Anticipated regret means that individuals anticipate the disagreeable effects of what they might feel in case of regret and seek to protect themselves before they make a decision. Anticipated regret is defined as beliefs about feelings of regret or upset followed from inaction (Abraham and Sheeran, 2003), which refers to the negative emotional reaction that individuals experience when they imagine and compare the expected outcome of their decision of not to act with the outcome they would have experienced if they had acted (Loewenstein and Lerner, 2003). Anticipated regret is involved in behavioral regulation strategies and recognized as a motivational factor that expresses the feeling of motivating individuals to take action and hence to avoid experiencing regret arising from their failure to act. If entrepreneurs experience anticipated regret, they are encouraged to regulate this negative emotion via entrepreneurial activity to ameliorate the adverse effects of enduring regret on their psychological and physical health (Bouderbala, 2020).

The role of anticipated regret is drawn from the regret regulation theory (Pieters and Zeelenberg, 2007), which proposes that “regret is an cognitive emotion that people are motivated to regulate in order to maximize outcomes in the short-term and learn to maximize them in the long run” (Neneh, 2019). Accordingly, regret is subject to temporal reversal. In the short term, individuals regret their actions. But in the long term, they are likely to feel more intense regret over inaction if action is the norm (Pieters and Zeelenberg, 2007). Zeelenberg (1999) found that people always tend to make the choice of minimizing regret rather than minimizing risk. The result of this study solves why sometimes people tend to choose safety but sometimes they tend to take risks. The decision-maker is more willing to live in the status quo when making decisions. If the utility is reduced when choosing other schemes, an individual will regret more than reducing the same utility due to living in the status quo. As for the entrepreneurial process, entrepreneurs are more inclined to adhere to the choice of continuing the entrepreneurial process rather than giving up entrepreneurship, even if they may face the risk of entrepreneurial failure in the future.

Since the 1980s, scholars have combined the theories of cognitive psychology and social cognition to focus on the impact of entrepreneurial cognition on entrepreneurial behaviors (Jun et al., 2015). The entrepreneurial cognitive theory explains the reasons for entrepreneurial success from the perspective of cognitive psychology and holds that entrepreneurs’ emotions will affect their cognitive process and then affect their entrepreneurial behavior (Gendolla, 2000; Mitchell et al., 2002, 2007; Baron, 2004; Bird et al., 2012). Entrepreneurs’ positive or negative emotions affect entrepreneurs’ cognitive structure and cognitive process (Byrne and Shepherd, 2015). At present, the studies on entrepreneurial cognition take “situation-thinking behavior” as the basic framework to explore “how the situation affects entrepreneurs’ cognition and decision-making process and then leads to different behavior results” (Mitchell et al., 2007). Entrepreneurial cognition reflects the rational thinking process and predicts entrepreneurial behavior effectively (Buckley and Halbesleben, 2017; Randolph-Seng et al., 2020).

The researchers of entrepreneurial cognition are different from psychology and behavioral science because the uniqueness of entrepreneurial cognition comes from entrepreneurial situations rather than entrepreneurs. The theory focuses on exploring the applicability and dynamics of existing knowledge and styles about cognitive processes in entrepreneurial situations (Baron, 2004). Mitchell et al. (2002) considered that entrepreneurial cognition is the knowledge structure for entrepreneurs to make evaluation, judgment, and decision in the process of opportunity evaluation and entrepreneurial enterprise growth, including three scripts, namely, preparation script, willingness script, and ability script. The script is a highly developed ordered knowledge (He et al., 2021), which can shape individual action-based knowledge structure. Preparation script refers to the necessary factors to form a new enterprise, such as relationships, resources, and property. Willingness script refers to the commitment to risk and the perception of entrepreneurship ideas. Ability script refers to the skills, knowledge, norms, and attitudes required for entrepreneurial success (Doern and Goss, 2013; Jun et al., 2015; Randolph-Seng et al., 2015). Entrepreneurs with rich entrepreneurial cognition can encode the perceived information in a more complex way. Entrepreneurs who have sufficient entrepreneurial preparations, greater willingness to start a business, and higher levels of entrepreneurial ability can identify entrepreneurial business opportunities by piecing useful information and resources together from the experience (Mitchell et al., 2002) and respond to the problems in the process of starting a business actively.

