The relationship between hotel staff’s organizational justice perception, relationship quality and job performance

Abstract With the increasing competition among hotel companies, how to improve the job performance of employees is the key to enhancing the competitiveness of hotel companies. Based on the theory of organizational justice and the theory of relationship quality, this paper develops an integrated model of the relationship between hotel employees’ organizational justice perception, relationship quality and job performance. The model is examined using a valid sample of 303 hotel employees in OYC Hotel, Star lake Hotel, Vienna Hotel and Goldnugget Hotel located in Zhaoqing, one of China’s top tourist cities.The basic data is acquired and the Structural Equation Model (SEM) method is adopted to conduct empirical tests of the integrated model. According to the analysis procedures and methods, the organizational justice scale is tested and the data is suitable for exploratory factor analysis. Exploratory factor analysis (EFA) shows that four common factors of organizational justice were more adequate for the interpretation of the construct. Based on exploratory factor analysis, confirmatory factor analysis (CFA) was conducted to further verify the dimensional structure of organizational justice. The result suggests that organizational justice is consisted by four dimensions, i.e. distributive justice, procedural justice, international justice, and informational justice. Four dimensions of organizational justice have significantly positive effects on employee satisfaction, and indirectly affect job performance through employee satisfaction. Except informational justice, the other three dimensions of justice significantly positive influence employee trust. At the same time, international justice directly affects contextual performance, and informational justice impacts task performance directly. In two dimension of relational quality, employee satisfaction has more significant impact on job performance than employee trust. Employee satisfaction has a directly positive impact on contextual performance and task performance at the same time, but employee trust has not a significant effect on each of them. This study can theoretically enrich the current research results of organizational justice, relationship quality and work performance, and in practice, it can provide theoretical guidance for hotel management levels to improve the performance level of employees.


PUBLIC INTEREST STATEMENT
With the increasing competition among hotel companies, how to improve the job performance of employees is the key to enhancing the competitiveness of hotel companies. Based on the theory of organizational justice and the theory of relationship quality, this paper develops an integrated model of the relationship between hotel employees' organizational justice perception, relationship quality and job performance. The model is examined using a valid sample of 303 hotel employees in OYC Hotel, Star lake Hotel, Vienna Hotel and Goldnugget Hotel located in Zhaoqing, one of China's top tourist cities. The basic data is acquired and the Structural Equation Model (SEM) method is adopted to conduct empirical tests of the integrated model. The result includes that employee satisfaction has a directly positive impact on contextual performance and task performance at the same time. Our finding can theoretically enrich the current research results of organizational justice, relationship quality and work performance, and in practice, it can provide guidance for hotel management levels.
conducted to further verify the dimensional structure of organizational justice. The result suggests that organizational justice is consisted by four dimensions, i.e. distributive justice, procedural justice, international justice, and informational justice. Four dimensions of organizational justice have significantly positive effects on employee satisfaction, and indirectly affect job performance through employee satisfaction. Except informational justice, the other three dimensions of justice significantly positive influence employee trust. At the same time, international justice directly affects contextual performance, and informational justice impacts task performance directly. In two dimension of relational quality, employee satisfaction has more significant impact on job performance than employee trust. Employee satisfaction has a directly positive impact on contextual performance and task performance at the same time, but employee trust has not a significant effect on each of them. This study can theoretically enrich the current research results of organizational justice, relationship quality and work performance, and in practice, it can provide theoretical guidance for hotel management levels to improve the performance level of employees.

