The mediating effect of leader–member exchange in relationship with emotional intelligence, job satisfaction, and turnover intention

The purpose of this research is to study the mediating effect of leader– member exchange quality variable in the relationship between emotional intelligence and turnover intention and job satisfaction of employees. In terms of methodology, this is a survey correlational research. Statistical population included the staff of Sayeh Sabz Oloum Company in Tehran-Iran that all members of the population were studied. Research data were collected through questionnaires. Reliability was accepted by Cronbach’s alfa coefficient and validity was accepted by confirmatory factor analysis. Data analysis and research hypothesis testing was performed through structural equation modeling and hierarchical multiple regression. Two softwares (SPSS and LISREL), also were used for statistical processing. The results showed the mediating effect of leader–member exchange in quality variable did not have any effect on relationship between emotional intelligence and job satisfaction and turnover intention. Emotional intelligence skills had positive relationship with job satisfaction, but had negative relationship with turnover intention. In other words, emotional intelligence improves staff’s attitude toward decreasing turnover and increasing job satisfaction. Job satisfaction also causes reduction in turnover. *Corresponding author: Ali Gholipour Soleimani, Faculty of Accounting &Management, Rasht Branch, Islamic Azad University, Rasht, Iran E-mail: Gholipourbahman@yahoo.com Reviewing editor: Len Tiu Wright, University of Huddersfield, UK Additional information is available at the end of the article


PUBLIC INTEREST STATEMENT
The purpose of this research is to study the mediating effect of leader-member exchange quality variable in the relationship between emotional intelligence and turnover intention and job satisfaction of employees. The findings of this study can help to increase managerial skills and organizational progress. The results indicate that the managers and their employees should improve their level of emotional intelligence and their ability to reduce the intention to leave work. Thus, it reduces the probable costs of personnel relocation and increases the organization efficiency, which can increase employees' job satisfaction. So it makes employees work more motivationally. Also managers should improve the quality of their relationship with their employees, as they can increase job satisfaction by reducing the intention to leave work. The use of this action for organizations is reducing stress, increasing the morale and confidence, and reducing costs of leaving work.

Introduction
Although psychologists have studied various aspects of emotional intelligence in organizations without using emotional intelligence term itself, emotional intelligence is a relatively new concept that is being used in management literature in the late 1990s. The importance of emotional intelligence in leadership researches has been acknowledged over and over. Over the last six decades, researchers have used emotional intelligence's constructs to explain the relationship between positive moods and charismatic leadership, emotional management and transformational leadership, expressing emotions and leader-member exchange, as well as leader's emotional displays and followership behaviors (Jordan & Troth, 2011).
The importance of emotional intelligence arises from the fact that it is not only an objective and abstract ability, but it also includes social abilities (Duckett & Macfarlane, 2003). According to its nature, emotional intelligence is a combination of different abilities that assists a person in interpersonal interactions. Emotional intelligence is the pivot of leadership since it affects leader-follower relationships (Humphrey, 2002).
While researchers have studied the strong impact of emotional intelligence on the satisfaction of staff's job (Ealias & George, 2012), reduction of staff's job desertion (Jeswani & Dave, 2012), leadership specialties (Caruso, Mayor, & Salovey, 2002), and its role on self-managing teams (Wolff, Pescosolido, & Druskat, 2002) but the main goal of this article is to review and answer the following questions: • Does the quality of the leader-member exchange as a mediator affect (1) the relationship between emotional intelligence and staff's resigning intention? (2) the relationship between emotional intelligence and staff's job satisfaction?
• Is there any relationship between emotional intelligence and staff's attitude toward job satisfaction or resignation?

