Can Co-Creation and Crowdfunding Types Predict Funder Behavior? An Extended Model of Goal-Directed Behavior

Crowdfunding is an emerging means for financing by small and medium-sized enterprises or individuals to attract capital from investors who look to obtain products, services, and/or equity in the future. Co-creation in crowdfunding projects substantially influences sponsors’ behavior, playing a critical role in crowdfunding performance. Despite the significance of co-creation in crowdfunding, research from the leisure and tourism fields has been largely neglected in terms of theory-based models of co-creation. To address this gap, the goal of this work is to study the effects of co-creation on the extended model of goal-directed behavior, along with the moderator of crowdfunding types. To do this, an online survey was conducted on crowdfunders in South Korea and partial least squares (PLS)-structural equation modeling (SEM) was applied to analyze the collected data. Results reveal that co-creation considerably leads to attitude towards, as well as behavior in relation to, crowdfunding participation. Funders’ attitude and positive and negative anticipated emotion also significantly influence desire to participate in crowdfunding. Behavioral intention is highly affected by perceived behavioral control as well as desire. Reward and investment types significantly moderate eight relationships in the research model. Hence, this study contributes to crowdfunding research and stakeholders in the visitor economy sectors.


Introduction
Crowdfunding is "a growing means to finance startups and small businesses, including in the visitor economy (tourism, leisure, sports, creative media, arts, and culture)" also referred to here as leisure and tourism [1] (p. 1). Co-creation in leisure and tourism-related areas is a key factor in enhancing consumer behavior and experience since tourism and hospitality rely on close interactions between service provides and consumers simultaneously at the same places [2][3][4]. Co-creation is also identified as significant in crowdfunding in terms of open innovation [5], value identification for sustainability [6], wisdom of the crowd [7], e-commerce [8], and psychological ownership [9]. However, despite the growing prominence of crowdfunding and co-creation as important concepts in which includes direct consumption of goods and services paid by people who visit a destination, activity generated indirectly from supply chains and services to the industry, construction, etc., and additional induced activity from what people who work, directly or indirectly, serving visitors spend in the local economy" [28] (p. 1). Similarly, the Australian state of Victoria uses the term visitor economy to refer to "the production of goods and services for consumption by visitors, which includes the industries that directly serve visitors, such as hotels, transport providers, tour companies and attractions, as well as intermediaries and those involved indirectly, such as retail and food production" [29] (p. 6). In particular, the visitor economy focuses on "the direct contribution of tourism activities (i.e., the value added generated by the provision of tourism-characteristic goods and services) and also takes into account indirect effects (via the supply chain), as well as the impact of capital investment and collective government expenditure on behalf of the tourism industry" [30] (p. 13). Based on these definitions, this work utilizes the terms of leisure/tourism and visitor economy interchangeably to comprehensively describe consumer behavior in the tourism-related crowdfunding context.
Scholars from leisure and tourism-related fields have started to develop substantial interest in crowdfunding [31][32][33][34][35]. Crowdfunding is used for marketing implements for advertising as well as progress for tourism-related innovative ventures [31,32]. The use of Twitter for the crowdfunding campaigns of tourism projects has been examined in terms of the number of different rewards and the relationship to the chance of tourism crowdfunding success [33]. Leisure and cultural projects in a touristic location have been found to have the best funding success rate [34]. For crowdfunding performance, the national tourism system needs to have a strong relationship with innovative Internet-technologies and socially active stakeholders [35].
Crowdfunding has also been used to help develop more sustainable forms of tourism. For example, civic crowdfunding has been used to raise funds from a community for the fulfilment of civic initiatives for clean energy, such as cycle tourism [36] as well as rural tourism development [37]. In terms of influencing consumer participation in crowdfunding, venture quality and uncertainty level have been found to have significant positive impacts on crowdfunding participation which, in turn, highly influences word-of-mouth and re-participation [1]. However, despite the significance of co-creation in tourism consumer research and the growing importance of crowdfunding as a source of investment for small and medium-sized enterprises (SMEs) and micro-enterprises, little research has been conducted from tourism perspectives on crowdfunding decision-making process particularly related to co-creation. In order to bridge the gap, this study aims to examine an EMGB on visitor economy related crowdfunding including co-creation and moderator of crowdfunding types.

Co-Creation
Co-creation is defined as "high-quality interactions that enable an individual customer to co-create unique experiences with the company are the key to unlocking new sources of competitive advantage" [38] (p. 7). This approach to customer engagement, taking into account the main underlying propositions of service-driven logic, explicitly presents the customer to the same level as the company in value co-creation [39]. In the context of tourism, businesses interested in providing a high value tourism experience benefit from significant co-creation with travel consumers [4]. The development and adoption of information and communication technology has had a profound impact on the tourism industry [26]. Operators have been found to achieve superior performance in terms of synergy with other members of the network, value co-creation, and information technology readiness [2]. Since tourism is characterized by high-contact services in which business co-creation with customers has a key role, the extent of co-creation is critical for company backing and customer service expenditures [3].
Recent research has examined co-creation in the crowdfunding context [5,[7][8][9]40]. Crowdfunder types have been found to reflect the nature of crowdfunding in the e-commerce context as a new form of co-creation [8], although several different types of value co-creation (co-ideation, co-evaluation, co-design, co-evaluation, co-launch, co-evaluation, co-financing, and co-consumption) have been identified [40]. Consumer role-identity and psychological ownership have been documented as Sustainability 2019, 11, 7061 4 of 23 a significant factor in co-creation during crowdfunding campaigns [7], although the effects of co-creation are coordinated by the entrepreneurial activity and the sponsor's social connections [9]. Co-participation, diverse ideas, and degree of effort all play important roles in value co-creation in nurturing crowdfunding [5]. Even though co-creation is significant in crowdfunding performance over a range of different sectors, research on fundraisers' co-creation with funders, platforms, employees, and other partners has been largely overlooked in leisure and tourism-related crowdfunding projects. In order to bridge the gap, this study considers co-creation with investors, sites, internal customers, and other system components as a main construct of the open innovation process in leisure and tourism-related crowdfunder behavior.

