulz, Greenley and Brown, 1995) among others.Organizational burnout has some severe results on both organizations and employees (i.e., increased job turnover, absenteeism, low rates of employee involvement, job satisfaction or commitment, and other mental and physical issues) (Brewer and Clippard, 2002).
Organizational cynicism, as highly occurring in other types of organizational settings, is also present in public institutions (Karadağ, Kılıçoğlu and Yılmaz, 2014), including universities (Akın and Ulusoy, 2016).Dean, Brandes and Dharwadkar (1998) evaluate the organizational cynicism as the extremely negative att itudes that employees have against their organizations.However, organizational cynicism is an employee behavior, not necessarily confi ned to one object, but can relate itself with regard to multiple objects (Johnson and O'Leary-Kelly, 2003, p. 629).Organizational cynicism does have various att ributes, ranging from low level of satisfaction to complete leave from organization (Chiaburu et al., 2013).
Leadership support is an important notion for employees, particularly for extraordinary cases in organizational environment which involve a dispute among employees over procedures or positions.Perceived leader support is crucial since it is 'proposed to be a key feature of the work environment for creativity' (Amabile et al., 2004, p. 5).
For theoretical background we employ leader-member exchange (LMX) theory in such a way that interactions between leaders and members have an impact on administrative employees on a reciprocal base.LMX is preferred as theoretical base since it is asserted that 'impact of change is felt by employees in their day-to-day interactions with others at work, and LMX theory and measurement capture the quality of such relationships' (Ritz et al., 2012, p. 161).Although theoretical content and its dimensionality have changed over years (Schriesheim, Castro and Cogliser, 1999), LMX theory is proved to be quite explanatory for subordinate and superior relationships.LMX theory argues that a supervisor has one of a kind relationship with each of the subordinates (Graen and Uhl-Bien, 1995) that is subject to negotiations over time (Liao et al., 2017).A positive and aff ective relationship based on reciprocal exchanges between leader and employees has numerous positive impacts (Liao et al., 2017).According to LMX theory, a high-quality relationship between leaders and employees might yield positive outcomes (i.e., trust, being valued and eff ective) due to mechanisms of reciprocity and social exchange, while low-quality relationships refer to sole economic exchange (Tummers and Knies, 2013).In sum, as Vigoda-Gadot and Meiri (2008, p. 116) cite, 'the process of reciprocity and met expectations contributes to a bett er fi t and a stronger congruence between the values of the employees and the organization' as stipulated by LMX theory.
We argue that employees believe that their leaders should support them for what they have done for the sake of their organizations in light of LMX theory.If this is not the case, then the employees feel cynicism towards the leaders, then might fall in stress, which turns into depression or emotional burnout if not recognized and alleviated in time.We know that 'burnout is a syndrome of emotional exhaustion and cynicism that occurs frequently among individuals who do 'people-work' of some kind' (Maslach and Jackson, 1981, p. 99).Since 'organizational cynicism is an att itude composed of beliefs, aff ect, and behavioral tendencies towards an organization' (Dean, Brandes and Dharwadkar, 1998, p. 341), perceived leadership support is of great importance (Netemeyer et al., 1997) for the employee to relieve and feel valued.

Research methodology
In this study we employ structural equation model (SEM) to analyze whether organizational cynicism has a mediating role between leadership support and emotional burnout.Thus, the goal of this paper is to examine the role played by leadership support in organizational cynicism that might turn out in emotional burnout.We conducted a fi eld research between March-July 2015, which gathered data from administrative personnel in six public universities and one foundation university in Turkey.

Research strategy
We have chosen systematic literature review for its capacity to reach the precise knowledge accumulation in a particular domain (Watt s and Robertson, 2011).We have researched many databases, such as Google Scholar, Sage, Routledge, Springer, Ulakbim, Blackwell, Palgrave and Scopus.We have used (public) university, burnout, leadership support, organizational cynicism, administrative personnel as keywords for making various combinations to search the databases.