Introduction
With the globalization of the hotel industry and the intensification of international competition, it is the key factor for the success of the hotel industry to manage all kinds of resources, retain employees and improve their work performance (Alamir et al., 2019;Nadiri & Tanova, 2010). Especially since 20 th century, the management method has been constantly updated, the application of new technology is constantly emerging, and the competition of the hotel industry is becoming more and more intense. So the improvement of staff performance has become the top priority of hotel management. Therefore, some scholars have developed studies on work performance and its influential factors (Alamir et al., 2019;Choi, 2011;Colquitt, 2001;Frazier et al., 2010). In these related studies, based on the theory of organizational justice and social exchange theory, they explore the influence of organizational justice on work behavior and attitude as well as job performance, and find that organizational justice has a significant impact on the attitude and performance of employees, which is an important predictor of work attitude and behavior and also is inevitable requirement for the realization of effective organization and management (Cohencharash & Spector, 2001;Colquitt, 2001;Hur & Ha, 2019). Although the researchers are trying to find the relationship between organizational justice and job performance, and the results have been obtained, the results are not satisfactory, as reflected in the inconsistent or undefined research conclusions (Alamir et al., 2019;Hur & Ha, 2019;Nadiri & Tanova, 2010). For example, some studies show that organizational justice can predict job performance, however, the influential conclusion of the dimensions of organizational justice on job performance is quite different. The functional mechanism of organizational justice on job performance is still very vague. In particular, most of these studies are based on the Western cultural background, and in the non-Western cultural context, the relationship between organizational justice and job performance may be different (Ekmekcioglu & Aydogan, 2019;Elanain, 2010). Therefore, it is of great significance to further explore and verify the relationship between organizational justice and work performance in the typical Chinese oriental cultural background.
In addition to the impact of organizational justice on job performance, with the development of the relationship quality theory, the relationship between the relationship quality and the work performance also arouses the extensive interest of the researchers, and it is found that the relationship quality is an important factor affecting the employees' attitude, behavior and performance. Satisfaction and trust are the two basic dimensions of relationship quality (Frazier et al., 2010;Hon & Lu, 2010;Nadiri & Tanova, 2010;Sanchez-Franco et al., 2019). In this study, it is believed that job satisfaction is significantly related to job performance, because satisfactory employees will provide better services, which will bring loyalty to customers, and will also affect customer confidence and word-of-mouth publicity (Hafeez, 2019;Nadiri & Tanova, 2010). Therefore, in highly competitive environment, promoting employee satisfaction helps to improve customer satisfaction and achieve customer loyalty, thereby reducing cost and establishing market share, and ultimately improving the performance of the hotel. The trust of the staff in the hotel management can make them establish a good working relationship with the hotel, which may improve the employee's perception level (Hon & Lu, 2010), and produce high work performance level. Therefore, improving employee satisfaction and enhancing the sense of trust in the organization will not only help the employees to conduct extra-role behavior,but also help to take positive work attitude and behavior, and create teamwork and coordination. This will be conducive to increase the competitiveness of the hotel enterprises and ultimately improve the development performance of the hotel enterprises (Hafeez, 2019;Nadiri & Tanova, 2010;Sanchez-Franco et al., 2019).In the hotel industry, the research results about the relationship between organizational justice and work performance are relatively less (Hafeez, 2019;Nadiri & Tanova, 2010). The research results that integrate organizational justice, relationship quality and work performance are still quite insufficient (Frazier et al., 2010;Yeh, 2019). According to the above cognition, this paper, based on the theory of organizational justice and the relationship quality theory, constructs the organizational justice (result justice, procedural justice, interpersonal justice and information justice) as the intermediary variable of the relationship quality (employee satisfaction and employee trust), which has integrated model of influential mechanism on work performance (task performance and contextual performance), and through the survey of staff of top hotels in Zhaoqing city, the empirical test was conducted on the integrated model. Generally, this article has expanded the current research literature in the following aspects: Firstly, it integrates the western theoretical framework and applies it to the Chinese cultural background to analyze the commonness and differences. Secondly, it explores and examines the dimension structure of the justice perception of the organization in the hotel industry; and thirdly it analyzes and compares the specific dimension's functions of the hotel organizational justice. Fourthly, the nature of relationship between organizational justice and job performance is explored by introducing relationship quality as an intermediary variable. By this study, it can theoretically enrich the current research results of organizational justice, relationship quality and work performance, and in practice, it can provide theoretical guidance for hotel management levels to improve the performance level of employees.

