Factors determining the entrepreneurial intentions among Chinese university students: the moderating impact of student internship motivation

Entrepreneurs play a critical role in economic development and youth job creation. Entrepreneurship encompasses opportunity recognition, venture launch, and post-launch management. Higher education institutions are key venues for cultivating entrepreneurial mindsets and skills, with entrepreneurship education (EE) fostering creativity, problem solving, and innovation. Entrepreneurial intention (EI) reflects the motivation to engage in entrepreneurial behavior. Motivations for entrepreneurship include autonomy, wealth, and freedom; student internships are highlighted as a mechanism to develop relevant competencies and bridge classroom learning with real-world practice. Internships help students acquire professional habits and values beneficial for employment and entrepreneurship. This study investigates Chinese college students’ entrepreneurial intentions using the Theory of Planned Behavior (TPB), testing how student internship motivation (SIM) moderates relationships between entrepreneurial attitude (EA), subjective norms (SN), perceived behavioral control (PBC), EE, entrepreneurial self-efficacy (ESE), and student entrepreneurial intentions (SEI). The COVID-19 pandemic disrupted internships globally, including in China; with restrictions easing, on-site internships have resumed, potentially offering stronger benefits than virtual formats. Prior work shows SIM directly relates to SEI, but the moderating role of SIM on TPB constructs and related factors has been underexplored. This study proposes and tests a novel model in which SIM moderates the effects of EA, SN, PBC, EE, and ESE on SEI, addressing a gap in the literature and informing the design of sustainable internship programs that foster entrepreneurship.

The study is grounded in the Theory of Planned Behavior (TPB), which posits that attitudes toward the behavior, subjective norms (SN), and perceived behavioral control (PBC) shape behavioral intentions and behavior. TPB extends the Theory of Reasoned Action by incorporating PBC to account for non-volitional control. Prior research in entrepreneurship using TPB has shown: values influencing sustainability-driven EI via attitudes and PBC; EE often influencing EI indirectly through TPB components and ESE; mixed evidence on the role of SN; and moderating/mediating roles for constructs such as psychological capital and cyber EE. Entrepreneurial attitude (EA) reflects evaluative beliefs about entrepreneurship and is linked to determination, confidence, adaptability, and passion; strong EA is associated with higher EI (H1: EA positively influences EI). Subjective norms (SN) capture perceived social pressure from family, friends, and significant others; in collectivist contexts like China, SN are salient and have been linked to EI (H2: SN positively influences EI). PBC reflects perceived capability to perform entrepreneurial behaviors and predicts EI and, in some cases, behavior directly (H3: PBC positively influences EI). Entrepreneurship education (EE), through formal programs, enhances skills and attitudes conducive to entrepreneurship and has been widely found to encourage EI (H4: EE positively influences EI). Entrepreneurial self-efficacy (ESE) is confidence in one’s entrepreneurial capabilities, distinct from PBC, and is positively related to EI (H5: ESE positively influences EI). Student Internship Motivation (SIM) is defined as students’ motivation and perceptions toward undertaking internships (rather than internship outcomes). Internships enhance employability, provide applied learning, and have been associated with higher EI. However, prior studies have not tested SIM as a moderator of the relationships between EA, SN, PBC, EE, ESE, and EI. The study proposes SIM as a moderator: H6 (SIM moderates EA→EI), H7 (SIM moderates SN→EI), H8 (SIM moderates PBC→EI), H9 (SIM moderates EE→EI), and H10 (SIM moderates ESE→EI).

Design: Quantitative survey based on TPB, testing SIM’s moderating role. Instrument: Constructs adapted from prior literature—EA and SEI (Lu et al., Ozaralli & Rivenburgh), SN (Muliadi & Mirawati), PBC (Ajzen; Vamvaka et al.), EE (Liñán et al.; Turker & Selcuk), ESE (Shahab et al.), SIM (Mensah et al.). Each construct measured with 3 items on a 5-point Likert scale (1=strongly disagree to 5=strongly agree). Items were shortened to three per construct for improved reliability and agility. The questionnaire had two sections: construct items and demographics (age, gender, education, future plans, school). Translation: English to Chinese with back-translation to ensure semantic equivalence. Pilot: Pre-tested with 100 students; feedback used to refine items; pilot data not included in final analyses. Ethics: Institutional approval obtained; participation voluntary with informed consent; anonymity and confidentiality assured. Sampling and data collection: Convenience sampling of college students at Jiangxi University of Science and Technology (Ganzhou, Jiangxi, China). Online survey distributed via WeChat and QQ. Data collection period: March–April 2022. Sample: 478 valid responses (adequate given an estimated population 30,000; 95% CI; 5% margin; required minimum ~380). Demographics: 56.7% female; 43.3% male; predominantly undergraduates (60.5%) and master’s students (36.2%); age mostly 18–21; future entrepreneurship likelihood: 44.6% very likely, 32.6% unsure, 22.8% not likely. Analysis tools and procedures: SPSS 26 used for descriptive statistics, reliability (EFA via principal components, composite reliability, Cronbach’s alpha, AVE), Pearson correlations, multiple linear regression (H1–H5), and common method bias (Harman’s single-factor). Moderation analyses conducted with Hayes PROCESS for SPSS (Model 1), 5,000 bootstrap samples, 95% bias-corrected CIs; simple slopes and Johnson–Neyman tests applied. Common Method Bias: Harman’s single-factor explained variance = 20.824% (<50% threshold); no high inter-construct correlations (r>0.90), indicating CMB unlikely. Reliability and validity: All constructs met thresholds: factor loadings generally >0.77; composite reliability ≥0.832; Cronbach’s alpha ≥0.73; AVE ≥0.625 (meeting thresholds ~0.7 for reliability and ~0.5 for AVE).

