Commitment refers to the attachment between workers and their careers and the relationship to a particular organization (Meyer et al., 1993). It should be noted that regarding commitment literature, several terms, including occupation (Young et al., 2023; Collie et al., 2020; Lin et al., 2021), profession (Ilyas et al., 2023), and career (Chen and Zhang, 2023; Son and Kim, 2021), have been used somewhat interchangeably. This study would like to explore and assess commitment to a particular work field—early care education (ECE). Therefore, the term occupation is more appropriate, and occupational commitment will be discussed in this research. Lee et al. (2000) defined occupational commitment as the psychological link between an individual and their occupation based on an affective reaction to that occupation. Thus, we can imagine that workers or employees with higher occupational commitment have positive and affective feelings and active pursuit for their professional development (Liu et al., 2021), which could be regarded as a crucial motivation for individuals’ positive organizational behavior (Bakker and Schaufeli, 2008) and superb work performance (Meyer et al., 1989). Prior studies have suggested that occupational commitment is not only associated with a variety of positive outcomes, including low rates of absenteeism and higher work engagement (Freund, 2005) but is also inversely linked with intentions to leave a profession (Antony et al., 2024; Hackett et al., 2001). Additionally, the study of Hackett et al. (2001) also demonstrated that intention to leave the occupation is empirically distinguishable from the level of occupational commitment; theoretically, the two share a causal relationship, with commitment preceding quitting intention. Moreover, occupational commitment is also reported to be a crucial factor in decreasing actual turnover in existing studies (Antony et al., 2024; Callado et al., 2023; Lee et al., 2000; Meyer et al., 2002).

Prior studies and research literature reflect little consensus about definitions of resilience, with substantial variations in operationalization and measurement of critical constructs (Luthar et al., 2000; Bonanno et al., 2015; Kalisch et al., 2015). Psychological resilience may be viewed as a measure of stress-coping ability and, as such, could be an essential target of treatment in anxiety, depression, and stress reactions (Connor and Davidson, 2003), which represents one such variable that has been found to reduce feelings of burnout (Johnson and Brown, 2013; Richards et al., 2016). It is measured in terms of both emotion regulation and coping strategies (Compas et al., 2017; Gross, 2015; Lazarus, 2000). Some scholars argue that psychological resilience is domain-specific, which may manifest differently in different domains (Todt et al., 2018). Therefore, it is vital to understand how psychological resilience manifests in teachers. Fortunately, understanding teachers’ life and career decisions from the perspective of psychological resilience has attracted wide attention (Lee and Sung, 2023; Zhi and Derakhshan, 2024; Chen, 2024; Brunetti, 2006; Castro et al., 2010; Gu and Day, 2007; Howard and Johnson, 2004; Kirk and Wall, 2010; Oswald et al., 2003). Additionally, more scholars and researchers are paying attention to teachers’ psychological resilience to understand teachers’ job satisfaction (Wang et al., 2024; Han, 2022; Brunetti, 2006; Kitching et al., 2009), teacher burnout and stress (Begum et al., 2024; Colomeischi and Ursu, 2024; Liu et al., 2021; Howard and Johnson, 2004), and career decision-making (Lee and Sung, 2023; Gu and Day, 2007). Several studies also suggested that psychological resilience may be the key to understanding teachers’ turnover, given identical and often dissatisfying and stressful working conditions (Liu et al., 2021; Beltman et al., 2011) For teachers, psychological resilience could explain coping with the less satisfying aspects of their work-life to become more committed to teaching over the long term (Beltman et al., 2011; Hong, 2012; Mansfield et al., 2012; Tait, 2008).