Theory of organizational justice
Justice theory, also known as social comparison theory, was first proposed by American psychologist Adams (1965). Justice theory believes that people's enthusiasm for work is not only influenced by individual rewards, but also closely related to whether people feel justice about rewards. People always compare their labor costs and their income with the labor costs and income paid by others as a criterion for whether they are treated fairly. When the ratio of income to labor cost is greater than or equal to others, it produces justice perception and vice versa. This kind of justice perception directly affects their work motivation and behavior. Therefore, the activated process of people's motivation is actually compared with others, checking if the judgment of justice or not and then producing the corresponding behavior. Justice theory believes that when people encounter unjustice, they may reduce their input and output, reduce their enthusiasm for work, and interrupt relations and so on. Later, researchers gradually introduced the justice theory into organizational behavior and formed the theory of organizational justice. In the late 1990s, justice theory was applied to the service industry (Smith et al., 1999). Earlier studies regarded organizational justice as a single dimension concept, and with the deepening of research, the academic community basically agreed that organizational justice is a multidimensional concept (Weber, 2008). At present, four types of organizational justice have been identified: distributive justice, procedural justice, interpersonal justice and information justice (Greenberg, 1990;Weber, 2008).
Distributive justice is the justice that researchers initially focus on, refers to the result of a decision or a transaction. Due to the neglect of the distribution process, Thibaut and Walker proposed interactional justice, which refers to the process of producing results (Thibaut & Walker, 1975). Distributive justice and procedural justice constitute the two basic dimensions of organizational justice. Later, some researchers found that interpersonal exchange and interaction are also an important part of justice perception, so Bies and Moag proposed interactional justice, which refers to the employees' perceptions of employers' attitude, behavior and manner to employees during the implementation of organizational processes (Bies & Moag, 1986). In a study conducted by Bies and Shapiro, firstly it separated informational justice from interactional justice. Then, Greenberg formally divided interactional justice into two parts: interpersonal justice and informational justice (Bies & Shapiro, 1988). Interpersonal justice refers to the degree to which a superior is perceived by an employee to respect and treat themselves in the execution of the relevant procedures and decisions. Informational justice refers to the reasons for a program and the extent to which the results are assigned to the employees in the process of executing the relevant procedures and making results. Later, more and more studies confirmed the four dimensions' structure of organizational justice (Hafeez, 2019;Liu et al., 2003;Sanchez-Franco et al., 2019;Xie et al., 2007;Yeh, 2019). Based on these findings, this paper argues that hotel employee's perception of organizational justice also includes four dimensions, such as distributive justice,procedural justice, interpersonal justice and informational justice,and tested by empirical analysis.

Theory of relationship quality
In the marketing literature, in order to win the competitive advantage, the enterprise expects to establish a good relationship with the customer in order to improve the development performance of the enterprise. The marketing researchers have put forward the concept of the relationship quality (Guo et al., 2017;Rauyruen & Miller, 2007;Sanchez-Franco et al., 2019). Huntley defines the relationship quality as "the satisfaction and trust of the buyer over the whole relationship as time goes on", which reflects by evaluating the quality of the product, the quality of service and the justice perception of the price and the transaction (Huntley, 2006). Although there is no consistent understanding of the definition and operationalized relationship quality in the current literature, researchers seem to agree that the concept of relationship quality is a higher order construct, consisting of several different but related components or dimensions (Huntley, 2006;Rauyruen & Miller, 2007;Sanchez-Franco et al., 2019). In marketing literature, satisfaction and trust are usually used to reflect the relationship quality (Ciavolino et al., 2019;Huntley, 2006). Morgan and Hunt believe that in the long-term relationship between development and customers, in addition to achieving customer satisfaction, the role of trust should be emphasized in particular (Morgan & Hunt, 1994). In the study of organizational behavior, some scholars regard employees as the internal customer of an enterprise. In addition to the contractual relationship, enterprises should treat employees as customers, provide them with various services, establish an emotional relationship with them, so as to keep the long-term relationship between the both parties, retain employees and change their work attitude and behavior, improve work enthusiasm, produce higher work efficiency, achieve win-win strategy of enterprise and staff (Ciavolino et al., 2019;Guo et al., 2017;Harris et al., 2009). Based on these research results, this paper regards hotel employees as their internal customers, and regards employee satisfaction and trust as the two dimensions of the relationship quality between hotel enterprises and employees.

