Abstract
BACKGROUND:
In recent years, crowdfunding platforms have gained importance in the financing industry. An emerging issue is that only a fraction of investors decide to reinvest after the initial investment experience. One key factor that might be responsible for low reinvest rates, which ultimately results in lower crowdfunding success and unsustainable crowdfunding platforms, is dissatisfaction.
OBJECTIVE:
This paper aims to shed light on the repercussions of dissatisfying investment experiences. First, the authors investigate to what degree individuals’ dissatisfaction with a crowdfunding platform’s founders has a direct negative impact on their decision to continue investing on the platform. Second, this study examines whether the emergence of individuals’ distrust in a platform’s founders is a consequence of dissatisfaction and how this influences their decision to reinvest in projects on the platform.
METHODS:
The authors develop a modified version of the extended valence framework (cf., Kim et al. [1]) to investigate the relationships among dissatisfaction, distrust, perceived benefit, perceived risk, and intention to continue investing. The research model was analyzed using a partial least square analysis approach based on a sample of 233 respondent that participated in an online questionnaire.
RESULTS:
Our findings show that dissatisfaction, directly and indirectly, affects individuals’ intention to continue investing. Distrust is found to be an essential mediator between dissatisfaction and intention to continue investing. Furthermore, the effect of distrust on backers’ intention to reinvest is fully mediated by perceived benefit and risk, which are both found to be important determinants of their intention to reinvest, while perceived benefit exhibits a stronger absolute effect.
CONCLUSIONS:
This study provides significant contributions to the related literature. It explains why platforms might suffer from low reinvestment rates. Furthermore, it sheds light on the crucial role of distrust in the crowdfunding context. Theoretical and managerial implications are discussed.
Introduction
Crowdfunding has been established as a mechanism for social entrepreneurs to raise funds [2]. By initiating an ‘open call’ to the crowd, usually through the Internet, founders ask for monetary contributions from anyone interested in supporting the underlying cause [3]. Depending on the specific crowdfunding category, the compensation to contributors could be a share of the entrepreneurs’ company (i.e., equity-based), monetary repayment (i.e., lending-based), non-monetary rewards (i.e., reward-based), or nothing (i.e., donation-based) [4]. This study focuses on reward-based crowdfunding. Founders, who are creators in this context, collect money from funders (i.e., backers) to realize projects and offer rewards, often a copy of the outcome of a project, in compensation for monetary contributions (i.e., pledges).
The Massolution [5] report reveals that the market grew from US$6 billion in 2013 to US$34 billion in 2015, which indicates the increasing importance of crowdfunding. This circumstance attracted researchers’ attention to examine factors that influence crowdfunding projects’ success from the creators’ perspective (e.g., Ahlers et al. [6], Agrawal et al. [7], Clauss et al. [8], Colombo et al. [9], Davis et al. [10], and Vismara [11], Xinmin et al. [12]), backers’ perspective (e.g., Kang et al. [13], Liang et al. [14], Strohmaier et al. [15], and Zhao et al. [16]) or intermediaries’ perspective (e.g., Cumming and Zhang [17], Rossi et al. [18], and Russi and Vismara [19]).
The reward-based crowdfunding market can be observed by examining Kickstarter’s funding statistics since it is the world’s second-largest reward-based crowdfunding platform [5]. From 2011 to 2018, the per annum pledging amount grew from US$100 million to over US$4 billion. Interestingly, only 32% of all backers decide to continue using Kickstarter after their first pledging experience [20], which indicates that most first-time users do not become loyal backers. Several scholars found that backers willingness to contribute to crowdfunding projects decreases after their first pledging experience (e.g., [21–24]). Hence, it can be assumed that once a particular point of saturation is reached, the crowdfunding market may begin shrinking. Only few studies have investigated factors that determine individuals’ intention to continue pledging to a platform’s projects (e.g., Vismara [25]). Therefore, it is crucial to gain further insights into the factors that deter backers from continuing to support crowdfunding projects, and the measures platforms can take in order to build a loyal backer base.
In this context, Gerber and Hui’s [26] interviewed 81 backers and found that distrust in creators is a major deterrent to individuals’ participation in crowdfunding activities. This is an interesting finding since individuals initially tend to trust each other, while distrust emerges for a reason (e.g. a negative experience with the other party) [27]. Distrust deals with firm beliefs about another entity’s negative characteristics [28] and is associated with negative emotions, such as fear, or wariness [29]. Since most backers do not want to continue pledging to projects after their first pledge, it is conceivable that distrust in a platform’s creators might emerge due to a dissatisfying pledging experience. Another indication for this relationship can be found in prior research, which shows that dissatisfaction is an antecedent of negatives emotions [30]; thus, it will be essential to examine whether dissatisfaction causes distrust, which, in turn, negatively affects backers’ willingness to continue pledging. This will provide new insights into the reason why a crowdfunding platform suffers from a low repledging rate that ultimately lowers the platform’s success and may undermine the sustainability of the platform.
Studies in related contexts, such as e-commerce (e.g., Boadi et al. [31], Valvi and West [32], and Yen and Lu [33]), show that consumers base their decision to continue purchasing products from a specific online marketplace on their satisfaction or dissatisfaction with the purchase. Hence, it is interesting to examine the direct effect of dissatisfaction on backers’ intention to continue pledging to projects on the platform. This approach allows the authors to reveal new theoretical insights regarding to what degree distrust is an essential mediator between dissatisfaction and continuance intention.