Research context
Before going any further, we have to note that our research is conducted mainly in public universities in a specifi c country.Supported by many laws and regulations, universities in Turkey, including foundation universities, are run by a particular code numbered 2547.The Turkish Constitution does not allow universities to be established for-profi t purposes.Foundation universities are also allowed to be established for non-profi t purposes.
Thus our discussion and arguments are to be limited by many contextual factors that include particular cultural, legislative, and institutional conditions.These contextual factors also limit the abilities of management except for tools devoted by institutional autonomy provided by the law since universities in Turkey are among the institutions of local administration.However, according to the constitutional mandate, all public institutions in Turkey constitute an entity irrespective of them being a centralized or decentralized public institution.Therefore, some issues pertaining to all public employees are regulated at central government level.For example, any raise in wages of administrative personnel is regulated by the central government policies.In addition, any other issues pertaining to employees' status are also designed at central level.Except for tools regulated at central level, management of public universities has specifi c tools to design the organizational setup for all employees.Irrespective of all contextual diff erentiations, we raise the question on how administrative employees feel about leadership support provided by their superiors, organizational cynicism that they feel, or emotional burnout which could generate decreased satisfaction, or temporary or permanent leave of the employee.

The research model
The research model (as shown in Figure 1), formed in light of previous relevant studies, was empirically tested.The research model is built on linear and mediatory relationship among variables in which leadership support and emotional burnout are the dependent variables, while organizational cynicism constitutes the mediating variable.

Organizational Cynicism
Emotional Burnout

Sampling and data collection
The research was conducted on administrative personnel working in seven Turkish universities, and data was collected through a questionnaire.The choice of the administrative personnel of universities was limited by some conditions.First of all, we concentrated on organizations holding a dual status regarding the employees.We argue that support staff is in a position of suff ering organizational cynicism and emotional burnout.Secondly, as the core personnel of universities, authors have their own personal experiences regarding their own staff colleagues.Thirdly, we assumed that non-existence of leadership support, or a low level of this type of support over administrative staff in bilateral institutions, as universities in our case, might have worse repercussions than their equivalents in the public sector.
We distributed 504 questionnaires to universities to a convenience sample of administrative personnel, and we had a return rate of 0.96 (485/504).Excluding 22 questionnaires that returned with some defi ciencies, we analyzed 463 questionnaires.We needed a sample size that had a ratio of at least 5:1 when compared to free parameters under normal and elliptical theory (Bentler and Chou, 1987, p. 91).For the reliability concern, the sample size was proposed to be around 200-500 where the reliability increases when the sample size reaches to 500 (Kline, 1995).The sample size in this study is close to be rather reliable.

Scales
During the fi eld research we employed a scale, comprised of three subscales and demographical questions, which had 27 items (questions) in total.One of the subscales that we employed was the one developed by Netemeyer et al. (1997) in order to measure leadership support dimension.In this subscale the questionnaire had fi ve items measured with a 5-point Likert scale.For the organizational cynicism subscale we used the scale developed by Brandes, Dharwadkar and Dean (1999).We also gave att ention to its Turkish validation by Karacaoğlu and İnce (2012).The questionnaire has thirteen items having 5-point Likert scale.The organizational cynicism subscale have three dimensions as behavioral cynicism (4 items), aff ective cynicism (4 items) and cognitive cynicism (5 items) that were designed following the model developed by Dean, Brandes and Dharwadkar (1998).For the emotional burnout subscale we employed the burnout scale developed by Maslach and Jackson (1981).In this subscale the questionnaire has three dimensions as emotional exhaustion (9 items), depersonalization (5 items) and low personal accomplishment (8 items).As regards reliability and validity of the research scale, particularly for internal consistency, we found a Cronbach's alpha coeffi cient ranging from 0.878 to 0.900 for any item through the scale (Table 1).The reliability and validity tests showed that any item in the scale is above 0.60.We know this is the base level for reliability (Hair, Bush and Ortinau, 2000).As the second for validity of the scales we have conducted a factor analysis for construct validity where we have checked the factor loadings of the items to decide whether the items are collected under true scale.Excluding the items in the demographical section, we have also calculated Eigen values for 27 items whose Eigen values proved to be greater than 1 in 5 factors, explaining the total variance more than 70.435%.The factor loadings of the items varied between 0.551 and 0.935.The result of KMO test for sampling adequacy proved to be 0.941 and Bartlett 's test of sphericity is also highly signifi cant (p<0.001).Thus, we argue for a high level of correlation among the variables and the data set is truly adequate for the analysis.Components revealed after factor analysis and factor loadings are as presented in Table 1.For the construct validity of the scales as the last step, we have performed a confi rmatory factor analysis using AMOS 20 program.Of the ratios for the goodness of fi t criteria, we argue that our results are confi rmative of the ratios as assumed in the literature (Anderson and Gerbing, 1984;Hu and Bentler, 1999;Cheung and Rensvold, 2002).In Table 2 we present the regarded ratios for the goodness of fi t criteria for the confi rmatory factor analysis.We also witnessed, after the results of confi rmatory factor analysis, that the subscale related to organizational cynicism fi tt ed to a 3-factor construct while the other two subscales as leadership support and emotional burnout are fi tt ed in one factor.