In entrepreneurship, it is better to understand the impact of regret not as a consequence of not following plans to create a venture but as a counterfactual and anticipatory negative emotion that drives entrepreneurs to regulate their behavior and is more likely to move to the later phase of the entrepreneur’s process. The regret regulation theory postulates that individuals are regret averse and try to regulate their regrets to maximize outcomes in the long run. Anticipated regret plays an important role in transforming the original entrepreneurial actions into follow-up entrepreneurial persistence. Thus, anticipated regret manifests as a feeling for doing by motivating individuals to take action to avoid experiencing regret arising from a failure to act (Zeelenberg and Pieters, 2007). In fact, in contrast with other negative emotions, anticipated regret can help people gain insights into themselves (Kuhnle et al., 2014) and act as a behavioral push toward engaging in business activities.

Entrepreneurs with higher anticipated regret are more likely to conduct counterfactual behavior analysis and give full consideration to the consequences of giving up entrepreneurship (Hatak and Snellman, 2017). Reducing the regret over not following plans for an entrepreneurial venture drives entrepreneurs to regulate their behavior in such a way that they are more likely to proceed to persistence. By engaging in actual start-up behavior, the nascent entrepreneur regulates the regret that would arise from a failure to act, thus establishing grounds for maximizing the uncertain but socially important start-up outcomes in the long run (Hatak and Snellman, 2017). Conner et al. (2015) recognized a predictive role associated with the anticipated affect that can account for up to 18.9% of intention and 9.9% of behavior. In other words, individuals who think about entrepreneurial action positively (latent entrepreneurs), while at the same time perceiving entrepreneurial inaction as regrettable (Hird, 2007), will engage in business gestation activities to minimize future regrets, whereas those who do not associate entrepreneurial inactivity with anticipated regret are less likely to persistence. Based on the abovementioned analysis, we proposed the following hypothesis:

Entrepreneurs’ emotions affect entrepreneurial actions by adjusting their cognitive abilities and cognitive processes (Liguori et al., 2018). Anticipated regret stimulates to compile information and to orient the cognitive process. Anticipated regret constitutes a non-rational belief in a future prospect. It simultaneously triggers regretful thinking, i.e., the automatically activated cognitive representation of alternative scenarios. The regret regulation theory holds that entrepreneurs will pay more attention to the consideration of entrepreneurial details and self-improvement to reduce the anticipated regret caused by giving up starting a business, thus optimizing the entrepreneurial cognitive structure. These entrepreneurs will make more efforts to integrate all kinds of knowledge, collect all kinds of information about their industries, markets, and competitors to understand entrepreneurial risks, and seek human support, financial support, and even emotional support related to entrepreneurship in a variety of ways. In turn, these will optimize the composition of entrepreneurial cognition from three aspects, namely, preparation script, willingness script, and ability script (Mitchell et al., 2002; Randolph-Seng et al., 2015), and then better identify new products or services to ensure the sustainability of entrepreneurial activities. Accordingly, we proposed the following hypotheses:

Based on the entrepreneurial cognitive theory, entrepreneurial cognition affect entrepreneurial behaviors (Heilman et al., 2010), and different entrepreneurial cognitive structures will lead to different entrepreneurial behaviors (Baron, 2004). Entrepreneurs who have sufficient entrepreneurial preparations, greater entrepreneurial willingness, and higher levels of entrepreneurial ability will be easier to identify entrepreneurial opportunities and deal with difficulties in the entrepreneurial process (Oyson and Whittaker, 2015). With better entrepreneurial cognition, entrepreneurs are more likely to make correct entrepreneurial decisions, gain more entrepreneurial support, and obtain higher entrepreneurial performance (Foo et al., 2014), which will lead entrepreneurs to persist in starting a business (Walsh and Elorriaga-Rubio, 2019). Gong and Yang (2021) hold that the richer the knowledge structure of various resources, willingness, and abilities that exist in the minds of entrepreneurs, the higher the cognitive level of entrepreneurs, so they are more likely to carry out persistence behavior automatically. We believe that in the context of entrepreneurial persistence, the three dimensions of entrepreneurial cognition are closely related and play their respective roles. The preparation script explains the resource conditions of persistence behavior, the willingness script explains the internal driving force of entrepreneurship persistence, and the ability script explains the ability conditions of entrepreneurs to persist in entrepreneurship. Therefore, a higher level of entrepreneurial willingness, more entrepreneurial preparation, and a higher level of entrepreneurial ability will help entrepreneurial persistence (Cardon and Kirk, 2015). Accordingly, we propose the following research hypotheses:

