How to Reduce Employees’ Turnover Intention from the Psychological Perspective: A Mediated Moderation Model

Abstract

Background

The hospitality industry is deemed a great generator of global GDP and employment. However, high rates of voluntary turnover have gradually undermined global service organizations and brought huge losses to them. Nowadays, the hotel sector continues to be plagued by high turnover rates.

Purpose

A research model investigating job attitudes (job satisfaction and organizational commitment) as mediators of the impact of psychological capital (PsyCap) on turnover intention and also examining position as a moderator between job attitudes and turnover intention was proposed and tested.

Methods

This study collected data from 406 employees selected from four-star and five-star hotels in the southwest region of China. Online survey questionnaires and a purposive sampling technique were employed in this study. Structural equation modeling was utilized to evaluate the direct, mediating, and moderating effects.

Results

The results showed that organizational commitment and job satisfaction fully mediated the association between PsyCap and turnover intention. Moreover, position played a moderating role on the effect of the aforementioned two job attitudes on turnover intention.

Conclusion

The findings implied that hoteliers should focus on employees’ PsyCap and job attitudes in order to mitigate serious turnover issues in the hotel sector in China. Besides, the fact that position resulted in disparity impacts in the formation of turnover intention was evidenced.

Keywords:

• psychological capital
• organizational commitment
• job satisfaction
• turnover intention
• position
• four-star and five-star hotels in China

Introduction

The rapidly changing business world is confronting unprecedented challenges and business organizations are gradually beginning to recognize the significance of sustainability issues which will bring benefits to the global economy, the environment, and the organizations themselves.¹ From a business perspective, sustainability has been regarded as an organization’s ability to achieve its business targets, which may include lowering operating costs and risks, increasing their appeal to skilled talents and competitiveness. However, it is impossible to achieve these business goals without highly qualified employees.²

The hospitality industry is deemed as a primary source of foreign currency earnings, a generator of personal and corporate income, a creator of employment, and a contributor to government revenue.³ As the World Travel & Tourism Council (WTTC)⁴ reported, the direct contribution of the hospitality industry to Global GDP was 2,570 billion USD (accounting for 3.2% of global GDP) in 2017 and it rose by 4.0% to 2,674 billion USD in 2018. In addition, the hospitality industry’s direct contribution to global GDP is expected to grow to 3,890.0 billion USD (accounting for 3.6% of global GDP) by 2028 (see Figure 1). From the perspective of employment, WTTC⁴ data also showed that the hospitality industry generated 118,454,000 jobs directly (accounting for 3.8% of global employment) in 2017 and 121,356,000 jobs in 2018. By 2028, the hospitality industry will supply 150,139,000 jobs for the global labor market (see Figure 2).

Figure 1 Direct contribution of hospitality to global GDP. Source: Economic Impact of Global Travel & Tourism (2018).

Figure 2 Direct contribution of hospitality to global employment. Source: Economic Impact of Global Travel & Tourism (2018).

Despite the hospitality industry being significant in many countries, survey data has shown an endemic voluntary turnover rate in the hotel sector, which was estimated to range from 30–300%, and was far higher than other industries.⁵ For example, in China, data from Aon Hewitt⁶ indicated that between 2016 and 2017, the average turnover rate in all industries was only about 20%, yet, the voluntary turnover rate in hotels has reached approximately 40%, which was the highest in China.

All the above statistics show that the hospitality industry is indeed a great generator of global GDP and employment; on the other hand, high rates of voluntary turnover have gradually undermined global service organizations and brought huge losses to them.⁷ Nowadays, the hotel sector continues to be plagued by high turnover rates, that is, the causes of turnover have not been clearly and effectively explained yet in prior literature.⁸ Therefore, ensuring the stability of human resources in order to promote service quality and organizational performance is an important task for hotel managers. Generally, hotel management tend to focus on tangible motivation factors such as salaries, paid vacation, and working environment and neglect employees’ intangible demands. As Luthans et al⁹ highlighted, psychological capital (PsyCap) and attitudinal strengths have to be assessed for organizational success and sustainable development.

