Shane and Venkataraman (2000) have defined entrepreneurship as a process that includes the discovery, identification, measurement, and utilization of entrepreneurial opportunities. It is the result of the joint action of individual endogenous and external supporting factors (Lim et al., 2016). Independent entrepreneurship, as independently exploits the contents of entrepreneurship (Margolis, 2014), is widely used in entrepreneurship-related research (Clark and Drinkwater, 2010; Kwon et al., 2013; Falco and Haywood, 2016). For example, Ajzen (1991) has constructed a model of the correlation between entrepreneurial behavior and entrepreneurial willingness based on the theory of planned behavior. The model considers that the emergence of entrepreneurial behavior is influenced by the factors of individuals, such as ability and attitude, and external environment, such as social norms. In terms of internal factors, entrepreneurs identify new entrepreneurial opportunities based on their own knowledge accumulation and practical experience and then decide whether to conduct independent entrepreneurship (Zampetakis et al., 2017). For external factors, the implementation of entrepreneurial behavior is influenced by the entrepreneurial experience of family members (Giannetti and Simonov, 2009), the level of educational institutions (Hrsman and Daghbashyan, 2014), and the overall entrepreneurial atmosphere of society (Shirokova et al., 2016). In addition, Fayolle et al. (2006) believed that entrepreneurship education would affect the behavior of independent entrepreneurship of college students. For instance, entrepreneurship education can improve entrepreneurial willingness. Students specialized in science and engineering have better entrepreneurial abilities after receiving entrepreneurship education (Liu et al., 2019). The influence of different types of entrepreneurship education on entrepreneurship behavior implementation is also different. For example, there are significant differences in the tendency of choosing independent entrepreneurship between the students receiving entrepreneurship theory education and those who receive entrepreneurship practice training (Panagiotis and Dimo, 2014).

Entrepreneurship education refers to the theory that teachers pass entrepreneurial knowledge and skills to the students (Gorman et al., 1997) and practice training in the hands-on activities (Rideout and Gray, 2013). The impact of entrepreneurial education was and still is everywhere, from an entity such as enterprises (Dana et al., 2021) to an abstract concept such as economic transition (Rachwal et al., 2016). Theory-based entrepreneurship education plays a fundamental role in entrepreneurship education (Fiet, 2001; Salamzadeh et al., 2014; Gorostiaga et al., 2019), which is helpful to improve the rate of independent entrepreneurship of college students (Zhao et al., 2005; Vukmirović, 2019), entrepreneurial ability (Detienne and Chandler, 2004), and entrepreneurial willingness (Wannamakok and Liang, 2019; Botha, 2020). Entrepreneurial practice can avoid the disconnection between entrepreneurial knowledge and entrepreneurial practice (Salamzadeh et al., 2013; An and Xu, 2021). Besides, entrepreneurial practice has a significant impact on the improvement of entrepreneurial ability (Morris et al., 2013). Both types of education can be regarded as taking a class in person or online (Radović-Marković et al., 2009), while the practice training features various formalities such as street vending, entrepreneurial contest, and internship. Durations for theory-based courses and practice-based courses are not strictly stipulated, but theory-based courses last generally longer for most Chinese universities. The ultimate purpose of a successful course design for college entrepreneurship education should be at satisfying industrial expectations (Wu and Chen, 2019).

There has been an important and long-term debate over entrepreneurial opportunities whether they should be recognized as being objective or subjective. The discovery views believed that opportunity is an objective existence that needs to be discovered (Hayek, 1937; Casson, 1982; Krueger, 2000; Shane and Venkataraman, 2000). The creation views suggested that entrepreneurial opportunity was a subjective procedure generated in the entrepreneurial interaction (Schumpeter, 1934, 1942; Alvarez and Barney, 2013). Despite the disagreement between the discovery views and creation views, a consensus has been reached to define entrepreneurial opportunities, i.e., the phrasing may vary, and entrepreneurial opportunities should be competitive situations in which goods and services can be introduced and sold at greater than their cost of production (McBride and Wuebker, 2021). Under the perspective of the creation viewpoint, every opportunity was supposed to be constructed, which was spontaneously exploitable to the entrepreneur, regardless of the identification procedure. Thus, entrepreneurial opportunity identification could only be linked to the discovery or recognition. The core of entrepreneurial opportunity identification was associated with information processing that was shaped by personally relevant factors such as knowledge acquired, growth environment, and innovation sparks (Vaghely and Julien, 2010). When opportunity got perceived, the ability of entrepreneurial opportunity identification directly affects the result of entrepreneurship (Krueger, 2000). To a certain extent, they affect the achievement of entrepreneurship goals or entrepreneurial benefits (Garud and Giuliani, 2013).