Emotional intelligence
The suggested model of emotional intelligence (Mayor & Salovey, 1997) includes four connected emotional skills which are: perception, absorption, recognition and management of self and other's emotions. There is an important debate on the validity of emotional intelligence structures and specially its relationship with the leadership (Antonakis, 2004;Antonakis, Ashkanasy, & Dasborough, 2009).
Briefly, Antonakis claims that emotional intelligence is not necessary to a proper leadership and proper leaders can cope with emotional needs of leadership role by simple manuscripts.
Against the denial of the emotional intelligence by Antonakis (2003), another group, responding to his idea, considered communicational aspects of leadership as instances of emotions. Ashkanasy and Daus emphasize on emotional intelligence as a structure which is the relationship between emotions and recognitions. Therefore, they know emotional intelligence as an ability that should be measured by traditional intelligence tests. This point of view is in opposition to researchers who agree with Mayor and Salovey (Ashkanasy & Daus, 2005). Some researchers define emotional intelligence as a kind of personality variety. In this group's model, emotional intelligence is considered as a compound model, but most researchers have consolidate on Mayor and Salovey model as foundation of university research at the moment (Cherniss, 2010). Practical ability of emotional intelligence of staff has been shown in researches related to the use of emotional intelligence in work place. Jordan, Dasborough, Daus, and Ashkanasy (2010) mentioned that effective staff needs to be aware of their own and other staff's emotions at workplace. They presented a four-dimensional definition of emotional intelligence in other research which extensively was affected by Mayor and Salovey's studies. In this model, the four structures which are presented for testing the emotional intelligence are: the perception, assimilation, understanding, and management of own and others' emotions. Mentioned variables are measured by following indicators: (1) description of emotions and feelings of a member to other members of organization.
(2) Expressing feelings that are pleasant for a person to colleagues.  Mayor and Salovey's (1997) discoveries show that emotional intelligence abilities help people in recognition, perception, and management of self-emotions and emotions of other people and it plays key role in improving their relationships in workplace. Based on Jordan and Troth research (2004), emotional intelligence is considered as a basic principle in leadership in order to promote relationship at workplace. There is some evidence that indicates emotional intelligence effects on staff perspective.

Emotional intelligence and employee's attitude
For instance, Carmeli has found the fact that there is a connection between emotional intelligence and how the senior managers manage it in public. Managers who have higher emotional intelligence were not only able to manage stress but also they could solve the problems which might cause stress. Some researchers believe that although the lack of job security can increase staff's job desertion, their job desertion amount can be decreased by managing the lack of job security tensions (Carmeli, 2003;Wong & Law, 2002).
Another study reports staff's emotional intelligence skills and their desire to establish high-quality relations in workplace.
Employees, who have high rated emotional intelligence, are able to establish and maintain relations in a higher quality (Lopes, Salovey, & Straus, 2003). The concept of employee's outlook has been measured by two variables which are job satisfaction and desertion intentions in this research.

Job satisfaction
Job satisfaction is a state of complacency and positive sense which is caused by an individual assessment of his job and experiences. Job satisfaction includes emotions and attitudes of a person toward his job (Robbins, 2016). Job satisfaction, along with efficiency quality and turnover, is one of the key dependent variables that is considered and measured in organizational psychology researches (Roudposhti, Mahmoodzadeh, & Mahmoodzadeh, 2011). The variable had been measured with following indicators in this research: (1) Satisfaction of job security rate.
(3) Personal growth and development satisfaction.
(4) Feeling to be valuable in job affairs.

Turnover intention
Turnover Intention is decision of an employee to leave the organization and find alternative job. Staff job desertion is considered as an important anxiety for corporate managers and it is considered as a negative event in lower levels of organization. When good staff leaves organization, organization should replace others and it should shoulder replacement and pedagogical costs, losing productions and disaffiliation with staff (Sexton, McMurtrey, Michalopoulos, & Smith, 2005). Job desertion intention variable is measured by following characters: (1) Thought of leaving the job; (2) Finding job in New Year; (3) Leaving corporation in case of finding other job; (4) Quick job desertion (Lutchman, 2008).