Descriptives: All constructs (EA, SN, PBC, EE, ESE, SIM, SEI) were positively correlated (p<0.05 to p<0.001). Means ranged ~3.39–3.79. Common Method Bias not problematic (Harman’s 20.824%). Reliability/validity metrics were acceptable (e.g., EA AVE=0.875; CR=0.954). Main effects (Multiple Regression, H1–H5): All predictors had significant positive effects on SEI: EA (B=0.077, p<0.05), SN (β=0.094, p<0.05), PBC (β=0.099, p<0.001), EE (β=0.114, p<0.01), ESE (β=0.134, p<0.001). Model fit: R^2=0.148, Adjusted R^2=0.157, F=17.610, p<0.001. Moderation (H6–H10; PROCESS Model 1, 5,000 bootstraps): SIM significantly and negatively moderated: EA→SEI (coef=-0.1918, SE=0.0715, p=0.0076, 95% CI [-0.3324, -0.0513]); SN→SEI (coef=-0.2432, SE=0.0765, p=0.0016, 95% CI [-0.3934, -0.0929]); PBC→SEI (coef=-0.1369, SE=0.0494, p=0.0058, 95% CI [-0.2341, -0.0398]); ESE→SEI (coef=-0.1999, SE=0.0710, p=0.0051, 95% CI [-0.3393, -0.0604]). SIM did not significantly moderate EE→SEI (coef=-0.0089, SE=0.0683, p=0.8968, CI [-0.1431, 0.1254]). Simple slope analyses showed stronger positive slopes at low SIM levels than at high SIM, indicating that higher SIM attenuates the strength of EA, SN, PBC, and ESE effects on SEI. Johnson–Neyman analysis: SIM moderation remained significant up to SIM values of 3.9260 (EA), 3.8854 (SN), 3.8864 (PBC), and 4.0357 (ESE); beyond these, moderation became non-significant at p<0.05.

Findings confirm TPB-based determinants (EA, SN, PBC) and related factors (EE, ESE) as positive drivers of student entrepreneurial intentions in the Chinese university context. Enhancing students’ entrepreneurial attitudes can foster opportunity recognition, resilience, and proactive behavior. Supportive subjective norms—messages from family, friends, and key actors—can encourage entrepreneurship, especially in collectivist settings. Strong perceived behavioral control, cultivated through training and practice, bolsters intentions by increasing perceived capability. Quality entrepreneurship education positively influences intentions, underscoring the need for action-oriented, practice-integrated programs with skilled instructors and institutional support. Entrepreneurial self-efficacy strongly predicts intentions, emphasizing interventions that build confidence through mastery experiences, vicarious learning, feedback, and supportive environments. Moderation results indicate SIM significantly but negatively moderates the effects of EA, SN, PBC, and ESE on SEI, implying that at higher levels of SIM, the marginal impact of these predictors on SEI decreases. This suggests internship motivation may partially substitute or saturate the influence of attitudinal, normative, control, and efficacy beliefs when high, while at low SIM levels, these predictors are more salient for intention formation. SIM did not moderate the EE→SEI link, indicating that perceived internship motivation does not amplify or diminish the direct influence of EE on intentions, possibly reflecting a gap between EE content and real entrepreneurial experience in the context studied. Overall, results support integrating internship policies and programs with efforts to strengthen attitudes, perceived control, and self-efficacy, while recognizing nuanced interplay where strong internship motivation may attenuate reliance on these beliefs in predicting intentions.

The study demonstrates that entrepreneurial attitude, subjective norms, perceived behavioral control, entrepreneurship education, and entrepreneurial self-efficacy each positively influence Chinese college students’ entrepreneurial intentions. The novel contribution is showing that student internship motivation significantly and negatively moderates the effects of EA, SN, PBC, and ESE on intentions, but not the effect of EE. Theoretically, integrating SIM within the TPB framework extends understanding of how internship-related motivation interacts with key intention antecedents. Practically, universities and policymakers should design high-quality, standards-based internship programs in partnership with industry that complement entrepreneurship education and capability-building. Such programs should ensure applied learning, clear objectives, professional supervision, adequate resources, exposure to meetings and networks, and consideration of incentives. Future research should examine additional contextual and policy factors (e.g., family background, government entrepreneurship policies), replicate the model across countries and institutions, and explore mechanisms explaining the negative moderation (e.g., substitution or saturation effects).

Generalizability is limited due to a convenience sample from a single Chinese university; results may not extend to other populations or countries. The study focuses on a subset of determinants; other influential factors (e.g., family background, government policies) were not included. Self-report measures may introduce bias despite checks indicating limited common method variance. Cross-sectional design precludes causal inference. The authors suggest future work to investigate family background and government entrepreneurship policies as moderators, and to validate the model in other contexts.