PsyCap, with self-efficacy, hope, optimism, and resilience as its main manifestations, is a recognized personality construct in the field of organizational behavior and management.¹⁰ According to COR theory, four components of PsyCap are individual resources that enable employees to adapt to challenging circumstances, maintain positive job attitudes in working environment, and stay with their organizations.¹¹ This is because employees with high levels of individual resources can be thought of having “resource caravans”, which can lead to positive outcomes.¹²

In the current study, researchers attempted to fill in various gaps, which could contribute to extant hospitality literature. Firstly, given the fact that employees prefer to concentrate on intrinsic motivators, such as the importance to the organization, recognition of their own achievements, and responsibility rather than improved financial returns,¹³ it is worthwhile to make an in-depth research of hotel employees’ turnover issue from the perspectives of psychology. Secondly, PsyCap has been recognized as a significant predictive determinant of job attitudes^14–16 and turnover intention.^10,17 However, surprisingly, empirical studies of the impact of PsyCap on turnover intention with mediating effects of organizational commitment and job satisfaction among frontline employees in the hotel sector, especially in China, are still lacking. Thirdly, although previous studies have proved that position (line-level employee and supervisor) is correlated to organizational commitment, job satisfaction, and turnover intention,^18,19 to our knowledge the moderating effects of position on the relationship between organizational commitment and turnover intention as well as between job satisfaction and turnover intention have not been examined before, especially in the hotel sector.

Literature Review and Research Hypotheses

PsyCap and Job Attitudes

The concept of PsyCap was originally put forward by Luthans and Youseff²⁰ and was extended into the area of human resource management and organizational behavior. PsyCap is regarded as a key psychological element beyond social capital and human capital and it comprises four critical dimensions, namely hope, optimism, self-efficacy, and resilience.

Ambrose and Schminke²¹ suggest that organizational commitment and job satisfaction as job attitudes have been trending topics for several decades in the hotel sector. Job satisfaction (JS) is a type of positive emotional state obtained by evaluating one’s own work itself or previous working experiences.²² On the other hand, organizational commitment (OC) is a firm belief in organizational objectives and values.²³

There has been adequate evidence regarding the components of PsyCap as predictors of job attitudes in existing literature on scoping the boundaries of hospitality management. To be specific, resilience is found to influence job satisfaction positively,^24,25 and self-efficacy can positively influence organizational commitment.^26,27 Besides, previous studies have shown that PsyCap can positively influence organizational commitment^23,28 or job satisfaction.^29,30 But surprisingly there are rare studies which have investigated the combined effect of four components of PsyCap on two critical job attitudes simultaneously, especially in the hotel industry. Thus, based on the content discussed above, the following hypotheses are proposed:

Hypothesis 1: PsyCap has a positive impact on job satisfaction.

Hypothesis 2: PsyCap has a positive impact on organizational commitment.

Job Attitudes and Turnover Intention

Allen and Meyer³¹ suggest that organizational commitment is an emotional bond and it tends to prompt employees’ willingness to remain with their organizations. Commitment usually results in positive outcomes such as higher organizational performance, reduced absenteeism, and turnover.^32–34 In addition, employees who possess high level of organizational commitment often express their affiliation with their organizations.³⁵ However, some other scholars believe that only normative and affective commitments have a significant effect on turnover intention.^36–38 Different findings have been obtained between the aforementioned variables in the context of hospitality.