Leader-member exchange
Leader-member theory or leader-member relationship mostly deals with developing dynamic relations among managers and their staff. The mentioned pattern is completely different from other theories of individual characteristics and behaviors of leader in the case of its emphasis on relations between managers and inferiors.
Since the pattern doesn't match such a theory, that is, the leadership style, that is determined on behaviors of leaders, has huge differences from previous patterns. In other words, it was assumed that behaviors of a leader is the same to other staff in previous patterns, but leader-member exchange pattern is based on the assumption that leaders have unique relationship with every inferior and they behave each employee in specific way. Behavioral scientists call such a relation as "vertical Dyad". Based on findings, development of vertical Dyad is the format of processes that occur naturally. This type of relation also arises due to leaders' efforts in delegating business role to members. From expected results of mentioned process, two types of leader-member exchanging relations could be pointed. First type is called "In-group exchange". In this type of relations, leaders and members develop a partnership emerging through interpenetration, mutual trust, honor and mutual sentiment toward fate. Second type of relations is called "out-group exchange". In this type of relations leaders are considered as supervisors who have failed in creating mutual trust, honor or common fate (Krintner & Kiniki, 2008). In-group members can go beyond official duties and job descriptions while out-group members usually are weary and unwilling of working beyond their duties. Leaders represent more responsibilities, opportunities, occasions and protection in quid pro quo in -group members' efforts and devotion (Flemming, 2009).
Anyway, Afzalur Rahim et al. (2002) believe that both in-group and out group members should negotiate with leader by taking initiative in order to achieve aims of group and going beyond assigned duties. Although it's obvious that in-group members are more desirable than out-group members, in-group members or inferiors should steadily urge to accept different and new job responsibilities, otherwise, they will be changed to a part of out-group members.
Conceptually, leader-member exchange can be used to analyze management policies that show how members create leadership networks throughout organization to help them in performing their duties effectively. Leader-member exchange theory also is applied in various organizational structures such as traditional, educational and governmental institutes. Leader-member exchange theory has strength and weaknesses. First, it represents a full analyze from power foundation. Second, there are various researches that show how leader-member exchange theory can be connected to positive organizational consequences. On the other side, it is clear that leader-member exchange can create unnecessary prejudices among inferiors and members and can be an obstacle for values of justice. Such perceptions of justice can lead to discrimination among people or genders, age, and different cultures. Another weakness can be main thoughts of leader-member theory has not developed properly and it does not explain properly that how a high-quality relationship between leader and members can be established. In addition, although, there are deep researches about leadermember exchange, mentioned researches have not led to clear conceptions and defaults about theory (Saks, 2006). Leader-member exchange has been measured by following characteristics in this research: (1) Understanding staff problems and needs by manager.
(2) Identifying staff abilities and potentials by manager.
(3) Trust and confidence to manager.
(4) Effective business relations with manager.

Review of literature
According to the evidences emotional intelligence is effective on staff prospective in workplace (Wong & Law, 2002). Based on Kakaie research, factors such as lack of appreciation, lack of opportunity of promotion, multiple administration, dissociative disorder, employee features below the needed level and employee features upper the needed level can lead to staff desertion (Kakaei, 2011). Jordan et al. show that while job in securing can increase staff job desertion, emotional intelligence helps staff to manage negative tensions of feeling job insecurity. In addition, Staff who have high emotional intelligence, are able to create relationship in higher quality (Lopes et al., 2003). Jeswan and Dave proved that emotional intelligence is effective on deduction of staff desertion by studying effects of emotional intelligence foregrounds (Interpersonal relationship, intrapersonal relationships, compatibility, stress management, general feeling) on staff desertion intention (Jeswani & Dave, 2012).
This research focuses on staff emotional intelligence skills and their desire to establish high-quality relationship in workplace (H1).
Scientific evidence has shown effect of emotional intelligence on staff job experiences, thus, represents a positive relationship between emotional intelligence and higher rated job satisfaction. Wong and Law's proved that emotional intelligence promote staff job satisfaction. Sy, Tram, and O'Hara (2006) also have highlighted the relationship between job satisfaction and emotional intelligence. In their opinion, managers with high emotional intelligence can manage tensions, may lead to deduction of job satisfaction better. In other researches that focus on staff emotional intelligence, it's shown that higher levels of emotional intelligence lead to higher quality relationships and being environmental support can cause to job satisfaction (Lopes et al., 2003). (H2) Barbuto and Bugenhagen (2009), in their research, discovered that there is an important and semantic relationship between leader emotional intelligence and leader-member exchange quality. Jordan and troth proved that staff emotional intelligence has a positive relationship with leader-member exchange quality and job satisfaction and has a negative relationship with job desertion intention (Jordan & Troth, 2011).
While leaders' emotional intelligence is important, we claim that emotional intelligence of members also is effective and important in their relationship with managers (H3). Maslyn and Fedor (1998), have shown that there is negative correlation between leader-member exchange and staff desertion intention. Members, who have high-quality relations with leaders may show less desire to leave the organization. Schyns and Wolfarm (2008) research also shows that member's perceptions of leader-member exchange effects on staff outlook. According to Schyns, Torka, and Gössling (2007), negative correlation is between leader-member relationship and staff desertion intention and positive correlation is between leader-member relationship and readiness to exchange (H4).
According to Schyns and Wolfarm (2008) studies, there is relationship between staff outlook of leader-member exchange and staff perception. They found relationship between assessed quality of staff from leader-member exchange and job satisfaction. If job satisfaction is an effective reaction, thus, it would be increased based on experiences due to positive tensions. Therefore, so called, high-quality relationships will cause to appear positive tensions and cause to happen job satisfaction in higher levels (Weiss, 2002) (H5).
Managers, consistently, concentrate on job satisfaction. First, we can observe a direct relationship between member job satisfaction rate and productivity. Second, it seems that member's job satisfaction has reserved relation with job desertion rate. At last, it can be reasoned that managers feel to be responsible toward staff satisfaction (Robbins, 2016). Khilji and Wang show that satisfaction of human resources has negative relation with optional job desertion intention. It means that increasing job satisfaction will cause to deduction of job desertion (Khilji & Wang, 2007) (H6).
According to evidences, there is relation between emotional intelligence and having high-quality relationship with others (Lopes et al., 2003). Based on mentioned theory, members with high emotional intelligence can effect on quality of their interaction with their managers.
Specially, members, who manage their relations actively and are aware of their own tensions, can manage tensions during their interactions with their managers and cause to appear high-quality leader-member exchange. Having more qualified relations plays an important role in staff working and it affects on staff outlook of job desertion and job satisfaction (Mitchell, Holtom, Lee, Sablynski, & Erez, 2001) (H7, H8).

