Expectation-Reality Gaps and Talent Re-Mobility: An Adaptive Cycle in Shaoxing’s Regional Talent Ecosystem ()
1. Introduction
Talent mobility is the core link in the optimal allocation of human resources, exerting a direct impact on regional innovation capacity and the quality of educational development. In recent years, policies including the development of China’s “Double First-Class” universities, integrated regional development, and the Belt and Road Initiative (BRI) have driven profound changes in the pattern of talent mobility in China. Against the backdrop of increasingly fierce regional competition for talents, high-level talents have become the core driver of urban innovative development. As a key node city in the Yangtze River Delta (YRD), Shaoxing has attracted a large number of high-level talents in recent years through a series of measures such as the “Hometown of Celebrities” Talent Program and the “Golden Keqiao” Talent Policy. According to official government documentation, Shaoxing’s city-level talent attractiveness was reported to rank 23rd nationwide for two consecutive years (Shaoxing Municipal Human Resources and Social Security Bureau, 2025; original ranking source not specified in-text). Nevertheless, it still faces practical dilemmas reported in local talent work, including high self-reported turnover intention among newly-recruited high-level talents, structural shortage of professionals, and the predicament of “ease in attracting talents but difficulty in retaining them”.
Focusing on high-level talents with doctoral degrees who have migrated to Shaoxing City as the research subjects, this study collects first-hand qualitative data through semi-structured interviews and conducts a systematic analysis using Constructivist Grounded Theory. It aims to develop a multi-dimensional trade-off and dynamic feedback theory of high-level talents’ geographic mobility, to fully explain the decision-making logic underlying their job location selection and mobility paths, as well as their subsequent adaptation and re-decision-making processes, and further to uncover the nature of the dynamic evolution of regional talent agglomeration and self-reported attrition intention.
Existing studies on talent mobility in China have yielded a wealth of academic findings. In the research field of high-level talent mobility in higher education institutions, Duan et al. (2024) pointed out that the mobility of high-level talents including academicians of the Chinese Academy of Sciences and the Chinese Academy of Engineering, Changjiang Scholars, and National Outstanding Young Scientists in universities in central and western China exhibits a pronounced profit-seeking tendency, further exacerbating the imbalance in the regional talent structure and the growing polarization in the development of higher education institutions between eastern and western China.
By constructing an influence mechanism model, Yuan et al. (2024) found that the core factors influencing the mobility of high-level talents in western China fall into six dimensions: economic benefits, scientific research development, performance appraisal, service guarantee, social adaptation, and talent development environment. Among them, economic benefits serve as the direct driver, scientific research development and performance appraisal constitute the intrinsic core drivers, while service guarantee, social adaptation, and talent development environment form the internal and external contextual constraints.
Focusing on the mobility of young high-level talents in Chinese universities, Xu and Wang (2021) revealed findings inconsistent with Western research conclusions: the effect of gender on the mobility of young talents in China is significantly weakened, and there is an offsetting effect between institutional prestige and geographical attributes. Meanwhile, a significant “seniority penalty” phenomenon was identified, where young talents with less seniority show a stronger willingness to move, and income level and academic performance exert significant positive effects on their mobility behavior.
In terms of research on the spatial mobility patterns of highly educated talents, Shi et al. (2024b) found that the mobility network of China’s highly educated population exhibits a markedly unbalanced characteristics. Specifically, the eastern coastal areas are the core hotspots of this network, and highly educated talents have a higher frequency of mobility to megacities and regional central cities. The overall hierarchical structure of the network exhibits a pattern of “north-south differentiation” with Beijing and Shanghai as the dual cores. Geographical distance exerts a significant inhibiting effect on the mobility of highly educated talents, while similarities in culture, institutions and dialects exert a significant facilitating effect on it. Cities with a high level of economic development, concentrated educational resources, well-developed infrastructure, and a sound living environment are more likely to foster the agglomeration of highly educated talents.
From the dual macro and micro perspectives, Shi, Yang et al. (2024a) found that the mobility networks of highly educated talents exhibit strong isomorphism alongside a degree of heterogeneity. Talents from cities with high administrative levels and superior economic development continuously converge to Shanghai, forming a “core-periphery” hierarchical structure with Shanghai as the sole core hub. The influencing factors of talent mobility at the macro and micro scales are highly convergent: highly educated talents are more inclined to move between cities with developed economies, robust scientific research capacity, well-developed infrastructure, and a pleasant living environment, while geographical distance, as well as differences in regional culture and institutions, still have a non-negligible impact on their mobility behavior.
In summary, existing studies have conducted multi-dimensional analyses of the characteristics, influencing factors, and spatial patterns of high-level talent mobility in China, forming an extensive core consensus, laying a solid foundation for subsequent research. Nevertheless, five critical limitations remain in the current literature:
First, most studies adopt a predominantly static research perspective, focusing heavily on the single decision-making behavior of talent mobility, while paying insufficient attention to the dynamic processes of post-mobility adaptation, evaluation, and re-decision-making. Second, the geographical scope of existing research is concentrated on extreme cases, with most studies focusing on core first-tier cities such as Beijing, Shanghai, and Hangzhou, or underdeveloped areas in central and western China, while devoting limited attention to medium-sized node cities in the Yangtze River Delta (YRD). Third, the research process is largely fragmented: most studies separately examine the influencing factors or mobility characteristics of talents, lacking in-depth analysis of the complete chain from “influencing factors” to “decision-making mechanism” to “mobility behavior” to “subsequent evolution”. Fourth, the theoretical construction of existing research is fragmented and unsystematic. Most studies carry out confirmatory analysis based on established theories, while few systematic theoretical frameworks have been constructed from the bottom up based on China’s local context. Fifth, existing research relies predominantly on quantitative methods: most studies use panel data to conduct regression analysis, while in-depth qualitative research based on the subjective experiences of talents remains relatively scarce.