Job satisfaction is conceptualized as a pleasant or positive emotional state derived from an assessment of one’s job or work experiences.²² When an employee fits well into the organization, job satisfaction is usually reflected.³⁹ Many prior studies have shown a positive correlation between job satisfaction and reduced turnover.^40,41

According to Joo and Park,⁴² job satisfaction has been proved to be a strong predictive antecedent of turnover intention in their research, which also examines organizational commitment as an antecedent. Similarly, Chan et al⁴³ propose that organizational commitment is negatively associated with turnover intention. Accordingly, in order to clarify the correlation between job attitudes and turnover intention, we put forward the following two hypotheses:

Hypothesis 3: Job satisfaction has a negative impact on turnover intention.

Hypothesis 4: Organizational commitment has a negative impact on turnover intention.

Psychological Capital and Turnover Intention

PsyCap is deemed as a personal characteristic resource which can reduce turnover intentions.⁴⁴ Evidence, although limited, supports such relationships. Karatepe and Karadas¹⁰ have proved that PsyCap can negatively affect turnover intention. What’s more, Avey et al’s¹⁵ meta-analytic study also shows that PsyCap is negatively correlated with turnover intention (k=5, corrected r=−0.32 and SD=0.11). Echoing the results of the above studies, there is also sufficient evidence regarding the components of PsyCap as predictors of turnover intention in the hospitality literature.^45–47 However, Abbas et al⁴⁸ found that PsyCap could positively influence turnover intention. They explained that employees with high PsyCap, being more confident and skillful, may choose to seek a better working environment and begin to look for new options. So far, no consensus has been reached. Thus, hypothesis 5 is put forward:

Hypothesis 5: PsyCap has a negative impact on turnover intention.

Psychological Capital, Job Attitudes, and Turnover Intention

PsyCap is defined as an individual resource which increases employees’ awareness of their own resources and job-related outcomes.⁴⁹ Prior studies have shown that PsyCap can positively affect employees’ job attitudes, including job satisfaction^50–52 and organizational commitment,^29,53,54 as well as job-related variables such as organizational citizenship behavior^55,56 and turnover intention.^17,57 Therefore, previous empirical results demonstrate that PsyCap can influence individual satisfaction with work as well as one’s commitment to it. However, the relationship between PsyCap, attitudes and turnover intention among frontline employees must be specifically studied. Accordingly, hypotheses 6 and 7 are proposed:

Hypothesis 6: Job satisfaction mediates the relationship between PsyCap and turnover intention.

Hypothesis 7: Organizational commitment mediates the relationship between PsyCap and turnover intention.

Position, Organizational Commitment, Job Satisfaction, and Turnover Intention

Although hotel managerial staff such as supervisors and department managers are thought to present lower turnover rates than line-level employees, much research have shown that the differences in turnover rates between the two groups of employees were not quite as big (line-level employees: 50.74% vs managerial employees: 39.19%).⁵⁸ Regarding the difference in turnover rates between supervisors and line-level employees, Carbery et al³⁶ attribute supervisors’ lower turnover rates to their higher satisfaction with and stronger commitment to hotels (compared to line-level employees). Given that the majority of line-level employees regard their current jobs as stepping stones for a future career, they are less likely to be highly committed to and satisfied with their organizations.¹⁸ The entry-level tasks in hotels inevitably involve large amounts of repetitive work which possesses less professional skills and diversity (compared to managerial tasks) and therefore offer limited opportunities for career growth among line-level employees. In contrast, supervisors tend to engage in their work to pursue career targets, since they have more promotion opportunities than line-level employees. In addition, line-level employees receive relatively lower incomes, and less fringe benefits and paid vacation than supervisors.⁵⁹ Hence, line-level employees are more likely to leave their jobs than supervisors, and the following hypotheses are proposed:

Hypothesis 8: Position moderates the relationship between job satisfaction and turnover intention.

Hypothesis 9: Position moderates the relationship between organizational commitment and turnover intention.

Method

Participants and Procedure

A quantitative survey with research objectives to examine the postulated hypotheses was planned to be conducted after a thorough literature review. We collected data by conducting a survey of full-time frontline employees (line-level employees and supervisors) in several divisions (eg, front desk, food and beverage, and guest relations) at four four-star and four five-star hotels in China. The sampling locations were in Chengdu and Chongqing, because Chengdu is an important provincial capital city and Chongqing a municipality of China. Both of them are deemed as the most popular tourist cities in southwest China, with a booming hotel sector.