Model and research hypothesis
In Figure 1, conceptual model of research, expected relations between variables, and other research theories have been represented.
H1: There is relationship between staff emotional intelligence and turnover intentions.
H2: There is relationship between staff emotional intelligence and job satisfaction.
H3: There is relationship between staff emotional intelligence and leader-member exchange quality H4: There is relationship between leader-member exchange quality and turnover intentions.
H5: There is relationship between leader-member exchange quality and job satisfaction.
H6: There is relationship between job satisfaction and turnover.
H7: leader-member exchange quality acts as mediator between emotional intelligence and job satisfaction.
H8: leader-member exchange quality acts as mediator between emotional intelligence and turnover intentions.

Research methodology
This research is applied in terms of target and correlation survey in terms of data collection. For data collection, five-point Likert scale questionnaire is used. Research statistical population includes 280 staff of Sayeh Sabz Oloum Company (Tehran) which all of them have been studied. The main activity area of this company is to present technical consulting in setting print systems up, sales services, equipment, security documents and different types of labels printing. From all distributed questionnaires, 246 questionnaires were collected that 11 questionnaires set aside due to incompleteness, and 235 questionnaires were analyzed. After checking descriptive statistics, demographic characteristics of respondents is shown in Table 1. Table 2 shows number of questions related to research variables and reliability coefficient. The questionnaire contains 22 questions and since obtained confidence coefficient is more than 7% by Cronbach's Alpha method for all questions, therefore, all questions are reliable. In tentative implantation stage of questionnaire by 30 people sample, Cronbach's alpha coefficient for all questions of questionnaire had been determined "0.873" and total questionnaires in last stage of Cronbach's alpha coefficient had been determined "0.828" which indicates high reliability of questions.

Data analysis
For studying that whether data is normal or not, Kolmogorov-Smirnov test was used. The results showed that significant level for all variables is above 0.05. So it can be concluded that all research variables have normal distribution. We also used structural equation model to test hypothesis 1 to 6, hierarchical regression model to test hypothesis 7, 8 and LISREL and SPSS for data processing.

Validity and confirmatory factor analysis
General framework of questionnaire has been designed based on standard questionnaires, therefore, it possesses sufficient validity in high amount, but we got profited by professors' views to ensure more about validity of questionnaire and necessary reforms were applied. We also used confirmatory factor analysis model for studying explanatory assessment of each variable by related questions (validity) which represents confirmation of research tools validity. Table 3 expresses results of factor analysis model.  In Figure 3, standardized coefficient (Factor loadings), explained variance, and significant effects of each variable and related question have been mentioned. The significant effects of all questions are more than 1.96, therefore, it can be concluded that all questions of questionnaires have enough validity for evaluating related variables. According to explained variance of each factor, we can find out the relative importance of observed factors as a measurement index.

