The Impact Mechanism of University Graduates’ Organisational Commitment on Turnover Intention in Early Career Stage: The Moderating Role of Career Resilience ()
1. Introduction
1.1. Background to the Research
The Report on the 20th National Congress of the CPC points out that “Employment is the most basic component of the people’s wellbeing”, and achieving fuller and higher quality employment is an important basis for promoting common prosperity for all the people. With the number of university graduates increasingly growing in recent years, the employment situation tends to be grim. In the Chinese context, the turnover of university graduates in their early career stage has become increasingly frequent. As the early career stage is defined as the stage within ten years after graduation (Xiao, 2010) , the data significantly presented that the mobility period of these younger knowledge workers is much higher than that average mobility period, especially the period of fresh graduates. This trend has a huge impact on individuals, organisations and society. According to research, university graduates’ turnover occurs frequently in the early career stage. Those university graduates in the early career stage continued to be the largest group of job hopping in China (Xu, 2022) . Influencing factors include overestimating their ability to work, dissatisfaction with their current situation in the organization, commitment to family and organizational commitment. As an ability to recover in the face of challenges (Collard et al., 1996) , there are few literatures to university graduate’s career resilience. According to the university graduate’s current situation of turnover intention in their early career stage and the challenging surroundings, it is of significance to explore their relationships.
1.2. Research Question and Objective
The research question is: how career resilience moderates the relationship between university graduates’ organisational commitment and turnover intention in their early career stage. The objectives of this research are:
1) To verify the relationship between university graduates’ organisational commitment and turnover intention in their early career stage;
2) To explore whether and how career resilience serves a moderating role in the relationship between university graduates’ organisational commitment and turnover intention in their early career stage;
3) To provide theoretical and practical implications in relation to the relevant findings.
In short, this study will present the relationship between university graduates’ organisational commitment and turnover intention in their early career stage and the moderating role of career resilience under the background of China. The focus of this study adds new value to turnover intention research as well as university graduates employment.
2. Literature Review and Hypotheses Development
2.1. Turnover Intention
Turnover refers to the cross-boundary movement of a member to another organization (Price, 1977) . Generally speaking, turnover is classified into voluntary and involuntary turnover, which have different impacts on the organizations (Wells and Peachey, 2011) . This study focuses on the intention based on voluntary turnover, since voluntary turnover tends to cause a series of loss to the current organization while individual further commitment to the next organization is also affected (Lee et al., 2004) . Turnover intention refers to an employee’s consideration to quit a job (Firth et al., 2004) . Since knowledge workers are more loyal to their career rather than the organizations they work for (Holland et al., 2002) , university graduates tend to have turnover intention and even actual turnover in pursuit of personal growth and additional value. As voluntary turnover has challenged many managers, it is of importance to reduce university graduates’ turnover intention in order to retain them.
2.2. Organisational Commitment
Organizational commitment refers to the degree to which an employee identifies and wishes to keep its membership in the particular organization and manage to achieve goals (Mowday et al., 1982) , which reflects that at least part of employees with strong organizational commitment tend to actively engage in the organizational activities. Organizational commitment is about the work-related attitude, which is quite crucial to the study of career or work behavior such as withdrawal (Greenhaus and Parasuraman, 1986) . It was proven that organizational commitment and job satisfaction, have a significant direct effect on turnover intention (Lantican, 2021) . There are varying classifications of organizational commitment dimensions in the relevant studies. Firstly, some organizational behavior scholars argue that organizational commitment should be single dimensional. Secondly, Meyer and Allen (1984) argue that organizational commitment should be two-dimensional, including continuance commitment and affective commitment. Thirdly, Wiener (1982) believes organizational commitment should include normative commitment, since employees feel obligated to organization according to their past education and help from organizations. Lastly, Ling et al. (2001) put forward a five-dimension organizational commitment consisting of affective commitment, normative commitment, ideal commitment, choice commitment and economic commitment, which is more adaptable to the Chinese context.