Against this backdrop, this study takes Shaoxing, a medium-sized node city in the Yangtze River Delta (YRD), as the research case. It is grounded in the policy and practical context, the foundation laid by existing research, the case’s regional representativeness, and methodological suitability. It makes academic innovations in five dimensions: theoretical, perspective-based, regional, content-related, and methodological. The multi-dimensional trade-off and dynamic feedback theory of high-level talents’ geographic mobility constructed in this study not only fills the research gaps in the existing literature and enriches the theoretical system of research on high-level talent mobility, but also provides scientific theoretical and empirical support for the work of talent introduction, cultivation, retention and utilization in Shaoxing and similar medium-sized node cities across China, thus facilitating the optimal allocation of regional human resources and innovative development.
2. Research Methods and Data Collection
2.1. Research Methods
This study adopts Charmaz’s (2014) Constructivist Grounded Theory approach for data analysis. Centred on the constant comparative method, this approach deconstructs and abstracts qualitative data step by step through the threefold logic of induction, deduction, and abduction. It aims to construct an explanatory theory and conceptual framework for the research phenomenon, with its core connotation being “encompassing the interrelationships between concepts and categories to interpret and explain the patterns or processes of social psychological phenomena”.
Three core reasons justify the selection of this method. First, compared with classical grounded theory, constructivist grounded theory is more compatible with the highly subjective and personalised accounts of complex social phenomena, and aligns well with the subjective and situational characteristics of high-level talent mobility decision-making. Second, this method more oriented toward practical application orientation, effectively bridging theoretical construction and policy optimisation, and aligning with the practical objectives of this study. Third, this method acknowledges the participatory role of the researcher and the subjectivity of theory construction. In line with constructivist grounded-theory epistemology, the researcher’s academic background in higher-education and talent-policy research shaped pre-understandings of talent-mobility mechanisms. To mitigate unchecked bias, interview protocols remained open-ended; the researcher bracketed prior assumptions during initial open coding while maintaining theoretical dialogue with existing literature at later analytical stages. All participants were university colleagues within the same municipal higher-education system; this shared institutional background facilitated trust-building during interviews yet risked encouraging socially desirable responses. The researcher kept reflexive memos throughout data collection and coding to document how positional standpoints may have shaped question framing, participant responses, and category interpretation. It enables the informed use of the researcher’s professional expertise and existing literature to engage in theoretical dialogue, and thereby avoiding the risk of reducing theoretical construction to a mere descriptive presentation of data.
The research design and data analysis procedures of this study strictly adhere throughout to the eight core criteria for excellent qualitative research proposed by Tracy (2010), namely a worthy topic, rich rigor, sincerity, credibility, resonance, significant contribution, ethics, and meaningful coherence, to ensure the standardisation and scientific rigour of the research process and findings.
2.2. Data Collection
2.2.1. Sampling Strategy and Participant Recruitment
The first-hand data of this study are derived from semi-structured interviews with high-level talents who have migrated to Shaoxing City. The sample selection criteria are as follows: 1) high-level talents who have migrated to and taken up stable employment in Shaoxing within the past five years (i.e., after the official launch of the Shaoxing “Hometown of Celebrities” Talent Programme in October 2020); 2) aged below 45 years; 3) holding a doctoral degree and a professional technical title of Lecturer or above; and 4) available and willing to participate in face-to-face semi-structured interviews.
The core reason for focusing on talents who migrated to Shaoxing within five years is that the Shaoxing Municipal Government officially issued the Shaoxing “Hometown of Celebrities” Talent Programme in October 2020, elevating the strategies of “strengthening the city through talent development” and “strengthening the city through innovation” to top-priority strategies, and making talent recruitment and introduction the core strategy for the city’s core development strategy. Talents who migrated within this time frame can more accurately reflect the actual effects of Shaoxing’s talent policies and the true level of the city’s talent attractiveness.
Given the inherent scarcity of high-level talents with doctoral degrees, this study employed snowball sampling. Snowball recruitment was initiated by purposively selecting three newly-arrived doctoral faculty members from the municipal public undergraduate university as seed participants. After completing each interview, participants were invited to refer other eligible doctoral-level colleagues who met inclusion criteria. All seven participants were ultimately recruited from this single municipal public university. This sampling boundary may amplify the salience of academic-network influences, institutional organisational experiences, and academic-career-related family trade-offs within the dataset; findings may not generalise to doctoral-level talents employed in enterprises, hospitals or research institutes. Using doctoral teachers from the only municipal public undergraduate university in Shaoxing as the initial sample, a total of seven eligible interviewees were finally recruited, including one Professor, three Associate Professors, and three Lecturers. All respondents hold doctoral degrees and work as university faculty, with their academic titles spanning three hierarchical levels, thereby capturing the decision-making logic of doctoral talents at different career development stages. The birthplaces of the respondents cover three major regions of China—Northeast China (Liaoning, Heilongjiang), Central China (Shandong, Anhui), and Western China (Sichuan)—and all migrated to Shaoxing across provincial boundaries, fully meeting the core eligibility criteria for the research subjects.
2.2.2. Interview Protocol
The interview guide was designed in a semi-structured format, centred on the full life course of high-level talents, focusing on key career milestones including their basic education, higher education, further study, employment choice, geographic mobility, and urban integration. Open-ended questions were used to explore in depth the influencing factors behind their major decisions, the underlying logic of job location selection, their adaptation experiences after migrating to Shaoxing, and future development plans. Specific questions included, but were not limited to:
“Could you describe your family background and early educational experience?”
“How did you decide on your university major and whether to pursue further studies (master’s degree, doctorate, or study abroad)?”
“What factors influenced your choice of employment location and employer?”
