Development of the Job-Related Uncertainty Stress Scale for Platform Workers

Abstract

Objectives: This study reports the development of the Job-Related Uncertainty Stress Scale for Platform Workers (JUSSPW) and examines its reliability and validity.

Methods: The research was conducted in 2 phases. In Study 1, item analysis and exploratory factor analysis were conducted on data from 343 platform riders (males: 321; females: 22; mean (SD) age: 27.03 (6.67) years) in Guangzhou, China. In Study 2, an additional 391 platform riders (males: 328; females: 63; mean (SD) age: 30.36 (4.49) years) were recruited. This phase involved conducting confirmatory factor analysis (CFA) and assessing criterion-related validity by using the Uncertainty Stress Scale (USS-4), Maslach Burnout Inventory (MBI), and Job Satisfaction Inventory (JSI).

Results: The JUSSPW scale comprises 8 items under a unidimensional structure, covering 4 perspectives: work environment, interpersonal relationships, industry-specific characteristics, and personal development prospects; it explained 71.07% of the total variance. CFA results indicated that this 1-factor model provided a good fit (χ²/df = 2.681, Root Mean Square Error of Approximation (RMSEA) = 0.066, Comparative Fit Index (CFI) = 0.987, Incremental Fit Index (IFI) = 0.987, Goodness of Fit Index (GFI) = 0.964, Tucker-Lewis Index (TLI) = 0.982). This scale also demonstrated good convergent (Average Variance Extracted (AVE) = 0.668, Composite Reliability (CR) = 0.941) and criterion validity (area under the curve = 0.935). The total score of JUSSPW was significantly positively correlated with the USS-4 and Maslach Burnout Inventory-Emotional Exhaustion (MBI-EE) scores, and remarkably negatively correlated with the JSI scores. Cronbach α and split-half reliability were .939 and 0.935, respectively.

Conclusions: These results suggest that this scale shows good reliability and validity and can be used as a sound measure to capture platform workers’ job-related uncertainty stress. Limitation and implications are discussed.

1. Introduction

In recent years, the digital economy has witnessed a vigorous surge, giving rise to new forms of economic activities, commonly known as the “gig economy,” “digital work,” or “online work.” Workers provide services through online platforms and generally carry out tasks offline, constituting approximately 1%-3% of total global employment and experiencing rapid growth.¹ Typical representatives of this group include ride-hailing drivers, food delivery riders, and household service providers. Current research underscores the prominent features of instability and uncertainty that characterize platform work conditions, coupled with various risks,² leading to psychological stress on platform workers. Previous studies have made efforts to measure the psychological health of platform workers, such as psychological stress and negative emotions like anxiety and depression,³ as well as their occupational health.⁴ Research into job-related stress measurement has also progressed. For instance, Tatar’s Job Stress Scale-20 for general jobs,⁵ Yang et al’s Sources of Pressure Scale for government employees,⁶ and Useche et al’s adaptations for transportation workers⁷ highlight the need for occupation-specific tools. However, these tools may not be fully suited to capturing the specific nature of job stress in environments characterized by high levels of uncertainty and instability. Therefore, the measurement of job-related stress among platform workers becomes particularly important.

1.1. Uncertainty as a work stressor of platform drivers

Compared with traditional industries, platform work has emphasized the flexibility of working hours, locations, and job content. Workers have the freedom to control their working schedules and space. However, this work arrangement or work mode is associated with different types of uncertainties. To begin with, the employment forms in the gig economy are diverse, with workers often being self-employed or working part-time,⁸ lacking fixed labor contracts, and departing from the traditional institutional framework that guarantees employment relationships. Relatedly, a lack of sufficient labor protection means that workers may face greater risk of income insecurity and greater financial risks.³ The nature of their work often exposes them to different work stressors, such as handling of challenging customers,⁹ and higher risk of occupational injuries.¹⁰ Because their job might entail frequent interaction with a large number of customers, the heightened risk is exacerbated by inadequate protective measures and constant exposure to various environments, resulting in a higher disease infection rate—including COVID-19 infection and other infectious diseases—compared with other occupations that have more stable or isolated working conditions.¹¹ Finally, some platform workers find it difficult to advance in their positions and feel uncertain about career prospects.¹² In summary, workers in the gig economy face uncertainties in various aspects of their work, such as interpersonal relationships, working environment, and career development.