The relationship between organizational justice and the quality of relations
Since the introduction of justice theory to the study of organizational behavior and the theory of organizational justice, the relationship between organizational justice and job satisfaction has been a well-concerning focus by the researchers. Although the influence of each dimension of organizational justice on job satisfaction may be different, for example, some research findings show that distributive justice produces stronger effect on employee satisfaction than procedural justice and interpersonal justice while other research findings demonstrate procedural justice and interpersonal justice may produce stronger effect on employee satisfaction than distributive justice (Folger & Konovsky, 1989;Lind & Tyler, 1988;Lu et al., 2016),most scholars believe that organizational justice perception can improve employee satisfaction with work and there is a significant positive correlation between organizational justice and employee satisfaction (Cohencharash & Spector, 2001;Pablo et al., 2013). Satisfactory employees will make a positive evaluation of the organization, and then encourage employees to produce a stronger commitment to the organization. In the context of the hotel industry, the empirical study of Nadiri and Tanova (2010) found that distributive justice has significantly positive impact on employee satisfaction. It was found that organizational justice (distributive result, procedural justice and interpersonal justice) were positively correlated with employee satisfaction (Palaiologos et al., 2011). Therefore, the author puts forward the following hypotheses: H1a: Distributive justice significantly and positively affect employee satisfaction.
H2a: Procedural justice significantly and positively affect employee satisfaction.
H3a: Interpersonal justice significantly and positively affect employee satisfaction.
H4a: Informational justice significantly and positively affect employee satisfaction.
In social exchange relations, trust is the core concept of procedural and interpersonal justice theory. Blau (1964) pointed out that "in establishing a trading relationship, it involves investing continuously in commitment to the other party. Since social transactions need mutual trust, the key is to enhance their credibility" (Blau, 1964). Fair treatment by one trading partner (superior or organization) to the other party (employee) can improve their own credibility and strengthen social trading relations (Aryee et al., 2002;Lu et al., 2016). According to the theory of organizational justice, when employees believe that a fair policy is used in the organization, they will generate a higher level of trust for the leaders to show more support for the policy (Lind & Tyler, 1988). Conversely, when employees perceive the unjustice of the organization's resource allocation process, they may develop negative attitudes, such as low trust and commitment to higher levels and organizations, and the behavior of opposition to the organization. Some empirical studies show that procedural justice and interpersonal justice are significantly related to employees' trust in organizations and superiors (Aryee et al., 2002;Lu et al., 2016). Perceived procedural justice has a significant impact on employees' attitude towards the overall management of the organization (such as positive evaluation of management) (Ciavolino et al., 2019;Pearce et al., 1994). Interpersonal justice is an important factor affecting the attitude of employees, such as the level of trust to the superior, and thinks that procedural justice is more closely than the justice of the results and the relationship between organizational commitment and the relationship of upgrading trust (Konovsky, 1987). Zhang and Yao (2005) found that organizational justice had the greatest impact on employee trust. Therefore, the author puts forward the following hypotheses: H1b: Distributive justice significantly and positively affects employee trust.
H2b: Procedural justice significantly and positively affects employee trust.
H3b: Interpersonal justice significantly and positively affects employee trust.
H4b: Informational justice significantly and positively affects employee trust.
2.2.2. The relationship between relationship quality and job performance Borman and Motowidlo (1997) divide work performance into two dimensions of task performance and contextual performance. Task performance is closely related to the content of the specific work, but also closely related to human ability, proficiency, and work knowledge. The perimeter performance refers to a broader organizational, social and psychological environment, which is closely related to the organizational characteristics of the performance. This behavior, although the maintenance and service of the organizational technology core, is closely related. There is no direct link, but this performance is more important when viewed from the broader business environment and the long-term strategic objectives of the business. Bolman and Motowatti's method of dividing job performance into task performance and peripheral performance has been widely accepted by scholars ( (Arshadi, 2010). This paper also uses this division method to divide hotel staff performance into two dimensions: task performance and job performance. On the relationship between satisfaction and job performance, Arshadi has found that job satisfaction has a significant positive impact on work performance (Arshadi, 2010).
In the hotel industry, Kim and Brymer did not directly examine the relationship between job satisfaction and job performance, but they found that job satisfaction was conducive to reducing turnover intention and thus enhancing the performance of the hotel (Kim & Brymer, 2011). According to the theory of relationship quality and the dual factor theory of "incentive health", there is a positive correlation between employee satisfaction and job performance, that is, employee satisfaction leads to high job performance, and dissatisfaction leads to low performance. Later, scholars made an empirical test of this thesis and confirmed its correctness. (Arshadi, 2010;Ciavolino et al., 2019) The attitude-behavior view of social psychology further believes that human behavior is largely determined by its attitude, and attitude can lead to behavioral tendencies, so employees' satisfaction will affect the size of their work performance. Therefore, the author puts forward the following hypotheses: H5a: Employee satisfaction significantly and positively affects task performance.
H5b: Employee satisfaction significantly and positively affects contextual performance.
In the study of the relationship between trust and performance, based on the relationship quality theory, Ramaswami and Singh (2013) conducted the empirical study on the employees of the retail industry and found that the trust of employees to higher levels was positively affecting their work performance. Compared with employees with lower trust in their superiors,those with higher trust in superiors have higher job performance higher commitment and more restraint in opportunistic behavior. Specifically, in terms of job performance, employees with high level of trust are more willing to cooperate, because they want to build a working relationship and increase the value for the working relationship. The purpose of increasing value for employees is to enhance their performance, and it helps to improve their performance (Ciavolino et al., 2019;Costa et al., 2001;Rich, 1997). The increase in employees' sense of organizational trust also raises the perceived value of long-term relationship with the organization, resulting in high commitment and job performance. Therefore, the author puts forward the following hypotheses: H6a: Employee trust has a significant and positive impact on task performance.
H6b: Employee trust has a significant and positive impact on contextual performance.
According to the theory of organizational justice, the theory of relationship quality and the above research hypotheses, the theoretical model constructed in this paper is illustrated in Figure 1.