Model of Goal-Directed Behavior (MGB)
Including anticipated emotions and desire, a model of goal-directed behavior (MGB) shows significantly greater amounts of variance that are broadening and deepening the TPB [41]. An extended MGB (EMGB) that includes key variables from leisure and tourism has been extensively studied in order to better understand travel consumers' decision-making processes [13,16,17,19]. For instance, by adding environmentally friendly perceptions, the EMGB has proven to be vital to understanding the perceptions and behaviors of festival participants based on nature [19], while consumers' social embeddedness and level of knowledge were found to significantly influence sustainable consumption [16]. By including behavioral interventions, an EMGB with constructs of non-volitional, volitional, and emotional aspects has well predicted consumer intention to travel during a pandemic [17]. An EMGB adding a frequency of past behaviors was found to provide a powerful framework for demonstrating that gender plays a decisive role in overseas travel behavior decision making processes [13].
In recent years, scholars have intensively utilized the EMGB to identify travel consumers' decision-making processes [14,15,18]. For example, by incorporating authenticity-related constructs into the MGB, the factors of EMGB have shown significant prediction for slow tourists' behavior [18]. In tourism information and communication technology, the EMGB has helped to explain the use of mobile devices in travel consumer behavior [15] as well as the role of perceived ethics in predicting consumer behavior in responsible tourism [14]. From the perspective of crowdfunding, the TPB has shed substantial light on funder behavior [10][11][12]42]. However, studies on the EMGB have been not conducted on crowdfunder behavior in a tourism context, thereby extending understanding of potentially crucial factors such as co-creation. To address this research gap, this study aims to apply the EMGB to know funder behavior to crowdfunding participation, including co-creation and funding types.

Co-Creation, Attitude, and Behavioral Intention
Attitudes are commonly defined as an evaluative dimension, "a person's attitude represents his evaluation of the entity in question" [43] (p. 889). Since crowdfunding projects are offered only on one platform, success requires not only the positive attitude of funders toward the project itself, but also a positive attitude toward the Internet platform from potential contributors [44]. Crowdfunders with strong attitudes help others identify with the crowdfunding community [45]. This is potentially significant for a fundraising project as its success clearly depends on people's contribution.
Behavioral intention refers to a proxy act in which "a single act is predictable from the attitude toward that act, provided that there is a high correlation between intention and behavior" [43] (p. 889). Perceived ethics was found to be a strong behavioral intention for responsible tourism in the EMGB [14]. Intention to participate in crowdfunding is also a key dependent variable in predicting actual funder behavior [10,22]. For example, social presence, trust, attitude, and perceived behavioral control have significant effects of behavioral intention to donate time and money for crowdfunding [10]. Also, commitment, platform, leader, and market have positive impacts on behavioral intention for Sustainability 2019, 11, 7061 5 of 23 crowdfunding investment [22]. Hence, this study regards attitude and behavioral intention as critical variables in leisure and tourism-related crowdfunding.
Crowdfunding stakeholders facilitate multi-party value co-creation in order for projects to succeed [5], with different stakeholders having different roles to play in funder behavior [46]. As online community providers crowdfunding platforms play a key role in co-creation processes in Internet crowdfunding communities [6], with co-creation on crowdfunding projects contributing to an increase in consumers' positive attitude and behavioral intention. In addition, sponsor co-creation has been shown to have highly significant effects on perceived control, self-rated investment, and the intimate knowing that influence funders' commitment to crowdfunding projects [9], suggesting that co-creation of fundraisers with funders, platforms, internal consumers, and external partners leads to positive attitudes and behavioral intentions of supporters. Such co-creation activities in crowdfunding are undertaken to raise money for projects run by entrepreneurs or artists by soliciting money in small amounts from the general public (i.e., sponsors) mainly through E-commerce platforms [8], implying that co-creation in crowdfunding projects is a key factor in project success. Drawing upon this literature, this study proposes the following hypothesis: Hypothesis 1 (H1). Co-creation in crowdfunding project positively influences attitude toward participation in crowdfunding in the leisure and tourism sector; Hypothesis 2 (H2). Co-creation in crowdfunding project positively influences behavioral intention to participate in crowdfunding in the leisure and tourism sector.

Attitude, Subjective Norm, Emotion, Perceived Behavioral Control, and Desire
A subjective norm is defined as "a perceived social pressure to perform or not to perform a particular behavior" [47] (p. 188). In the context of crowdfunding, a subjective norm refers to "the individual's beliefs about whether significant others think he or she should engage in the behavior and are assumed to capture the extent of perceived social pressures exerted on individuals to engage in a certain behavior" [42] (p. 58). Anticipated emotions to a specific behavior can be defined as "important determinants of a range of intentions toward goals and actions with salient affective outcomes" [48] (p. 1553). Positive anticipated emotion with goal attainment and negative anticipated emotion with goal failure have a role in predicting desire that eventually leads to goal pursuing behavior [41]. Perceived behavioral control as a non-volitional dimension is conceptualized as "an individual's confidence or ability to perform a specific behavior and it is considered to be an imperative factor of behavioral intention and actual behavior" [19] (p. 1420). In the context of extended TPB, perceived behavioral control significantly influences intention to crowdfund [11]. Desire is defined as "a critical factor in explaining a person's decision formation" [19] (p. 1419). It has also been suggested that incorporating the desire as a mediator into the MGB between antecedents (attitudes, subjective norms, perceived behavioral control, anticipated emotions, and frequency of past behaviors) and motivational aspects of human behavior improves the model's predictive ability [41].
In the EMGB context, it has been shown that students' attitude toward negative anticipated emotions, subjective norms, and perceived behavioral control have a significant impact on the desire to study: positive anticipated feelings, subjective norms, and perceived behavioral control have a significant impact on the desire for weight control [41]. In EMGB studies in a tourism context consumer subjective norm, attitude, and positive anticipated emotion have also been found to positively influence desire for revisiting festivals [19] and desire for participation in UNICEF's pro-environmental campaigns [16]. While, the addition of negative anticipated emotion to the EMGB has also been significant in studies of the desire for international travel [13,17]. Responsible tourists' subjective norm, positive and negative anticipated emotions, and perceived behavioral control lead to their desire to travel overseas [14]. Therefore, this work suggests the five subsequent hypotheses for leisure and tourism crowdfunding: Hypothesis 7 (H7). Perceived behavioral control positively influences desire to participate in crowdfunding in the leisure and tourism sector.