Limitations of the study
We have some limitations of the study.First of all, the data collected through the study is not longitudinal, but cross-sectional.Thus we have no opportunity to set a causal relationship yielding exact interactions, instead of talking about potentialities.The study is based on the data collected from administrative employees excluding any academicians, thus we have no generalizations for other sectors, or professionals in universities.We have no control group, so any prospective studies might need to employ a sum of various methods, including diff erent sectors and diff erent types of data collection.

Demographical fi ndings
In the study, there are some items to identify the descriptive statistics of the participants in terms of demography.In Table 3 we present the results of these demographical items.
Before proceeding further to test the hypothesis, we have checked the signifi cant diff erence among demographic characteristics and variables.According to T-test results we have found no signifi cant diff erence between variables and sex and marital status of the participants (p>0.05).For other relations between characteristics and variables, we have employed analysis of variance test (ANOVA).According to the ANOVA results, there is no signifi cant diff erence between variables and participants' work places, or their educational statuses (p>0.05).On the other hand, we have found some signifi cant diff erences between scale variables and age, work duration and positions of the participants (p<0.05).For the clearance of this signifi cance, we employed Tukey's multiple comparison (TMC) test.According to this test, we have noticed that participants who are between 31-40 years old have higher perceptions for leadership support when compared to those whose ages are between 41-50 (p<0.05).According to results of TMC test, we have also found that chiefs (=23, 20) have more sensitivity in terms of emotional burnout when compared to division directors (=16, 97).