“Why did you choose Shaoxing? What were the key attractions or concerns?”
“How have you adapted to working and living in Shaoxing? What has met or disappointed your expectations?”
“Do you intend to stay in Shaoxing long term? Why or why not?”
The interviews were conducted from March to August 2025. Prior to each interview, participants were fully informed of the aims of the study, data usage protocols, and confidentiality principles via a standardised verbal informed consent script. Interviews were fully audiorecorded only after explicit verbal informed consent was obtained from each participant. Each interview lasted 30 to 60 minutes, with no leading questions asked at any point. All participants took part in the study on a purely voluntary basis and received no financial compensation.
2.2.3. Data Processing
In total, seven interview recordings were collected. Each recording was anonymised (with all personal identifiers removed, including names, specific departmental information, and other details that could reveal their identity) and then transcribed verbatim. The transcription process followed strict rules: all utterances were transcribed exactly as spoken, including hesitations, repetitions, and colloquial expressions, but personal identifiers were replaced with codes (e.g., “Respondent 1”, “Respondent 2”). The final transcribed corpus comprised approximately 82,000 Chinese characters of original text, serving as the core analytical dataset for this study.
2.2.4. Participant Demographic Characteristics
Table 1 presents the basic demographic characteristics of the seven interview participants. Notably, female respondents accounted for 71.4% of the sample, and the middle-aged group (40 - 49 years) comprised 71.4% of the sample. All respondents hold doctoral degrees and work as university faculty, with academic titles spanning Lecturer, Associate Professor, and Professor. Their birthplaces cover five provinces across Northeast, Central, and Western China, and all have migrated to Shaoxing across provincial boundaries. This sample structure adequately captures the diversity of career stages, disciplinary backgrounds, and regional origins, providing a solid empirical foundation for the grounded theory analysis.
Table 1. Demographic characteristics of interview participants (N = 7).
Demographic Indicator |
Category |
Frequency (n) |
Proportion (%) |
Gender |
Male |
2 |
28.6 |
Female |
5 |
71.4 |
Age Group |
30 - 39 years |
2 |
28.6 |
40 - 49 years |
5 |
71.4 |
Degree |
Doctorate |
7 |
100 |
Academic title |
Professor |
1 |
14.3 |
Associate Professor |
3 |
42.9 |
Lecturer |
3 |
42.9 |
Occupation |
University faculty |
7 |
100 |
Place of birth |
Anhui Province |
1 |
14.3 |
Heilongjiang Province |
1 |
14.3 |
Liaoning Province |
1 |
14.3 |
Shandong Province |
2 |
28.6 |
Sichuan Province |
2 |
28.6 |
Years since moving to Shaoxing |
Within 5 years |
7 |
100 |
Source: Compiled by the author based on interview data.
2.2.5. Research Area
The research was conducted in Shaoxing City, Zhejiang Province, a city adjacent to Shanghai and bordering Hangzhou. In 2024, Shaoxing’s total economic output ranked 33rd nationwide and 4th in Zhejiang Province, and the total number of high-level talents introduced in the preceding five years ranked first in Zhejiang Province (Shaoxing Municipal Bureau of Statistics, 2025). Shaoxing is therefore a representative case of talent introduction and cultivation among medium-sized node cities in the Yangtze River Delta (YRD), making it an appropriate case for studying the influencing factors and mobility paths of high-level talent agglomeration.
3. Analysis and Theoretical Framework Construction
This study strictly follows the analytical procedure of Constructivist Grounded Theory (Charmaz, 2014) and conducts a systematic analysis of the seven raw interview transcripts in accordance with the progressive logic of open coding → axial coding → selective coding. The constant comparative method was applied throughout the entire research process to refine categories and identify interrelationships. To clarify coding metrics (see Tables 2-4): 1) Initial open codes (N = 49): discrete conceptual labels attached to individual meaningful text segments; multiple open codes can be applied to one text excerpt. 2) Frequency (Freq): number of coded text segments linked to each category. 3) “No. of Open Codes” in Table 4: total distinct initial open-code labels subsumed under each main category, not text-segment counts. Initial-code labels and text-segment frequencies are distinct metrics, explaining numerical divergence between “No. of Open Codes” and frequency values. The results of each coding round were subjected to multiple rounds of cross-validation and theoretical saturation tests to ensure that all categories, interrelationships, and theoretical constructs were fully grounded in the raw data, free from subjective presuppositions. Theoretical saturation assessment was performed on two randomly selected full interview transcripts that had not contributed to initial category generation. Analysts conducted full open and axial coding for these two transcripts. No substantively new conceptual categories, properties, or relational dimensions emerged; only additional instances of already-established categories were identified. This supported preliminary theoretical saturation for the present dataset.
3.1. Open Coding
Open coding is the foundational step of grounded theory analysis. It deconstructs raw interview transcripts into discrete analytical units through three sequential steps: labelling, conceptualisation, and initial categorisation.
The study followed a strict, pre specified procedure. First, the researcher conducted a line by line review of the 82,000 character anonymised interview transcripts, excluding greetings, repetitive statements, and topic irrelevant content. A total of 359 valid statements related to high level talents’ career choice, geographic mobility, agglomeration in Shaoxing, and urban integration were screened. Second, the researcher assigned initial labels to each valid statement, strictly conforming to the original semantic meaning and avoiding subjective interpretation. After consolidating semantically identical labels, 49 initial codes were retained. Third, through constant comparison, these 49 initial codes were further abstracted into 17 initial conceptual categories based on semantic relevance and logical subordination (see Section 3.2). Fourth, these 17 categories were merged into 8 main categories (axial coding) through iterative comparison. Supplementary coding of two randomly selected transcripts generated no new initial concepts or categories, confirming that the data met preliminary theoretical saturation criteria.