2.3. Early Career Stage and Organizational Commitment
Many well-known career stage models show that career development involves the process that an individual engages in professional activities from the beginning to the complete withdrawal (Dalton et al., 1977; Schein, 1978) . Specifically, two indicators should be considered in the career stage in terms of age and tenure (Cohen, 1991) . While age seems to be the most common indicator of career stages, some researchers prefer to choose tenure to develop the organizational commitment models (Cohen, 1991) . Mowday et al. (1982) employ the Reichers’s (1986) career stages to develop the organizational commitment in terms of early, mid and late career stages. Xiao (2010) argues that the early career stage is within 10 years after graduation, which is adopted and used in this study. Specifically, in the initial career stage, the individual should be a trainee with improved performance after training (Xiao, 2010) . Career stages, including early career stage are considered to be the moderator in the relationship between organizational commitment and outcomes (Cohen, 1991) . However, it is not valuable to use career stage to measure some outcomes such as turnover intention (Cohen, 1991) , since all the literatures tend to support the moderating effect. This study tends to explore the career resilience, since there are few existing researches exploring the career resilience together with early career stage and organizational commitment.
2.4. Career Resilience
Career resilience as an independent concept, is first used by London (1993) . Previous literature offers three different definitions of career resilience. According to one definition, career resilience is the ability to adapt to dynamic environment, which includes the positive attitude to job and organizational changes, the optimistic expectation to others, maintaining self-confidence as well as willingness of taking risks (London, 1993) . In terms of the second definition, career resilience is regarded as an ability to adapt to organizational need and recover in the face of difficulties (Collard et al., 1996) . In order to have effective career development, Collard et al. (1996) believe individual should be dedicated to acquiring personal excellence in accordance with organizational goals. As to the last definition, the ability to be a well combination of ability to take risks as well as recover is put forward, applying to the “portfolio thinking” and daily effort to deal with challenges in both work and life (Willis and Wilkie, 2009) . As career resilience tends to have a significant impact on career development, individuals with low resilience are likely to feel tough and have turnover intention.
2.5. Relationships between Organisational Commitment and Turnover Intention
A series of studies show that organizational commitment has a certain relationship with turnover intention. There is a clear negative correlation between organizational commitment and turnover intention in both overall and partial aspects. It should be acknowledged that the development of university graduates in organizations is actually a kind of equivalent or unequal exchange between their labor and their gains (Homans, 1951) . As university graduates are in their early career stage, they are likely to pay attention to their personal growth and development and focus on the career planning based on the high mobility characteristics. In this way, when the pay of university graduates is unequal exchange in their thoughts, university graduates tend to have turnover intention in this period. However, the organizational commitment as a binding, can not only stabilize the employee’s psychological situation, but also play a guiding role to the individual. It is found that perceived organizational commitment has inhibitory effect on the turnover intention of university graduates. In light of this, it is assumed that organizational commitment of university graduates has a negative impact on the turnover intention in their early career stage. The hypothesis is provided below.
H1: University graduates’ organizational commitment has a negative correlation to turnover intention in their early stage.
According to the organizational commitment scale of Ling et al. (2000) , organizational commitment is divided into five dimensions consisting of affective commitment, normative commitment, ideal commitment, economic commitment and opportunity commitment. According to the definition of these five dimensions, each dimension has negative influence on the turnover intention of university graduates in their early career stage. Hence, assumptions are made as follows.
H1a: In the early career stage of university graduate, affective commitment has a negative correlation to turnover intention.
H1b: In the early career stage of university graduate, normative commitment has a negative correlation to turnover intention.
H1c: In the early career stage of university graduate, ideal commitment has a negative correlation to turnover intention.
H1d: In the early career stage of university graduate, economic commitment has a negative correlation to turnover intention.
H1e: In the early career stage of university graduate, opportunity commitment has a negative correlation to turnover intention.
2.6. The Moderating Role of Career Resilience
As career resilience is the ability of employees in the face of the pressure, it is available to accept the situation and deal with the challenges. Career resilience plays a moderating role in this situation, which can reduce the turnover intention of employees and avoid actual turnover. Long (2012) states that there is a certain relationship between career resilience and organizational commitment. Career resilience tends to have the impact on moderating the relationship between the pressure and turnover intention. Additionally, university graduates’ career resilience can be higher than the average employees (Yin and Liu, 2013) . Therefore, high career resilience may lead to low turnover intention. Hence, assumptions are made as follows.
H2: In the early career stage of university graduate, career resilience plays a moderating role between the relationship of organizational commitment and turnover intention.