Uncertainty is a significant component of stress, closely intertwined with an individual’s physiological and psychological well-being.¹³ When a vague or unpredictable event or situation arises, it generates a certain degree of uncertainty stress. The Stress and Adaptation Theory highlights that individuals, when confronted with environmental stressors, may undergo a series of physiological and psychological reactions. In response, individuals may employ various coping strategies, including avoidance, confrontation, and adaptation, to address the stressors in their environment.¹⁴ Research has shown that, compared with general life stressors, uncertainty stress is more strongly linked to both short-term and long-term illnesses and has a more pronounced adverse impact on health.¹⁵ Numerous large-scale cross-sectional studies have indicated that psychological stress induced by uncertainty not only negatively affects psychological health such as mental disorders,¹⁶ but also is more likely to lead to problematic behaviors such as problem alcohol use and deliberate self-harm.^17,18

Uncertainty stress is closely linked to the social and working environments, affecting individuals’ physiological and psychological health in unique ways. Research has shown that uncertainty in the workplace can lead to heightened anxiety and stress levels because employees face unpredictable changes in their roles, job security, and organizational expectations.^19,20 This stress undermines job satisfaction and increases the sense of isolation especially when social support systems are lacking in the workplace.²¹ In broader societal contexts, such as rising unemployment rates, employees may increasingly doubt the continuity of their jobs, depleting psychological resources and leading to emotional burnout.²² Therefore, the development of a specific job-related uncertainty stress scale is crucial and adds significant value to the field. As discussed earlier, the gig economy, and platform riders in particular, represents a new type of employment relation. On the one hand this provides a greater sense of autonomy for job incumbents, but on the other hand exposes these workers to different uncertainty risks. To understand the scope of the impact, established scales, such as the Hilton Uncertainty Stress Scale for the assessment of illness-related uncertainty,²³ as well as Yang’s Chinese versions of the Uncertainty Stress Scale, available in both 4-item and 10-item formats, for uncertainty in the general society,²⁴ may not be optimal as these scales do not consider the unique job characteristics of platform work. Meanwhile, reviewing previous studies on occupational stress measurements, many of them focus on traditional professions such as health care workers and teachers,^25,26 with limited attention given to this new occupational group. Thus, a new industry-specific uncertainty stress scale would provide important insight on how these uncertainty stressors influence the psychological well-being of the platform workers.

Building upon this, the present study aims to develop the Job-Related Uncertainty Stress Scale for Platform Workers (JUSSPW) and assesses its reliability and validity. The goal is to explore the current state of uncertainty stress of platform workers and enrich research content in the field of uncertainty stress.

2. Study 1

2.1. Purpose

Study 1 aimed to develop an initial item pool of uncertainty stress among platform workers. These items were generated by semi-structured interviews, expert evaluations, and referencing existing stress scales (eg, the Uncertainty Stress Scale). Subsequently, an item analysis was conducted to assess the discrimination and homogeneity of the scale’s items. The factor structure of the new JUSSPW scale was examined using exploratory factor analysis (EFA).

2.2. Methods

2.2.1. Development of the item pool

A thorough review of the existing literature indicated that the work stress experienced by delivery riders stems from various aspects of their work environments, such as customer relationships and risks encountered during the delivery process. Based on the relevant literature regarding the current working conditions of delivery riders, as well as stress measurement and stress coping,^5-7 a framework for semi-structured interviews was established. Sample interview questions included: “Have you experienced uncertainty in this job? If so, under what circumstances?” “What uncertainty factors or situations do you believe create significant pressure for you?” and “How do you typically respond when faced with such uncertainties?”. Prior to the interviews, participants were given a comprehensive introduction to the concept of uncertainty in their job, including specific examples. Throughout the interviews, they were encouraged to express their viewpoints, experiences, and responses to uncertainty.

Second, the objective sampling method in “non-probabilistic sampling” was employed, aiming to selectively identify research subjects that offer the most comprehensive information in alignment with the research purpose.²⁷ The sampling process was ongoing until saturation, where no new responses to the interview questions emerged. All participants were over 18 years old and had been working as delivery riders for at least 6 months. Finally, 12 platform riders, comprising 2 females and 10 males aged between 22 and 30 years, took part in semi-structured interviews. Among them, 3 had less than 1 year of experience as delivery riders, 5 had been in this industry for 1-3 years, and the remaining 4 had worked for more than 3 years.