Table 2 presents a representative sample of open codes, their frequencies, and illustrative raw data excerpts. The full list of 49 codes is available in the supplementary materials.
Table 2. Representative open codes with frequencies and raw data anchors.
Open Code |
Frequency |
Illustrative Raw Data Anchor |
Personal trait: decision-making autonomy |
3 |
“My parents wanted me to study medicine, but I didn’t listen. I decide my own life.” |
Significant others (teacher) guiding role |
2 |
“That teacher was very important in my life… he gave me advice on college applications.” |
Growth environment: quality basic education |
2 |
“In our school, the top two students in Mianyang city were both in my grade.” |
Career planning: lack of long-term planning |
2 |
“I never had a clear plan. I just took things one step at a time.” |
Career choice: professional match |
2 |
“The main reason was the major. I wanted to be an English teacher, so I chose a teaching job.” |
Household registration (Hukou) system restricting education choice |
1 |
“There were no private schools. To study in the city, you needed an urban Hukou.” |
Seeking new environment motivation |
1 |
“I had stayed in Hefei for seven years. I thought it was time to move on.” |
Urban livability influencing job location choice |
1 |
“The city feels comfortable, not too crowded, much better than Hangzhou’s crowded buses.” |
Career development stagnation and unfulfilled promises |
1 |
“After I got my PhD, the university didn’t give me a permanent position. They just kept delaying.” |
Expectation-reality gap: actual salary much lower than expected |
1 |
“The salary was far lower than I imagined—a huge gap.” |
Source: Compiled by the author based on interview data.
3.2. Axial Coding
Axial coding builds on the results of open coding. The study applied constant comparison to identify inherent logical correlations among the 17 initial conceptual categories and constructed a hierarchical category system of main categories and sub categories. To ensure groundedness, each sub category was assigned one or two condensed original interview utterances as data anchors.
The analysis yielded 8 main categories and 17 sub categories. To save space in the main text, Table 3 presents only representative coding entries.
Table 3. Axial coding: Main categories, sub categories, frequencies, and raw data anchors.
Main Category |
Sub-category |
Freq. |
Core Connotation |
Theoretical Role |
Sample Raw Data Anchor |
Early Life Foundation |
Personal growth background and educational experience |
45 |
Family origin, urban-rural household registration, basic education quality, early migration experience |
Initial shaping variable of human capital and geographical vision |
“There was no primary school in our village… it is 15 km from the county.” |
Educational Advancement Path |
Higher education and career decisions |
45 |
University and major choices at undergraduate, master’s and doctoral levels, motivation for further study |
Capital accumulation mechanism determining professional eligibility and development boundaries |
“I really disliked the Business English major and planned to drop out and go home in October.” |
Personal Orientation |
Personal traits and values |
19 |
Decision autonomy, risk preference, work-life balance values |
Subjective moderating mechanism for differentiated interpretation of opportunities |
“I never had a proper plan.
I came to Zhejiang in a muddle
and have stayed here for ten
years.” |
Relational Tug |
Family and emotional factors |
23 |
Children’s education, spousal relocation, parental care, marital status |
Core constraint variable narrowing
the scope of employment location choices |
“The child is absolutely the most important factor… that’s why we
came to
Shaoxing.” |
|
Social capital and interpersonal networks |
20 |
Information and opportunity support from alumni, teachers and colleagues |
Boundary-shaping resource affecting job search scope
and post-entry sense of belonging |
“Several of my classmates have worked in Shaoxing for many years and we are very close.” |
Source: Compiled by the author based on interview data.
3.3. Selective Coding
Selective coding is the final phase of grounded theory analysis. Its core objectives are to identify the core category from the eight main categories, clarify the inherent logical relationships between the core category and the remaining categories, and construct a theoretical model that systematically explains the research phenomenon.
3.3.1. Identification of the Core Category
Following the three criteria of grounded theory—centrality, intercorrelation, and explanatory power—this study identifies Career Opportunity Choice as the core category. Centrality: All other seven main categories revolve around this core. Early Life Foundation and Educational Advancement Path serve as its initial conditions; Personal Orientation as its moderating variable; Relational Tug as its boundary intervention; Local Attraction Conditions and Policy & Economic Guarantee as its spatial and feasibility conditions; and Post Entry Retention Experience as its outcome and feedback. Intercorrelation: Career Opportunity Choice has direct logical correlations with all other seven categories. Raw data contain 57 utterances linking it to other categories—the highest among all categories. Explanatory power: It directly answers the core research question: why and how high level talents choose Shaoxing, and whether they stay after choosing it. Table 4 presents the eight selective coding categories (main categories) identified thorough axial coding, together with their frequencies, number of open codes, proportions, definitions, and the core category relationships established through selective coding.
Table 4. Selective coding: Main categories, frequencies, definitions, and associated axial sub categories.
Selective Code (Main Category) |
Freq. |
No. of Open Codes |
Proportion (%) |
Definition |
Associated Axial Sub-categories |
Early Life Foundation |
45 |
44 |
12.53 |
Family background, urban-rural origin, basic education quality, early migration, Hukou constraints |
Personal growth background and educational experience |
Educational Advancement Path |
75 |
73 |
20.89 |
Processes of accumulating human capital through bachelor, master, doctoral, overseas training, and professional interests |
Higher education and career decisions; Overseas experience and cross-cultural perception; Professional interest and major choice |
Personal Orientation |
19 |
17 |
5.29 |
Stable psychological tendencies, values, decision styles, and life attitudes formed during growth |
Personal traits and values |
Relational Tug |
62 |
61 |
17.27 |
Pulling, limiting, or supporting effects of family members, partners, supervisors, teachers, peers on talent mobility |
Significant others’ influence; Family and emotional factors; Social capital and interpersonal networks |
Career Opportunity Choice (Core) |
57 |
59 |
15.88 |
Process of identifying, comparing, accepting, or rejecting different job and city opportunities |
Career paths and mobility; Employment decision criteria; Chance and contingency factors |
Local Attraction Conditions |
41 |
41 |
11.42 |
Attractiveness or repulsiveness of the employment location as a spatial place, including livability, transport, public services, climate, culture, and regional advantages |
Job location attributes and geographical considerations |
Policy & Economic Guarantee |
14 |
19 |
3.90 |
Economic support and practical guarantees provided by government, employers, and local talent schemes |
Policy and economic incentives; Economic status and salary perception |
Post-Entry Retention Experience |
46 |
52 |
12.81 |
Continuous assessment of city life, work atmosphere, career development, policy fulfillment, and future mobility intentions |
Post-entry adaptation and experience; Career environment and development perception;
Post-decision evaluation and
future outlook |
Source: Compiled by the author based on interview data.