Description of the study
The data collection of this study was completed in a survey of several hotels in Zhaoqing,a famous tourist destination in China, noted for its picturesque scenery all year round and cultural and historic attractions linked with Taoism,Buddhism and Confucianism. Zhaoqing lies on the west bank of theXi jiang River and was ranked state-list famous historical and culture cities at the first round of public appraisal in 1984 and in 1994 and a national AAAAA scenic area in 2019. The location has been a sightseeing destination especially since the Song Dynasty.
The surveyed hotels,situated around the star lake and center of the city, included OYC Hotel, Star-lake Hotel, Vienna Hotel and Goldnugget Hotel of the same staff size which are the top four hotels selected and nominated by the local tourism administration. Four Hotels are set against a group of lakes surrounding the limestone peaks in Zhaoqing, Guangdong Province, China. They enjoys a view of blue lakes and mountains which comprise a splendid panoramic landscape to the south of the Five Ridges and Seven Star Crags. As internationally-verified five-star hotels, hotels surveyed integrate the functions of conference, leisure and resort. Each hotel offers over 700 luxury rooms and suites.All the guestrooms are equipped with the centralized control system and broadband communication network so as to provide complete and quality service to the guests.

Research design
The questionnaire design of this study consists of four parts. The first part is the basic situation of hotel staff, including age, gender, working age length of service, etc. The second part is the measurement of organizational justice; 20 measurement questions were designed from the four dimensions, mainly from the research results of Ramaswami and Singer, Greenberg (Greenberg, 1990;Ramaswami & Singh, 2013). The third part is the measurement of relationship quality, including two dimensions of Note: Potential variables in the set model are represented by corresponding English letters to save space. Among them: DJ represents distributive justice; PJ represents procedural justice; IPJ represents interpersonal justice; IFJ represents informational justice; ES represents employee satisfaction; TP represents task performance; ET represents employee trust; CP represents contextual performance. In the following Figure 2, the potential variables are also represented by the corresponding English letters, whose meanings are completely consistent with the set theoretical model. employee satisfaction and employee trust, in which employee satisfaction includes 3 measurement questions and employee trust includes 4 measurement questions, mainly from the research results of Ramaswami and Singh (2013), Morgan and Hunt (1994). The fourth part is the measurement of job performance, including task performance and contextual performance, each having three measurement questions, mainly from the research results of Borman and Motowidlo (1997).

Data collection
Before the formal sampling, four hotels' human resource directors were assembled,briefed about the goals of the current study, provided background information on the study and given instructions on how to request participation in the study from potential respondents. An agreement was reached on the sampling rules that potential respondents consisit of first-line, middle and top managers from room division,front office,housekeeping department, security department, engineering department,human resource department, finance and accounting department, food & beverage department,,sales and marketing division, and purchasing department participated in the study. A drop-off and pick-up method was used to reduce nonresponse bias (Jennifer et al., 2009).A total of 100 questionnaires were designated to the human resource director of each of the hotels surveyed.The questionnaire was left in the hotels and was distributed by the hotel management staff to each employee, returned after filled in, and then handed over by the hotel management personnel collectively. The survey was started on 8 November 2017 and ended on 9 January 2018. A total of 400 questionnaires were distributed in this survey and 358 were recovered, with the recovery rate of 89.5%. The questionnaires were sorted out and the invalid questionnaires caused by incomplete information were removed. The final valid questionnaires were 303, with an effective rate of 84.6%.
Respondents were slightly more likely to be female (51.5%) and under the age of 30 years (84.9%). A wide range of monthly incomes were reported, with most respondents reporting having acquired a bachelor or master degree (48.5%). The 3-5 years length of service accounted for 45.9% of the participating respondents. More complete sample characteristics are provided in Table 1.On one hand, education background is in direct proportion to post,which is closely linked with age and monthly income;on the other hand, length of service and number of times of changing jobs may mirror the job satisfaction and the perception of hotel staffs toward organizational Justice and relationship Quality.