The questionnaire is originally written in English and, in order to avoid errors and ensure validity, a linguist who is good at both Chinese and English is in charge of translating and back-translating questionnaire. Prior to distribution, we carried out a pilot study with test–re-test method for the reliability of the questionnaire,⁶⁰ and the content validity was examined by three professional practitioners to ensure appropriate items which were suitable to the context of the research.⁶¹ Online survey questionnaires and purposive sampling technique were employed in this study. After obtaining hotels’ approval, the hyperlink of the survey website was delivered to every potential respondent and they could complete questionnaires during their latest shifts. All of the respondents were told that participation was voluntary, but they were encouraged to complete the questionnaire with reference to their experiences. In addition, the survey included an instruction emphasizing that all responses were completely confidential and anonymous to minimize any possible bias.

With the utilization of AMOS 24.0 software, structural equation modeling (SEM) can examine the hypothesized relationships with 406 hospitality full-time line-level employees and supervisors. Three approaches were discussed in order to confirm the above sample size. Firstly, the rule of thumb for sample size, especially using Structural Equation Modeling (SEM), was considered, that is, the sample is normally larger than 200.⁶² Secondly, Cochran⁶³ formula of sample size for categorical and continuous variables was considered. Thirdly, subject-to-item ratio was also taken into account. A general rule is that there should be at least five subjects for each item (5:1) and ratios higher than this are generally better.⁶⁴

Measurement Tools

Four existing tools were employed in this study. Firstly, a 24-item PsyCap questionnaire (PCQ) measure proposed by Luthans et al⁶⁵ was applied to evaluate PsyCap. This scale consists of four dimensions and each dimension includes six items. Before the utilization of PCQ, permission was obtained from the author. Secondly, in terms of turnover intention, a three-item scale developed by Singh et al (1996) was utilized. This scale has been used in many hotel empirical studies.^10,66,67 Thirdly, the tool of organizational commitment consists of three dimensions and 18 items (three negatively worded items).³¹ The validity and reliability of this scale has been verified iteratively by many hospitality related studies.^68,69 Lastly, the measurement of job satisfaction was derived from Weiss et al's⁷⁰ Minnesota studies on Vocational Rehabilitation. This valid questionnaire has been widely used in hotel-related research.^71,72

Results

Sample Characteristics

The basic demographic characteristics of the participants are as follows. Among the respondents, 187 (46.1%) were male, and 219 (53.9%) were female. Regarding marital status, 240 (59.2%) respondents reported that they were single, and 166 (40.8%) were married. In terms of age, 61.3% were 30 years old and below and 31.6% were 31–40, while the age group of 41 and above contributed 7.1% to the study. Concerning the highest level of education, 2.1% of the respondents achieved a master’s degree and above, and 29.3% attained a bachelor’s degree, while respondents with a college degree and lower accounted for 68.6%. The majority of respondents (73.2%) were line-level employees, while others were supervisors (26.8%).

Full Measurement Model Validation

In order to evaluate the fit of full measurement model, goodness-of-fit measures were adopted. Bentler and Bonett⁷³ argue that the model will be acceptable if the chi-square is not significant. However, a number of scholars have disregarded this theory because chi-square is often reported as significant, mainly due to the sample size limitations and its sensitivity to the likelihood test ratio. In view of chi-square’s limitations, the CMIN/DF fit (X² divided by the degrees of freedom) becomes a more appropriate fit index. If the value of CMIN/DF is less than 3, an acceptable fit will be achieved.⁷⁴ The model in the current study showed an acceptable CMIN/DF of 1.172. In the light of other fit indexes, the model yielded a RMSEA of 0.021 (below the acceptable threshold of 0.070),⁷⁵ a CFI score 0.979, and a TLI score 0.978, which are within the ideal ranges (>0.90) for acceptable fit.⁷⁶

