29 June 2015 Mr. Dr. J. Sol
Group: 25 – Crowdfunding. Study year 2014/2015
Faculty of Economics and Business
Bachelor thesis: Motivations for investment in crowdfunding projects Bachelor’s Thesis and Thesis Seminar Business Administration (6013B0347)
Name student: Student number:
Statement of originality
This document is written by student John van Haaster, who declares to take full responsibility for the contents of this document. I declare that the text and the work presented in this
document is original and that no sources other than those mentioned in the text and its references have been used in creating it. The Faculty of Economics and Business is responsible solely for the supervision of completion of the work, not for the contents.
Abstract
This thesis aims to analyse which motivations drive crowdfunding investors to invest
financially in a crowdfunding project and how reward systems can influence these investment decisions. Crowdfunding can be defined as the online request for resources from a distributed audience, often in exchange for a reward. This research contributes to theory by offering new insights into the motivations for investment in crowdfunding projects and thereby extending the body of knowledge in this field. In addition, for practical reasons it is, especially for crowdfunding platforms and project initiators, valuable to know what motivates the crowd to invest financially in a project, since investments are their main goals. It was hypothesized that all eight of the motivations, which were identified in the existing literature, had a significant positive effect on the likelihood to invest and the investment amount. It was also examined if there were any differences between three different reward systems: reward-, equity-, and lending-based systems. A semi-fictional crowdfunding project and an associated survey, which was completed by 67 participants, were used to test the hypotheses. Significant values for the likelihood to invest were found for the motivations indirect identification and
recognition, but only when demographic factors were not taken into account. For the investment amount, only the motivation return showed a significant positive effect under a particular regression. In addition, only reward-based crowdfunding deviated significantly from the other crowdfunding forms.
Table of contents 1.#Introduction# 4! 2.#Theoretical#framework# 6! 2.1$Crowdfunding$and$crowdsourcing$ 6! 2.2$Motives$and$motivations$ 7! 2.3$Motivations$for$crowdfunding$ 9! 3.#Methodology# 13! 3.1$Research$design$ 13! 3.2$Procedure$and$strategy$ 14! 3.3$Analyses$ 15! 3.4$The$sample$ 16! 3.5$Reliability$and$Correlations$ 16! 4.#Results# 16! 4.1$Normality$check$ 16! 4.2$Correlations$ 17! 4.3$Ordinary$least$squares$regressions$ 19! 4.4$Tobit$regressions$ 22! 4.5$TJtests$ 23! 4.6$NonJparametric$tests$ 25! 5.#Discussion# 26! 5.1$General$discussion$ 26! 5.2$Managerial$implications$ 28! 5.3$Contributions$and$limitations$ 28! 5.4$Recommendations$for$future$research$ 29! 6.#Conclusion# 30! References# 32! Appendices# 36! Appendix$1:$The$survey$ 36! Appendix$2:$Operationalization$of$motivations$ 41! Appendix$3:$Demographic$characteristics$of$sample$ 43! Appendix$4:$Descriptions$of$variables$ 44! Appendix$5:$Testing$for$normality$ 45!
1.#Introduction#
Crowdfunding is nothing new. What most people don’t know is that the Statue of Liberty was crowdfunded. What’s different today is that you have access to so many more people than you
otherwise would. - Erica Labovitz, director of marketing of IndieGoGo (Mehlhaff, 2011). The term crowdfunding is fairly recent and was first mentioned by Michael Sullivan in 2006 (Gobble, 2012). Hui, Greenberg & Gerber (2014) define crowdfunding as the online request for resources from a distributed audience often in exchange for a reward. But as Erica
Labovitz said, it is not something new. The construction of the Statue of Liberty was financed in France by a combination of public fees and lottery sales, and the statue's pedestal was paid for by a campaign that solicited small donations from the American people (Gobble, 2012). But only recently, because of the development of the Internet and the changed global
economic conditions is crowdfunding becoming an important financing alternative. Financing from venture capitalists and banks is usually available only in the later development phases of startups (Berger & Udell, 1998). In the early phases of a company’s life cycle, a substantial amount of their funding are provided by insiders; the entrepreneur, other members of the team, family and friends (Moritz & Block, 2014). Now crowdfunding could be a good alternative next to those sources, to fill the funding gap.