Three psychology graduate students transcribed and analyzed interview content. Based on the interviews and the framework of Ecological System Theory,²⁸ they identified 4 themes associated with job-related uncertainty experienced by platform riders, including uncertainty in the work environment (ie, struggling with unfamiliar surroundings and risks from uncontrollable factors), uncertainty in interpersonal relationships at work (ie, dealing with unstable relationships with colleagues and clients), industry uncertainty (ie, facing challenges due to incomplete industry regulations), and personal development uncertainty (ie, encountering challenges in personal career planning). An initial item pool comprising 20 items was formed. To ensure the accuracy and effectiveness of each item, a panel of 6 psychology professors and graduate students first discussed the items and put forward suggestions for revision. Additionally, this study invited another 12 delivery riders to rate the importance of the items and provide their understanding and feelings on the content. Items about related factors were consolidated based on their feedback. Consequently, a preliminary questionnaire comprising 10 items was developed. Furthermore, 4 subject matter experts from the fields of psychology, sociology, and public health were invited to discuss and revise the items. For instance, an original item stating, “the work environment is often unpredictable” was revised to “there is no fixed workplace, and the work environment is complicated and unfamiliar” to improve specificity. Similarly, the phrase “the income situation is often unstable” was changed to “the salary system is unreasonable, and the work income is unstable” for greater clarity. This helped eliminate redundancy and enhance clarity. After 3 rounds of revisions, 8 items were retained, addressing uncertainties in the work environment (2 items), interpersonal relationships (2 items), industry management (2 items), and personal development prospects (2 items). Responses were graded using a 5-point Likert scale (1 indicating very little stress and 5 indicating extreme stress). The final score was calculated by adding up each of the items, ranging from the minimum of 8 to the maximum of 40. Higher scores indicated a greater level of uncertainty stress.

2.2.2. Participants and sampling procedure

This study employed a mixed sampling approach to recruit platform riders in China. Firstly, a random sampling method was adopted to select 3 districts among the 7 main districts in Guangzhou, specifically Baiyun District, Haizhu District, and Yuexiu District. Within each of these districts, 4 communities were randomly chosen for the research. The sampling frame for districts and communities was based on official administrative data. Then, a convenience sampling method was adopted to invite delivery riders from each community to participate in our study through street intercepts.

The study was conducted in August 2022 based on the Wenjuanxing Platform, an online questionnaire tool that facilitates the collection of responses. Twenty-two university students majoring in psychology were recruited from Guangzhou and Shenzhen, 2 major cities in Guangdong Province, to disseminate the survey link via street interception and social media such as platform workers’ WeChat work groups, which greatly enhanced the diversity of perspectives in our research and improved the efficiency of our data collection process, especially against the background of the COVID-19 pandemic. To ensure the authenticity of the participants, this study also implemented strict identity verification measures before the data collection. Prior to participating in the survey, all participants needed to fill out a consent form to ensure their informed participation and agreement to follow the study’s guidelines. The study received approval from the Ethics Committee of Shenzhen University (Approval Number: 202200044). All procedures adhered strictly to the relevant ethical and research standards.

A total of 346 participants were recruited in Study 1, and after screening out data from participants who had not completed the questionnaire or completed the questionnaire in under 300 s, 343 valid responses were obtained. The individuals in the sample were aged between 18 and 51 years, with a mean age of 27.03 years (SD:6.67). Among them, 321 (93.58%) were male, and 22 (6.42%) were female. Regarding experience as delivery riders, 231 (67.35%) had been working for 1 year or less, 73 (21.28%) had worked for 1 to 3 years, and 39 (11.37%) had worked for 3 years or more. In terms of education, 122 (35.57%) had only completed junior high school or even not, 121 (35.28%) had completed high school, 95 (27.64%) held an associate or bachelor’s degree, and 5 (1.46%) possessed a master’s degree or higher.

2.2.3. Statistical analysis

The data were analyzed using SPSS 26.0 software, and involved employing descriptive statistics, conducting item analysis, and performing EFA. In the item analysis, the total score of the scale was first calculated and ranked in ascending order. The top and bottom 27% scores were used as thresholds to categorize participants into high- and low-scoring groups. An independent sample t test was further conducted to assess the critical ratio value of all items to indicate whether the item discriminated the level of subject response adequately. Statistically significant items with critical ratio absolute values <3.50 should be eliminated.²⁹ Subsequently, correlation coefficients between items and the total score were calculated. Items with correlation coefficients <0.4 were eliminated from the analysis.³⁰ Finally, the initial questionnaire’s internal consistency was assessed using the Cronbach α coefficient. Any item that resulted in an increase in the Cronbach α coefficient upon deletion was considered for potential removal.