Perceived Behavioral Control and Behavioral Intention
In studies applying norm activation theory and TPB perceived behavioral control has a significant influence on intention to crowdfund for both time and money donations [10]. Hierarchical multiple regression undertaken in the extended TPB (ETPB) context on intention to crowdfund for sustainable rural development has shown that perceived behavioral control influences intention to donate [11]. In addition, in ETPB studies perceived behavioral control leads to financial contribution intention in reward crowdfunding contexts [42]. In the TPB context, the variables confidence in ability (i.e., perceived behavioral control) to donate to a crowdfunding campaign best predicts an individual's intention to donate to a given crowdfunding campaign [12]. In line with previous research findings, this study therefore suggests the following hypothesis in visitor economy crowdfunding: Hypothesis 8 (H8). Perceived behavioral control positively influences behavioral intention to participate in crowdfunding in the leisure and tourism sector.

Desire and Behavioral Intention
In an EMBG study of slow tourists, the desire to travel by slow means of transport has a highly positive effect on behavioral intention to undertake slow travel [18]. Similarly, in tourism information and communication technology, consumer desire for using mobile devices positively influences behavioral intention to use mobile devices for tourism-related purposes [15], while tourists' desire for participation in responsible tourism substantially influences behavioral intention to participate in responsible tourism [14]. Pro-environmental visitors' desire to participate in an environmentally friendly festival was found to lead to their behavioral intention to participate in the festival [19]. Therefore, this study postulates the following hypothesis in visitor economy crowdfunding: Hypothesis 9 (H9). Desire positively influences behavioral intention to participate in crowdfunding in the leisure and tourism sector.

Moderating Role of Reward and Investment Crowdfunding Types
In reward crowdfunding, funders receive goods or services in exchange for investment and these rewards may be provided in other forms, including public approval, product pre-sale, or limited edition products. [21]. In equity crowdfunding, funders receive a stake in the capital of the company the funders invest in, and in loan crowdfunding, funders invest in the form of loans, allowing the funders to potentially regain investment with interest [21]. As an innovative financial technology (fintech), investment crowdfunding (e.g., equity, lending) disrupts traditional financial intermediation in various ways. For example, the implicit trust in investment crowdfunding eliminates the need for brokers in Sustainability 2019, 11, 7061 7 of 23 some sectors [49,50]. In recent years, the investment crowd financing mechanism has gained popularity as an alternative form of investment to corporate funding, particularly for startup financing [51]. The international development of investment crowdfunding for profit has also increased the chances of success by providing entrepreneurs with opportunities to access new sources of capital [52]. In this study, equity and lending crowdfunding are categorized as investment types, whereas return based products and services crowdfunding are considered as reward types.
In the context of investment crowdfunding, communication, values, benefits, and risks have a significant impact on crowdfunding investor confidence in the fundraiser of crowdfunding projects [22]. Among funders in investment crowdfunding, uncertainty level, and venture quality have considerable effects on crowdfunding behavior that substantially lead to word-of-mouth and re-participation in visitor economy crowdfunding [1]. Investment crowdfunding of new methods for entrepreneurial firm financing is also quite different from reward crowdfunding in terms of the cause and effect of funders' attitudes, emotions, and behaviors [49][50][51][52]. Since investment crowdfunding shows unique characteristics, we anticipate differences between reward and investment types of crowdfunding in leisure and tourism fields. Hence, this work suggests the subsequent hypothesis in the crowdfunding context: Hypothesis 10 (H10). Reward/investment crowdfunding types significantly moderate the relationships between co-creation, attitude, subjective norm, positive anticipated emotion, negative anticipated emotion, perceived behavioral control, desire, and behavioral intention to participate in crowdfunding in the leisure and tourism sector.
Based on the hypotheses above, this work proposes a research framework to investigate the relationships between co-creation, attitude, positive anticipated emotion, subjective norm, negative anticipated emotion, desire, perceived behavioral control, and behavior, along with the moderating role of reward and investment crowdfunding types as shown in Figure 1. allowing the funders to potentially regain investment with interest [21]. As an innovative financial technology (fintech), investment crowdfunding (e.g., equity, lending) disrupts traditional financial intermediation in various ways. For example, the implicit trust in investment crowdfunding eliminates the need for brokers in some sectors [49,50]. In recent years, the investment crowd financing mechanism has gained popularity as an alternative form of investment to corporate funding, particularly for startup financing [51]. The international development of investment crowdfunding for profit has also increased the chances of success by providing entrepreneurs with opportunities to access new sources of capital [52]. In this study, equity and lending crowdfunding are categorized as investment types, whereas return based products and services crowdfunding are considered as reward types.
In the context of investment crowdfunding, communication, values, benefits, and risks have a significant impact on crowdfunding investor confidence in the fundraiser of crowdfunding projects [22]. Among funders in investment crowdfunding, uncertainty level, and venture quality have considerable effects on crowdfunding behavior that substantially lead to word-of-mouth and re-participation in visitor economy crowdfunding [1]. Investment crowdfunding of new methods for entrepreneurial firm financing is also quite different from reward crowdfunding in terms of the cause and effect of funders' attitudes, emotions, and behaviors [49][50][51][52]. Since investment crowdfunding shows unique characteristics, we anticipate differences between reward and investment types of crowdfunding in leisure and tourism fields. Hence, this work suggests the subsequent hypothesis in the crowdfunding context: Hypothesis 10 (H10). Reward/investment crowdfunding types significantly moderate the relationships between co-creation, attitude, subjective norm, positive anticipated emotion, negative anticipated emotion, perceived behavioral control, desire, and behavioral intention to participate in crowdfunding in the leisure and tourism sector.
Based on the hypotheses above, this work proposes a research framework to investigate the relationships between co-creation, attitude, positive anticipated emotion, subjective norm, negative anticipated emotion, desire, perceived behavioral control, and behavior, along with the moderating role of reward and investment crowdfunding types as shown in Figure 1.