Job satisfaction, simply put, refers to what people think about their jobs (Spector, 2022) or an attitude toward the job resulting from a cognitive evaluation of specific job aspects (Spector, 2022; Weiss, 2002). Across a variety of occupations, job satisfaction has been found to have an impact on a variety of outcomes for it is not only crucial for employees’ wellbeing and productivity but also plays as a pivotal role in the health of the employer-employer relationship (Jogi et al., 2025; Redondo et al., 2021; Hsiao et al., 2020). The link between job satisfaction and turnover gets the most attention. Many researchers have theorized that job satisfaction is a crucial antecedent of employee turnover (Jogi et al., 2025; Mobley et al., 1979; Williams and Hazer, 1986), and job satisfaction is found to be negatively associated with turnover intention (Chiho, 2022; Redondo et al., 2021; Hsiao et al., 2020; Hom and Kinicki, 2001; Lee et al., 1996; Michaels and Spector, 1982; Mobley et al., 1979). That is, when employees are dissatisfied with their work, they will have lower work motivation, which will decrease the organization’s overall performance. Moreover, higher job satisfaction can positively sustain great work productivity (Akosile and Ekemen, 2022) and push them to make new accomplishments (Mardanov, 2021). Conversely, a lack of satisfaction can lead to opposite outcomes, such as work stress and burnout. Additionally, Ostroff (1992) found that organizations with more satisfied employees tend to be more effective than those with dissatisfied employees. In the area of education, previous studies have found a negative correlation between teacher turnover and job satisfaction (Shah and Jumani, 2015). For teachers, job satisfaction predicts retention and performance of teachers (Chen et al., 2020) and higher job satisfaction could reduce the turnover (Ren et al., 2024). In this regard, it is vital to consider the role of job satisfaction in the teaching profession (Wartenberg et al., 2023), which could link to the different outcomes such as turnover intentions ((Li and Yao, 2022; Madigan and Kim, 2021).

Burnout is a set of symptoms or physical conditional of emotional exhaustion, cynicism, depersonalization, and personal accomplishment that occurs frequently among individuals when they feel they can no longer give of themselves at a psychological level in response to working stress (Santi et al., 2020; Maslach and Jackson, 1981). Burnout is categorized into three dimensions: emotional exhaustion, depersonalization, and reduced personal accomplishment (Maslach and Jackson, 1981; Schaufeli et al., 2009). Burnout can mean unnecessary and demanding tasks for employees (Wazir and Jan, 2020). Maslach and Leiter (2016) have pointed out that burnout has been recognized as an occupational hazard in various human-centered occupations, such as human services, education, and health care. According to prior research, job burnout has a strong positive relationship with turnover intention (Srivastava and Agrawal, 2020) whereas it has a negative relationship with job satisfaction (Skaalvik and Skaalvik, 2020; Wong and Spence Laschinger, 2015; Lu and Gursoy, 2013) and occupational commitment (Kresna and Putra, 2020). In the context of education, teaching is an emotionally draining and stressful profession (Kyriacou, 2001), and burnout appears to be a relatively common experience in education occupations (Parmar et al., 2022). Most studies and researchers also regard burnout as a state of emotional, physical, and attitudinal exhaustion that may develop in teachers who have been unsuccessful in coping effectively with stress over a long period (Guglielmi and Tatrow, 1998), which will lead to different outcomes: decreased job satisfaction (Skaalvik and Skaalvik, 2010), reduced teaching self-efficacy (Roohani and Iravani, 2020) and an increased decision or desire to leave the job or profession and reduced job performance (Parmar et al., 2022; Perrone et al., 2019). Burnout of kindergarten teachers would inevitably increase the kindergarten teacher turnover (Chen et al., 2023). Organizations should make strategies and policies to reduce burnout and to prevent burnout to be able to perform better (Tanculescu-Popa, 2020).