The full measurement model presented a satisfactory fit with the data. To be specific, all existing standardized factor loadings ranged from 0.641–0.871, meaning that all items could measure their corresponding constructs effectively. In addition, with SPSS 24.0, constructs’ Cronbach alpha coefficients range in value from 0.773–0.942. If all the Cronbach alpha coefficients are larger than 0.70, items will reflect corresponding constructs well.⁷⁷ With regard to composite reliability values for all the constructs, all were larger than the minimum threshold of 0.7.⁷⁸

Many measurements were performed to evaluate the construct validity of the proposed model. The construct validity comprises convergent and discriminant validity. Convergent validity refers to the proportion of shared variance by items of a specific construct and it can be satisfied based on average variance extracted (AVE) larger than 0.50.⁷⁹ In this study, the latent variables’ AVE values were larger than the threshold value of 0.50, which meant convergent validity for all constructs were achieved.

Discriminant validity was assessed with the following approach. If values of AVE are larger than the corresponding squared correlation coefficients, the discriminant validity will be achieved.⁸⁰ As is shown in Table 1, all constructs’ squared correlation coefficients are smaller than corresponding AVE values, which strongly proved that discriminant validity for all constructs were achieved.

Table 1 The Results of Discriminant Validity Test

Variables	AFV	COV	NOV	SEV	HOV	OPV	REV	EXV	INV	TIV
AFV	0.779
COV	0.588**	0.714
NOV	0.522**	0.516**	0.752
SEV	0.225**	0.304**	0.220**	0.747
HOV	0.219**	0.227**	0.172**	0.498**	0.760
OPV	0.190**	0.200**	0.168**	0.557**	0.550**	0.711
REV	0.167**	0.146**	0.076	0.512**	0.522**	0.505**	0.749
EXV	0.247**	0.276**	0.205**	0.344**	0.307**	0.320**	0.245**	0.762
INV	0.282**	0.294**	0.213**	0.359**	0.287**	0.278**	0.238**	0.582**	0.765
TIV	−.499**	−.539**	−.471**	−.423**	−.417**	−.339**	−.273**	−.506**	−.465**	0.792

Notes: **P<0.01; The square root of AVE is shown in italics.

Abbreviations: AFV, affective commitment; NOV, normative commitment average; COV, continuance commitment average; SEV, self-efficacy average; HOV, hope average; OPV, optimism average; REV, resilience average; EXV, extrinsic satisfaction average; INV, intrinsic satisfaction average; TIV, turnover intention average.

Common Method Bias

In the current study, two recommended approaches have been utilized to reduce common method bias. Firstly, respondents all emphasized the survey’s confidentiality and anonymity to avoid social desirability or respondent acquiescence.⁸¹ In addition, researchers also asked respondents to answer each question as honestly as they could because there was no right or wrong answers. Secondly, we utilized Harman’s one-factor test, which consists of a factor analysis of all the items. Finally, the first factor accounted for only 21.413% of the variance in the unrotated factor analysis, which indicated common method bias was not a serious issue in the current study.⁸²

Structural Model Estimation and Hypotheses Testing

As the reliability and validity of full measurement model had been verified, the structural model was going to be evaluated and the hypotheses were going to be tested in the next step. After statistical analysis, fit indices from the following structural model (Figure 3) demonstrated that the proposed structural model was satisfactory. In addition, in order to reduce the complexity of structural model, item parceling technique was used for three second-order attitudinal constructs.

Figure 3 Proposed structural model.

Abbreviations: PsyCap, psychological capital; OC, organizational commitment; JS, job satisfaction; TI, turnover intention; df, degree of freedom; GFI, goodness-of-fit index; AGFI, adjusted goodness-of-fit index; CFI, comparative fit index; RMSEA, root mean square error of approximation.