Measurement
This study applies multi-measurement questions to address the disadvantages of solo item measures (e.g., [53]). After an extensive literature review on reward and investment crowdfunding, co-creation, and the EMGB [1][2][3][4]7,15,[17][18][19]24,43], the questionnaire initially comprised 32 items for measuring the eight concepts. The constructs included co-creation, attitude, subjective norm, positive anticipated emotion, negative anticipated emotion, perceived behavioral control, desire, and behavioral intention to visitor economy crowdfunding. Questions are measured by 7-point Likert-type scales. In addition, nine items concerning crowdfunding participant behavior (financial characteristics of project, involvement in crowdfunding outside of South Korea (hereafter Korea), length of participation, experience, frequency of visiting platforms, average investment amount, primary reason to invest, projects participated in, and platforms used) were utilized in this study based on prior literature [1]. Seven questions associated with socio-demographics were also included (e.g., age, gender, education, monthly family income, marital status, occupation, residential district).

Co-Creation
The construct of co-creation on crowdfunding project has been operationalized with four items as recommended by previous studies [5,7,8,40]. These items included co-creating with funders, platforms, employees, and others by the crowdfundraiser in the visitor economy sector (e.g., "The visitor economy crowdfundraiser encourages co-creation with funders") (see Table 2).

Attitude
The construct of attitude toward visitor economy crowdfunding participation has been operationalized with four items as suggested by [41,43,45,47]. These questions comprise funders' thinking on crowdfunding participation as an affirmative, beneficial, valuable, essential behavior (e.g., "Participating in visitor economy crowdfunding is an affirmative behavior").

Subjective Norm
The construct of subjective norm on crowdfunding participation has been operationalized with four items utilized in previous literature [10,12,19]. These items are relevant to how people who are important to funders think visitor economy crowdfunding participations (e.g., "Most people who are close to me agree with my participation in visitor economy crowdfunding").

Positive Anticipated Emotion
The construct of positive anticipated emotion on crowdfunding participation has been operationalized with four questions from earlier literature [13,15,17,41]. The question items incorporated funders' positive feeling on visitor economy crowdfunding participation, such as being excited, glad, happy, and satisfied (e.g., "If I participate in visitor economy crowdfunding, I will be excited").

Negative Anticipated Emotion
The construct of negative anticipated emotion on crowdfunding participation was operationalized based on four questions from prior literature [13,15,17,41]. These items included funders' negative feeling on visitor economy crowdfunding participation, such as being disappointed, sad, sorry, and so on (e.g., "If I fail to participate in visitor economy crowdfunding, I will be disappointed"). The construct of perceived behavioral control on crowdfunding participation was operationalized based on four questions from earlier literature [14,18,42]. The question items refer to funders' perceptions of their ability to participate in visitor economy crowdfunding (e.g., "I am financially able to participate in visitor economy crowdfunding").

Desire
The construct of desire to participate in crowdfunding was operationalized with four questions based on items recommended in [14,18,19]. The items refer to funders' desire to participate in visitor economy crowdfunding, such as hoping, eager, enthusiastic, and so on (e.g., "I hope to participate in visitor economy crowdfunding").

Behavioral Intention
The construct of behavioral intention to crowdfunding participation was operationalized by four questions as proposed in prior literature [10,14,22,43]. The items incorporate willingness to invest regularly as well as sooner or later and encouraging people around them to invest in visitor economy crowdfunding (e.g., "I have a willingness to invest in visitor economy crowdfunding").

Content Validity and Pre-Test
The questions were originally written in English. Questions were then converted into Korean by three university language professionals who are proficient in both languages. The survey questionnaire was then translated back into English so that incongruities could be rectified [54]. Three academics knowledgeable about this area of research assessed the content validity of the questionnaire. Five practitioners of crowdfunding also evaluated whether the questionnaire appropriately assessed consumer behavior of visitor economy crowdfunding. As a result of these steps, four items were deleted from the questionnaire because of overlapped meaning (i.e., one item from each of subjective norm, negative anticipated emotion, perceived behavioral control, and desire). As well, one item was added to attitude in order to capture the concept more clearly (i.e., "Participating in visitor economy crowdfunding is a legitimate behavior"). In addition, the screening question was modified from "In the past year, have you participated in any tourism-related crowdfunding project of reward, equity, bond, or lending?" to "In the past 12 months, have you participated in any crowdfunding project for reward, equity/bond, or/and lending within the visitor economy sector?" A pilot study using seven Ph.D. students majoring in tourism studies who had experience of visitor economy crowdfunding further assessed the survey instrument. These candidates were chosen because they understand both the visitor economy and crowdfunding fields. Based on their comments, several items on co-creation, attitude, and behavioral intention were reworded. In particular, the students said that general information questions were confusing in terms of the visitor economy sectors and crowdfunding types. Thus, specific examples and explanations were provided at the beginning of the questionnaire. Fifty university staff and faculty who had experienced visitor economy crowdfunding in the prior year completed a pretest. We asked these respondents to provide feedback and whether, in their opinions, any changes, revisions, additions, or deletions were necessary. This method resulted in further minor revisions of some questions on the five main constructs of the model of goal-directed behavior in crowdfunding to assure clarity (i.e., subjective norm, positive anticipated emotion, negative anticipated emotion, perceived behavioral control, and desire). After completing these changes, the adjusted questionnaire was utilized for the main survey.

