College students' entrepreneurship policy, regional entrepreneurship spirit, and entrepreneurial decision-making

The study investigates how national and regional entrepreneurship policies influence college students’ entrepreneurial decision-making, focusing on China’s context of mass entrepreneurship and innovation. It highlights the importance of decision-making in entrepreneurial success and notes recent policy efforts by the Chinese government to support student entrepreneurship. Despite policy emphasis, misalignments and short-termism in implementation have been reported. The research addresses gaps concerning the lack of large-scale quantitative evidence on the effects of entrepreneurship policies on students’ decisions and on gender differences in these effects. The research question asks: What is the internal mechanism linking college students’ entrepreneurial policies and entrepreneurial decision-making? The study proposes that regional entrepreneurship spirit mediates the policy–decision link and that gender moderates these relationships.

The theoretical basis integrates policy tool theory and regional entrepreneurship dynamics. Entrepreneurship policies are categorized following Rothwell and Zegveld (1981) into supply-based (e.g., education investment, facilities, training), demand-based (e.g., competitions, mentoring, government procurement, market shaping), and environmentally-oriented tools (e.g., tax relief, simplified registration, finance and regulatory environment). Prior research shows entrepreneurs’ decisions are shaped by institutional and environmental conditions. The study theorizes that entrepreneurial policies positively affect student decision-making (H1a–H1c). Regional entrepreneurship spirit is conceptualized (drawing on GEM) as a regional-level construct reflecting entrepreneurial awareness (knowing recent founders), perceived opportunity, and entrepreneurial self-efficacy; it is expected to mediate policy effects on decisions (H2a–H2c). Gender differences in risk tolerance, information processing, and decision strategies suggest moderation in policy effects and in the mediation via regional spirit (H3a–H3b).

Design and sample: National cross-sectional survey (2017–2018) of undergraduate and postgraduate students with entrepreneurial practice experience (excluding first-year students in 2018). 35,340 responses collected; 1,680 removed for quality issues; 33,660 valid responses (95.25%). Data collected via online Star software.

Measures: 5-point Likert scales (1–5). Independent variables—policy tools: Supply-oriented (X1: integrated entrepreneurial practice services; pioneer park; off-campus practice bases; integration with professional learning). Demand-based (X2: usefulness of competitions; internal/external tutors; free societal training). Environmental (X3: tax reductions/exemptions; simplified registration; interest-free startup loans). Mediator—Regional entrepreneurship spirit (M: peers/classmates who started businesses in past year; provincial opportunity perception; self-assessed knowledge/skills/experience). Dependent—Entrepreneurial decision-making (Y: policies increase willingness to start; policies practically help launch).

Descriptives (means, 1–5): X1 items 3.66–3.77; X2 items 3.76–3.81; X3 items 3.76–3.82; M items 3.09–3.27; Y items both 3.89. Correlations positive among X1–X3, M, and Y.

Reliability/validity: Overall Cronbach’s α=0.928; KMO=0.957. EFA loadings >0.6; CR>0.7; AVE>0.5; discriminant validity supported (square root AVE > inter-factor correlations). CFA fit indices: RMSEA=0.072, CFI=0.971, GFI=0.946, AGFI=0.919, IFI=0.971, NFI=0.971. Common method bias: Harman’s single-factor 47.86% (<50%). Multicollinearity: VIF<10 (mostly <5).

Analysis: Moderated mediation (Wen & Ye, 2014) using SPSS with bootstrapping. Model: M = Σ a_i X_i + e; Y = Σ c_i X_i + c’_i X_i + b M + interaction terms (XW, MW) + controls. Controls: permanent residence, family entrepreneurial experience, educational resources, family entrepreneurial resources. Gender (W) moderates second stage (M→Y) and policy→Y paths. Separate tests for X1, X2, X3 with bootstrapped CIs for conditional indirect effects by gender.

• Correlations: Entrepreneurial decision-making (Y) significantly and positively correlates with supply (X1), demand (X2), environmental (X3) policies, and regional entrepreneurship spirit (M). M correlates positively with X1–X3. Men scored higher than women on relationships among policies, M, and Y.