Moreover, it will be worthwhile to study whether individuals’ distrust directly affects their continuance intention or if this effect is mediated by other factors. Prior empirical studies have provided evidence that distrust has direct and indirect effects on online users’ behavior. McKnight et al. [34] identified perceived risk as an important mediator between these two constructs. In this context, Kim et al.’s [1] extended valance framework postulates that individuals’ decision to perform a behavior is based on trust, perceived benefit, and perceived risk. The current study proposes a modified version of this framework and postulates that distrust, perceived benefit, and risk are determinants of continuance intention, while perceived benefit and risk are also mediators between distrust and continuance intention. In the crowdfunding literature, researchers have examined the role of perceived risk regarding backers’ intention to invest in crowdfunding (e.g., Zhao et al. [16]), while little attention has been paid to a comparative examination of perceived benefit and risk as they relate to backers’ intention to continue pledging. Studying whether losses have a greater impact than gains or vice versa will be worthwhile as it would provide key insights for practitioners and highlight whether managing backers’ perceptions of risks or benefits are more critical.
To summarize, the major objective of this study is to provide insights into the reasons why crowdfunding platforms suffer from low reinvest rates. Dissatisfaction with a platform’s creators is assumed to both, directly and indirectly, affect backers’ intention to continue pledging to projects on the platform. The indirect effect lies in the possibility that dissatisfaction causes distrust, which is— based on the modified version of Kim et al.’s [1] extended valance framework— associated with perceived benefit, risk, and continuance intention. Therefore, this article addresses the following research questions:
Is distrust in a platform’s creators an important mediator of dissatisfaction and continuance intention in crowdfunding?
To what extent do perceived benefit and risk mediate the effect of distrust on continuance intention?
Does backers’ perceived benefit or risk have a greater absolute impact on their intention to continue pledging to a crowdfunding platform’s projects?
The results show that dissatisfaction has negative direct and indirect effects on backers’ continuance intention. The indirect effect is partially mediated by distrust. Furthermore, backers’ perceptions of potential gains (benefits) have an absolute higher impact on their intention to continue pledging than perceptions of potential losses (risks). Finally, the results provide empirical evidence that the relationship between distrust and continuance intention is fully mediated by perceived benefit and risk.
Based on these results, the study provides new insights pertinent to researchers and practitioners in the crowdfunding context. From a theoretical viewpoint, crowdfunding is an emerging issue [35]. While most studies have examined crowdfunding success from the creators’ viewpoint, this is one of the first studies to examine continuance intention from the backers’ perspective. This study sheds light on how individuals’ dissatisfaction influences their intention to continue pledging and provides empirical support for the hypothesis that dissatisfaction is an essential antecedent of distrust. It extends our knowledge of the role of distrust in the context of crowdfunding by providing an examination of how distrust influences individuals’ decision making. From a managerial perspective, this study is of interest to crowdfunding platform proponents and platform managers since the findings provide guidelines for increasing backers’ intention to continue pledging to a platform’s projects, which ultimately leads to higher reinvestment rates and increases the platform’s success.
This paper is structured as follows. The next section provides an overview of the theoretical foundations, followed by a section that introduces the conceptual model and hypotheses. The fourth and fifth sections discuss the methodology and data analysis. The last section presents the findings, implications, and limitations.
Theoretical foundations
This section discusses the theoretical background of the present research. First, we explain the institutional logic (cf., Thornton and Ocasio [36]) that backers apply when deciding to pledge to projects. Second, we review literature related to dissatisfaction, followed by a discussion of distrust. Finally, we introduce the extended valence framework [1], and the theory of reasoned action (TRA) [37] as background frameworks for this study.
The logic behind backers’ decision to invest in crowdfunding
Institutional logics are defined as “socially constructed, historical patterns of material practices, assumptions, values, beliefs, and rules by which individuals produce and reproduce their material subsistence, organize time and space, and provide meaning to their social reality.” [36] (p. 804). Therefore, individuals’ of different contexts are dominated by different institutional logics because each context has different values, beliefs, and rules that individuals’ follow. Fisher et al. [38] identified institutional logics for audiences of new technology-based ventures and found that backers of reward-based crowdfunding platforms are dominated by community logic. Backers characteristics are that they support projects for community advancement, have beliefs in trust and reciprocity, and are emotionally connected to projects they pledged money for and their creators [38].
From a community logic point of view, it is conceivable that negative emotions and beliefs regarding project creators might deter backers from perusing to support projects. In this context, it is conceivable that distrust, which deals with firm beliefs about another entity’s negative characteristics [28] and is associated with negative emotions [29] (see section 2.2 for a detailed explanation), might be a key deterrent for backers to continue pledging. Therefore, the current study aims at shedding light on whether dissatisfaction causes distrust and whether distrust decreases backers’ intention to continue pledging.
Dissatisfaction
Scholars identified key factors that influence customer satisfaction (e.g., [39, 40]). The expectation-disconfirmation theory (EDT) postulates that individuals’ satisfaction or dissatisfaction with a purchased good or service depends on the positive or negative disconfirmation of their prior expectations (i.e., their beliefs about the product prior to the transaction) after consuming the purchased good or service and evaluating its performance [41]. Thus, negative disconfirmation occurs when the product is worse than expected, which leads to dissatisfaction, while a product that performs better than expected results in satisfaction. In general, dissatisfaction deals with the degree to which a product or service provides unpleasant consumption-related fulfillment [42]. Dissatisfied individuals are in a psychological state of cognitive and affective discomfort [43].
In the marketing literature, satisfaction or dissatisfaction has been used to evaluate a person’s intention to continue using an information system. In studies conducted by Lu et al. [30], Bhattacherjee [44], Valvi and West [32], satisfaction or dissatisfaction is defined based on the usefulness of an information system or e-shop, while in a study by Dabholkar et al. [45], satisfaction is related to customers’ perceptions of service quality. Other studies (e.g., Anderson and Srinivasan [46]) have investigated how individuals’ satisfaction or dissatisfaction with the purchase process affects e-loyalty. The current study focuses on backers’ dissatisfaction with creators who operate on a specific crowdfunding platform because the implementation of projects and the associated delivery of rewards depend entirely on the projects’ creators [4, 47]. Thus, when backers decide to pledge to projects, they have expectations of how the creators will perform. Consequently, dissatisfaction emerges when the creators’ performance is worse than expected. Following the related literature, the current study defines dissatisfaction with a platform’s creators as backers having cognitive and affective discomfort resulting from prior transactions with project creators operating on a specific crowdfunding platform.