Data Collection
Since Internet surveys result in rapid responses, access a variety of populations, and reduce expenses, online surveys are commonly used [55]. Since the study examines crowdfunder behavior via platforms, Internet surveys are deemed particularly appropriate for data collection for this research. To administer the survey instrument, an Internet survey firm having over three million panelists [56] was hired. Using a quota sampling method based on mobile Internet users' age and gender, subjects were drawn from their panel. To ensure response validity, the company adheres strictly to protocols. To verify respondent's personal information, the protocols the firm uses are also carefully designed. A screening question was placed at the beginning of the survey so subjects who did not qualify for participation in the survey were deleted from the sample frame. Respondents who completed the screening question successfully were required to name a visitor economy crowdfunding project they had recently experienced. On each subsequent item, the project name stated by each subject was then presented to that subject. To avoid response bias, the multiple-choice items were rotated [57]. An automated procedure removed respondents who answered too fast or used repetitive patterns, therefore these questionnaires were not included in the subsequent analysis.
The sample profile matched the age and gender profile of Korean mobile Internet users [58]. The subjects were all Koreans, who had experienced visitor economy crowdfunding in the previous year and were 18 or more years old. The Internet survey was conducted between April 1 and 11, 2019. The survey company generally gets complete responses of approximately 5% from initial invitations. Since our comprehensive model of eight constructs and 29 indicators with multiple group analysis (MGA) needs a sample size of around 400 cases, the letter to request participation in this survey was emailed to 9,710 panelists [59]. The email invitation was opened by 2616 individuals and resulting in 2353 respondents clicking through to the questionnaire. Each subject was presented with the screening question. Of those who were presented with the screening question 1489 panelists responded in the affirmative and, hence, were qualified to answer to the questions. Accordingly, a total of 485 respondents completed the questions. After outliers, subjects not saying the crowdfunding project name, and respondents investing under KRW 10,000 were eliminated, 442 completed responses were used for the analysis. The response rate represents 32.6% (485 completed surveys divided by 1489 respondents who successfully passed the screening question) [60].

Data Analysis
This study used partial least squares-structural equation modeling (PLS)-(SEM) to investigate the proposed research framework. To verify the study framework by bootstrapping resampling techniques of non-parametric methods, PLS needs a minimum criterion for measurement scale, residual distribution, and sample size compared to traditional covariance-based (CB)-SEM [59]. PLS-SEM is more suitable than CB-SEM for MGA and complicated models [61]. Accordingly, SmartPLS 3.2.8 was [62] to validate the measurement and structural models in the research because the data did not have normal distribution and this study conducts MGA.
Since subjects were requested to assess all survey items at the same time with the same subjects, common method bias could be a potential problem. Hence, protections were taken applying various methodical remedies to control for common method variance drawing upon [63]. A selection inquiry was applied for choosing subjects who had experienced leisure and tourism-related crowdfunding within the past year. Along with a declaration promising privacy of every subject, the introductory section of the survey included the study goal. In order to decrease subject anxiety, the survey guidelines stated that questions have no correct or incorrect answer. The key terms of crowdfunding, funding types, visitor economy sectors, and leisure and tourism-related fields were explained, along with their examples. The questionnaire had three sections: the first section incorporated definitions on key terms and general questions; the second included items associated with the research framework; and the third comprised socio-demographic characteristics. The sequence of questions related to the research framework was switched randomly in order to reduce response bias.
For determining if the resulting data set was influenced by common method variance, Harman's single-factor test is performed as a post hoc statistical [60]. With every self-administrated questionnaire item, exploratory factor analysis (EFA) was conducted. Common method bias can be a concern once a single factor reveals. EFA results showed seven variables (eigenvalue > 1) indicating 74.4% by factors containing the first factor (44.5%), and follow-up factors (7.3%, 6.1%, 5.1%, 4.4%, 3.6%, and 3.5%). Since Harman's single-factor checking includes weaknesses, the comparison of EFA and confirmatory factor analysis (CFA) has been applied [62]. According to Korsgaard and Roberson [64], a comparison of the hypotheses (e.g., CFA) and only one factor specified (e.g., EFA) was conducted. The hypothesis framework is more suited to the data than the single factor framework. In addition, compared to the single factor framework, the hypothesis model was statistically significant. Specifically, the difference in chi-square statistics between the hypothesis and single factor frameworks is highly significant (χ 2 (2391.7)/df (27) = 84.9, p < 0.001). Therefore, common method bias is not a problem in this work as the two statistical examinations show.

Grouping Check
The reward or investment type of visitor economy crowdfunding was assessed applying a question (e.g., "What is your primary reason for participating in visitor economy crowdfunding?"). The result shows that reward type (71.5%) (product reward 39.4%, n = 174; non-product reward 32.1%, n = 142), investment type (28.8%) (equity 17.4%, n = 77; lending 10.4%, n = 46; other type 0.7%, n = 3) as the primary reason for participating in leisure and tourism-related crowdfunding. To conduct a multi-group analysis, the respondents were divided into two clusters of crowdfunding types: reward crowdfunding group (n = 316) and investment crowdfunding group (n = 123), removing the other crowdfunding type (n = 3).