H2a: In the early career stage of university graduate, high career resilience plays a significant moderating role between the relationship of organizational commitment and turnover intention.
H2b: In the early career stage of university graduate, low career resilience plays an insignificant moderating role between the relationship of organizational commitment and turnover intention.
Figure 1 presents the research model that aims to test the relationships between organizational commitment and turnover intention and the moderation role of career resilience.
3. Methodology
3.1. Sampling and Procedure
Virtual snowball sampling using social media WeChat was adopted. The sample presumably focused on Shanghai, Hangzhou and Wenzhou in terms of large-, medium- and small-sized cities in China and university graduates within ten years after graduation were sampled. As a first-tier city in China, Shanghai is the representative of big cities. Hangzhou, as a new first-tier city, represents medium-sized cities, while Wenzhou represents smaller cities. Within the scope of our study, it is typical and representative to take these three cities as sampling sites. In the procedure of snowball sampling, the initial participants were selected by non-probability method which could meet the screening criteria of university graduates within ten years after graduation (Acharya et al., 2013) . A pilot study of 40 university graduates was invited to complete the questionnaire online to test the reliability and validity of the initial questionnaire and give some corresponding feedbacks including the comprehension and fluency of each item. The formal questionnaire was formed after the modifications and adjustments of items. This structured questionnaire was used for collecting data to test the hypotheses, consisting of four major sections. After a brief introduction explaining the purpose and anonymity of this study, the first three sections were presented including organisational commitment scale (Ling et al., 2000) , career resilience scale (Song, 2011) and turnover intention scale (Mobley et al., 1978) . The last section covered individual profile, consisting of gender, education background, age category, marital status, years after graduation, career, current organisation and turnover history. The formal questionnaire was formed in the Qualtrics and distributed online through WeChat. All the participants engaged in this survey voluntarily. 405 participants completed the questionnaire and 368 samples were valid with a response rate of 90.9%. In order to focus on the research of university graduates graduating within ten years, the participants who graduated ten years ago were omitted before data processing.
3.2. Measure, Reliability and Validity
Each response was captured on 5-point Likert scale ranging from 1 = strongly disagree to 5 = strongly agree. Each scale was chosen based on the literatures in the corresponding field. Reliability analyses of three scales were employed in SPSS 19.0 for each scale. Validity analysis was also conducted to ensure the measured results could reflect the expected content. Three variables were measured and explained below (Tables 1-3).
1) Organisational Commitment: Organisational commitment was an independent variable, which was measured by organisational commitment scale in this study (Ling et al., 2000) . Typically, a scale with a Cronbach’s Alpha above 0.6 could be regarded to be reliable (Pallant, 2005) . After analyzing the data, all the items of organisational commitment could be kept to conduct following analyses.
2) Career Resilience: The career resilience scale was employed from Song (2011) . And given individual characteristics, career characteristics and indigenous culture, this career resilience scale ultimately consisted of three dimensions of cognition, emotion and behaviour (Song, 2011) .
3) Turnover intention: Turnover intention of university graduates was assessed by three-item measure of Mobley et al. (1978) . This scale was quite widely used with good internal consistency coefficients of turnover intention (Yin-Fah et al., 2010) .
![]()
Table 1. Summary of organisatonal commitment scale reliability.
![]()
Table 2. Summary of career resilience scale reliability.
![]()
Table 3. Summary of turnover intention scale reliability.
Following the measure of reliability in relation to three scales, validity should also be considered. Validity related to questionnaire refers to the ability of questionnaire to measure the intended content (Saunders et al., 2012) . In light of this, the questionnaire tends to have high validity when the results reflect the required content (Bryman and Cramer, 1994) . Measuring content validity was quite important to ensure construct validity and support the instruments (Yaghmaei, 2003) . According to Burns and Grove (1993) , content validity could be gained from literatures, representativeness of respondents and experts. A series of previous researches also showed these three scales had high validity (Yin-Fah et al., 2010; Mu, 2007; Benson, 2006; Miller et al., 1979) . In addition, virtual snowball sampling was adopted. As discussed above, the online distribution method and the screening of respondents based on the scope of university graduates’ characteristics could improve the representativeness of sample (Baltar and Brunet, 2012) . Moreover, the ambiguous items were modified in the pilot study according to the feedback of respondents. Therefore, this survey had good validity to some extent.