3.3.2. Logical Relationships among Main Categories
Centred on Career Opportunity Choice, the study identified six core logical relationships, each supported by at least two original utterances.
Causal/Conditional (Early Life Foundation → Career Opportunity Choice)
Anchor: “I held a rural Hukou until university… there were no private schools, and you had to have an urban Hukou to study in the city.” Moderating (Personal Orientation → Career Opportunity Choice) Anchor: “I never seemed to have a proper plan… I came to Zhejiang in a muddle.” Constraint (Relational Tug → Local Attraction Conditions) Anchor: “The child is absolutely the most important factor… that’s why we came to Shaoxing.” Deterministic Causal (Significant Others’ Influence → Career Opportunity Choice) Anchor: “This job was not my own idea… in the end, it was my doctoral supervisor who decided.” Mediating (Employment Decision Criteria ↔ Local Attraction Conditions → Career Opportunity Choice).
Anchor: “I looked for universities with Spanish language or Latin American research directions—that was the best match.” Feedback Loop (Career Opportunity Choice→ Post Entry Retention Experience → Employment Decision Criteria) Anchor: “If I had another chance, I would probably stay in Zhejiang, but definitely not at Shaoxing University.”
The identified core category and logical relationships collectively form the basis for a dynamic theoretical model of high-level talents’ geographic mobility. The full theoretical model, along with its visual representations and empirical findings, is presented in Section 4.
3.3.3. Construction of the Theoretical Model
Through iterative comparison among the eight main categories and their logical relationships identified above, a core theoretical logic emerged: high-level talents’ geographic mobility is not a one-time static choice but a sequential and dynamically cyclical decision-making process. This process can be analytically disaggregated into five interlocking stages: 1) Initial Endowment—shaped by Early Life Foundation and Personal Orientation; 2) Multi-dimensional Trade-off—matching Employment Decision Criteria with Local Attraction Conditions; 3) Relational Boundary Intervention—imposed by Relational Tug (family constraints and significant others’ influence); 4) Career Opportunity Choice Realisation—the actual mobility decision; and 5) Post-Entry Evaluation & Dynamic Feedback—where Post-Entry Retention Experience reshapes future decision criteria. This feedback mechanism resonates with scholarship on organisational entry and psychological expectation-reality gaps, whereby unmet pre-entry expectations reshape subsequent behavioural intentions (Morrison, 2002; Wanous et al., 1992). The theoretical model is visually summarised in Figures 1 and 2 (see Section 4). The following section reports the core empirical findings that support and elaborate this model.
4. Results
Based on the three-level coding analysis and theoretical model construction, this study yields four core findings that fully reveal the influencing factors, decision-making logic, and dynamic evolution of high-level talent agglomeration in Shaoxing. Before detailing these findings, Figure 1 presents the overall theoretical conceptual framework, illustrating the central role of “Career Opportunity Choice” and its relationships with other main categories.
This framework centres on “Career Opportunity Choice” as the ultimate core category. The left side shows Early Life Foundation, Educational Advancement Path, and Personal Orientation, representing pre-decision life-course conditions, capital accumulation, and subjective moderating factors. The top shows Relational Tug, emphasising how supervisors, family, partners, and social networks reshape choice boundaries. The right side shows Local Attraction Conditions and Policy & Economic Guarantee, representing Shaoxing’s spatial attractiveness and institutional support. The bottom-right shows Post-Entry Retention Experience, indicating that initial choice must be continuously tested through work, life, policy, and relational integration. Solid arrows represent main influence paths, and the dashed arrow represents the feedback loop from Post-Entry Retention Experience to subsequent Career Opportunity Choices.
![]()
Figure 1. Theoretical conceptual framework: Multi-embedded career-place matching. Source: Compiled by the author based on interview data.
4.1. The Mobility of High-Level Talents to Shaoxing Is a
Three-Stage Progressive Process of “Initial Endowment
Shaping → Multi-Dimensional Trade-Off → Relational
Network Constraint”
The agglomeration of high-level talents in Shaoxing is not a one-time random decision but a structured process with progressive layers.
Stage 1: Initial endowment shaping (45 utterances, 12.53%): Early urban-rural background, basic education quality, and Hukou constraints not only determine human capital levels but also profoundly shape geographical vision and career preferences. Talents from rural backgrounds with restricted early mobility are more inclined to move to economically developed regions such as the Yangtze River Delta. Shaoxing’s location advantage precisely fits this demand.
Stage 2: Multi-dimensional trade-off (41 utterances on local attraction, 57 on career choice): With Personal Orientation as the moderating variable, talents seek the optimal match between five dimensions of employment decision criteria (economy, life, policy, career, interpersonal relations) and five job location attributes of Shaoxing (urban livability, location advantage, public services, industrial environment, cultural atmosphere). The matching degree is the core prerequisite for choosing Shaoxing.