Findings of hypotheses testing
For testing the hypotheses we employed SPSS and AMOS package programs.For this purpose, we fi rst examined the arithmetic averages, standard deviations and correlations (Table 4).We argue for signifi cant relations between the dependent and independent variables.After fi nding some signifi cant relationships among variables, we conducted path analysis (PA) and mediation analysis (MA) under SEM for further analyzing these relations.SEM is the collection of statistical techniques allowing us 'a set of relations between one or more independent variables (IVs) and one or more dependent variables (DVs) to be examined.Both IVs and DVs can be either measured variables (directly observed), or latent variables (unobserved, not directly observed)' (Ullman, 2006, p. 35).Since SEM allows us to see both directly and indirectly observed variables, we preferred to use SEM instead of regression analysis.
It is an important assumption in SEM that the data is to have a multivariate normal distribution (Schermelleh-Engel, Moosbrugger and Müller, 2003; Lee, Poon and Bentler, 1995), thus, we confi rmed that the data in our study have multivariate normal distribution (Table 1).For a practical knowledge, skewness and kurtosis need to have values between +/-1.0 and +/-1.5 (Tabachnick and Fidell, 2013).
We have formed two diff erent models to study the mediating role of cynicism over leadership infl uence on emotional burnout.Mediating variable is defi ned as a variable to play a mediating role in the process of independent variable aff ecting the dependent variable under three conditions as such: 'First, the independent variable must aff ect the mediator in the fi rst equation; second, the independent variable must be shown to aff ect the dependent variable in the second equation; and third, the mediator must aff ect the dependent variable in the third equation' (Baron andKenny, 1986, p. 1177).Applying the conditions on our case study variables, we have to see fi rst that leadership support as an independent variable must have an eff ect on cynicism as mediating variable, we also have to see that leadership support is shown to aff ect the emotional burnout as the dependent variable, and we need to present that cynicism as the mediating variable has eff ects on emotional burnout, while leadership support has less eff ects on the dependent variable in the third equation when compared to its eff ect in the second equation (as cynicism having partial mediating role), or has no eff ect at all when the cynicism is controlled, implying cynicism to have the perfect mediating role.
First construct (Model 1) aims to show the eff ects of leadership support on emotional burnout.Goodness-of-fi t (GOF) indices pertaining to this model (Table 5) show that some of the values of the model (i.e., relative chi-square-CMIN/DF, chi-square-χ2, Degrees of Freedom-df, normed fi t index-NFI, comparative fi t index-CFI, goodness-of-fi t index-GFI, and adjusted goodness-of-fi t index-AGFI) are in bett er fi t than some others (Root mean square error of approximation -RMSEA) (Hair et al., 1998).
In Model 1 we have found some test result values (non-standardized regression coeffi cient is -0.372; standard error is 0.047; standardized regression coeffi cient is -0.415 and p value is 0.000).Thus, we found that leadership support signifi cantly but negatively aff ects emotional burnout (β= -0.415; p<0.001).From here we argue that the administrative employees in Turkish universities will suff er less from emotional burnout if their directors increase their leadership support.Conclusively, we positively confi rmed our fi rst hypothesis (H 1 ).In Model 2, we examined our variables as if they were not interrelated for the presence of second and third conditions stated in Baron and Kenny's study (1986).We see that our structural equation model is acceptable if goodness-of-fi t values are taken for granted (Table 5).We present the results of the analysis and path coefficients in Figure 2 and Table 6, respectively.According to the analysis, we see that the leadership support negatively aff ects the employee att itudes towards cynicism (β= -.589; p<0.001).Additionally, the level of cynicism att itudes of employees positively aff ects the level of their emotional burnout (β= .760;p<0.001).Thus we have confi rmed second (H 2 ) and third (H 3 ) hypotheses.When we add the cynicism to the model we see that the explanatory power of the model signifi cantly rises from 17% to 55%.We observed that signifi cantly negative aff ection of leadership support on employee burnout in Model 1 (β= -.42; p<0, 01) seem to be insignifi cant (β= .03;p> 0.05) in Model 2. We confi rmed that our models seem to meet all the conditional require-ments set by Baron and Kenny (1986).We have employed Sobel's test of mediation to validate the signifi cance of any decrease in β value, for which we have found 3.88027, p<0.001.Thus we have confi rmed additional tests for mediating relationship among variables.So, we have also confi rmed our fourth hypothesis (H 4 ) arguing that cynicism plays a mediating role between leadership support and emotional burnout.We present the results of hypotheses testing in Table 7.