Respondents' Profile
As shown in Table 1, the sample of the reward type has 52.5% female representation, while the investment type has 63.4% male. The reward group is younger than the investment group. The majority of respondents attended a university and/or higher (reward 69.3% and investment group 79.7%). Marital status was similar in both groups. More reward subjects earned a monthly family income of Korean Won (KRW) four million and over (72.5%) compared to investment type respondents (66.6%). The majority of respondents had full time jobs (reward: 66.5%; investment: 78.2%). More than a half of the sample were living in metropolitan areas (reward: 60.4%; investment: 60.2%). More than a half of the sample also participated in profit crowdfunding project (reward: 58.8%; investment: 71.5%).
A little more than a half of reward group (53.0%) and two-thirds of investment group (70.6%) had more than seven months or more experience with visitor economy crowdfunding prior to completing the survey. Around a half of respondents recently participated in visitor economy crowdfunding for product reward (59.5%) from the reward group and for equity from investment group (67.5%). Over a third of participants visited crowdfunding platforms monthly or more frequently (reward 37.9%; investment 44.0%). A majority of respondents invested an average funding amount of KRW 100,000 and over in visitor economy crowdfunding (reward 47.8%; investment 87.0%). A majority of respondents had a product reward as the primary reason from reward group (55.1%) and equity from investment group (62.6%). Regarding the type of visitor economy crowdfunding that respondents participated in, travel and leisure was the highest project from reward group (31.7%), while other (food/beverage, events, hobbies) was the highest projects from investment group (35.8%). The most frequently visited platform was Wadiz from reward group (42.1%) and investment group (40.7%). Some subjects were involved in crowdfunding outside of Korea (reward 13.6%; investment 7.3%).

Measurement Model
With regard to the measurement model, this study has conducted CFA [59]. All items for eight constructs appeared to have over 0.7 factor loadings and thus, the analysis has been applied to the total 29 items ( Table 2). Valuation of reliability, discriminant, and convergent validity were performed [65]. Cronbach's α, composite reliability, and Rho_A of all concepts are larger than 0.70, confirming the reliability as well as sufficient internal consistency (Table 3) [60]. Each concept's average variance extracted (AVE) was larger than 0.5 and all items' factor loading was greater than 0.7, resulting in confirming convergent validity [66]. Discriminant validity has been confirmed with a Heterotrait-Monotrait Ratio (HTMT) that demonstrates its superior performance by means of a Monte Carlo simulation study [61]. For example, the highest value between positive anticipated emotion and desire for crowdfunding participation is 0.793, which is lower than the criterion value of 0.85 so that the discriminant validity was established in this model. In addition, Q2 values are larger than zero for three endogenous variables, indicating acceptable predictive relevance [67].   Figure 2 shows the results from PLS-SEM that estimated the proposed study framework [62]. All the endogenous variables have sufficient variance explained as the R-square (R2) for co-creation (26.7%), desire for (63.3%), and behavioral intention for crowdfunding participation (38.9%). Both path coefficients and t-statistics have been assessed for testing the hypotheses by utilizing a PLS bootstrapping approach of 5,000 re-samplings [61]. According to results, the relationship between co-creation and attitude (γ = 0.517, t-value = 11.943) as well as between co-creation and behavioral intention (γ = 0.210, t-value = 3.721) were highly significant. Also, the relationships between attitude and desire (γ = 0.236, t-value = 4.248), positive anticipated emotion and desire (γ = 0.336, t-value = 4.248), negative anticipated emotion and desire (γ = 0.298, t-value = 7.861), and perceived behavioral control and behavioral intention (γ = 0.404, t-value = 7.687) were significant. Moreover, behavioral intention to crowdfund was significantly influenced by desire (β = 0.160, t-value = 2.819). Thus, Hypotheses 1, 2, 3, 5, 6, 8, and 9 have been supported in this study. However, the relationship between subjective norm and desire (γ = 0.069, t-value = 1.567) as well as perceived behavioral control and desire (γ = 0.059, t-value = 1.184) were insignificant, and thus, H4 and H7 were not supported in this work.  Figure 2 shows the results from PLS-SEM that estimated the proposed study framework [62]. All the endogenous variables have sufficient variance explained as the R-square (R2) for co-creation (26.7%), desire for (63.3%), and behavioral intention for crowdfunding participation (38.9%). Both path coefficients and t-statistics have been assessed for testing the hypotheses by utilizing a PLS bootstrapping approach of 5,000 re-samplings [61]. According to results, the relationship between co-creation and attitude (γ = 0.517, t-value = 11.943) as well as between co-creation and behavioral intention (γ = 0.210, t-value = 3.721) were highly significant. Also, the relationships between attitude and desire (γ = 0.236, t-value = 4.248), positive anticipated emotion and desire (γ = 0.336, t-value = 4.248), negative anticipated emotion and desire (γ = 0.298, t-value = 7.861), and perceived behavioral control and behavioral intention (γ = 0.404, t-value = 7.687) were significant. Moreover, behavioral intention to crowdfund was significantly influenced by desire (β = 0.160, t-value = 2.819). Thus, Hypotheses 1, 2, 3, 5, 6, 8, and 9 have been supported in this study. However, the relationship between subjective norm and desire (γ = 0.069, t-value = 1.567) as well as perceived behavioral control and desire (γ = 0.059, t-value = 1.184) were insignificant, and thus, H4 and H7 were not supported in this work.

Moderating and Mediating Effects
With regard to the moderating role of reward and investment types of crowdfunding, nine hypotheses in the research model were tested (Table 4). In order to identify differences between reward and investment groups, this study has utilized the explained variances (R 2 ) [59]. The results indicate that explained variances achieved greater percentage points for the reward group with respect to co-creation (9.3%) than the investment cluster. On the contrary, the investment cluster achieved higher percentage points of R 2 with desire (13.7%) and behavioral intention (18.6%) than the reward group. The outcomes of two sub-models showed that the paths of eight hypotheses for the reward and investment groups were significantly different. Thus, H10a, H10c, H10d, H10e, H10f, H10g, H10h, and H10i were supported. Furthermore, the magnitudes of the effects of co-creation on attitude (reward = 0.541 > investment = 0.446), attitude on desire (reward = 0.254 >