4. Findings and Analysis
This chapter focuses on analysing the data and testing the relationship among three variables.
4.1. Profile of Respondents
In order to present a clear reflection of university graduates’ demographic distribution, the data in the final sample was processed. Invalid data had been omitted before reliability test. In this final sample, 119 respondents were male (32.3%) and 249 respondents were female (67.7%) (M = 1.68, SD = 0.468). In regarding to educational level (M = 2.02, SD = 0.481), the totals of 286 respondents (77.7%) were bachelor’s degree holder level, followed by 44 post-graduate’s degree holder (12.0%), 37 associate degree holder (10.1%) and 1 respondent who had doctor’s degree or above (0.3%). The age of 280 respondents were under 25 (76.1%), followed by 57 respondents aged 26 to 30 (15.5%) and 31 respondents aged 31 to 45 (8.4%) (M = 1.32, SD = 0.623). In terms of marital status (M = 1.40, SD = 0.700), 328 respondents were married (89.1%), 37 respondents were single (10.1%) and 3 respondents were divorced or separated (0.8%). The distribution of years after graduation in descending order was less than 2 years (268 respondents, 72.8%), 2 to 5 years (54 respondents, 14.7%) and 5 to 10 years (46 respondents, 12.5%) (M = 2.75, SD = 0.854). As to the current organisation, 212 respondents worked in enterprise (57.6%), 57 respondents worked in institution (15.5%), 45 respondents worked in the government and 54 respondents worked in other organisation (14.7%). In terms of historical turnover behavior (M = 1.61, SD = 0.488), 225 respondents did not change their jobs (61.1%) while 143 respondents had turnover experience (38.9%) (see Table 4). According to Wu and Tu (2014) , the data needed to be further processed to identify the outliers
only if the standard deviation was greater than the mean. As presented above, each mean of demographic factors is greater than standard deviation, which indicated that there was no need to further consider the extreme data (Table 4).
4.2. Principal Component Analysis
In this study, the data of 33 items in relation to three variables were processed to conduct KMO and Bartlett’s test. According to Kaiser and Rice (1974) , a KMO values should be greater than 0.6 and a significant Bartlett’s test should be less than 0.05 in order to continue the analysis. Field (2009) holds that KMO values of 0.5 - 0.7 are mediocre, 0.7 - 0.8 are good, 0.8 - 0.9 are great and 0.9+ are superb. The KMO value in this study was 0.890, which indicated great sampling adequacy (Table 5). Bartlett’s test of Sphericity was 5764.598 and 528 degrees of
freedom with a significance level of 0.000, which indicated that the data were suitable to conduct principal component analysis.
In order to explore the relationship between organisational commitment and turnover intention overall, a principal component analysis was applied to five dimensions of organisational commitment including affective commitment, normative commitment, ideal commitment, economic commitment, choice commitment. Griffith (1997) states that only eigenvalues greater than 1 are important. Two components emerged with eigenvalues greater than 1, accounting for 81.97% of the overall variance (Table 6). Although a small fraction of information contained five dimensions was lost, most of the information was retained, which indicated effective data compression and feature extraction (Jolliffe, 1986) . The weighted average value of component 1 and component 2 was formed as organisational commitment, which could be used in regression test.
4.3. Regression Analysis
In order to test the impact mechanism of five dimensions of organisational commitment on turnover intention, a multiple linear regression analysis was conducted with the control of gender, educational background, age, marital status, years after graduation, current organisation and historical turnover behavior. Based on the results in this multiple regression, the impact mechanism of five dimensions was at a significance level (p < 0.01) (Table 7). When each dimension in terms of affective commitment, normative commitment, ideal commitment, economic commitment and choice commitment increased by per unit, turnover intention correspondingly declined by 0.597, 0.533, 0.532, 0.403, 0.196 units respectively (Table 7). Affective commitment had a significant negative impact on turnover intention, which supported the hypothesis H1a. Normative commitment had a significant negative impact on turnover intention, which supported the hypothesis H1b. Ideal commitment had a significant negative impact on turnover intention, which supported the hypothesis H1c. Economic commitment had a significant negative impact on turnover intention, which supported the hypothesis H1d. Choice commitment had a significant negative impact on turnover intention, which supported the hypothesis H1e.