In EFA, the Kaiser-Meyer-Olkin (KMO) value and the P value of Bartlett χ² were calculated first. A KMO value >0.60 and a P value from the Bartlett test <.05 indicated that the data were appropriate for factor analysis. Principal component analysis and orthogonal varimax rotation were then applied for factor extraction, considering eigenvalues and scree plots. Any items that exhibited factor loadings <0.40 were subsequently removed.

2.3. Results

2.3.1. Item analysis

Table 1 displays the results of the item analysis. Scores in the “high” group in the initial questionnaire were significantly higher than those in the “low” group (P < .01), and the critical ratio values evaluated by independent sample t test were all above 3.5 and were statistically significant (P < .01), indicating excellent discrimination for all items. Further, a substantial positive relationship was observed between the scores of each item and the overall score, with correlation coefficients ranging from 0.750 to 0.836 (all P < .001). The initial questionnaire had a Cronbach α coefficient of .941. Removing any individual item did not result in a significant increase in the Cronbach α coefficient. Therefore, all 8 items of the initial questionnaire were retained.

2.3.2. Exploratory factor analysis

The initial questionnaire was assessed with a KMO value of 0.916, and the Bartlett test of sphericity yielded a value of 2284.021 (P < .01), confirming the data’s suitability for EFA. Among the 8 items, 1 factor was extracted with an eigenvalue >1, contributing to a cumulative variance explanation of 70.794%. Table 2 displays the factor loadings of the 8 items on 1 factor, exhibiting strong values from 0.812 to 0.884.

3. Study 2

3.1. Purpose

Study 1 led to the development of a formal scale with 8 items following item analysis as well as EFA. The objective of Study 2 was to further evaluate the scale’s reliability and validity, involving 3 main components: First, to assess the fit of a single-factor model through confirmatory factor analysis (CFA). Second, to determine the reliability of the scale by calculating split-half reliability and Cronbach α coefficient. Third, to conduct tests for criterion validity and convergent validity.

3.2. Methods

3.2.1. Participants and sampling procedure

Similar to Study 1, Study 2 used a random sampling method to choose 4 communities from each of the 3 districts, specifically Tianhe District, Baiyun District, and Yuexiu District, among the 7 main districts of Guangzhou. The sampling frame for districts and communities was firmly grounded in official administrative data. Subsequently, a convenience sampling approach was employed to invite delivery riders from these communities to participate in the study.

The study was conducted 3 months after Study 1, in November 2022, based on the Wenjuanxing Platform. A total of 391 participants were randomly selected from the sampling frame to participate in our study. All participants consented prior to beginning the survey. In this study, the participant cohort comprised 328 males (83.88%) and 63 females (16.12%). The age range of the participants was 21 to 43 years, with a mean (SD) age of 30.36 (4.49) years. Within this sample, 87 (22.25%) had been engaged in platform delivery for 1 year or less, 205 (52.43%) for 1 to 3 years, and 99 (25.32%) for 3 years or more. The distribution of educational levels was as follows: 63 (25.32%) had a junior high-school education or below, 161 (41.18%) had a high-school education, 153 (39.13%) had a college or undergraduate degree, and 14 (3.58%) had a graduate degree or higher.

3.2.2. Measurements

Uncertainty Stress Scale (USS-4): The Chinese version of the USS-4 was developed by Yang et al³¹ to assess individuals’ responses to general uncertainty stress in life. The scale comprises 4 items and uses a 5-point rating scale ranging from 1 (very little stress) to 5 (extremely stressful), with higher scores indicating higher perceived levels of stress. A total score >12 or an average item score >3 indicates severe uncertainty stress,²⁴ serving as an important cutoff point. In this study, the Cronbach α coefficient for this scale was .908.

Maslach Burnout Inventory (MBI): The MBI is currently the most widely used tool for measuring job burnout globally. Developed by Maslach and Jackson,³² this scale comprises 3 dimensions: emotional exhaustion, depersonalization, and personal accomplishment. In this study, the emotional exhaustion subscale was selected to measure platform riders’ feeling of emotional fatigue, consisting of 5 items rated on a scale from 1 (never) to 7 (every day). Higher scores indicate a greater level of emotional exhaustion. The Cronbach α coefficient for this scale in the present study was .935.