Entrepreneurial experience refers to the entrepreneurial activities carried out by entrepreneurs before the establishment of a new enterprise (Stuart and Robert, 1990). It not only helps individuals to acquire entrepreneurial knowledge and skills from past experience (Tornikoski and Newbert, 2007), as well as master valuable market and product information (Clarysse et al., 2013), but also helps entrepreneurs to identify entrepreneurial opportunities in related fields (Yu et al., 2021). Entrepreneurial experience can affect the behavior of entrepreneurs (Tihula and Huovinen, 2010), for example, it can affect the individual judgment on the feasibility of entrepreneurial opportunities (Canavati et al., 2021), give full play to the ability of entrepreneurial opportunity identification (Ucbasaran et al., 2003), and improve entrepreneurial willingness (Graevenitz et al., 2010).

The sample used in this study was obtained by the China Institute of Innovation and Entrepreneurship Education through online surveys and offline interviews for 6,424 graduates with a postgraduation period not more than 1 year of graduation. An online invitation with the quick response (QR) code was sent out to the 5,000 graduates provided by the collaborating universities, outlining the brief purpose of the study. Corresponding data could be accumulated directly. The offline interview has followed the same pattern and the same questionnaire, and the interviewees were recruited by the administrative staff on campus during alumni gatherings. A total of 1,424 pieces of written questionnaires were recovered. The questions were elaborately generated, combining the coverage of core variables and the convenience of understanding and answering.

The data collection was performed within 31 provinces across China between 2017 and 2018. The output was shown as follows. For gender, 45.3% were men and 54.7% were women. For entrepreneurial experience, 34.4% of the respondents had an experience of establishing a start-up during their studentship period. They recognized its positive inspiration. Notably, 16.3% of the respondents had declared their willingness to start their own business, as the most favorable postgraduation choice; 10.7% of sampled graduates were from the provincial capital cities or province-level municipalities; 24.2% of the surveyed sample confirmed that their parents or other immediate family members have experienced entrepreneurship; and 27.0% of samples specialized in engineering and 13.1% graduated from “Double First-class” top universities. Refer Tables 1, 2.

On the basis of a well-established scale among existing studies, combined with the results of in-depth semi-structured interviews of scholars in the field of entrepreneurship education, the questionnaire for this study was designed after three rounds of modification and testing. The variables used in the research and analyses were derived from elements that have been recognized in the Pedagogy and widely applied. These variables were constructed based on the fundamentals in the mainstream field of interest to ensure validity, attempting to offer a panorama for the study. The content has covered entrepreneurial spirit, entrepreneurship theory-based course, and entrepreneurship practice training. Except for certain demographic characteristics and options requiring special clarification, the scales used in this study were in Likert 5-score measurement. Score 1 represented strongly disagree, while Score 5 was strongly agree. All the data were statistically processed using SPSS.

Entrepreneurship theory-based course questions were designed with reference to certain publications (Hosseini and Pouratashi, 2011; Wiley and Berry, 2015). Three measurement items were included, which were “diversified types of entrepreneurship education courses,” “closely integrated with major,” and “closely integrated era trends.”

Entrepreneurship practice training questions were designed with reference to certain publications (Greefs, 1998; Dos-Santos and Spann, 2011; Zou and Zhao, 2014). Five measurement items were included, which were “exclusive funding provided,” “integrated entrepreneurial practice services provided,” “independent entrepreneurship park provided,” “exclusive off-campus practice base provided,” and “practice projects highly integrated with professional learning.”

Entrepreneurial opportunity identification question was designed with reference to the 2017 GEM report “U: whether the respondent thinks there are good opportunities for starting a business in their local area.” Entrepreneurial opportunity identification was indicated by “good entrepreneurial opportunities in the local province.”

Virtual variables were used to measure entrepreneurial experience. For the question “among all the entrepreneurial practice activities you have participated in during school, which has helped you more,” digit 1 was marked for response “start a company outside school,” and the rest was recorded as 0.

Virtual variables were also used to measure independent entrepreneurship. For the question “the most favorable postgraduation choice,” digit 1 was marked for response “independent entrepreneurship,” and the rest was recorded as 0.

In terms of variable control, referring to the relevant research on college student independent entrepreneurship, four variables were selected, including gender, family entrepreneurship experience, major type (limited to engineering), and university type (limited to “Double First-class” top universities).

The overall results derived from the test for reliability and validity were described as α = 0.921 and Kaiser-Meyer-Olkin (KMO) = 0.945, indicating that the overall scale has demonstrated decent reliability and validity. The value of α was larger than 0.7, leading to the fact that the reliability test and validity test results of each factor (refer to Table 3 for details) have all passed the internal consistency test, indicating that the reliability of each factor scale was satisfying. The KMO measure and Bartlett sphere test showed that all variables pass the test (KMO > 0.5), which met the factor analysis standard. The results of exploratory factor analysis showed that the factor load after item rotation was greater than 0.6, the combined reliability (CR) of all factors was greater than 0.7, and the average variance extracted (AVE) value of all factors was greater than 0.5. Each of the abovementioned three standards has clarified that the scale designed then utilized in the study had outstanding convergence validity.