Causal Effects and Hypotheses Testing

As shown in Table 2, there existed a positive correlation between PsyCap and job satisfaction (β=0.561, P<0.001), which supported Hypothesis 1. Besides, there was also a positive association between PsyCap and organizational commitment (β=0.401, P<0.001), which supported Hypothesis 2. In addition, both job satisfaction (β=−0.426, P<0.001) and organizational commitment (β=−0.557, P<0.001) had a negative impact on turnover intention, which supported Hypotheses 3 and 4. However, PsyCap (β=−0.123, P=0.054) had no direct effect on turnover intention due to a P-value larger than 0.05. Hence, Hypothesis 5 was not supported. That is to say, organizational commitment and job satisfaction played a full mediating role in the relationship between PsyCap and turnover intention.

Table 2 Summary of Hypotheses (Causal Effects) and Results

Hypotheses	Std. Path Coefficient	T-value	P-value	Results
H1: PsyCap→ (+) JS H2: PsyCap→ (+) OC H3: JS→ (-) TI H4: OC→ (-) TI H5: PsyCap→ (-) TI	0.561	7.637	***	Accepted
	0.401	6.139	***	Accepted
	−0.426	−6.440	***	Accepted
	−0.557	−9.388	***	Accepted
	−0.123	−1.929	0.054	Not accepted

Note: ***P<0.001.

Abbreviations: PsyCap, psychological capital; OC, organizational commitment; JS, job satisfaction; TI, turnover intention.

Mediating Effects and Hypotheses Testing

As can be seen in Table 3, to identify the mediating role of job satisfaction and organizational commitment, bias-corrected and percentile bootstrapping methods were utilized.⁸³ If zero is not included in bootstrapped CI, it means that the mediating effect will not equal zero. In this research, 95% bias-corrected CI and Percentile CI were estimated with 2,000 bootstrapped samples. Because the P-value of direct effect is larger than 0.05, the direct effect of PsyCap on turnover intention is not significant and accordingly a partial mediation effect is excluded. To summarize, both bootstrapping approaches showed that organizational commitment and job satisfaction could significantly and fully mediate the effect of PsyCap on turnover intention. Therefore, Hypotheses 6 and 7 were also supported.

Table 3 Summary of Hypotheses (Mediating Effects) and Results

Effects	Path	Estimate	Bias-Corrected 95% CI		Percentile 95% CI		P-value
			Lower	Upper	Lower	Upper
Direct Effect	PsyCap –>TI	−0.123	−0.256	0.069	−0.254	0.075	0.165
Indirect Effects	PsyCap–>JS –>TI PsyCap–>OC–>TI	−0.239 -0.224	−.439 -0.337	−0.141 -0.138	−.410 -0.334	−0.132 -0.133	*** 0.001
Total Effect	PsyCap–>OC/JS–>TI	−0.586	−00.705	−0.458	−00.699	−0.454	0.001

Note: ***P<0 0.001.

Abbreviations: PsyCap, psychological capital; OC, organizational commitment; JS, job satisfaction; TI, turnover intention; CI, confidence interval.

Moderating Effects and Hypotheses Testing

The multi-group structural equation modeling within AMOS 24.0 software was utilized to investigate the differences in the strengths of the structural relationships by evaluating the moderating variable effects.^84,85 The aim of multi-group analysis is to identify whether the path coefficients for the associations between job attitudes and turnover intention were equal in both line-level employee and supervisor groups. For the moderation tests, the data was divided into two subgroups based on position, respectively, 297 line-level employees and 109 supervisors. A two-step approach was used for multi-group comparison test. Firstly, the appropriate structural parameters are constrained to be equal across groups to generate an estimated covariance matrix for each group and an overall Χ² value for the sets of sub-models as part of a single structural system. Secondly, a second Χ² value with fewer degrees of freedom is obtained by removing the parameter equality constraints. And then, the moderating effects are examined by evaluating whether there are significant differences between the two Χ² values. If the change in the Χ² value is statistically significant, the null hypothesis of parameter invariance is rejected and a moderating effect is indicated.⁸⁶ As is shown in Table 4, ∆Χ²/∆df=4.064 and P<0.05, which indicates that significant differences were found between employee and supervisor models. Therefore, position is a moderator for the aforementioned relationships, which supported Hypotheses 8 and 9.