Moderating and Mediating Effects
With regard to the moderating role of reward and investment types of crowdfunding, nine hypotheses in the research model were tested (Table 4). In order to identify differences between reward and investment groups, this study has utilized the explained variances (R 2 ) [59]. The results indicate that explained variances achieved greater percentage points for the reward group with respect to co-creation (9.3%) than the investment cluster. On the contrary, the investment cluster achieved higher percentage points of R 2 with desire (13.7%) and behavioral intention (18.6%) than the reward group. The outcomes of two sub-models showed that the paths of eight hypotheses for the reward and investment groups were significantly different. Thus, H10a, H10c, H10d, H10e, H10f, H10g, H10h, and H10i were supported. Furthermore, the magnitudes of the effects of co-creation on attitude (reward = 0.541 > investment = 0.446), attitude on desire (reward = 0.254 > investment = 0.204), subjective norm on desire (reward = 0.076 > investment = 0.025), and positive anticipated emotion on desire (reward = 0.365 > investment = 0.251) for the reward type were greater than for the investment type. On the other hand, the magnitudes of the effects of negative anticipated emotion on desire (reward = 0.258 < investment = 0.417), perceived behavioral control on desire (reward = 0.008 < investment = 0.198), perceived behavioral control on behavioral intention (reward = 0.384 < investment = 0.492), and desire on behavioral intention (reward = 0.133 < investment = 0.225) for the investment type were greater than for the reward type. However, the magnitude of the effect of co-creation and behavioral intention (reward = 0.215 < investment = 0.210) was insignificantly different between reward and investment types of crowdfunding.
Regarding the mediating roles of attitude and desire for crowdfunding participation, PLS bootstrap 5000 re-samplings were employed to test the mediating effects of co-creation, social norm, positive anticipated emotion, negative anticipated emotion, and perceived behavioral control in the proposed research model. The co-creation positively and indirectly influenced desire (γ = 0.122, t-value = 3.775) and behavioral intention (γ = 0.020, t-value = 2.203). Also, the attitude (γ = 0.038, t-value = 2.265), positive anticipated emotion (γ = 0.054, t-value = 2.359), and negative anticipated emotion (γ = 0.048, t-value = 2.815) positively and indirectly influenced behavioral intention. Thus, attitude and desire for crowdfunding participation partially played mediating roles in influencing behavioral intention to participate in crowdfunding for visitor economy sectors. In addition, multicollinearity of all variables was tested, applying the variance inflation factor (VIF). Since the constructs' inner VIF values ranged from 1.000 to 2.383, multicollinearity does not appear to be a problem in this study's data, which are all lower than the cutoff of five [59]. Furthermore, according to Cohen [68], 0.02, 0.15, and 0.35 of f2 (effect size) suggest small, medium, and large effects, separately (Table 5).  Note: *** p < 0.001; ** p < 0.01; * p < 0.05; ns = non-significant. a Variance inflation factor is collinearity statistics between constructs in hypotheses. b Effect sizes are the correlation between two variables, which is a quantitative measure of the magnitude of a phenomenon.

Discussion
Crowdfunding is a novel way for SMEs and micro-business entrepreneurs and individuals to raise capital for a wide variety of projects, including for purposes of promoting more sustainable environments [24]. Entrepreneur activity with respect to co-creation also has significant effects on sponsors' participation and plays an important role in the promotion and success of crowdfunding [9]. Despite the importance of co-creation on crowdfunding, little research on co-creation between investors, platforms, internal customers, and others has been conducted in leisure and tourism crowdfunding. Accordingly, the purpose of this study is to build and test a theoretically comprehensive EMGB research model to understand funders' decision-making process in terms of the effect of co-creation on crowdfunding.
Results reveal that co-creation has the greatest effect on attitude toward crowdfunding participation in this study, implying that entrepreneurs' co-creation with investors, platforms, internal customers, and others, highly encourages funders' positive attitude toward crowdfunding participation in the visitor economy. Given the significant effect of co-creation on behavioral intention, fundraisers' active co-creation with stakeholders can encourage consumers' actual investment in crowdfunding projects. Attitude, positive anticipated emotion, and negative anticipated emotion all have positive impacts on desire for crowdfunding participation. These imply that consumers' affective and emotional aspects play important roles in enhancing desire. Importantly, behavioral intention was substantially influenced by perceived behavioral control and desire, suggesting that crowdfunders' decision-making process is substantially explained by their goal-directed behavior. On the other hand, the relationships between subjective norm and desire as well as perceived behavioral control and desire were insignificant, showing that volitional and non-volitional elements are not critical to desire for visitor economy crowdfunding participation.
Reward and investment categories significantly moderate eight relationships out of nine hypotheses in this research model. Specifically, the relationships between co-creation and attitude, attitude and desire, subjective and desire, and positive anticipated emotion and desire are greater in the reward crowdfunding group than in investment crowdfunding group. This implies that affective factors may be impactful for the reward group in deciding their participation in crowdfunding. On the other hand, the relationships between negative anticipated emotion and desire, perceived behavioral control and desire, perceived behavioral control and behavioral intention, and desire and behavioral intention have stronger impacts in the investment crowdfunding group than in the reward crowdfunding group. It infers that volitional dimensions are more likely to be influential for the investment crowdfunding group to decide their goal-directed behavior such as participation in crowdfunding. However, the relationship between co-creation and behavioral intention is insignificant, revealing that co-creation is noticeably important for both groups of reward and investment crowdfunding types.