Hypothesis test
All research hypotheses were tested in error alpha level (0.5) and level of significance (0.95). Regarding acceptable error rate (0.5), table (t) is placed in range of (−1.96 to +1.96). In other word, if computational (t), that is determined by structural equations and LISREL software, is out of mentioned range (−1.96 to +1.96), hypothesis of H0 will be rejected and hypothesis of H1 will be confirmed. As we can find out from Figure (2), computational (t) coefficients for hypothesis 1-6, is out of (−1.96 to +1.96) range. Therefore, we can conclude that the first six hypotheses are confirmed. Figure 3 shows research model in standardized coefficients form. The results of mentioned model are compatible with significant state. Based on Table 4, emotional intelligence has the highest total effect of 0.72 on leader-member among all variables. On the other side, the highest indirect effect of emotional intelligence variable on job satisfaction variable has the effect coefficient of 0.38. In Table 5, the summary of research hypothesis test is presented to show results better.

Mediating variables
To study mediating effect of "leader-member exchange quality" among staff emotional intelligence, job satisfaction, and turnover intention variable (H7, H8), first, the regression between "emotional intelligence and turnover intention" was calculated without mediating interference and their   relationship was analyzed. Then, mediating variable was entered the model and regression was calculated again, in order to know whether results were changed by entrance of mediating variable or not. The summary of obtained analyze is shown in Tables 6 and 7.
The results of estimating hierarchical regression are shown in Table 6. Standardized coefficient (β) in the relationship between emotional intelligence and job satisfaction, before entrance of mediator variable, was 0.714. After entrance of leader-member exchange variable to the model, the figure decreased to 0.326. Therefore, it can be concluded that H7 is rejected because of decreasing effect coefficient of emotional intelligence on job satisfaction after entrance of mediator variable. In Table (7), the results of estimating hierarchical regression show that the effect of standardized coefficient (β) in the relationship between emotional intelligence and turnover intention, before entrance of mediator variable, was −0.569. But after entrance of leader-member exchange variable to the model, the figure increased to −0.352. So, it can be concluded that H8 is also rejected because there isn't any improvement in the relationship after entrance of mediator variable.

Conclusions
Based on results of research hypothesis test, we can mention following results and comparison of them with similar researches: • According to the results of first to third research hypothesis test, it seems that if staff emotional intelligence improves, turnover intention will be decreased, job satisfaction will be increased and leader-member exchange quality will be improved. Regarding emotional intelligence description and its components in research questionnaire, we can declare that if members' skills in identifying, recognizing, understanding and managing their own tensions and others increase, (that is because of job type, workplace and relationship between manager and colleagues), their attitude toward job desertion and job satisfaction may be improved. So they may establish relationships with their managers in higher quality. Therefore, sharing emotions with other organization members, sharing pleasant emotions with others, respecting the ideas of other colleagues, overcoming negative emotions and anger, understanding and finding out emotions and feelings of others, encouraging and giving spirit to disappointed colleagues and totally having positive energy can help to increase emotional intelligence ability. Mentioned finding is favorable with research result of Jeswani and Dave (2012), Jordan and Troth (2011), Ealias and George (2012), and Barbuto and Bugenhagen (2009).
• Based on results of hypothesis 4 and 5 tests, it seems that if quality of leader-member exchange increase, not only job desertion intention will decrease but also amount of job satisfaction will increase. Therefore, according to leader-member variable structures, if staff feel that their managers understand their problems and issues, recognize their abilities and talents, trust their managers and have effective business relationships with their managers, they will help their managers, so the quality of relationship between manager and staff will be increased. It also seems that leader-member exchange quality may increase in case of being job satisfaction among staff.
Result of this research is consistent and in line with research result of Jordan and Troth (2011).
• Regarding hypothesis 6 test, it seems that job satisfaction can decrease staff job desertion intention. Therefore, according to job satisfaction structures, staff will have positive attitude toward their job, if they have job security feeling, satisfaction about salaries and wages, they feel individual promotion and development, and they have the feeling of being valuable. The result of this research is consistent and in line with research result of Troth (2011), McNabb (2009) and Khilji and Wang (2007).
• Based on the hypothesis 7 and 8, the mediating role of the leader-member exchange is not confirmed in the present study, and this finding is not consistent with the research of Jordan and Troth (2011). It seems that the cause of inconsistency is the difference between the statistical population of the present research and the mentioned research.

Managerial implications
The findings of this study can help to increase managerial skills and organizational progress. The results of this research indicate that the managers and their employees should improve their level of emotional intelligence and their ability to reduce the intention to leave work. Thus, it reduces the probable costs of personnel relocation and increases the organization efficiency, which can increase employees' job satisfaction. So it makes employees work more motivationally.