Each variable was paired to the rest of the variables for the calculation of variance dilation factor (VIF). Every VIF was less than 5, representing that multicollinearity did not happen. The results are shown in Table 4. Entrepreneurship theory-based course, entrepreneurship practice training, entrepreneurial opportunity identification, and entrepreneurship policy were positively correlated with independent entrepreneurship.

Considering the insufficiency of Harman’s single factor test, confirmatory factor analysis was performed to exclude the homology bias. The results are shown in Table 5. Among them, the single factor latent variable model with each item had a poor fitting, which indicated that common variance deviation was not detected among the variables used in this study. By comparing the fitting degree of the five-factor model and the four-factor model, it was found that the five-factor model used in this study was better than the four-factor model. In addition, the common method variance (CMV) was tested by setting nonmeasurable potential method factors. The output was found that the fitting statistics of the model did not appear obvious optimization, which further showed that the homologous method problem of the data used in this study was controlled, and the fitting degree of the model was satisfying.

The regression coefficient of X1 Entrepreneurship theory-based course and Y Independent entrepreneurship was significantly positive (β = 0.141, p < 0.05). The effect of X2 Entrepreneurship practice training on Y Independent entrepreneurship was not significant (β = 0.053, p > 0.05). Derived from such results, hypothesis H1a was corroborated, and H1b was not supported. Refer Table 6.

By substituting each index into formula (1), formula (2), and formula (3), the analysis results of Model 2, Model 3, and Model 7 were obtained. The direct effect of X1 Entrepreneurship theory-based course on Y Independent entrepreneurship was not significant (β = 0.073, p > 0.05). X1 Entrepreneurial course had a significantly positive impact on M Entrepreneurial opportunity identification (β = 0.323, p < 0.01). M Entrepreneurial opportunity identification had a significantly mediating effect on the relationship between X1 Entrepreneurial course and Y Independent entrepreneurship (β = 0.217, p < 0.01). Consequently, both hypotheses H2a and H4a were corroborated. As hypothesis H1b was not supported, it was impossible to test the mediating effect through the step-by-step method. However, X2 Entrepreneurial practice training had a significantly positive impact on M Entrepreneurial opportunity identification (β = 0.160, p < 0.01). It was revealed that hypothesis H2b was corroborated, while H4b was falsified. Refer Table 7.

To ensure the accuracy of the test, according to the study by Zhao et al. (2010), the product of coefficients was directly tested by bootstrap analysis with 1,000 samples (Felsenstein, 1985). The results showed that the interaction between X1 Entrepreneurial course and M Entrepreneurial opportunity identification was significant (95% CI = 0.035–0.077); thus, hypothesis H4a was supported.

By substituting each index into formula (4), the analysis results of Model 4 showed that M Entrepreneurial opportunity identification had a significantly positive impact on Y Independent entrepreneurship (β = 0.259, p < 0.01). Thus, hypothesis H3 was supported. By substituting each index into formula (5), the results of Model 5 demonstrated that the interaction between M Entrepreneurial opportunity identification and W Entrepreneurial experience had a significantly positive regression (β = 0.250, p < 0.05). The hypothesis H5 was corroborated. Refer Table 8.

According to the analysis results of Models 2 and 8, each index was substituted into formula (6) to get the analysis results of Model 5. Under the moderation of W Entrepreneurial experience, the direct effect of the X1 Entrepreneurial course on Y Independent entrepreneurship was not significant (β = 0.070, p > 0.05). M Entrepreneurial opportunity identification had a significantly mediating effect on the relationship between X1 Entrepreneurial course and Y Independent entrepreneurship (β = 0.151, p < 0.01). The interaction between M Entrepreneurial opportunity identification and W Entrepreneurial experience had significantly positive regression (β = 0.247, p < 0.05). Through the stepwise test, there was a moderating mediating effect emerged, and it showed positive regulation, namely, hypothesis H6a was supported. Hypothesis H1b, which should be a prerequisite for testing the moderating mediating effect using the stepwise method, was falsified. It is naturally transpired that H6b was also falsified.

To ensure the accuracy of the test, a 1,000-sample bootstrap analysis was conducted on the coefficient product of each variable with moderated mediating effect. It was found that the interaction items of X1 Entrepreneurship theory-based course, M Entrepreneurship opportunity identification, and W Entrepreneurship experience were significant (95% CI = 0.030–0.096). Hypothesis H6a was firmly corroborated. Refer Table 9.