Entrepreneurial persistence will be affected by entrepreneurial cognition, and anticipated regret will affect entrepreneurial persistence through entrepreneurial cognition. Based on the theory of entrepreneurial cognition, entrepreneurial cognition plays a mediating role between entrepreneurial emotions and entrepreneurial behaviors (Liguori et al., 2018). The negative emotions such as anticipated regret will enhance the cognitive ability of individuals to think about entrepreneurship in a complex way, constantly promote entrepreneurs to break the original mental models, and make entrepreneurs better understand the positive and negative information in the process of entrepreneurship. As entrepreneurs face many complicated entrepreneurial tasks or attractive business opportunities in the middle and late stages of in-depth entrepreneurial activities, according to the regret adjustment theory, after choosing entrepreneurship, entrepreneurs tend to adjust the composition of entrepreneurial cognitive structure through information collection, resource acquisition, and ability improvement, so as to adhere to entrepreneurial activities. Anticipated regret makes the entrepreneur consider the different details of the start-up process carefully and adhere to the entrepreneurial behavior rationally. In order to reduce the anticipated regret caused by giving up starting a business, many entrepreneurs will adopt effective strategies to regulate their behaviors (Xuejun and Chunguo, 2020), such as collecting a large amount of positive entrepreneurial information, improving entrepreneurial skills, and changing the entrepreneurial cognitive structure (Adomako et al., 2016). Those with lower anticipated regret are unlikely to obtain more entrepreneurial information and resources to persist in entrepreneurship (Hatak and Snellman, 2017). Based on this, this study puts forward the fourth hypothesis as follows:

Whether entrepreneurial activities can be carried out smoothly or not will be affected by a series of external environments such as infrastructure, natural ecology, policy support, and entrepreneurial atmosphere. The values of the entrepreneurial environment perceived by entrepreneurs are important to adhere to entrepreneurship (Peng et al., 2021). When the Internet, communication equipment, and transportation facilities are complete, the ecology is excellent, the entrepreneurial policy is supported strongly, and the entrepreneurial atmosphere is good, it will be easier for entrepreneurs to identify entrepreneurial opportunities and get more entrepreneurial materials and financial supports. Social cognitive theory holds that the entrepreneurial environment will affect entrepreneurs’ mood and behaviors (Luthans et al., 2000; Gomezelj and Kusce, 2013), and the perception of the entrepreneurial environment will directly stimulate their evaluation of the current situation and future of entrepreneurship. The richer and better the entrepreneurial environment is, the more satisfied with the external environment the entrepreneurs will be. The better the entrepreneurial environment is, the stronger the anticipated regrets are, and the more reluctant to give up continuing to start a business they will be. Therefore, even if the entrepreneurial process is complicated and difficult, in the context of a good entrepreneurial environment, they are more willing to choose to persist in entrepreneurship. The stronger their anticipated regrets are, the more willing they are to continue their entrepreneurial activities. Based on this, the following assumption is proposed:

Based on the abovementioned analysis, this study builds a theoretical model of anticipated regret, entrepreneurial cognition, entrepreneurial persistence, and entrepreneurial environment as shown in Figure 1.

A field survey approach was employed in this study. Our survey was performed in the training class of the Fujian Entrepreneurship Center in 2019 and 2020. The entrepreneurs who participated in the training course were all financially supported by the “Xing Chuang Tian Di” project. During our investigation, we found that many of the entrepreneurs are farmers or entrepreneurs who worked in cities and then returned to their hometown to start a business. These entrepreneurs played an important role in the implementation of the strategy of Rural Revitalization and the tackling of poverty in China. This study analyzes their entrepreneurial persistence, which is of great significance to help rural economic development.

The questionnaires were distributed to entrepreneurs during the classes with instructors’ help. A detailed explanation of the research purposes was given to the participants. On completion, the questionnaires were immediately collected. First, a pilot test was conducted prior to the formal investigation, and 151 respondents completed the questionnaires, 146 were returned, and 101 valid questionnaires remained after questionnaire screening to verify the readability and clarity of all the measurement items. Outliers include those whose entrepreneurial enterprises were established for more than 8 years. This ensures the internal consistency of our sample. Second, a total of 325 were distributed for the formal investigation. Of the 325 cases, 248 responses were used for data analysis after the removal of outliers and invalid data. Finally, the data were screened for the characteristics of unengaged responses. Cases with an SD less than 0.5 were subject to an investigation for potential unengaged responses.