Job Satisfaction Inventory (JSI): Developed by Price and Mueller,³³ this inventory consists of 7 items. For this study, 3 items most applicable to platform workers were selected. Respondents use a 5-point scale, ranging from 1 (strongly agree) to 5 (strongly disagree), to indicate their level of agreement, measuring their job satisfaction. After reverse scoring, higher total scores indicate higher job satisfaction. The Cronbach α coefficient for the scale in this study was .892.

3.2.3. Statistical analysis

Data from Study 2 were subjected to CFA using Amos 24.0, and reliability and validity tests were conducted using SPSS 26.0. First, for the CFA, the model fit was considered good when meeting the following conditions: χ²/df <3, Root Mean Square Error of Approximation (RMSEA) <0.08, and Comparative Fit Index (CFI), Incremental Fit index (IFI), Goodness of Fit Index (GFI), and Tucker-Lewis Index (TLI) >0.90.³⁴ Second, the reliability of JUSSPW was evaluated using Cronbach α coefficient and split-half reliability. When the reliability coefficient exceeded 0.7, the scale exhibited high reliability.³⁵ Third, Pearson correlation analysis was performed to assess the correlation between JUSSPW and the criterion variables (ie, general uncertainty stress, job satisfaction, and emotional exhaustion). The average variance extracted (AVE) was used to evaluate the convergent validity. To determine how well the USS performed in identifying individuals with high uncertainty stress, sensitivity and specificity were further evaluated by receiver operating characteristic (ROC) curve analysis.

3.3. Results

3.3.1. Reliability test

The Cronbach α coefficient for the total JUSSPW scale was 0.939, and the split-half reliability coefficient was 0.935, both exceeding 0.7 and thus indicating good reliability.

3.3.2. Confirmatory factor analysis

The results of the CFA indicated a good model fit for this unidimensional structure scale, with χ² = 53.624, df = 20, χ²/df = 2.681, RMSEA = 0.066, CFI = 0.987, IFI = 0.987, GFI = 0.964, and TLI = 0.982.

3.3.3. Criterion validity test

As shown in Table 3, Pearson correlations indicated that the total score of JUSSPW was positively correlated with USS-4 and MBI-EE scores, whereas it was negatively correlated with JSI scores.

Consistent with prior research, this study dichotomized the USS-4 total score using a threshold of 12 points, and the JUSSPW total score was used as the state variable to create a ROC curve. The area under the ROC curve (AUC = 0.935; 95% CI, 0.924-0.965) exceeded the commonly accepted threshold of 0.70 for good diagnostic accuracy,³⁶ indicating that JUSSPW had similar effectiveness to other uncertainty stress measurement tools. In addition, the sensitivity and specificity of JUSSPW were 0.77 and 0.95, respectively (Figure 1).

3.3.4. Convergent validity test

Table 4 demonstrates the loadings of 8 items on a single factor, with loading values ranging from 0.646 to 0.921. Factor loadings >0.50 were typically considered acceptable in indicating item-factor congruency.³⁷ Furthermore, the AVE (0.668) exceeded the recommended threshold of 0.50, indicating that a substantial amount of variance in the items was explained by the latent factor.³⁸ Additionally, the composite reliability (CR = 0.941) surpassed the commonly accepted level of 0.70, confirming the internal consistency and reliability of the scale.³⁷ These results collectively indicated good convergent validity.

4. Discussion

This study describes the process of the development of a new scale intended to assess the uncertainty stress among platform workers. This new scale considered uncertainty stress in 4 perspectives: work environment, interpersonal relationships, industry-specific characteristics, and personal development prospects. The findings align with the challenges faced by platform workers both domestically and internationally.^2,12

Item analysis showed a significant difference in uncertainty stress between high- and low-score groups, confirming the scale’s discrimination ability. Both exploratory and confirmatory factor analyses revealed a single factor, with a cumulative variance of 71.074% and good fit indices. The scale also exhibited strong convergent validity. A higher AVE value indicates that the items share a substantial amount of variance, which is preferable for the scale’s overall reliability and validity.³⁸ Furthermore, the results of the criterion validity test indicated that the total score of JUSSPW was positively correlated with the general uncertainty stress and emotional exhaustion, whereas it was negatively correlated with job satisfaction, aligning with findings from previous studies.^21,22