Based on the Based on the principal component of organisational commitment, the regression between organisational commitment and turnover intention was developed with the control of gender, educational background, age, marital status, years after graduation, current organisation and historical turnover behavior. The impact mechanism of organisational commitment on turnover intention was at a significance level (p < 0.01) (Table 8). Also, the regression
![]()
Table 6. Coefficients for organisational commitment of university graduates.
![]()
Table 7. Regression between different commitments and turnover intention.
Standard errors in parentheses. ***p < 0.01, **p < 0.05, *p < 0.1.
![]()
Table 8. Regression between organisational commitment and turnover intention.
Standard errors in parentheses. ***p < 0.01, **p < 0.05.
revealed 0.282 units of turnover intention’s reduction when organisational commitment increased by per unit (Table 8). The Beta (−0.282) should not be compared to previous regression results, since information loss existed and could not be avoided in data compression and feature extraction (Jolliffe, 1986) . Taking the significance level (p < 0.01) and Beta (−0.282) into account, organisational commitment still had a significant impact on turnover intention, which could sufficiently support the hypothesis H1.
4.4. Moderating Effect Analysis
Before exploring the moderating role of career resilience, the correlations among three variables were processed. Career resilience had a positive relationship with organisational commitment and had a negative relationship with turnover intention; career resilience had a negative relationship with turnover intention (Table 9). Therefore, further regression should be conducted in order to investigate the moderating role of career resilience in the relationship between organisational commitment and turnover intention.
In order to explore the moderating effect of career resilience in the relationship between organisational commitment and turnover intention, this study conducted the moderating effect test with control variables including gender, educational background, age, marital status, years after graduation, current organisation and historical turnover behavior. The career resilience as a moderator, organisational commitment as an independent variable, organisational commitment * career resilience as interaction terms emerged and turnover intention as a dependent variable emerged in Table 10. R-squared = 0.231 and p < 0.01 presented reasonable and fitting processed results. Based on the results of regression, the Beta of organisational commitment * career resilience was positive (0.161) while Beta of organisational commitment was negative (−0.942), which demonstrated that career resilience could moderate the relationship between organisational commitment and turnover intention, which supported the hypotheses H2. The higher career resilience was, the insignificant impact of organisational commitment on turnover intention was. In Figure 2, it was clearly presented that high career resilience plays an insignificant moderating role between the relationship of organisational commitment and turnover intention, which indicated the hypothesis H2a was not supported. Also, low career resilience played a significant moderating role between the relationship of organisational commitment and turnover intention, which indicated that the hypothesis H2b was not supported (Figure 2).
![]()
Figure 2. Moderator role of career resilience.
![]()
Table 9. Correlations between organisational commitment, career resilience and turnover intention.
Standard errors in parentheses. ***p < 0.01, **p < 0.05, *p < 0.1.
5. Discussion and Conclusion
Based on the findings and analyses in the previous chapter, this chapter discusses the results as well as literature reviews to highlight the research question and research purposes and then come up with the limitations, contributions and conclusions.
The results of this study show that university graduates’ organisational commitment including five dimensions of affective commitment, normative commitment, ideal commitment, economic commitment and choice commitment has a significant negative impact on turnover intention in their early stage. The results can confirm those literatures’ perspectives of the relationship between organizational commitment (Steel and Ovalle, 1984; Tett and Meyer, 1993; Lee et al., 2004) . According to Table 8, the impact mechanism of each commitment on turnover intention is different. To be specific, the affective commitment has a strongest negative impact on the turnover intention, which present that it is reasonable that organizational commitment used to be considered an affective attachment to the organization (Allen and Meyer, 1990) . The normative commitment and ideal commitment also have a strong impact on the turnover intention (Beta > 0.5). As all the respondents of this study are university graduates, the result in relation to ideal commitment confirms the university graduates’ concern about personal growth and ideal achievement (Benson, 2006) . As to the choice commitment refers to the attachment based on lack of choices and techniques Ling et al. (2001) , university graduates with unique skills and knowledge may not highlight it when deciding their turnover intention, which matches the result. Since the economic commitment mainly addresses the potential economic loss, university graduates do not so expect it. Thus, the economic commitment’s impact on turnover intention is not strong.