This study devoted three theoretical contributions to the mechanism of entrepreneurship education on college student independent entrepreneurship.

This research expanded new ideas of entrepreneurship education research. Previous studies on entrepreneurship education had shown that both entrepreneurship theory-based course and entrepreneurship practice training had exposed a positive role in promoting independent entrepreneurship. However, based on the data analysis of the questionnaire survey, this study confirmed that entrepreneurship theory-based course and entrepreneurship practice training had different impact directions. Entrepreneurship theory-based course had a significantly positive impact on independent entrepreneurship, while entrepreneurship practice training had not. But both entrepreneurship theory-based course and entrepreneurship practice training could generate a positive impact on the entrepreneurial opportunity identification. Therefore, this new research discovery expanded new ideas for the research of entrepreneurship education.

This research demonstrates that entrepreneurship education played an important role in entrepreneurial opportunity identification and indirectly promoted independent entrepreneurship. Previous studies on entrepreneurship education had majorly focused on the direct promotion of entrepreneurship education for college students (Fayolle et al., 2006). However, based on the data analysis of sampled graduates, this study finds that entrepreneurial opportunity identification played a complete mediating role between entrepreneurship curriculum education and independent entrepreneurship. Entrepreneurship theory-based course indirectly affects independent entrepreneurship through entrepreneurial opportunity identification, rather than directly promoting college students to choose independent entrepreneurship.

This research enriches the horizon on the process mechanism of entrepreneurship education affecting independent entrepreneurship. Previous studies on the mechanism of entrepreneurship education on independent entrepreneurship were mostly based on the mediating role of opportunity identification (Nidal and Norashidah, 2017) or the moderating role of entrepreneurial experience (Shirokova et al., 2015), but studies in the field of mediating mechanism were severely inadequate (Wei et al., 2019). Through investigation and exploration, this study analyzes the correlation between entrepreneurship education, entrepreneurial opportunity identification, entrepreneurial experience, and independent entrepreneurship in different dimensions and demonstrated the existence of moderated mediating effect model with entrepreneurship theory-based course as the antecedent variable, entrepreneurial opportunity identification as the intermediary variable, entrepreneurial experience as the moderating variable, and independent entrepreneurship as the dependent variable, which enriched the affecting mechanism of entrepreneurship education on independent entrepreneurship.

This study puts forward three practical implications for the mechanism of entrepreneurship education on independent entrepreneurship.

The teaching approaches of entrepreneurship theory-based course can be improved. By deepening the existing diversified entrepreneurship theory-based courses such as case analysis, exclusive lecture, group discussion, and simulation practice, online teaching of entrepreneurship theory-based course should be appropriately promoted to help different types of students find their own teaching content and obtain the relevant knowledge needed for entrepreneurship, so as to help them carry out independent entrepreneurship.

The form of entrepreneurship practice training should be expanded. On the basis of improving the existing entrepreneurial contest, on-campus entrepreneurial park practice, entrepreneurial simulation training camp, and other practice activities within the campus, off-campus practice activities should be appropriately developed, enterprise management internship, for instance. Enriched forms of practical education provided conditions for college students to carry out practical activities and help them test whether it could be plausible for entrepreneurship at the minimum cost.

The support system for college students to start their own businesses should be improved. Cooperation between the university and local government/enterprise should be further strengthened to build a set of support systems, in terms of venues, funds, personnel, and policies, capable for the student to establish off-campus start-ups, as a method of willingness promotion.

Due to limitations such as funds, time, and other factors, the proportion of samples collected in this study was not satisfactorily abundant. Since this is a cross-sectional design study, the sampled data represented only the situation of college students who have received entrepreneurship education within 1 year after graduation instead of a more common period of 3–5 years. Although the variables used in the research and analyses were dedicated generated within the field of Pedagogy, they could not completely avoid the deviation caused by individual subjective opinions. Besides, there were no controlled variables, which possibly led to the deviation even when three rounds of test modification had been carried out.

In the future, this study can further discover the relationship between variables at different time points or periods. Top priority should be given to the in-depth comparative analysis in the follow-up studies. The collected data can be further compared with general university graduates in China to conduct a comparative analysis of data grouping and subspecialty. Moreover, diverse datasets to reexamine and multiexamine the relationship of the factors on the various background graduates can further be utilized to justify the entrepreneurial opportunity results. It is also practicable to explore the impact of other personal characteristics or situational factors as moderators on independent entrepreneurship, in order to find more moderators that can affect the relationship between entrepreneurship education, entrepreneurial opportunity identification, and independent entrepreneurship. It is worthy of considering absorbing other entrepreneurial capabilities that are not involved in this study and then exploring whether entrepreneurial self-efficacy, entrepreneurial awareness, and other mediating variables are applicable in the model constructed.