Discussion
Our fi ndings point out to some indications about organizational issues that administrative employees experience in universities.The study, as shown by other studies, fi nds that leadership support has signifi cantly negative eff ects on emotional burnout.Studies in literature found that organizational cynicism, not necessarily an external phenomenon brought to the work environment together with employees, is an internal emotional att itude, which is created within the organization through employees' experiences ( Johnson and O'Leary-Kelly, 2003).If this att itude is reinforced with non-existence of leader support, we might expect employees to go through an emotional burnout syndrome among others.Burnout and organizational cynicism have really close ties and form an intertwined phenomenon (Özler and Atalay, 2011).Tüzün, Çetin and Basım (2014) cited that social support that nurses had from their supervisors not only decreased burnout levels but also lessened their intentions to leave.Employees may perceive leader support as a kind of perceived job security, irrespective of work nature, though public employees are guarded by the rule of law when compared to others.Moreover, it was found that job insecurity increases the employees' emotional burnout levels (Westman, Etz ion and Danon, 2001).Leadership support, in the same direction with organizational support (Chiaburu et al., 2013;Byrne and Hochwarter, 2008) has negative eff ects on organizational cynicism and emotional burnout.Stordeur, D'Hoore and Vandenberghe (2001) have found that leadership styles were partially aff ecting emotional burnout in nursing context since there might have been many other factors sourcing from the nature of the nursing job.Skakon et al. (2010) state that there is an association between leader's behavior and employee stress in organizational contexts in their review paper of empirical research conducted for 30 years.Moreover, van Emmerik, Bakker and Euwema (2009, p. 594) argue that 'support from the supervisor buff ered the negative relationship between emotional demands and favorable evaluations of organizational change'.We have to cite Reichers, Wanous and Austin (1997) who argue that organizational cynicism of the employees would be closely associated with the behaviors of supervisors, or top managers as the agents, or representatives of the organization.Byrne and Hochwarter (2008, p. 67) state that 'it is plausible that a lack of support promotes cynicism'.
In another study, it is argued that 'burnout is characterized by a combination of exhaustion (low activation) and cynicism (low identifi cation)' (Schaufeli et al., 2002, p. 74).It is clear that there is strong and signifi cant correlation between cynicism and burnout.It is demonstrated in our study as well.As a second fi nding of the study, the level of organizational cynicism of admin employees is positively aff ecting the level of emotional burnout, also as proven in the literature.In similar vein, cynicism was found as a key dimension for burnout resulting in employee turnover (Leiter and Maslach, 2009, p. 331).
Another relevant fi nding of the study is that cynicism has a mediating role on leadership support aff ecting the emotional burnout.This dimension is well studied by employing various perspectives.For the relationship between leaders' behaviors and traits with their refl ections on subordinates, Ronen and Mikulincer (2012, p. 841) argue that leaders 'being att entive, responsive and supportive may be a key for managers to promote subordinates' job satisfaction and prevent job burnout'.Yürür and Sarikaya (2012) found that social support from supervisors aff ected workers' emotional exhaustion in negative terms, and personal accomplishment in positive direction.In a similar line, Neves (2012, p. 965) found that 'organizational cynicism interferes in the relationship employees develop with their supervisors, with consequences to performance'.Cole, Bruch and Vogel (2006, p. 476) found that 'perceived supervisor support and hardiness to have indirect relationships with cynicism through employees' aff ective reactions'.For a direct aff ection of leadership support on cynicism, Laschinger, Wong, and Grau (2013, p. 541) found that 'higher levels of authentic leadership and structural empowerment were associated with lower emotional exhaustion and cynicism'.

Conclusion
According to Hofstede (1983), one of the key issues in studying organizations is the eff ect of contextual culture on the management.Such an approach assumes ethnocentric factors as crucial as others, his studies arguing for the cross-cultural diff erentiations in management business act as seminal in literature, though.Keeping in mind that we have no aim to show the eff ects of Turkish type of ethnocentric administration on public employees, rather we may benefi t from Hofstede's conceptualization of dimensions of national culture (1980; 1983; 1984).We argue for strong uncertainty avoidance to fi t best for administrative employees in Turkish public universities for at least two reasons.First, it is a public institution framework where all the discourses and actions are strictly and clearly designed by law under the umbrella of hierarchy.Second, studies in relevant literature in Turkish context (Wasti, 1998;Pellegrini and Scandura, 2006; Pellegrini, Scandura and Jayaraman, 2010; Keleş and Aycan, 2011) point out to some particular type of societal culture composed of paternalistic, au-thoritarian and representing a high level of power distance elements, even predominant in the private sector, in Turkey.Thus, we expect that leadership support would have more eff ects on decreasing the emotional burnout levels in the Turkish context since it inherently entails collectivist values when compared to other contexts where individualistic values are more intrinsic.Thus we strongly assume that cultural characteristics of Turkish society have impact on the behaviors of all employees irrespective of sector.Since the study relies upon the data collected in Turkish universities, we might wait that the results of the study could be used for a comparison in prospective intercultural studies.
Following the results of the study we propose managers in universities to increase the leadership support to administrative employees.We argue that by increasing leadership support to the att itudes of administrative employees in conducting their duties might help decreasing the emotional burnout levels.This also helps administrative employees to feel more loyal to their institutions, which could not be provided at full by material gaining, or legal measures proactively taken otherwise.

Figure 1 :
Figure 1: The research model

Table 1 :
Results for factor analysis, reliability and descriptive statistics

Table 2 :
Results of scales' goodness of fi t after confi rmatory factor analysis

Table 3 :
Results for the Demographic Information of the Participants

Table 5 :
Results of GOF indices for Structural Equation Models

Table 6 :
Standardized and unstandardized regression coeffi cients for Model 2

Table 7 :
Results of hypotheses testing