To the best of the authors’ knowledge, few studies have empirically investigated the role of dissatisfaction in the crowdfunding context. The current study aims to examine the effect of backers’ dissatisfaction with creators that operate on a specific platform on their intention to continue pledging to projects on this platform. We also focus on shedding light on repercussions that emerge due to individuals’ dissatisfaction and analyze how they affect backers’ intentional behavior.
Distrust
Distrust is defined as “confident negative expectations regarding another’s conduct” [29] (p. 439) and is associated with various emotions, such as wariness, skepticism, and anxiety. It amplifies certain behaviors, such as observed defensiveness, watchfulness, and vigilance [29, 48], and is related to negative beliefs about a person’s characteristics [28, 49]. Therefore, distrust consists of affective and cognitive aspects that influence individuals’ behavior [49]. This is underpinned by numerous empirical studies. In business-to-consumer e-commerce, distrust is identified as a key determinant of consumers’ buying behavior [50, 51]. In business-to-business data exchange, it influences the proponents’ willingness to continue using the data exchange system [34]. It should be noted that individuals initially do not distrust another entity (e.g., a person or group), but distrust can emerge due to unpleasant events, such as a dissatisfying experience [27].
Pavlou and Gefen’s [52] findings show that individuals’ decision to use services offered by proponents of an online platform depends on their trust or distrust in the proponents as a group. Individuals are unwilling to further interact with specific members and build dyadic trust or distrust unless they have trust in the community. These findings have been confirmed in later studies (e.g., Hong and Cho [53], Verhagen et al. [54], and Yang et al. [55]). The present study aims to examine the determinants of backers’ intention to continue pledging on a specific platform, which, in most cases, means that they support projects from different creators. This intention might not depend on distrust in a specific creator since the decision to continue pledging needs to be made before selecting a specific project. Therefore, this study focuses on distrust in a platform’s creators as a group and how it affects the backers’ behavior.
In this study, distrust in creators is identified as a unidimensional construct composed of four negative beliefs about creators’ characteristics. The basis for this construct is Moody et al.’s [28] trust-distrust model, which postulates that distrust has the opposite valence of trust but contains the same components. Distrust occurs when one believes that the other party is incompetent, malevolent, or deceitful. We define these beliefs in the crowdfunding context following Moody et al. [28]. Incompetence is related to the creators’ lack of the required technical and business-related skills to develop and realize projects. Malevolence refers to the degree to which creators are suspected of being likely to harm their backers intentionally. We define deceit as the belief that creators will behave insincerely and post fabricated project data. Finally, the fourth belief is related to backers’ general attitude regarding whether project creators of a specific platform should be distrusted.
The extended valence framework and theory of reasoned action
The valence framework [56] is based on the idea that individuals’ decision to perform a behavior depends on subjective beliefs about possible gains (i.e., perceived benefit) and losses (i.e., perceived risk) while trying to maximize the net valence (i.e., gains minus losses). Empirical studies examining offline or online consumer decision-making have found that perceived benefit and risk are key antecedents of consumer behavior (e.g., Kim et al. [1], Peter and Tarpey [56], Forsythe and Shi [57], and Kuo and Feng [58]). Following Peter and Tarpey [56], in this study, perceived benefit is defined as a person’s subjective beliefs about potential gains as a consequence of pledging to crowdfunding projects, while perceived risk is defined as a person’s subjective beliefs about potential losses as a consequence of pledging to crowdfunded projects (see section 3.3 for a detailed explanation).
Kim et al. [1] developed the extended valence framework (Fig. 1, upper part), which is based on the theory of reasoned action (TRA) [37] and the Web Trust Model proposed by McKnight et al. [59]. Trust, perceived benefit, and perceived risk are antecedents of purchase intention, which, in turn, determines actual purchasing behavior. Furthermore, trust indirectly affects purchase intention via benefit and risk. Theoretically, the model is built upon the TRA that postulates behavioral intention is the best predictor of actual behavior. Behavioral and normative beliefs are key constructs that influence individuals’ intentional behavior. Depending on the developed model, these beliefs can be mediated by the individual’s attitude towards the behavior and subjective norms [26]. Numerous meta-analyses have provided evidence for the TRA’s predictive utility (e.g., Armitage and Conner [60], Sheeran and Taylor [61], and Webb and Sheeran [62]). Since the current study aims to examine the role of distrust in individuals’ decision to continue pledging on a specific platform, the authors propose a modified version of Kim et al.’s [1] extended valence framework. Distrust (i.e., negative beliefs about the creators’ characteristics) is postulated to be partially mediated by perceived benefit (i.e., beliefs about possible gains) and perceived risk (i.e., beliefs about possible losses). Based on the TRA, intention to continue pledging is the key determinant of actual behavior (i.e., a pledge; see Fig. 1, lower part). It should be noted that numerous studies provide empirical evidence for the strong relationship between behavioral intention and actual behavior (e.g., [63–65]).

The extended valence framework in the e-commerce context (adapted from [1]; upper part) and a modified version that is based on distrust in the crowdfunding context (lower part).
The research model is illustrated in Fig. 2. Based on the theoretical foundations, dissatisfaction is hypothesized to have a direct impact on continuance intention (H1) and distrust (H2). The remainder of the research model follows a modified version of Kim et al.’s [1] extended valence framework. Distrust is assumed to affect continuance intention directly (H3) and through perceived benefit (H4 and H6) and perceived risk (H5, and H6) indirectly. To gain insights into how dissatisfaction affects individuals’ behavior, we evaluate the direct and indirect effects of dissatisfaction on individuals’ continuance intention.