Table 4 Chi-Square Value and Degree of Freedom for the Constrained and Unconstrained Models

Model	Χ^2	df	∆Χ^2	∆ df	P-value
Constrained	209.762	148	4.064	1	0.044
Unconstrained	205.698	147

Abbreviations: Χ², Chi-square; df, degree of freedom.

Furthermore, Table 5 shows the results of a multi-group comparison test between line-level employees and supervisors with regard to job satisfaction, organizational commitment, and turnover intention. As we can see, for line-level employees, both organizational commitment (β=−0.547, P<0.001) and job satisfaction (β=−0.426, P<0.001) have significantly negative impacts on turnover intention. On the other hand, for supervisors, organizational commitment (β=−0.669, P<0.001) and job satisfaction (β=−0.441, P<0.05) can also significantly and negatively influence turnover intention.

Table 5 Results of the Multi-Group Comparison Test

Path	Line-Level Employees			Supervisors
	Estimate	SE	CR	Estimate	SE	CR
Organizational Commitment → Turnover Intention	−0.547***	0.068	−7.934	−0.669***	0.192	−4.724
Job Satisfaction → Turnover Intention	−0.426***	0.089	−5.801	−0.441*	0.291	−2.388

Notes: ***P<0.001; *P<0.05.

Abbreviations: SE, standard error; CR, critical ratio.

Discussions

Scholars have investigated the association of PsyCap and job satisfaction^50,87 and the relationship between PsyCap and organizational commitment,^54,88 however, scarce studies have examined the predictive effect of PsyCap on the aforementioned two job attitudes simultaneously.²⁹ In accordance with the above findings, our study also revealed significant positive associations between PsyCap and job satisfaction (β=0.561, P<0.001) and organizational commitment (β=0.401, P<0.001).

In line with Siu et al,⁵⁷ the results revealed that job satisfaction and organizational commitment could mediate the correlation between PsyCap and turnover intention. Furthermore, in contrast with many hotel sector studies,^10,89 our findings could not support that there existed a direct negative association between PsyCap and turnover intention (β=−0.123, P=0.165). Hence, it can be seen, although PsyCap has been highlighted before as a component of reducing turnover intention,^17,90 job attitudes have a more fundamental function as a cause and antecedent of turnover intention.

In addition, the results showed the indirect effect through job satisfaction (β=-0.239, P<0.001) was a little bit higher than organizational commitment (β=-0.224, P=0.001). Thus, this research offers significant and interesting findings about linking the role of job attitudes as they associate with PsyCap and turnover intention.

Another vital point produced by current research was the role of position as a moderator, and position could lead to significant difference in the association of job attitudes and turnover intention. To be specific, the negative effects of both organizational commitment and job satisfaction on turnover intention were higher for supervisors as compared to line-level employees. This finding evidenced the fact that position resulted disparity impacts in the formation of turnover intention.⁹¹ Therefore, current research breaks new ground by proposing a totally new conceptual model and conducting a rigorous empirical study which linked psychological capital, job attitudes, turnover intention, and position in the hotel sector in China.

Theoretical Implications

From a theoretical point of view, this study contributes to the theoretical field of hotel employees’ turnover intention. Scholars who aim to study hotel employees’ turnover intention can benefit from this study through a replication in different context or industry that they focus on. With theoretical directions, the results of this study will help to demonstrate how to reduce employees’ turnover intention from a psychological perspective in the hotel sector in China.