For most individuals, uncertainty is a powerful source of stress.¹³ In the gig economy, workers face significant stressors such as unstable income, irregular working hours, and a lack of job security, which can exacerbate job insecurity and life uncertainty, contributing to substantial job-related pressures.³⁹ Stress is a prevalent negative factor affecting job satisfaction across various industries, and high-stress environments can lead to emotional exhaustion and subsequent job burnout.⁴⁰ Therefore, job-related uncertainty stress is closely linked to emotional exhaustion and job satisfaction. The ROC curve analysis yielded an AUC of 0.935, further confirming the scale’s strong criterion validity. Reliability test results indicated that the Cronbach α coefficient of the total scale in this study was 0.939, and the split-half reliability was 0.935, demonstrating strong internal consistency for the scale. Based on a series of psychometric tests, the new 8-item USSPW has demonstrated satisfactory reliability and validity and can serve as a tool for measuring occupational uncertainty stress among platform workers.

The successful development of JUSSPW contributes significantly to the understanding of uncertainty in the workplace. Existing research highlights the social, physiological, and psychological impacts of uncertainty stress,^15,19-21 as it can be detrimental to both personal and professional well-being. Compared with traditional professions such as doctors and teachers, platform work involves greater uncertainty due to the flexibility and variability of its working environment and patterns.⁹ Unlike other job-related stress scales, the JUSSPW is uniquely tailored to capture this specific characteristic of platform workers by assessing uncertainty stress across 4 critical dimensions. This comprehensive approach allows JUSSPW effectively to capture the unique stressors experienced by platform workers, making it more targeted and relevant to this group than general job stress scales and enriching the research on uncertainty stress. Practically, it also provides a reliable tool for understanding and addressing the psychological stress of platform workers, which is crucial for increasing awareness of their psychological health and promoting better support, ultimately enhancing the well-being of this rapidly growing occupational group.

5. Limitations and future research

Some limitations remain in this study. Firstly, as the study used a convenience sampling method, the generalizability of the findings is limited. Future research should adopt more rigorous sampling methods and consider the influence of multiple factors on the results. Secondly, the study was conducted exclusively among platform riders, so there are certain constraints when generalizing the results to other groups. Future research should expand the sample scope to include more workers from various gig platforms, such as ride-hailing drivers and home service providers, to enhance the representativeness of the study. Thirdly, the current sample was predominantly male. Although this aligns with the existing gender distribution in the delivery industry in China,¹⁰ it is necessary for future research to strive for a more balanced gender ratio, further improving the external validity of the scale. Finally, the item set in the study was constructed during the COVID pandemic, and therefore with this background, some items (eg, item 1) deliberately referred to evaluate COVID-related uncertainty. However, this item set is flexible and generic, and future research could modify the original items by eliminating the reference to the COVID-19 pandemic to fit different study purposes.

6. Conclusion

Platform workers face significant uncertainty in their work. As the gig economy scales up and the number of platform workers increases, it is necessary to develop an assessment tool with good reliability and validity to measure the job-related stress levels of this group specifically derived from uncertainty stressor. The JUSSPW is a new domain-specific scale that has demonstrated excellent psychometric properties. It can be used as a valid tool for assessing stress levels related to job uncertainty among platform workers.

Acknowledgments

We thank research participants for their participation in this study. The study was approved by the Ethics Committee of Shenzhen University (Approval Number: 202200044). All the participants provided informed consent. All procedures were performed in accordance with relevant guidelines.

Author contributions

The authors confirm contributions to the article as follows: Y.Z. and D.W.: study conception and design; Y.Z., C.L., and H.L.: data collection; Y.Z.: analysis and interpretation of results; Y.Z. and D.W.: draft manuscript preparation; Y.Z., D.W., F.C., and H.X.: manuscript revision. All authors reviewed the results and approved the final version of the manuscript.

Funding

This study was partially supported by Shenzhen Natural Science Fund (the Stable Support Plan Program 20 220 811 090 420 002), Guangdong Philosophy and Social Science Planning Project (Grant No. GD23YSH09), Shenzhen University-Lingnan University Joint Research Programme (Grant No. 202202005), and Ministry of Education Humanities and Social Sciences Research Youth Fund Project (Grant No. 24YJC840036).

Conflicts of interest

The authors declare that they have no conflicts of interest to report regarding the present study.

Data availability

The datasets used and analyzed during this study are available from the corresponding author upon reasonable request.

References