Research model.
As mentioned in section 2.1, dissatisfaction with a platform’s creators is described as backers having cognitive and affective discomfort resulting from prior transactions with project creators. Satisfaction or dissatisfaction is an essential part of the marketing literature. Various studies have found that satisfaction has a positive impact on continuance intention (e.g., Bhattacherjee [44], Valvi and West [32], Bhattacherjee and Premkumar [66], Hong et al. [67], and Yen and Lu [33]). For example, in the context of e-commerce, Yoon [68] found that satisfaction with shopping websites has a positive effect on consumers’ purchase intention. It is, therefore, reasonable to assume that this link also exists in the crowdfunding context. Since satisfaction and dissatisfaction are two ends of the same continuum [42], dissatisfaction is supposed to have a significant negative impact on continuance intention. Hence, we hypothesize the following:
The link between dissatisfaction and distrust can be considered from the affective and cognitive perspectives. In general, distrust can be seen as a construct that is activated by affective processes, such as negative emotions [69]. Prior studies provide empirical evidence that dissatisfaction triggers negative emotions [30] and shed light on the influence of dissatisfaction on distrust from an affective perspective. From a cognitive perspective, McKnight et al. [27] postulated that individuals initially trust each other, while distrust emerges for a reason. Thus, it is reasonable to a assume that backers’ distrust in a platform’s creators is triggered by dissatisfaction because it reflects the creators’ negative characteristics, such as being incompetent, malevolent, or deceitful (cf., Moody et al. [28]).
Distrust’s effects on continuance intention and perceived risk and benefit
The TRA [37] postulates that individuals’ positive or negative beliefs about whether a certain behavior will lead to the expected outcome directly or indirectly affect their behavioral intention, which, in turn, determines actual behavior. Based on Moody et al. [28], in this study, distrust is related to individuals’ negative beliefs about another’s entity’s characteristics. An individual who believes that an entity is incapable, malevolent, or deceitful (distrust beliefs, cf., Moody et al. [28]) will suspect that the entity is not capable of fulfilling promises and, given the opportunity, will behave opportunistically. Thus, these beliefs may play an essential role when individuals decide whether to continue pledging to a platform’s project. It is, therefore, reasonable to assume that individuals who believe that creators are incapable, malevolent, or deceitful are likely to lower their intention to continue pledging to projects since it increases the perceived possibility that creators will fail to deliver the promised outcome. Previous studies have provided empirical evidence for the negative impact of distrust on behavioral intention (e.g., Ou and Sia [50]). Based on these empirical findings and logical conclusions, we propose the following:
Perceived risk is related to an individual’s subjective beliefs about potential losses as a consequence of (re)pledging to crowdfunded projects. Since distrust is associated with various feelings, such as fear, suspicion, vigilance, and watchfulness [29], it is conceivable that individuals who distrust an entity may be more sensitive to perceived negative consequences. McKnight et al.’s [70] study revealed that the disposition to distrust has a significantly strong impact on high-risk perceptions. Furthermore, McKnight et al. [34] found that in business-to-business relationships, distrust in the supplier significantly increases the buyer’s perceived risk. For these reasons, we developed the following hypothesis:
In the current study, perceived benefit is related to an individual’s subjective beliefs to obtain certain gains when (re)pledging to crowdfunded projects. The outcome of crowdfunding projects and related gains backers hopes to obtain (e.g., collecting a special reward) depends solely on the creators. Thus, backers who believe in obtaining certain benefits from pledging to projects must also believe that the creators are capable, honest, and benevolent. Conversely, backers who distrust the community of creators are confident that creators will behave incompetently, malevolently, or deceitfully, which might mitigate their perceptions of possible gains. From an affective perceptive, individuals’ perceived benefits rely on positive feelings, such as hope and faith, and has, therefore, a positive valence (cf., Izard [71]). However, the process of fostering distrust is accompanied by negative feelings, such as fear, skepticism, wariness [29]. Based on these arguments, the authors posit the following hypothesis:
Effects of perceived benefit and risk on continuance intention
In the marketing literature, several studies have provided empirical evidence that perceived risk is a key determinant of behavioral intention (e.g., Forsythe and Shi [57], Jiuan Tan [72], Jarvenpaa and Todd [73], and Crespo et al. [74]). Regarding consumers’ purchase-related behavior, Jiuan Tan [72] found that perceived risk is more important online than offline. Online transactions are associated with a certain amount of information asymmetry due to the absence of face-to-face conversations and lack of physical observation of the seller’s actions, which can cause uncertainties regarding the seller’s behavior and trigger concerns about potential losses [75]. In this context, the frequent potential losses individuals confront when conducting online transactions are identified as performance, financial, time, psychological, social, and privacy loss [74].