In summary, the existing research provides a preliminary understanding of the relationship of turnover intentions with four factors, respectively. In the current study, we investigated the relationship between occupational commitment, psychological resilience, job satisfaction, burnout, and teachers’ turnover intentions to figure out the validity and reliability of these four factors together and turnover intentions among kindergarten teachers in rural China. Based on the current analysis, several hypotheses are made:

The survey procedures begin with a stratified sampling framework based on regional GDP levels. First, counties are categorized into high-, medium-, and low-GDP tiers using official economic data, ensuring systematic representation of diverse socioeconomic contexts. Within each tier, representative counties are randomly selected to mitigate geographic bias. Subsequently, a multi-stage sampling approach is implemented: rural kindergartens are randomly chosen from these counties, followed by the recruitment of teachers within each kindergarten, with attention to demographic diversity in experience, position, and educational background. To accommodate regional disparities in infrastructure, a hybrid data collection strategy (online and offline mix) is employed in this study. This study strictly adhered to academic research ethics guidelines. All participants were fully informed in writing and electronically about the research purpose, procedures, and data usage, and approval was obtained from the university’s ethics committee. Finally, we selected kindergarten teachers in the rural area of Hebei Province, China, as the research object, and these teachers were from 54 public kindergartens, 5 public-private kindergartens, and 20 private kindergartens. The total number of distributed questionnaires was 1,020, and 956 questionnaires were recovered. However, 36 participants were excluded after the data collection because of (a) a regular pattern of responses; (b) significant missing data; and/or (3) contradictory responses to relevant items. Thus, 920 valid questionnaires were included in the subsequent analyses.

The demographic descriptive statistics of the sample are shown in Table 1. Within our sample of teachers, 58% were married, 71.1% were from public kindergartens, 46.1% of teachers had no establishment, and the average income of the sample was about RMB 1,600 yuan. Regarding the education variable, teachers with technical secondary degrees accounted for 45.6%. The average number of years working in the early care education industry was more than 7 years, and the average length of service in current kindergartens was about 4 years. The demographic descriptive statistics of the sample were generally consistent with the findings from the Educational Statistics Yearbook of China (2022), except our sample had a higher education background than the national average level. The demographic descriptive statistics show that the surveyed sample was reasonable with a strong explanatory power.

Given the relevant research conducted by scholars in China and abroad, we determined the research framework and developed the research tools. At the level of job requirements, we chose several indicators such as organizational management, labor reward, work itself, emotional regulation, and coping ability. The level of job resources included several indicators, such as remuneration and incentives, self-development, policy support, and emotional status. Finally, this study determined 29 first-level indicators and 53 determinants, and the principal components were extracted by principal component analysis. According to the steep slope diagram and the structure of determinants after turning the axes (see Table 2), this study finally determined four common factors that were consistent with the constructs and items of the survey.

Based on the mentioned above, the questionnaire included three parts. The first part was demographic variables, mainly including teachers’ marital status, length of service in the current kindergarten, years working in the early care education industry, establishment, type of kindergartens, income, and education. The second part was the questionnaire on teachers’ turnover intentions. The third part was the determinants scale, which consisted of the occupational commitment scale, psychological resilience scale, job satisfaction scale, and burnout scale. The total number of items was 62, and the details of the questionnaire and each scale were as follows:

This questionnaire aimed to determine the teachers’ turnover intentions; thus, we designed the following question.

“Which is more in line with you?”

Based on the three-component conceptualization of Meyer and Allen (1991), this scale was compiled from three dimensions, namely, emotional commitment (liking the current occupation), normative commitment (being constrained by social norms), and continued commitment (leaving the current occupation will lose the corresponding benefits). This scale included 10 items, each rated on a Likert-type scale ranging from 1 (strongly disagree) to 5 (strongly agree). After the analysis, the Cronbach alpha coefficient of this scale concerning this study was 0.816, which showed good reliability and validity.

Psychological resilience scale in this study was measured in terms of both emotion regulation and coping strategies. Based on the Emotion Regulation Scale developed by Gross (1998), this scale was considered at two levels: cognitive reappraisal and expressive inhibition. Additionally, given the Brief COPE developed by Carver (1997), this scale was evaluated at three levels: planning, active coping, and instrumental support. After exploratory factor analysis, this part determined 14 items, and participants rated their consistency on a score of 1 (strongly disagree) to 5 (strongly agree) on a Likert-type scale. It showed good internal consistency with a Cronbach alpha coefficient of 0.804.