Scale reliability test
The reliability of the organizational justice scale was tested against the Cronbach alpha value. The SPSS18.0 analysis showed that the Cronbach alpha value of the scale was 0.923, which was much larger than the standard of 0.70 (Fornell & Larcker, 1981), indicating that the internal consistency of the measurement questions of the scale is strong and has sufficient reliability (Anderson & Gerbing, 1988).

KMO value and Bartlett spherical value test
According to the analysis procedures and methods of the Exploratory Factor Analysis (EFA), KMO value and Bartlett spherical value tests should be performed first to determine whether the data in this study are suitable for exploratory factor analysis (Kelloway, 1998). The analysis by SPSS18.0 showed that the KMO value was 0.905, the Bartlett spherical value was 3428.728, the degree of freedom was 105, and the corresponding adjoint probability was 0.000, indicting that the data of this study is suitable for exploratory factor analysis (Kelloway, 1998).

Common factor extraction
In this paper, the principal component method was used to extract the common factor, and the variance-maximizing orthogonal rotation method was used to rotate the common factor to be extracted, so that the load of each measurement question on its common factor is the largest, while the load on other common factors is smaller. The number of common factors was obtained according to the criterion that the eigenvalue is greater than 1 (Hu & Bentler, 1999).
From the results of exploratory factor analysis, the eigenvalues of four common factors were greater than 1, and the cumulative equation contribution rate was 78.540%, indicating that the four common factors of organizational justice were more adequate for the interpretation of the construct of organizational justice. However, from the factor load of each question, the result justice, procedural justice, and interpersonal justice each had a factor load less than 0.4, and the information justice had two questions with factor loads less than 0.4, so the five measurement questions were deleted, the final result is shown in Table 2.

Fit index test
To further verify the dimensional structure of organizational justice, this paper conducts Confirmatory Factor Analysis (CFA) based on Exploratory Factor Analysis (EFA). The fit index of the Confirmatory Factor Analysis is shown in Table 3. In comparison with values suggested by Hu and Bentler (1999), the overall fit of the proposed model to the data is acceptable (X 2 /df = 2.111,RMSEA = 0.061, GFI = 0.929, AGFI = 0.899, NFI = 0.949, RFI = 0.937, IFI = 0.9731, TLI = 0.966, CFI = 0.972). From the fit index of the measurement model in Table 3, it can be deduced that each of the fit indexes basically reached an excellent level and was a very excellent model (Cohen, 1988).

Reliability and validity test
It can be seen from Table 4 that the Cronbach alpha value of each dimension of organizational justice was between 0.875 and 0.942, and the combination reliability was between 0.874 and 0.943, which were far greater than the standard of 0.70 (Hu & Bentler, 1999), indicating that the internal consistency of each measurement item was very good and the dimensions of organizational justice were measured well, thus showing that the data of this study had high reliability.
It can be seen from Table 4 that the average extraction variance (AVE) of each dimension of organizational justice was between 0.635 and 0.864, both of which were greater than 0.50 (Fornell & Larcker, 1981), indicating that each measurement question reflected their respective constructs well, thus reflecting the fact that this study had a good aggregation validity for the measurement of the dimensions of the organizational justice.
It can be seen from Table 5 that the correlation coefficient of each dimension was between 0.369 and 0.702, the square root of their average extraction variance was between 0.797 and 0.929, and the correlation coefficient is smaller than the square root of the average proposed variance, thus indicating that there was sufficient discriminant validity between dimensions of the organizational justice.

Empirical test of the theoretical model
The exploratory factor analysis of organizational justice aforesaid shows that the initial judgment of organizational justice consists of four dimensions: result justice, procedural justice, interpersonal justice and informational justice. The confirmatory factor analysis further shows that the fourdimensional division method of organizational justice has sufficient reliability and validity, thus confirming the four-dimensional structure of organizational justice. To verify the relationship between organizational justice, relationship quality and job performance, this paper constructed a structural model to empirically test the theoretical model.