Participants’ demographic data were as follows: 97 (39.1%) were women and 151 (60.9%) were men. In terms of entrepreneurs’ age, 22 (8.9%) were aged 20 years and below, 106 (42.7%) were aged 21–30 years, 89 (35.9%) were aged 31–40 years, 29 (11.7%) were aged 41–50 years, and 2 (0.8%) were aged above 51 years. No entrepreneurs were educated in elementary school and below, 38 (15.3%) were educated in junior high school, 89 (35.9%) were educated in high school (or middle school), 25 (25%) were educated in university, and 59 (23.8%) were educated as graduate students. Regarding work years before entrepreneurship, 13 (5.2%) at least had 1 year of working experience, 37 (14.9%) had 2–5 years of working experience, 88 (35.5%) had 6–10 years of working experience, 74 (29.8%) had 11–15 years of working experience, and 36 (14.5%) had above 15 years of working experience. In terms of the years of these enterprises, 43 (17.3%) were established at least 1 year, 95 (38.3%) were established 1–2 years, 91 (36.7%) were established 3–4 years, 17 (6.9%) were established 5–6 years, and 2 (0.8%) were established 7–8 years.

In this study, the measures for study variables were adopted from the existing literature. Because some of the original questionnaire items were written in English, we need to make sure that the items should be translated appropriately and understandable for Chinese respondents. Before translating the questionnaire items to English version, we formed an editing group that consists of five members. Among whom, three members are university professors with the same research field of entrepreneurship. Then, two overseas students were invited to participate in the group. They are both native English speakers and also have a good command of Chinese. Hence, their participation was beneficial for the translation and revision of the questionnaire items. First of all, the authors had a discussion with the two overseas students about the aim of the research and the requirements for the translation of questionnaire items. Then, they were asked to translate the questionnaire items into a Chinese version separately. After that, their translations were thoroughly checked by the three professors in order to ensure the content validity of the items. They also offered suggestions to add, remove, or revise inappropriate items. Finally, the checked translation was further discussed by the whole editing group. The revised version could not be accepted until the whole editing group had reached an agreement on it. The questionnaire designing procedure is shown in Figure 2.

Also, a 5-point Likert-type scale ranging from 1 (extremely disagree) to 5 (extremely agree) was utilized to measure anticipated regret, entrepreneurial cognition, entrepreneurial persistence, and entrepreneurial environment.

With regard to anticipated regret, which is considered as an independent antecedent of intention, it is recommended to measure in relation to not performing a behavior, whereas attitudes, subjective norms, and perceptions of control are assessed in relation to performing it (Ajzen and Sheikh, 2013). Combined with the situation in China and the suggestions of three entrepreneurship experts, we added an item based on the measurement of anticipated regret in the previous literature research (Abraham and Sheeran, 2004; Ajzen and Sheikh, 2013). Hence, it was measured by three questionnaire items, namely, “If I don’t continue to engage in entrepreneurial activities in the next 12 months, I will regret it,” “If I don’t continue to engage in entrepreneurial activities in the next 12 months, I will be frustrated,” and “If I don’t continue my entrepreneurial activities in the next 12 months, I will feel anxious.” All measurement items have passed the readability and clarity in the pilot test.

Entrepreneurial cognition was measured with 18 items from the classification proposed by Mitchell et al. (2002). Of which, five items were for the preparation script. The example items are “I have the resources to start a business” and “I can easily enter the field ready to start a business.” Five items were for the willingness script, and the example items are “I can quickly adapt to the new environment” and “I’m always ready for change.” Five items were for the ability script, and the example items are “I can accurately identify potential business opportunities” and “I have a good entrepreneurial knowledge reserve.” All measurement items have passed the readability and clarity in the pilot test.

To measure entrepreneurial persistence, we used six items from the study by Baum et al. (2001), such as “I will continue to work hard to start a business even if others oppose me,” “No matter how challenging and difficult my business is, I will not give up,” and “Compared with most other people, I put more effort into entrepreneurship.” Five measurement items have passed the readability and clarity in the pilot test, but one item failed, so we deleted the item. Finally, five measurement items left.