Currently, Hirschfeld (2000) dimensions of intrinsic (e.g., a chance to try out one’s ideas and feelings of accomplishment) and extrinsic (e.g., tangible aspects such as pay and working conditions) subscales are internationally recognized. Still, there is yet to be an agreement on classifying internal dimensions. To avoid indicator crossover with other determinants, through the discussion with experts, we designed the scale to include five dimensions, namely, job satisfaction (job match, occupational characteristics, autonomy); internal satisfaction (interpersonal relationships, working conditions, support system); reward satisfaction (remuneration and incentives, internal development); organizational management satisfaction (system construction, leadership behavior); and occupational environment satisfaction (policy support, social status, parental relationship). The scale included 24 items, and a Likert-type scale was used to rate the participants from 1 (strongly disagree) to 5 (strongly agree). The higher the score, the higher the job satisfaction. After the exploratory factor analysis, this scale deleted two items, and 22 remained. A revised Cronbach alpha coefficient was 0.862, indicating good internal consistency.

The most widely used questionnaire for measuring burnout was MBI (Maslach et al., 1997), containing three different versions: MBI-Human Service Survey (MBI-SS), MBI-Educators Survey (MBI-ES) and MBI-General Survey (MBI-GS). The MBI-ES with good reliability and validity was used to investigate teacher burnout in this study, which is divided into three dimensions, namely emotional exhaustion (depletion of emotional resources, exhaustion, occupational stress), depersonalization (indifference in the relationship between the teacher and the client) and self-efficacy (the value of the individual to the job, competency). The MBI-ES consists of 14 items on a 5-point Likert-type scale ranging from 1 (strongly disagree) to 5 (strongly agree). The current study showed good internal consistency with a Cronbach alpha coefficient of 0.83.

We conducted statistical analyses to examine the collected data comprehensively. We utilized SPSS 14.0 for descriptive statistics and correlations to offer an initial understanding of the dataset. In the preliminary analysis, we analyzed the data descriptively to summarize the relevant characteristics. Subsequently, we tested binary correlations between the main predictors and turnover intentions to explore the correlations between the variables. To clarify the risk rate of the predicted outcomes, we also conducted multinomial logistic regression analyses on the teachers’ turnover intentions (staying/moving/leaving). Meanwhile, we also utilized the logistic regression model and OLS regression of two turnover intention outcomes (staying or non-staying) to predict the direction and probability of teachers’ turnover intentions, assessing the predictive power of the regression models through standard errors. Finally, we also developed data models and trend plots based on the results of the above analyses.

There were 194 (approximately 21.1%) teachers intending to leave, 466 (about 21.1%) teachers intending to move, and the number of teachers who chose to stay was 466 (approximately 50.6%). Table 3 shows the mean levels of the four predictors and corresponding internal indicators. Psychological resilience had the highest mean value of 3.8. Occupational commitment was the next highest level, with a mean value of 3.68. The mean job satisfaction and burnout values were 3.61 and 3.06, respectively. The value of the four dimensions remained above the mean level.

To investigate the correlation between the predictors and teachers’ turnover intentions, we further predicted the correlation coefficients among the predictors. Table 4 reflected the correlation coefficients among the predictors, indicating the linkage between the two variables to some extent. We could also observe the significance level according to the P-value. For example, the correlation coefficient between teachers’ occupational commitment and teachers’ turnover intention (leaving) was −0.438, which was significant at the 0.01 significance level. According to Table 4, there was a substantial correlation between predictors, which can be further analyzed.

The first four columns in Table 4 compared turnover intentions: Move vs. Stay, Leave vs. Stay, Leave vs. Move, and Move/Leave vs. Stay. The first four columns describe the logistic regression analysis results, relative risk ratios describe the first three, and odds ratios represent the fourth. Finally, the fifth column in Table 4 describes the results of the OLS regression analysis, analyzing the correlation between predictors and teachers’ turnover intention after adjusting for covariates.