Supply-oriented policy (Table 4):

• X1→M: a1x1=0.2694 (95% CI 0.2611–0.2776), positive.
• M→Y: b1x1=0.1520 (95% CI 0.1265–0.1775), positive.
• Direct X1→Y after M: c’1x1=0.7769 (95% CI 0.7521–0.8017), positive (partial mediation).
• Moderation: X1W→Y: −0.0407 (95% CI −0.0564 to −0.0251), negative; MW→Y: −0.0272 (95% CI −0.0430 to −0.0114), negative.
• Moderated mediation index (X1MW): 0.0151 (95% CI 0.0119–0.0179), significant. Conditional indirect effects: male 0.0336 (95% CI 0.0296–0.0376); female 0.0263 (95% CI 0.0224–0.0302). R²: M=0.4673; Y=0.5679.

Demand-based policy (Table 5):

• X2→M: a1x2=0.2747 (95% CI 0.2666–0.2828), positive.
• M→Y: b1x2=0.0931 (95% CI 0.0710–0.1152), positive.
• Direct X2→Y after M: c’1x2=0.8065 (95% CI 0.7852–0.8278), positive (partial mediation).
• Moderation: X2W→Y: −0.0036 (95% CI −0.0172–0.0099), not significant; MW→Y: −0.0195 (95% CI −0.0332 to −0.0059), significant negative.
• Moderated mediation index (X2MW): 0.0122 (95% CI 0.0092–0.0151), significant. Conditional indirect effects: male 0.0202 (95% CI 0.0168–0.0237); female 0.0148 (95% CI 0.0113–0.0183). R²: M=0.4711; Y=0.676.

Environmental policy (Table 6):

• X3→M: a1x3=0.2500 (95% CI 0.2419–0.2582), positive.
• M→Y: b1x3=0.1294 (95% CI 0.1082–0.1507), positive.
• Direct X3→Y after M: c’1x3=0.8273 (95% CI 0.8067–0.8479), positive (partial mediation).
• Moderation: X3W→Y: −0.0190 (95% CI −0.0320 to −0.0059), negative; MW→Y: −0.0242 (95% CI −0.0374 to −0.0111), negative.
• Moderated mediation index (X3MW): 0.0126 (95% CI 0.0096–0.0155), significant. Conditional indirect effects: male 0.0263 (95% CI 0.0231–0.0295); female 0.0202 (95% CI 0.0172–0.0233). R²: M=0.4597; Y=0.6891.

Overall: All three policy types positively affect entrepreneurial decision-making; regional entrepreneurship spirit partially mediates these relationships. Gender moderates the M→Y path (stronger effects for men) and moderates X1→Y and X3→Y but not X2→Y. Evidence suggests a potential “double-skin phenomenon” where some policy effects may not fully align with students’ needs, especially across genders.

The study confirms that a supportive policy environment enhances college students’ entrepreneurial decision-making and that regional entrepreneurship spirit explains part of this effect. Embedding entrepreneurship within regional social norms and networks strengthens awareness, opportunity perception, and self-efficacy, which in turn translate policy inputs into individual decisions. Gender differences show that men derive stronger decision-making benefits from supply-oriented and environmental policies, while the indirect path through regional spirit is also stronger for men. Demand-based policies show less gender-specific direct moderation but still operate through gendered differences in regional spirit effects. These findings address the research question by identifying the mediating role of regional entrepreneurship spirit and the moderating role of gender in the policy–decision nexus, underscoring the need for context-aware, inclusive policy designs tailored to different student groups and regional ecosystems.

This study provides large-scale empirical evidence (n=33,660) that supply-, demand-, and environmentally oriented entrepreneurship policies positively influence college students’ entrepreneurial decision-making, with regional entrepreneurship spirit acting as a partial mediator. Gender moderates both the direct (for supply and environmental policies) and indirect (via regional spirit) effects, revealing heterogeneous policy impacts across student groups. Contributions include integrating policy tool theory with regional entrepreneurship spirit and gender into a moderated mediation framework, offering insights into how policies translate into individual decisions. Practically, the study recommends more precise, coordinated policy implementation; streamlined administrative processes; and targeted support—especially for female students—to strengthen self-efficacy and participation. Future research should employ longitudinal designs, broaden policy dimensions, and deepen gender-differentiated analyses to optimize policy effectiveness.

The study is cross-sectional, capturing policy and decision-making relationships at one time point, which limits causal inference and temporal dynamics. Many variables rely on self-reported perceptions, raising potential response bias. Generalizability may be bounded by the Chinese higher education and policy context. Future work should include longitudinal data, diversified policy measures, stronger objective indicators, and more granular gender-differentiated analyses.