Fit index test
The fit index of the theoretical model is shown in Table 3. In comparison with values suggested by Hu and Bentler (1999), the overall fit of the theoretical model to the data is acceptable (X 2 / df = 1.903,RMSEA = 0.055, GFI = 0.871, AGFI = 0.844, NFI = 0.915, RFI = 0.904, IFI = 0.958,  Note: The factor load was obtained by the maximum orthogonal rotation, and the common factor was extracted by the Principal Component Analysis (PCA).  Table 3, it can be deduced that the fit index of the theoretical model basically reached an excellent level, indicating that the theoretical model fitted well and was an acceptable model (Hu & Bentler, 1999).

Hypothesis test results
Seen from the results of hypothesis testing, distributive justice, procedural justice and interpersonal justice all had a significant positive impact on employee satisfaction and employee trust, so that the study hypotheses H1a, H1b, H2a, H2b, H3a, H3b were verified. Information justice posed significant impact on employee satisfaction, but posed less obvious impact on employee trust, so study hypothesis H4a was verified, and H4b was rejected. Employee satisfaction had significant impact on task performance and contextual performance, and employee trust has no significant impact on them, thus study hypotheses H5a and H5b were verified, and H6a and H6b were rejected.

Theoretical model correction
To better fit the data and the theoretical model, the theoretical model was modified according to the Modified Index (MI) (Kelloway, 1998)without violating the theoretical basis. The modification was carried out by gradually removing the path that did not pass the test and adding the suggested new path.
From the explained variance of each construct in the modified model in Figure 2, the explained variances of the four endogenous variables were above 60%, among which the explained equation of employee satisfaction reaches 80.2%, the employee trust and contextual performance reached 72.1%, and the task performance also reached 68.3%. This fully demonstrates that the path relationships in the theoretical model were very stable and well reflected the relationship between the variables. Note: The value below the diagonal is the correlation coefficient between the latent variables, and the value above the diagonal is the square root of the average extracted variance.

Discussion and conclusions
In the context of the hotel industry, based on the theory of organizational justice and the theory of relationship quality, this paper constructs an integrated model of the relationship between the justice perception, the relationship quality and the work performance of the hotel staff, and through the survey of the employees of the hotels in a very famous scenic city of Zhaoqing, China, the empirical test of the integrated model has been carried out, and the important conclusions are obtained.

Dimension of organizational justice
Organizational justice dimension structure is an important part of organizational justice research.
In the early studies, most of the studies used only single dimensional structures. Later, with the depth and refinement of the research, it was found that organizational justice was a multidimensional structure.
At present, the most common organizational justice of three-dimensional structure i.e. the distributive justice, procedural justice and interpersonal justice (Kim & Brymer, 2011;Weber, 2008). But Bies and Shapiro (1988) separated informational justice from interpersonal justice, which was accepted and used by some researchers (Liu et al., 2003;Xie et al., 2007).
In the study of organizational justice in the hotel industry, the three dimensions of organizational justice, for example, distributive result, procedural justice and interpersonal justice (Ha & Jang, 2009;Kim & Brymer, 2011;Nadiri & Tanova, 2010), are mostly involved in the study of organizational justice, and the informational justice is less concerned by the three dimensions. In this paper, with the expansion of the scale of the hotel enterprise, whether the related information of the hotel is fair and open is also very important. Therefore, the informational justice is regarded as a dimension of the justice of the hotel organization. Through exploratory factor analysis and confirmatory factor analysis, after a rigorous test of reliability and validity, the organizational justice of hotel industry includes four dimensions of distributive justice, procedural justice, interpersonal justice and informational justice. Therefore, in addition to achieving distributive justice, procedural justice and interpersonal justice, hotel managers should establish a smooth channel of information to acquire informational justice. The discovery of the dimension of informational justice enriches the research results of organizational justice in the current hotel industry context. Note: *** indicates that the path is significant at the level of 0.01, indicating that the path is not significant.