In the crowdfunding context, backers might fear that rewards will not perform as promised or might not be delivered (financial loss) or delivered with a delay (time loss) due to the creators’ opportunistic behavior. They might also fear the loss of social status amongst their friends and family by pledging money for projects with an unpredictable outcome. Furthermore, the ever-present fear of the possibility that creators’ might behave opportunistically may result in feelings of anxiety or uneasiness (i.e., psychological loss). Finally, because crowdfunding is conducted online, backers may also be confronted with the fear of losing control of their data (i.e., privacy loss). Since the backers’ perceived risk is composed of these potential losses, we posit the following hypothesis:
As defined above, perceived benefit is related to individuals’ subjective beliefs about potential gains as a consequence of pledging to projects. Therefore, it provides incentives for backers to pledge to crowdfunded projects. In this context, (re)pledging to crowdfunding projects can be motivated by a variety of possible gains (i.e., performance, financial, time, safety, psychological, and social gains), which is in line with prior studies examining factors that motivate backers to pledge to projects (i.e., collecting rewards, being part of a community, helping others, supporting a cause that is in line with one’s identity; cf., Gerber and Hui [26, 76]). First, by pledging to a project, a person might believe in obtaining a performance gain due to specific features or design rewards may offer (cf. [77]). Second, pledging to a unique project may result in time gain, since the person receives the reward (a final product) before other people can buy it on the market. Third, often the incentive for backers to pledge to a project is to obtain the final product (the outcome of the project, e.g., a smartwatch) at a lower price than when the product is commercially available later on, resulting in financial gains. Fourth, a person might obtain a social gain from pledging to a project because he or she might rise in social status among friends by owning a product that is not on the market yet. However, backers that are motivated by supporting an important cause or helping others might also increase their status among their friends and family. Finally, individuals may feel a psychological gain by being part of a community that fits well to their self-image. Since backers’ perceived benefit is composed of these potential gains, the authors propose the following hypothesis:
Control variables
The authors included two additional control variables in the research model: a disposition to distrust (i.e., affecting distrust) and familiarity (i.e., affecting perceived risk, perceived benefit, and intention to continue pledging). An individual’s disposition toward being distrustful of others has been shown to increase distrust (e.g., McKnight et al. [34, 70], Wu et al. [78], and Bélanger and Carter [79]). Therefore, the authors expect that this relationship also holds in the crowdfunding context. Familiarity is related to a person’s experience in the crowdfunding context. Several studies have shown that familiarity influences a person’s intentional behavior (e.g., Gefen [80]) and perceived risk (e.g., Nepomuceno et al. [81]). Individuals who are more familiar with crowdfunding feel less uncertainty when using a platform to pledge than those who have less experience.
Methodology
Sampling and data collection
This study focuses on factors that influence backers’ intention to continue pledging to a platform’s projects. Because different platforms have different rules and policies that might influence a person’s pledging experience, this study collects data from individuals who has pledged to at least one project on Kickstarter and has had a dissatisfying experience with this platform. This choice is appropriate as Kickstarter is currently the second-largest reward-based crowdfunding platform headquartered in the United States [5] and has a sizable community of backers (∼15.6 million), who funded 156,138 projects in 2018 [20]. Because the authors do not have access to the database containing all Kickstarter users, a non-probabilistic sampling procedure was applied (i.e., consecutive sampling) to collect data.
In practice, an online survey was used to gather data. Data collection was conducted using Amazon Mechanical Turk (MTurk; www.mturk.com), a crowdsourcing marketplace, and SurveyMonkey (www.surveymonkey.com), an online survey tool. MTurk is deployed by Amazon Inc. and allows researchers (i.e., requesters) to recruit suitable participants (i.e., workers) for their surveys. Workers that participate in a survey are compensated with small amounts of money from the requesters [82]. SurveyMonkey provides web-based functionalities to create questionnaires and collect user responses. Thus, workers were linked from MTurk to SurveyMonkey to fill out the survey. It should be noted that many studies in various research fields have provided evidence that data collected through MTurk have the same degree of validity as data collected from other sources (e.g., Berinsky et al. [83], Simons and Chabris [84], Horton et al. [85], Holden et al. [86], Schütz et al. [87]).
Furthermore, the authors decided to follow Sheehan’s [88] suggestions regarding ways to avoid common problems using MTurk for data collection by (i) implementing screening questions, (ii) implementing attention checks, and (iii) limiting the workers’ maximum time for completion. The former was performed by adding several screening questions at the beginning of the survey to verify whether the participant was suitable. A suitable participant could answer questions regarding specific processes and terms on Kickstarter correctly, assured that he or she pledged to at least one project on Kickstarter and has had a dissatisfying experience in the context of reward-based crowdfunding. Second, the authors periodically included open-ended questions to check whether the participant was still attentive and, therefore, able to provide a meaningful answer. Finally, Sheehan’s [70] recommends setting the time limit to three times the pilot test’s average. Therefore, we set the time limit to 21 minutes because pilot participants needed 7 minutes on average to finish the survey.
On MTurk, the requester must determine the number of desired participants. Thus, it is necessary to calculate the minimum sample size required to perform a meaningful analysis beforehand. Since the authors decided to use partial least square (PLS) analysis (the reasons for this choice are provided in section 5), an a priori power analysis was performed, using the free statistics calculator V4 by Soper [89], as suggested by Chin and Newsted [90] to calculate the required minimum sample size. The number of predictors was set to six since it represents the maximum number of predictors (see perceived risk). The remaining parameters were set following Cohen et al. [91]: statistical power level (0.80), probability level (0.05), and effective size (0.10). The computed required minimum sample size was 142. The survey was hosted on MTurk on August 28, 2019, collecting data from 346 respondents. 113 individuals who were willing to participate were excluded due to the screening questions or they did not pass the attention checks— they did not answer the open-ended questions. After a preliminary data analysis, the sample contained 233 entries. It should be noted that each participant was compensated US$1.50.
Table 1 presents a demographic profile of the participants. It is consistent with other studies in the crowdfunding context (e.g., Liang et al. [14], Zhao et al. [16], and Cumming et al. [92]). Most of the participants were between 30 and 44 years old (>64%) and predominantly male (∼62%). This is in line with Cumming et al.‘s [92] study showing that the average age on other crowdfunding platforms is ∼42 (Crowdcube) and ∼46 (AIM) and that around 30% of the proponents are female. Furthermore, most of our participants were highly educated (∼80% of the participants have a bachelor degree), which is in accordance with the demographic profile of Liang et al. [14] and Zhao et al. [16]’s studies.