This finding is consistent with Eisenberger, Huntington, Hutchison, and Sowa’s (1986) argument that organizations wanting affectively committed employees must demonstrate their own commitment by providing a supportive work environment.
Not surprisingly, some meta-analytic researches hold that organizational commitment is closely associated with individual turnover as well as turnover intention (Steel and Ovalle, 1984) . As to the organizational commitment, the results demonstrate that career resilience moderates the relationship between organisational commitment and turnover intention. Long (2012) holds that career resilience has a relationship with organizational commitment, which is proved by the result of moderating effect. However, although career resilience has a positive relationship with organizational commitment, the previous sub-hypotheses H2a and H2b are not supported. High career resilience of university graduates in their early career stage has an insignificant moderating effect between the relationship of organisational commitment and turnover intention. Yin and Liu (2013) argue that university graduates focus on the career earlier in order to achieve long-term goal; this category of workers are likely to have high resilience leading to low turnover intention. According to the results, this assumption is unreasonable.
To rethink about the unsupported two sub-hypotheses, job embeddedness is adopted to compare. This study mainly focuses on the relationship between university graduates’ organisational commitment and turnover intention in their early career stage. Organisational commitment is a traditional predictor of turnover intention (Mobley, 1977; Price and Mueller, 1981) . Yet many researchers prefer to put forward different directions to studying turnover intention such as the new construct of job embeddedness (Mitchell et al., 2001) . Job embeddedness is increasingly and widely used to predict turnover intention and voluntary turnover because it is better predictive than other core variables such as job satisfaction, organisational commitment, job alternatives and job search (Mitchell et al., 2001; Crossley et al., 2007) . Job embeddedness can be divided into on- and off-the-job aspects. Off-the-job embeddedness was significantly predictive of subsequent voluntary turnover which shows the rising importance of off-job factors. To further research the turnover intention, new predictors should be taken into consider.
It should be acknowledged that there are several limitations in this study and some attempts to improve them. One limitation was snowball sampling. Snowball sampling is a non-probability sampling method (Fricker, 2008) . Biased results might exist because sampling units seem not to be independent and the projecting data beyond the range of sample was not justified (Acharya et al., 2013) . However, this study attempts to use snowball sampling online because expansion of virtual geographical scope can make it easier to reach particular populations (Baltar and Brunet, 2012) . In light of this, virtual networks enable this non-probability sampling to expand the sample size as well as increasing the representativeness to some extent (Baltar and Brunet, 2012) . The second limitation is the only focus on the impact of organisational commitment on turnover intention and the moderating role of career resilience. Some previous empirical studies hold that the increase of age, gender, employment history as well as length of service in current organisation will reduce the turnover intention of employees (Russ and McNeilly, 1995; Ghiselli et al., 2001; Yin-Fah et al., 2010) . This study attempts to control the demographic factors (e.g., age, gender, historical turnover behaviour) so these factors’ impacts on turnover intention of university graduates in their early career stage are not in-depth expounded and explored. As to the third limitation, this study focused on the early career stage of university graduates. This early career stage was defined by Xiao (2010) as period within ten years when individuals start their career. During the data analysis, university graduates with more than ten years career experience were screened out in order to meet the standard of early career stage. However, different researchers had different definitions of early career stage based on age or professional tenure (Reilly and Orsak, 1991) . Thus, potential bias might affect the results.
This study has theoretical contributions. This study originally explores the moderating role of career resilience in the relationship between organizational commitment and turnover intention. Specifically, previous studies on the university graduates’ turnover intention mainly tend to explore the relationships related to job satisfaction, perceived organizational support, oraganisational commitment and actual turnover (Shore and Martin, 1989) . This study adds value to turnover intention research, focusing on the moderating role of university graduates’ career resilience in the relationship between organizational commitment and turnover intention. This study also focuses on the perspective of university graduates’ turnover intention. In the context of China’s employment environment, this study provides a certain value for universities to carry out employment guidance work.
Acknowledgements
This paper is supported by Zhejiang Education Science Planning 2023 Common Prosperity Special Project (Project No: 2023GF053) and Zhejiang Soft Science Research Project (Project No: 2019C35013).