The entrepreneurial environment was measured with 5 items from the classification proposed by Liu (2011), e.g., “I think the financial service environment is good” and “I think the social and economic environment in my hometown is good.” All measurement items have passed the readability and clarity.

Considering the influence of the control variable, we took gender, age, previous working years (W experience), educational background, and entrepreneurial experience (E experience) as control variables.

We conducted reliability and confirmatory factor analyses (Table 1). All the Cronbach’s alpha values were above 0.90, and the factor loadings, composite reliabilities (CRs), and the average variance extracted (AVE) values were higher than their recommended coefficient weights (Fornell and Larcker, 1981).

As shown in Table 2, the bold numbers in the diagonal direction represent the square roots of AVEs. Because the square roots of AVEs in the diagonal direction are all greater than the off-diagonal numbers, the discriminant validity is satisfactory for all variables. In addition, there are significant correlations between anticipated regret, entrepreneurial cognition, entrepreneurial persistence, and entrepreneurial environment. The relationship between the four variables needs to be deeply explored.

In this study, SPSS software was used for regression analysis. The hierarchical regression method was used to verify relevant research hypotheses. First, the results of the regression test of anticipated regret and entrepreneurial persistence are shown in M1 in Table 3. It shows that anticipated regret has a significantly positive impact on entrepreneurial persistence (β = 0.434, p < 0.001), thus Hypothesis 1 is supported. Second, from M2, willingness script (β = 0.211, p < 0.01), preparation script (β = 0.231, p < 0.001), and ability script (β = 0.2, p < 0.01) significantly affect entrepreneurial persistence. From M3, it can be found that entrepreneurial cognition (β = 0.639, p < 0.001) has a significantly positive impact on entrepreneurial persistence, assuming that Hypothesis 2a, Hypothesis 2b, Hypothesis 2c, and Hypothesis 2 are supported. Then, from M4, M5, and M6 in Table 3, it can be noted that anticipated regret significantly affects willingness script (β = 0.187, p < 0.01), preparation script (β = 0.319, p < 0.001), and ability script (β = 0.15, p < 0.05), assuming that Hypothesis 3a, Hypothesis 3b, and Hypothesis 3c are supported. Also, the non-standardized regression coefficient of the anticipated regret on entrepreneurial cognition is 0.219 (p < 0.001), which shows that Hypothesis 3 is supported.

The analysis of mediating effect was conducted according to the processes proposed by Zhao et al. (2010). To verify the mediating effect of entrepreneurial cognition, it is necessary to regress the anticipated regret and entrepreneurial cognition to the entrepreneurial persistence at the same time. According to the results of the data in Table 4, compared with the corresponding results of the data in Table 3, the correlation coefficients of anticipated regrets and entrepreneurial cognition have decreased, refer to M8 (β = 0.376 < 0.434, p < 0.001), M9 (β = 0.37 < 0.434, p < 0.001), and M10 (β = 0.383 < 0.434, p < 0.001). It can be noted from M11 that the correlation coefficient of anticipated regret drops to 0.322 (p < 0.001), indicating that entrepreneurial cognition plays a mediating role between anticipated regret and entrepreneurial persistence. Hence, Hypothesis 4 is supported. Therefore, the more anticipated regret the entrepreneurs have, the more they tend to look for better entrepreneurial information, thus bringing more entrepreneurial activities.

The hierarchical regression method was used to examine the moderating effect of entrepreneurial environment between anticipated regret and entrepreneurial persistence. The regression results are shown in Table 5. The moderation effect of anticipated regret×entrepreneurial environment on entrepreneurial persistence is 0.15 (t = | 3.276| > 1.96, p < 0.01), indicating the existence of a moderation effect, representing that every unit increases in the moderation variable of entrepreneurial environment, the slope of anticipated regret on entrepreneurial persistence increases by 0.15 units, and Hypothesis 5 is supported. The simple slope analysis in Figure 3 shows that the entrepreneurial environment plays a significant moderating effect between anticipated regret and entrepreneurial persistence. Therefore, the better the entrepreneurial environment, the more anticipated regret the entrepreneurs have, and the more they tend to persist to achieve entrepreneurial success.