Demographic profile of respondents (N = 233)
Demographic profile of respondents (N = 233)
The survey instrument is illustrated in Table A1. Multiple measurement items were used for each construct to increase reliability, and all items were developed based on prior literature. The authors used a 5-point Likert scale from 1 (i.e., “strongly disagree”) to 5 (i.e., “strongly agree”) to measure all items. Items for dissatisfaction were derived from Hellier et al.’s [93], Oliver and Burke’s [94] and Yen and Lu’s [33] research. Four items were used to measure distrust. Three were adapted from McKnight et al.’s [34] study. Additionally, one item was added measuring the overall distrust belief. Measurement items of perceived risk and perceived benefit were developed based on Pavlou and Gefen’s [52] and Peter and Ryan’s [95] work, respectively. Intention to continue pledging was measured with items adapted from Ajzen’s [96] research. Furthermore, control variables were also adapted from prior research. Items regarding a disposition to distrust and familiarity were derived from McKnight et al.’s [70] and Gefen’s [80] research, respectively.
Results
The proposed model was examined using PLS analysis [97]. In practice, the SmartPLS 3.0 program [98] with bootstrapping set to 5000 subsamples was leveraged to conduct the PLS algorithm. We chose a component-based structural equation modeling approach (here, PLS) over covariance-based-structural equation modeling for two reasons: sample size and normality. First, PLS analysis is the appropriate choice when the sample size is small as long as it has adequate power [98]. As discussed in Section 4.1, although our sample is moderate (233), it possesses adequate power. Furthermore, it does not contain extreme outliers or missing data. Second, in contrast to covariance-based-structural equation modeling, PLS analysis is robust against slightly non-normal data (skew <1.1 and kurtosis <1.6) [99–102]. Goodhue et al. [100] stated that only very skewed data (skew> = 1.8 and kurtosis < = 3.8) might result in lower power. In our data, only 2 out of 27 items exhibited skewness and kurtosis values of 1.1 and 1.6, respectively, and all the skewness and kurtosis values were lower than 1.8 and 3.8, respectively. For these reasons, PLS is considered the appropriate choice.
Since the same instrument was used to gather data for dependent and independent variables, the author performed two tests to rule out common method bias (CMB). First, the authors applied Kock’s [103] full collinearity test and found that all factor-level VIFs were lower than 3.3, indicating that the model is free of CMB issues. Second, a model may be affected by CMB issues if correlations between pairs of constructs are greater than 0.90 [104]. All the correlations were smaller than.90 (see Table 2), which provides further evidence that the model does not suffer from CMB issues.
Correlation among constructs. The bold values are the square roots of the corresponding construct’s AVEs
Correlation among constructs. The bold values are the square roots of the corresponding construct’s AVEs
Based on Hair Jr et al.’s [105] and Ab Hamid et al.’s [106] suggestions regarding ways to evaluate the reliability and validity of a measurement model, the authors assessed indicator reliability, internal consistency, convergent validity, and discriminant validity. Indicator reliability assesses whether the items of a construct measure similar aspects [106]. Items with loadings lower than 0.70 should be dropped [105]. All of the item loadings are higher than the cut-off value of 0.70 (see Table 3, bold values), except for PR5, which was dropped. The authors used Cronbach’s alpha (CA), and Dillon-Goldstein’s rho (RHO) [107] to validate internal consistency. Each construct’s CA, and RHO values of the measurement model must exceed the threshold values of 0.6, and 0.7, respectively [99, 108]. Table 3 shows that internal consistency is acceptable since all constructs’ CA, and RHO value are greater than the recommended cut-off values.
Cross loadings, Cronbach’s Alpha (CA), D.G. Rho (D. G.), Average Variance Extracted (AVE)
Cross loadings, Cronbach’s Alpha (CA), D.G. Rho (D. G.), Average Variance Extracted (AVE)
The authors calculated the average variance extracted (AVE) to check the convergent validity of the model. Since the AVE of each construct is greater than the threshold value of.50, the convergent validity is confirmed [105, 108] (see Table 3). Finally, the Fornell-Larcker criterion [108] was used to validate discriminant validity. If the AVE’s square root of each construct is greater than the correlation between the considered and other constructs discriminant, validity is established. As shown in Table 2, the criterion is fulfilled.
The authors tested for multicollinearity by analyzing the variance inflation factors (VIFs). The fact that all the VIF values are below five provides support that the model is free of multicollinearity issues [101].
The results of the hypothesis testing and the effects of control on latent variables are presented in Fig. 3 and Table 4. All of the hypotheses are supported except for H3. The results are discussed in section 6.

PLS analysis results.
Results of Hypotheses Testing
*Significant at the 0.05 level; **Significant at the 0.01 level; ***Significant at the 0.001 level.
The authors computed the adjusted R2 values [109] (i.e., explained variance) to estimate the structural model’s predictive power. The findings indicate that 47% of the variance in intention to continue pledging was explained by dissatisfaction, perceived benefit, perceived risk, and familiarity. Distrust and familiarity explain 21% and 34% of the variance in perceived benefit and risk, respectively. Finally, dissatisfaction and disposition to distrust explain around 29% of the variance in distrust. We further examined two mediating effects: (i) distrust as a mediator between dissatisfaction and continuance intention and (ii) perceived risk and benefit as mediators between distrust and continuance intention. The mediation analysis was performed following Hayes’s [110] procedure and is summarized in Table 5.
Mediating effects
*Significant at the 0.05 level; **Significant at the 0.001 level.
Our study focused on the repercussions of backers’ dissatisfaction with a platform’s creators and, in particular, the direct and indirect effects of dissatisfaction on backers’ intention to continue pledging to a platform’s projects with the objective to shed light on why crowdfunding platforms suffer from low reinvestment rates. By investigating dissatisfaction from a distrust-based perspective, the authors posed the question of whether distrust is an essential mediator between dissatisfaction and continuance intention. Finally, based on a modified version of Kim et al.’s [1] extended valence framework, we investigated how distrust affects perceived benefit and risk, which are postulated to be determinants of continuance intention.