Relationship between organizational justice and relationship quality
Although the dimension of the relationship quality has not been agreed by the academic community, most researchers use the two dimensions of satisfaction and trust to measure (Huntley, 2006), which is used in this study. Organizational justice is the psychological needs and expectations of employees, and the realization of organizational justice is conducive to the establishment of good relations between hotel enterprises and employees. In the study of organizational justice, there are still differences between the dimensions of organizational justice and employee satisfaction (Folger & Konovsky, 1989;Sweeney & McFarlin, 1993).Folger and Konovsky's empirical results show that distributive justice explains employee satisfaction better than other dimensions (Folger & Konovsky, 1989). And Colquitt studies have shown that interpersonal justice has a stronger influence on employee satisfaction than distributive justice and procedural justice, because when distributive justice such as remuneration and promotion is perceived, employees will be satisfied with the remuneration and job (Colquitt, 2001). The study found that the each dimension of organizational justice has the great impact on employee satisfaction, and the procedural justice is the strongest, the interpersonal justice is the second, the results are fair third, the information justice is minimum. Thus, to improve employee satisfaction, hotel managers should pay attention to organizational justice, especially to achieve procedural justice and interpersonal justice.
In the relationship between organizational justice and employee trust, previous related studies have shown that organizational justice is beneficial to employees' sense of trust (Masterson et al., 2000). Hong and Lu's survey of Chinese hotel companies found that employee trust mediates in organizational justice and job performance (Hon & Lu, 2010). Previous studies have found that procedural justice has a stronger impact on employee trust than distributive justice. There are some differences between this study and previous ones. It is found that distributive justice plays the strongest impact on employee trust and interpersonal justice and procedural justice come to the second and third while informational justice produce no influence. The difference in industry may produce this difference. The impact of organizational justice on employee trust shows that hotel managers should pay special attention to the distributive justice and interpersonal justice in order to improve employee trust.

Relationship between relationship quality and job performance
In this study, it is found that employee satisfaction has a significant positive impact on task performance and contextual performance,which is in accordance with the conclusions of Kim and Brymer's findings (Kim& Brymer, 2011), indicating that the improvement of employee satisfaction is an important means to improve job performance. However, it is surprising that the impact of employee trust on task performance and contextual performance is not significant. This conclusion is inconsistent with existing research findings.The high mobility of the hotel staff may account for the difference and employee trust is not enough to affect the performance of the hotel. On the contrary, they pursue a kind of satisfaction more, so the job satisfaction has more significant impact on the job performance. On the whole, this conclusion is similar to the conclusions of Ramaswami and Singer's research findings on industrial workersthat the effect of employee satisfaction and trust on higher level is not significant (Ramaswami & Singh, 2013). The relationship between employee trust and job performance is less, so the relationship between employee trust and work performance remains to be further tested.

Relationship between organizational justice and job performance
Based on the theory of organizational justice and the relationship quality theory, this study believes that the relationship quality plays a complete intermediary role in the effect of organizational justice on job performance, but the result of the structural equation model shows that interpersonal justice has an indirect effect on contextual performance not only through employee satisfaction, but also has a significant and direct positive effect on contextual performance. Informational justice also produces a significant and direct positive impact on task performance, and therefore employee satisfaction plays a mediating role in the impact of distributive justice and procedural justice on job performance, while interpersonal justice and informational justice play a partial intermediary role in the impact of job performance. This finding is similar to that of Ramaswami and Singh (2013). Interpersonal justice is mainly concerned about the justice of interpersonal communication while contextual performance mainly examines the relationship between employees and organizations. Informational justice refers to the interpretation of the reasons for a certain procedure and the degree of distribution of outcomes while task performance is closely related to the specific content of the work, and also closely related to capability, proficiency, and work knowledge. As a result of significant correlation, informational justice directly affects task performance.

Research limitations and directions for future research
Based on organizational justice theory and relationship quality theory, this paper constructed an integrated model of organizational justice, relationship quality and job performance in the situation of hotel industry, and empirically tested the model, with some important conclusions obtained. However, limited by research funding and capabilities, there are still some research limitations that constitute important research directions in the future. First of all, this paper used the convenient sampling method to obtain the cross-section data. In the future research, the longitudinal data may be collected to test the integrated model. Secondly, the research literature on relationship quality showed that organizational commitment is an important dimension of relationship quality, and employee trust is a multi-dimensional construct (Morgan & Hunt, 1994). Therefore, future research may incorporate organizational commitment into the integrated model and conduct research on the employee trust by dimension division. Finally, some literature suggested that demographic characteristics of employees may have a regulatory effect on the relationships between variables in the model (Harris et al., 2009). The research in the future may use the demographic characteristics of employees as regulatory variables to analyze their impact on the relationships between variables in the model.