Stage 3: Relational network constraint (62 utterances on relational tug): The seemingly rational trade-off process is constantly subject to the rigid constraints of family and emotional factors (23 utterances) and the external intervention of significant others’ influence (19 utterances). Family factors often have “veto power”, and recommendations from authoritative figures such as supervisors can even directly determine the final decision. One respondent stated: “This job was not my own idea… in the end, it was my doctoral supervisor who decided.” The final mobility choice is the result of the mutual game between individual rationality and relational rationality.
4.2. Non-Economic Factors Are the Core Differentiated
Competitive Advantage of Shaoxing, Observed within This Interview Sample,
Existing studies mostly regard economic benefits and salary subsidies as the core drivers of high-level talent mobility. However, within the present interview data, non-economic attributes appear particularly salient in participants’ accounts, even though utterance frequency counts cannot establish definitive causal dominance relative to economic drivers.
On the one hand, non-economic attributes of Shaoxing—urban livability, core location advantage in the Yangtze River Delta, efficient public services, and profound cultural atmosphere as the “Hometown of Celebrities”—constitute its differentiated competitive advantages over core cities such as Hangzhou and Shanghai. Most respondents stated that the core reason for choosing Shaoxing was not higher salary but its “moderate urban scale, low living pressure, livable climate and environment, and close proximity to Hangzhou”, enabling a “balance between career development and quality of life”.
On the other hand, family and emotional factors serve as the core non-economic driver (23 utterances, 6.41%). More than half of the respondents prioritised children’s education, spousal employment, and parental support over economic benefits and career development when choosing an employment location. Shaoxing’s high-quality basic education resources, relatively lenient Hukou settlement policies, and stable living environment precisely fit the family needs of high-level talents and have become a key determinant of their choice.
Axial coding confirms that Local Attraction Conditions (41 utterances, 11.42%) and Relational Tug (62 utterances, 17.27%) far outweigh Policy & Economic Guarantee (14 utterances, 3.9%). One associate professor noted: “I gave up a higher-paying position in another city because Shaoxing offers a better balance between career development and family life.”
4.3. Influence of Significant Others Is an Important Implicit
Channel for High-Level Talent Agglomeration in Shaoxing,
and the Role of Academic Networks in Talent Introduction
Is Urgently to Be Tapped
The influence of significant others (19 utterances, 5.29% of total) operates as a powerful non-market channel for talent agglomeration—a factor that has received limited attention in existing studies.
Interview data show that most respondents were significantly influenced by teachers and supervisors at key career nodes such as further academic study and employment. For some respondents, the final employment location and employer were completely determined by their doctoral supervisors, where “trust-based authority guidance” outweighed the individual’s independent trade-off. One lecturer stated: “This job was not my own idea… in the end, it was my doctoral supervisor who decided.” Another respondent noted: “My master’s and doctoral supervisors both believed that staying in Zhejiang would be better for my African studies career.”
This finding reveals an “academic circle effect” in the mobility of high-level talents in universities. The academic networks of supervisors and professors in Shaoxing’s local universities serve as a critical implicit resource for talent introduction. Compared with market-oriented recruitment channels, the supervisor-recommendation approach has a higher success rate and stronger talent retention stability. However, at present, this implicit channel has not been fully utilised in Shaoxing, and a systematic incentive mechanism targeting academic networks has not yet been formed.
4.4. The Agglomeration of High-Level Talents in Shaoxing Is a
Dynamic Cyclical Closed-Loop Process, Where Post-Decision
Evaluation Directly Shapes Self-Reported Turnover Intentions, and the Expectation-Reality Gap Is the Core Trigger of Subjective Intentions to Leave
Existing studies have mostly focused on the “attracting talents” link while paying insufficient attention to the subsequent evolution of “retaining talents”. This study finds that the mobility of high-level talents to Shaoxing is not the end of decision-making but the starting point of a dynamic cyclical process.
After moving into Shaoxing, talents compare their actual work and life experiences with their prior decision-making criteria and psychological expectations, forming a Post-Entry Retention Experience (46 utterances, 12.81%). When the matching degree is high, talents develop high satisfaction and generate long-term retention intention. When a huge gap exists, talents experience disappointment and regret, which in turn breed new mobility intentions and initiate the next round of the decision-making cycle.
Interview data show that unfulfilled policy commitments, salary and benefits inconsistent with expectations, lower-than-expected career development space, and inefficient administrative services are the core causes of the expectation-reality gap, as well as the main triggers of participants’ self-reported intention to leave Shaoxing. One respondent stated: “The actual salary was far lower than I imagined—a huge gap.” Another associate professor concluded: “If I had another chance, I would probably stay in Zhejiang, but definitely not at Shaoxing University.”
This finding indicates that the core pain point of Shaoxing’s talent work lies not in “difficulty in introducing talents” but in “difficulty in retaining talents”. The “last mile” of policy implementation, as well as post-employment services and development support for talents, are the key determinants of talent retention.
To further visualise the sequential and cyclical nature of talent agglomeration, Figure 2 illustrates the complete mobility path from initial conditions to retention feedback.
Figure 2. Mechanism/Path diagram: From early conditions to retention feedback.
This diagram takes time and process as the main thread, showing the full path from early conditions to final retention. The path begins with initial conditions (family background, urban-rural origin, educational resources). It then proceeds to high-level employment capital formation (bachelor, master, doctoral, overseas experience, professional training). Next, personal orientation, family relationships, and social networks shape the employment choice boundary. The core action is Career Opportunity Choice—judging job matching, organisational platform, and development prospects. Afterwards, Shaoxing’s livability, transportation, and public services constitute local matching, while salary subsidies, housing guarantees, and children’s education constitute policy feasibility. Finally, the retention feedback stage: based on post-entry satisfaction, belonging, and development experience, talents decide to stay or re-mobilise. The diagram emphasises that talent agglomeration is not a one-time attraction but a continuous process of matching and feedback adjustment.