Dissatisfaction has been found to have a significant, direct, and negative impact on individuals’ intentions to continue pledging (H1 was supported). Thus, this finding supports the assumption that dissatisfaction will directly result in low reinvestment rates, which, in turn, lead to lower crowdfunding success. Moreover, the results show that dissatisfaction increases distrust (H2 was supported) and is found to partially mediate the relationship between dissatisfaction and an individual’s intention to continue pledging (see Table 5). McKnight et al. [27] postulated that individuals initially trust each other, while distrust emerges for a reason. The findings show that dissatisfaction is an important antecedent of distrust. A possible explanation is that dissatisfaction with a platform’s creators leads to negative emotions (cf., Lu et al. [30]), which, in turn, foster distrustful beliefs about the creators’ characteristics, for example, that they might be incompetent or have deceitful intentions. This might be due to the fact that distrust can be seen as an affectively activated construct [69], which makes negative emotions the ideal trigger for the emergence of distrust [111].
Based on the modified version of the extended valence framework, distrust was hypothesized to decrease continuance intention and perceived benefit, as well as increase perceived risk. The results provide some interesting findings. The effect of distrust on continuance intention was not significant (H3 was not supported), and the magnitude was small (–0.06). Thus, the direct impact of distrust on continuance intention is negligible. However, distrust significantly increased perceived risk (H4 was supported) and significantly decreased perceived benefit (H5 was supported). While some studies have analyzed the effect of distrust on perceived risk (e.g., McKnight et al. [34]), little attention has been paid to a comparative examination of perceived risk and benefit and how distrust affects them. Table 4 shows that the effect of distrust on an individual’s intention to continue pledging was fully mediated by perceived benefit and risk. A possible reason for the strong relationships among distrust, perceived benefit, and risk is that distrust is accompanied by negative feelings, such as wariness and fear [29]. Thus, individuals may be more sensitive regarding negative comments about fraudulent project creators and scams on crowdfunding platforms (i.e., ‘e-word-of-mouth’), which increases perceptions of potential losses and decreases perceptions of possible gains.
Finally, perceived risk and benefit were hypothesized to be determinants of intention to continue pledging. Both hypotheses were supported since the results show that perceived risk had a significant negative impact and perceived benefit a significant positive effect on continuance intention (H6 and H7 were supported). The most surprising aspect of the comparative examination is that the absolute impact of perceived benefit is more than twice as strong as perceived risk. Thus, backers’ intention to continue pledging to projects is driven more by the perception of possible gains than losses. An explanation can be found by taking the motivating factors of the reward-based crowdfunding community into account. In this context, Gerber and Hui [26] interviewed 81 backers and found that the greatest incentive for backers to pledge to a project is the possibility of collecting unique rewards that are impossible to obtain from another source (e.g., a crowdfunded album of a specific music band). The possibility of owning unique products that are impossible to obtain otherwise might outweigh fears about potential losses. In fact, experienced backers are aware of the fact that investing in crowdfunding is accompanied by certain risks, which may, therefore, play a tangential role regarding their decision to continue pledging when compared to the perceived potential benefits they hope to gain.
However, the findings show that perceived risk significantly decreases reward-based crowdfunding backers’ intention to continue pledging. Arguably, investors’ perceptions of risk depend on the pledging amount. The more money that is at stake, the more risk the backer might perceive [112, 113]. This is underpinned by results of studies in the context of donation-based crowdfunding showing that the absence of financial risk (donors mostly pledge small amounts of money) is a motivating factor for donors to participate (e.g., [114]).
Theoretical implications
The findings of this study contribute to the literature in several ways. First, crowdfunding remains an emerging issue [26] that lacks related research regarding the backers’ perspective. To date, most studies that take the backers’ perspective into account have examined factors that influence their pledging intention (e.g., Liang et al. [14], and Zhao et al. [16]). The current study analyzes factors that influence the backers’ intention to continue pledging to projects, thereby providing new, significant insights.
Second, this study examines dissatisfaction from a distrust-based perspective and, therefore, adds new knowledge to the marketing and crowdfunding literature by providing novel insights on how dissatisfaction, directly and indirectly, affects backers’ behavioral intention. The relationships among dissatisfaction, distrust, and continuance intention were examined. The findings provide empirical evidence that distrust is an essential mediator between dissatisfaction and continuance intention. This result also contributes to the distrust literature, which postulates that distrust only emerges when individuals believe that the other party is capable of behaving opportunistically [115], by showing that dissatisfaction is an antecedent of distrust. Thus, future research investigating backers’ backers’ post-campaign behavior should consider these relationships.
Third, this study contributes to the literature of behavioral change theories by presenting a modified version of Kim et al.’s [1] extended valence framework. This complements existing research on distrust, which has found that distrust has a negative impact on perceived risk (e.g., McKnight et al. [34]]) but neglected perceived benefit. Our findings show that the effect of distrust on continuance intention is fully mediated by perceived benefit and risk.
Fourth, based on valence framework [56], which is part of the extended valence framework, the authors examined the direct effects of perceived benefit and risk on backers’ intention to continue pledging to projects. Therefore, this study extends the valence framework to the reward-based crowdfunding context and contributes to the crowdfunding literature by showing that individuals’ perceived benefits have a greater impact on their intention to continue pledging than perceived risk. Thus, based on these findings, future studies in the crowdfunding context should allow modeling perceived benefit even though their main objective lies in examining perceived risk
Practical implications
Since this study provides insights into how dissatisfaction influences a backer’s willingness to continue investing money in other projects, it is considered to be of major interest from a managerial perspective. As pointed out in the introduction, two-thirds of backers on Kickstarter do not continue pledging to projects after their first pledging experience [20]. Based on the findings, it is highly recommended that platform proponents try to minimize backers’ dissatisfaction. Dissatisfaction with creators is the result of negative disconfirmation of individuals’ expected experience and perceived performance. In this context, the emergence of dissatisfaction may often lie in the insufficient communication of the creators’ actual performance. The primary issue might be that even if creators tried their best to implement projects, they failed; backers cannot distinguish whether the failures were due to the creators’ malevolent intentions or unforeseeable business issues during the project’s implementation. Thus, backers’ dissatisfaction might emerge due to a lack of third parties that could provide further information about creators’ behavior. In practice, third parties could analyze the creators’ behavior based on which measures were taken to implement the project and how the funding was spent and then publish an overall compliance rating of the creators’ performance. This process could reduce the divergence between individuals’ perceived and creators’ actual performance, which ultimately lowers their dissatisfaction and consequently, the emergence of distrust.