Principally, this study seeks to make an original contribution to the body of knowledge by investigating psychological determinants in affecting hotel employees’ turnover intention. In addition, the present study also extends COR theory by integrating job satisfaction and organizational commitment as mediating variables to provide better clarity on the interactions between all related variables and to explain the phenomenon of turnover intention.

Practical Implications

The results of the current research have implied that hotel management should adopt relevant strategies to develop employees’ psychological resources in order to advance their positive organizational attitudes and behaviors. In the first place, hotel management is supposed to realize the significance of PsyCap and proactively develop relating training programs to maintain their staff’s PsyCap at a high level.⁹² Through training, hotel staff can learn how to protect, maintain, and accumulate their PsyCap more effectively. Besides, more comprehensive training programs which cover PsyCap and positive job attitudes will be more effective. Therefore, line-level employees and supervisors in hotels who have received such trainings can master how to avoid the loss of resources. If they are able to maintain a high level of psychological resources, they will be highly satisfied with and more committed to their jobs. Simultaneously, they also feel a strong sense of belonging with their current jobs and have less intention to withdraw from their organizations.⁹³

Furthermore, we have chosen employees who work in four-star and five-star hotels as respondents because such hotels are the pioneers in human resource management and often make their efforts visible compared to other hotels with lower star ratings.⁹⁴ More importantly, four-star and five-star hotels of the same brand have a relatively uniform management system and standards. That is to say, the management problems existing in China (such as high turnover rate) may also exist in other regions or countries. Therefore, the conclusions of this research have great reference value for hotels of the same brand in other regions or countries.

Despite this research making some significant contributions to existing knowledge in human resource management in the hotel industry, we must acknowledge that there are still some specific limitations which can put forward feasible prospects for future studies.

Limitation and Direction for Future Studies

Firstly, since the process of data collection is cross-sectional, two recommended approaches have been carefully implemented to control potential common method bias (a) the design of the study’s procedures, and (b) statistical checks. Nevertheless, the data may still be at risk of common method bias. Hence, in subsequent research, the measurement of predictor and criterion variables should be separated by time to make a firm relationship of employee attitudes and behaviors.⁸¹

Secondly, only two cities were selected, which may limit the generalizability and representativeness. Future research can remove this limitation by sampling employees in various hotel settings in other regions or countries, and also a larger sample size is recommended.

Thirdly, this research has used an important individual psychological determinant called psychological capital. Future empirical studies should try to investigate more employees’ individual psychological determinants that can affect job attitudes, job performance, and even withdrawal behaviors in the hotel sector.

Finally, this study also reveals how the relationships between work-related variables vary across employees’ positions. In later research, more demographic factors such as gender, age, and tenure can be taken into account, so that hotel managers are able to derive a more in-depth understanding of employee’ turnover issues through the combinative role of individual differences.

Conclusions

This research has investigated organizational commitment and job satisfaction as mediators of the relationship between PsyCap and turnover intention in the hotel sector in China. As hypothesized, job attitudes were verified as significant and positive outcomes of PsyCap and also significant and negative antecedents of turnover intention. The role of job attitudes in this model could be explained via Conversation of Resource Theory.⁹⁵ Furthermore, this study has also examined the moderating effect of position on the association between job attitudes and turnover intention. The results imply that hotel line-level employees and supervisors who possess a high level of PsyCap and positive job attitudes have a tendency to stay in their organizations¹⁵ and position can generate a significant difference in the relationship between job attitudes and turnover.⁹¹ The findings demonstrate that training programs which include PsyCap and positive job attitudes should be widely promoted. As a result, line-level employees and supervisors in hotels who have received such trainings are able to avoid the loss of resources as much as possible. If hotel employees can maintain a high level of psychological resources, they will be highly satisfied with and more committed to their organizations. Besides, they also feel a strong sense of belonging and identity and have less intention to quit.

Disclosure

The authors report no conflicts of interest in this work.