The first three columns in Table 4 also presented the logistic regression analysis results regarding relative risk ratio (RRR), and a relative risk ratio described each predictor variable. The fourth column predicted the odds ratio (OR) between Move/Leave and Stay. The relative risk ratio (RRR) indicates the change multiple of the event occurrence ratio when the observed predictor variable increases by one unit under the condition that other predictor variables remain unchanged. For example, the RRR of occupational commitment in Table 3 was lower than 1 (RRR = 0.51), which indicated that teachers would have a higher risk of the second outcome (Stay) compared to the first (Move). Specifically, a one-unit increase in teachers’ occupational commitment was associated with 0.51 times the risk of Move versus Stay, indicating that teachers are more likely to stay in their jobs rather than change jobs after adjusting for all other key predictors and covariates. Conversely, RRR greater than one meant that teachers were more likely to have the first outcome than the second. A relative risk ratio (RRR) equal to one means that teachers with these characteristics have a comparable risk or likelihood of having either outcome.

Table 4 showed that teachers with higher occupational commitment had a higher risk or likelihood of Stay over Move/Leave (RRR = 0.51 and 0.83 respectively, p < 0.01), and teachers were more likely to Move than Leave (RRR = 0.13). Teachers with greater psychological resilience predicted a higher risk of Moving and Leaving (RRR = 2.45 and 1.062, respectively, p < 0.05) and a higher likelihood of moving (RRR = 0.02) of the independent variables, greater emotional regulation predicted a higher risk of Moving and Leaving (RRR = 3.53 and 1.90 respectively, p < 0.05), and according to the third column, the risk of Moving was greater than that of Leaving. Additionally, teachers with higher burnout were more likely to have the outcome of Leave (RRR = 3.10 and 1.02), and the likelihood of Leave was twice as high as the likelihood of Moving. Regarding the internal structure, the stronger the teachers’ emotional exhaustion, the greater the risk of Moving or Leaving (RRR = 1.09 and 1.25, respectively, p < 0.05), and there was no statistically significant difference between the risk of Moving and Leaving for teachers. Moreover, teacher self-efficacy predicted no significant relationship or effect on teachers’ turnover intention. Finally, teachers with higher job satisfaction reported a lower risk of Leave (RRR = 0.04 and 0.1, respectively, p < 0.01).

Beta (β) in column five of Table 4 illustrated the association between the predictors and teachers’ turnover intentions, i.e., when β was positive, the predictors positively affected teachers’ turnover intentions. For example, teachers’ occupational commitment associated with teachers’ turnover intention (β = −0.27, p < 0.01) signified that greater occupational commitment reported an upward trend in turnover intentions. Conversely, psychological resilience associated with turnover intentions (β = 0.25, p < 0.01) indicated that psychological resilience had a significant positive effect on teachers’ turnover intention, i.e., the stronger the psychological resilience, the stronger the turnover intention. In addition, among the teachers’ demographic variables, including marital status, education level, income status, type of position, years of experience in the kindergarten, years of experience in the industry, type of kindergarten and establishment, only the type of kindergarten had a significant positive effect on teachers’ turnover intentions, that is, the statement that only teachers in private kindergartens had a higher likelihood of turnover was statistically valid.

Nearly 60% of kindergarten teachers prefer to Move rather than Leave when having turnover intentions. In this research, nearly half (49.5%) of the kindergarten teachers surveyed had turnover intentions, consistent with existing studies. The proportion of kindergarten teachers intending to leave the field of early care education entirely accounted for 21%. Besides, 29% of kindergarten teachers had the intention to move and change jobs within the field of early care education, accounting for 59% of the teachers with turnover intentions. Based on the mentioned above, we could find that kindergarten teachers reported high turnover intentions, but 59% of them would choose to change to a new job in the same field of early care education rather than leaving this field entirely.