Furthermore, this study shows that backers’ intention to continue pledging to a platform’s projects is driven more by the perception of possible gains than losses. These findings may guide platform managers and proponents to adopt new fundraising and marketing strategies. Creators, therefore, are recommended to invest considerable effort into promoting the benefits they provide to their backers. To date, project creators mainly focus on promoting rewards, while they might unintentionally ignore communicating other benefits of interest to backers. For example, backers might perceive a psychological gain if they know how the rewards are produced (e.g., manufactured under fair working conditions) or a social gain if creators present long-term plans about how the project might influence the lives of people involved in the projects for the better. In this context, platform managers should refine their platforms’ rules and guidelines for creators to promote this behavior. Moreover, projects that individuals associate with possible high gains might also be those, which they connect with high possible losses. Based on our findings, platforms should adopt their recommender systems by trying to recommend projects to regular backers that perceive possible high gains, although they might be seen as risky to invest.
Limitations
Although this study produced theoretical and practical insights, it is subject to limitations that need to be discussed. This study collected data from a sample of Kickstarter backers., the pledging system of reward-based crowdfunding platforms operates on either the ‘keep-it-all’ or ‘all-or-nothing’ principle [116]. The former allows project creators to keep the money collected at the end of the pledging period, even if the pledging goal has not been reached. However, ‘all-or-nothing’ means that pledges will be returned to the backers if the project fails to reach the pledging goal. Since this study focused on Kickstarter, which utilizes the ‘all-or-nothing’ principle, the findings of this study may not be generalized to crowdfunding platforms following another pledging system. Furthermore, the variables used in this study, such as distrust in creators or perceived risk, might be highly influenced by the pledging system’s mode of operation. Finally, PLS-SEM was used to evaluate the research model, which has the disadvantage of not providing a global goodness-of-fit measurement [105].
Footnotes
Appendix A
Measurement items for constructs
| Constructs | Measurement items | |
| Dissatisfaction with creators operating on the crowdfunding platform | ||
| DIS1 | Regarding the projects I previously pledged to on Kickstarter, I am dissatisfied with the performance of the creators. | |
| DIS2 | Regarding the projects I previously pledged to on Kickstarter, I am displeased with the experience of transacting with the creators. | |
| DIS3 | Regarding the projects I previously pledged to on Kickstarter, I am unhappy with the creators’ behavior. | |
| Distrust in the creators operating on the crowdfunding platform | ||
| DC1 | I am not sure that project creators on Kickstarter would act in my best interest. [malevolence] | |
| DC2 | It is uncertain whether project creators on Kickstarter would keep their commitments [deceit] | |
| DC3 | I am skeptical about whether project creators on Kickstarter have the necessary skills to implement their projects after successful funding. [incompetence] | |
| DC4 | I believe that most project creators on Kickstarter should be distrusted. | |
| Perceived risk | ||
| PR1 | I believe that pledging to projects on Kickstarter is risky because the rewards may not meet my expectations. | |
| PR2 | I believe that pledging to projects on Kickstarter is risky because I am afraid that I will not get my money’s worth from the reward. | |
| PR3 | I believe that pledging to projects on Kickstarter is risky because it is likely that there will be a delay in reward delivery. | |
| PR4 | I believe that pledging to projects on Kickstarter is risky because it is likely that my personal information will be used improperly. | |
| PR5 | I believe that pledging to projects on Kickstarter is risky because my friends and family would think less of me. | dropped |
| PR6 | I believe that pledging to projects on Kickstarter is risky because it makes me feel uneasy. | |
| Perceived benefit | ||
| PB1 | I think that pledging to projects on Kickstarter would lead to a performance gain for me due to the special features of rewards. | |
| PB2 | I think that pledging to projects on Kickstarter would lead to a financial gain for me because the rewards are worth more than the money I pledged. | |
| PB3 | I think that pledging to projects on Kickstarter would lead to a time gain because I receive rewards before I can buy them on the market. | |
| PB4 | I think that pledging to projects on Kickstarter would lead to a social gain for me because my friends and relatives would think more highly of me. | |
| PB5 | I think that pledging to projects on Kickstarter would lead to a psychological gain for me because it would fit in well with my self-image. | |
| Intention to continue pledging to projects on the crowdfunding platform | ||
| ICP1 | I expect to pledge again to a project on Kickstarter in the near future. | |
| ICP2 | Given the opportunity, I intend to pledge again to a project on Kickstarter. | |
| ICP3 | I plan to continue pledging to projects on Kickstarter. | |
| Constructs | Measurement items | |
| Control variables | ||
| Disposition to distrust | ||
| DD1 | I believe that most people would tell a lie if they could gain by it. | |
| DD2 | I believe that most people pretend to care more about one another than they really do. | |
| DD3 | I believe that most people are usually out for their own good. | |
| Familiarity | ||
| FAM1 | I am familiar with various crowdfunding platforms. | |
| FAM2 | I am know the crowdfunding environment well. | |
| FAM3 | I am an experienced crowdfunding platform user. | |