5. Discussion and Conclusion
This study takes seven doctoral-level high-level talents who have migrated to Shaoxing City as research subjects. Using constructivist grounded theory, the study conducted three-level coding on semi-structured interview data and constructed the Multi-dimensional Trade-off and Dynamic Feedback Theory of High-Level Talent Geographical Mobility. The theory systematically explains the influencing factors, decision-making logic, and dynamic evolution of high-level talent agglomeration in Shaoxing, a medium-sized node city in the Yangtze River Delta (YRD).
5.1. Summary of Core Findings
It is found that high-level talents’ geographical mobility to Shaoxing is a staged, dynamically cyclical process that follows a five-stage closed-loop logic: Initial Endowment (Early Life Foundation + Personal Orientation) → Multi-dimensional Trade-off (Employment Decision Criteria vs. Local Attraction Conditions) → Relational Boundary Intervention (Relational Tug: Family Constraints + Significant Others’ Influence) → Career Opportunity Choice Realisation → Post-Entry Evaluation & Dynamic Feedback (Post-Entry Retention Experience feeding back into decision criteria).
Three core findings emerge from the analysis:
Non-economic factors constitute Shaoxing’s core differentiated competitive advantage within this sample. Local Attraction Conditions (41 utterances, 11.42%) and Relational Tug (62 utterances, 17.27%) far outweigh Policy & Economic Guarantee (14 utterances, 3.9%) in influencing talent agglomeration. Talents prioritise urban livability, work-life balance, family needs, and supervisor recommendations over pure salary incentives.
The influence of significant others operates as a key implicit channel for talent introduction. Supervisors’ and teachers’ recommendations—particularly from doctoral supervisors—can directly determine final employment locations. This “academic circle effect” represents an underutilised resource for Shaoxing’s talent attraction strategies.
The expectation-reality gap is the core trigger of self-reported talent-leaving intention. Post-Entry Retention Experience (46 utterances, 12.81%) reveals that unfulfilled policy promises, salary shortfalls, limited career development, and bureaucratic inefficiencies generate disappointment, leading to renewed mobility intentions. The core pain point shifts from “difficulty in attracting talents” to “difficulty in retaining talents”.
5.2. Theoretical Contributions
This study makes three key theoretical contributions to the field of high-level talent mobility research.
First, it breaks the static analytical perspective of extant talent mobility research. Most existing studies treat talent mobility as a one-time decision driven by economic incentives or city characteristics. By contrast, this study develops a five-stage dynamic closed-loop model (Initial Endowment → Multi-dimensional Trade-off → Relational Boundary Intervention → Career Opportunity Choice → Post-Entry Feedback Loop) that captures the full decision-making and re-decision-making cycle of talent mobility. The model explains not only why talents choose Shaoxing but also why they stay, report regret, or express intentions to leave after entry.
Second, grounded in China’s local context, this study bottom-up reveals the core role of non-economic factors and relational networks in talent mobility within the studied sample. While existing quantitative studies have emphasised salary, subsidies, and economic returns, our qualitative analysis demonstrates that family responsibilities (children’s education, spousal employment, parental care) and authoritative guidance from supervisors often override purely economic calculations. This finding challenges the over-emphasis on economic factors in current talent mobility literature and deepens scholarly understanding of Chinese high-level talents’ decision-making logic.
Third, this study focuses on a previously understudied geographical context—medium-sized node cities in the YRD. Most talent mobility research concentrates on either core first-tier cities (Beijing, Shanghai, Hangzhou) or underdeveloped western regions. Shaoxing, as a typical medium-sized city with economic prosperity but limited high-end academic resources, fills an important research gap. The theoretical model provides a reference framework for studying talent agglomeration in similar non-core cities across China.
5.3. Practical Implications for Shaoxing and Similar Medium-Sized
Cities
Based on the theoretical model and core findings, the study proposes four dimensions of policy optimisation for talent attraction, cultivation, retention, and utilisation.
5.3.1. Precise Talent Introduction: Building a Differentiated,
Non-Economic Competitive Brand
Consistent with the patterns observed in this interview sample, Shaoxing may shift from over-reliance on salary subsidies to fully leveraging its non-economic advantages as a differentiated competitive strategy. A differentiated talent introduction brand should be built around three pillars: 1) YRD location advantage—proximity to Hangzhou and Shanghai, enabling access to core resources while avoiding megacity congestion; 2) Jiangnan livability—moderate urban scale, low living pressure, pleasant climate, and efficient public services; 3) profound cultural heritage as the “Hometown of Celebrities”—a culturally rich yet affordable environment. This brand targets talents who prioritise quality of life, work-life balance, and family needs, forming genuine differentiation from megacities.
5.3.2. Systematic Talent Cultivation: Unlocking Career Bottlenecks
through Full-Cycle Support
Given that Post-Entry Retention Experience directly determines long-term retention, Shaoxing must establish a full-cycle development support system. First, implement a hierarchical training system for different academic ranks: for early-career doctoral lecturers, launch a “Special Research Start-up Program” with senior academic mentors to accelerate adaptation to university research and teaching; for mid-career associate professors and professors, build cross-disciplinary academic exchange platforms and support applications for national/provincial research projects. Second, leverage Shaoxing’s proximity to Hangzhou and Shanghai to establish collaborative talent training mechanisms with top YRD universities, supporting local talents to conduct academic visits, joint research, and doctoral co-supervision at Zhejiang University, Fudan University, etc., thereby compensating for the shortage of local high-end academic resources.