Theoretical Implications
The findings of this study offer several contributions to the literature in the visitor economy. Primarily, in utilizing the EMGB in crowdfunding this is the first work in leisure and tourism-related fields to identify funder decision-making process by goal-directed behavior. Applying co-creation in visitor economy crowdfunding helps better understand the impacts of interactions among stakeholders, such as entrepreneurs (fundraisers), investors (consumers), platforms (online sites), policy makers from governments and local communities, and others (employees, partners, or professionals from other areas). In addition, the moderating effect of reward and investment crowdfunding types adds new awareness to the literature on the study topic, providing a clearer understanding of a variety of types of crowdfunding.
The strong effects of co-creation on attitude and behavioral intention to participate in crowdfunding in the visitor economy extends prior findings on the relationships between sponsor co-creation, their mental state, and actual commitment to crowdfunding projects [9]. The significant effects of attitude, positive anticipated emotion, and negative anticipated emotion on desire for crowdfunding participation in leisure and tourism-related fields also expand previously observed relationships in airline sustainability programs [16] and environmentally friendly festivals [19]. In addition, the positive impacts of perceived behavioral control and desire on behavioral intention strengthen previously observed relationships in crowdfunding for sustainable rural development [11] and responsible tourism [14]. The moderating effect of funding types expands previous results on the characteristics of investment crowdfunding type in leisure and tourism-related areas [1] and differences among reward, equity, and lending crowdfunding types [21].

Practical Implications
This research provides some implications to crowdfunding stakeholders, such as small and medium-sized businesses, individual fundraisers, platforms, investors, and governments. First, the substantial impact of co-creation on attitude in this research model implies that fundraisers should put effort into enthusiastic interaction with funders, platforms, internal customers, and others to encourage potential consumers to have positive attitudes toward their crowdfunding participation. Second, the significant effect of co-creation on behavioral intention suggests that entrepreneurs should ensure active communications with investors, online sites, employees, and partners in order to enhance funders' actual investment in crowdfunding projects. Third, the positive influence of attitude toward desire indicates that businesses can increase the number of consumers' wanting to invest their money in crowdfunding by promoting messages about their project in terms that are affirmative, beneficial, valuable, essential, and legitimate. Fourth, the highly significant effect of positive anticipated emotion on desire suggests that crowdfunding projects could improve backers' desire to invest money by building their digital storytelling in terms of the excitement, happiness, and satisfaction of funders. Fifth, the positive impact of negative anticipated emotion on desire shows that enterprises need to stimulate sponsors' negative feelings when potential sponsors do not fund crowdfunding initiatives, by indications of disappointment, sadness, and sorrow.
Importantly, in order to boost investment in visitor economy crowdfunding, stakeholders must emphasize perceived behavioral control by highlighting volitional aspects. This can be done by delivering messages via social media so that consumers feel they have the ability to participate in leisure and tourism-related crowdfunding, perhaps by utilizing virtual reality technologies, because of the strong relationship between perceived behavioral control and behavior intention. Also, for improving funder participation, crowdfunding managers should encourage sponsors' desire by using mobile messages with hope, eager, and enthusiastic affections, using appropriate video and audio. In addition, if SMEs and micro-businesses want to launch reward crowdfunding, they might concentrate on designing the digital storytelling on websites with co-creation, attitude, subjective norm, and positive anticipated emotion aspects in mind because of the higher effect of these factors on the reward crowdfunding group. In contrast, if initiatives are sought to encourage investment crowdfunding in the visitor economy, those seeking crowdfunding support should focus on using elements of negative anticipated emotion, perceived behavioral control, and desire due to the higher impact of those elements on the investment crowdfunding group. Such insights are significant given the potential of crowdfunding to contribute to community and local economies [69], especially small ventures that may struggle to attract financial support through more traditional channels [70][71][72][73].

Limitations and Future Study Directions
Even though this work offers major contributions to better understanding leisure and tourism-related crowdfunding projects, there are some limitations that provide opportunities for future studies. This study examined entrepreneurs' co-creation with consumers, platforms, internal customers, and others so future research may need to investigate platforms' co-creation with other sets of stakeholders. Importantly, this project was undertaken in Korea and further studies are needed of crowdfunding in other cultural and regulatory contexts which may have different understandings of financial and risk and online and institutional trust in an investment context. In addition to international comparisons, research on crowdfunding in the visitor economy ideally needs to undertake more comparative work on crowdfunding between different sectors as well.
Since prior research found some implications from crowdfunding and sustainable development [74], future research on the crowdfunding of projects designed to promote sustainability, such as low-carbon developments, would be valuable to shed light on the risks and constraints of using crowdfunding to achieve finance solutions for the visitor economy and the sustainable development goals. Furthermore, because platform crowdfunding models (e.g., keep it all versus all or nothing) have a big impact on meeting fundraising goals [75], further multi-group analysis on different funding models could identify specific implications for successful projects. Cleantech crowdfunding projects are more likely to have higher capital goals, more photos, a video pitch, and longer text descriptions of the campaigns compared to no-cleantech ones [76]. Therefore, future research on crowdfunding in the visitor economy in relation to sustainability should focus attention on the differences between cleantech and non-cleantech crowdfunding in tourism-related industries, particularly in an Asian context where research is otherwise sparse.
Although co-creation in visitor economy crowdfunding is recognized as significant for local economies [69], research on co-creation in leisure and tourism-related crowdfunding has been substantially overlooked. Hence, this study built and verified a theoretically comprehensive research model including co-creation and the seven key MGB concepts (attitude, subjective norm, positive anticipated emotion, negative anticipated emotion, perceived behavioral control, desire, and behavioral intention), along with the moderator of investment and reward crowdfunding types. The findings of this study suggest that entrepreneurs need to highlight co-creation with consumers to better encourage support from potential investors. In addition, crowdfundraisers needs to focus on the six key constructs of MGB of attitude, positive and negative emotions, perceived behavioral control, desire, and behavior to obtain greater economic benefits. Finally, crowdfunding stakeholders need to develop their marketing strategies based on the two types of investment or reward crowdfunding in order to strengthen their appeal to potential sponsors, thereby potentially contributing to a more sustainable visitor economy.
Author Contributions: M.J.K. contributed to conducting the survey, collecting data, analyzing data, and writing draft; C.M.H. contributed to developing the questionnaire, conceptualization of research model, and writing and editing the manuscript.