5.3.3. Dedicated Talent Retention: Narrowing the Expectation-Reality
Gap
The expectation-reality gap is the core trigger of self-reported turnover intention among interviewed participants. Consistent with organisational-entry and met-expectation scholarship (Wanous et al., 1992; Morrison, 2002), unmet pre-entry expectations erode work-related satisfaction and trigger individuals’ re-evaluation of their current employment choices. To close this gap, Shaoxing must take two actions. First, strengthen rigid constraints on policy implementation. Establish a clear accountability system for talent policy delivery, specifying responsible departments, processing timeframes, and assessment standards for each policy (settlement allowance, housing subsidy, children’s school enrollment, spousal employment). Simplify application procedures, shorten fulfillment cycles, and eliminate unfulfilled promises and redundant bureaucratic trips. Second, establish a full-cycle talent service system and a regular talent communication mechanism (e.g., quarterly symposia) to timely identify and resolve work and life difficulties, enhancing talents’ sense of belonging and satisfaction.
5.3.4. Efficient Talent Utilisation: Building Person-Job Fit Platforms
To ensure that introduced talents can fully contribute to Shaoxing’s development, the study recommends building industry-driven talent development platforms. Focusing on Shaoxing’s leading industries (e.g., textile, chemical manufacturing, integrated circuit, bio-medicine), promote in-depth industry-university cooperation by establishing industrial research institutes and key laboratories. Guide university talents to conduct targeted research aligned with local industrial needs, achieving precise matching between talents’ professional strengths and regional demands. Additionally, break down institutional barriers to talent mobility and promote two-way flow between Shaoxing’s universities, research institutes, and enterprises—supporting university talents to take temporary enterprise positions and encouraging enterprise experts to serve as industry mentors in universities.
5.4. Research Limitations
This study has several limitations that should be acknowledged.
First, sample size and diversity. Limited by the inherent scarcity of doctoral-level high-level talents, our sample includes only seven respondents, all employed in a single municipal public undergraduate university in Shaoxing. While theoretical saturation was achieved, the sample does not include talents from other industries (e.g., enterprises, hospitals, research institutes) or different types of cities. Furthermore, snowball recruitment from a single municipal university is likely to have amplified the prominence of academic network influences, family-related mobility trade-offs, and university-specific organisational experiences in the findings. These patterns may not generalise to doctoral-level talents in enterprises, medical institutions or independent research institutes. Future research should expand the sample size and include diverse occupational sectors to improve generalisability.
Second, single-city case study. The theoretical model was developed solely from the Shaoxing context. Although Shaoxing is a typical medium-sized node city in the YRD, the applicability of the model to other city types—such as core megacities (Shanghai, Beijing), western underdeveloped cities, or specialised industrial cities—requires further comparative validation.
Third, purely qualitative methodology. This study relies exclusively on qualitative data and grounded theory analysis. While this approach is appropriate for theory building, it does not quantify the relative weights of different influencing factors or test causal paths statistically. Future studies should combine quantitative methods (e.g., questionnaire surveys, structural equation modelling) to verify and optimise the theoretical model. Additionally, aligned with constructivist grounded theory epistemology, the researcher’s academic background in talent policy research and shared institutional affiliation with all participants may have shaped interview rapport, response patterns, and interpretive framing. Reflexive memos were maintained throughout the process to document and mitigate this potential bias.
Fourth, lack of longitudinal data. The study captures talents’ decision-making at a single point in time. Although the dynamic feedback loop is theorised, the study does not have longitudinal data tracking talents’ actual retention or further mobility over three to five years. Future research could adopt a longitudinal design to empirically observe how post-entry satisfaction evolves and influences long-term retention.
5.5. Directions for Future Research
Building on the findings and limitations of this study, five future research directions are proposed.
First, cross-city comparative studies. Compare talent agglomeration mechanisms across different city types: core megacities (Beijing, Shanghai), medium-sized YRD node cities (Shaoxing, Jiaxing, Zhenjiang), western provincial capitals (Chengdu, Xi’an), and specialised industrial cities (Shenzhen, Suzhou). Such comparisons would test the boundary conditions of our theoretical model.
Second, mixed-methods validation. Design a large-scale questionnaire survey based on the eight main categories identified in this study, and use structural equation modelling to quantify the path relationships among Initial Endowment, Multi-dimensional Trade-off, Relational Tug, Career Opportunity Choice, and Post-Entry Retention Experience. This would combine the strengths of qualitative theory building and quantitative hypothesis testing.
Third, differentiated typology of talents. High-level talents are not homogeneous. Future research should distinguish among returnee PhDs, domestic PhDs, enterprise R&D talents, university faculty, and medical/educational professionals, constructing a typology-specific talent attraction model. For example, university faculty may prioritise academic networks and research platforms, while enterprise R&D talents may emphasise industry clusters and intellectual property protection.
Fourth, multi-actor perspective. The current study focuses solely on the talent’s individual decision-making perspective. Future research could incorporate the perspectives of employers (universities, enterprises) and local government talent bureaus, constructing a multi-actor interactive theory of talent agglomeration that explains how individual choices, organisational needs, and regional governance co-evolve.
Fifth, longitudinal tracking studies. Follow a cohort of introduced talents over three to five years, using repeated interviews or annual surveys to observe how policy implementation, career development, and social integration actually affect retention. This would directly test the dynamic feedback loop proposed in our theoretical model.
In summary, this study provides a dynamic, cyclical, and multi-embedded theoretical explanation for high-level talent agglomeration in Shaoxing. It demonstrates that talent attraction is not a one-time transaction but a continuous process of matching, feedback, and adaptation. For medium-sized node cities like Shaoxing, leveraging non-economic advantages, activating implicit academic networks, and closing the expectation-reality gap appear more salient than simply raising subsidy levels in participants’ decision-making accounts. The theoretical model and practical recommendations offered here serve as a foundation for both future research and evidence-based talent policy design.