The motivation of the project initiator is thus quite clear; crowdfunding could help them to overcome the funding gap. But the motivations of capital providers are less clear, because they tend to be quite heterogeneous (Lin et al., 2014). Still, it is important to know what motivates crowdfunding investors, because in this way intermediary crowdfunding websites can adjust their websites to match those motivations and gain more potential from the investors. For capital seekers it’s also quite useful to know what motivates crowdfunding investors and in what way differ, because then they can better assess their chances of success with the different types of investors and adjust their campaign to the right crowd. By gaining more understanding about the motivations of crowdfunding investors, it is also possible to link these motivations with the motivations of other informal investors, such as business angels.
Research on crowdfunding at the moment is largely lacking, because of its novelty. Moritz and Block, for example, shows that there are only 127 articles and working papers about crowdfunding at the moment of their literature review (2014). Nowadays, a lot more researchers are getting interested in crowdfunding. I am particularly interested in the capital providers and what motivates them to provide capital. Motivations of capital providers have been researched by a number of researchers. Allison et al. (2015) come to the conclusion that
social reputation and intrinsic motives play a significant role for participating in
crowdfunding. In addition, Lin et al. (2014) suggests that motives are heterogeneous and depend on the crowdfunding model. Both Allison et al. (2015) and Lin et al. (2014) used an experiment design for their research and the data was collected from crowdfunding platforms. Gerber et al. (2012) and Ordanini et al. (2011) both found that capital providers have some common characteristics: they are innovation-oriented, are interested in interacting with others, identify themselves with the company or the product, and are interested in the financial result. Gerber et al. (2012) used a qualitative exploratory study with semi-structured interviews, while Ordanini (2011) used a grounded theory approach with in-depth analysis of cases.
When looking at this literature, the conclusion is that there has already been done some research on the motivations of crowdfunding investors, but that there are still some gaps. One of these gaps, which is already addressed by Lin et al. (2014), is that of the motivations of different kind of crowdfunding investors. They used the platform Kickstarter to look for different investor types and which kind of strategies these different kind of
investors use. In my thesis, I also want to address the motivations of crowdfunding investors, but also look at the differences between the motivations linked to the different types of crowdfunding. Mollick (2014) identifies four different types of models: donation-, reward-, equity- and lending-based crowdfunding. It would be interesting to see how investors rate a project, which is identical, but have different rewards based on these models. In addition, some motivations of capital providers have been found, but not the ratio between them. In this way we can identify the most important motivations for crowdfunding investors to invest and the less important motivations. These gaps have led to the following main research question: ‘What motivates crowdfunding investors to invest financially in crowdfunding projects and
how do the different reward systems influence investment decisions?’
The structure of the thesis is as follows: First, relevant literature about crowdfunding and motivations are discussed, including a conceptual model which forms the basis of this
research. Also, hypotheses will be derived from the theory. In Section 3, the methodology will be discussed, which includes substantiations for choosing certain methods and it shows how the data is collected. After the method section, the results of the data will be presented and discussed. Finally, in the last section, conclusions can be drawn regarding the hypotheses and research questions.
2.#Theoretical#framework#
2.1#Crowdfunding#and#crowdsourcing#
The concept of crowdfunding comes from the broader concept of crowdsourcing, which involves using the “crowd” to obtain ideas, feedback, and solutions to develop corporate activities (Belleflamme et al., 2013). Kleeman et al. (2008) for example, states that
“Crowdsourcing takes place when a firm outsources specific tasks essential for the making or sale of its product to the crowd in the form of an open call over the internet, with the intention to make individuals give a contribution to the firm’s processes for free or for significantly less than that contribution is worth to the firm”. Crowdfunding differs from crowdsourcing in the way of help that is provided. Crowdfunding uses money or equity from the public instead of ideas, feedback and solutions. In addition, Schwienbacher and Larralde (2010) states that crowdfunding also differs from professional financing parties, such as banks venture capitalists or business angels, because the projects or the venture is financed by a group of individuals and occurs without any intermediaries. Nowadays this is a bit different, because different platforms have emerged which help to intermediate between crowdfunding investors and project initiators, such as Kickstarter, Sellaband or smaller platforms like geldvoorelkaar. On these platforms, project initiators put up their campaign with a certain goal of money and a time limit. Crowdfunding investors, also known as backers or funders, can invest in these projects. The crowdfunding platforms arrange the payments and get a fee from the total amount of money that is raised from the project (Gerber et al., 2012). Because crowdfunding covers so many current (and likely future) uses across many disciplines, a broad definition of crowdfunding is elusive (Mollick, 2014). As we are particularly interested in crowdfunding in the entrepreneurial context, the following definition from Mollick (2014) will be used:
“Crowdfunding refers to the efforts by entrepreneurial individuals and groups – cultural, social, and for-profit – to fund their ventures by drawing on relatively small contributions from a relatively large number of individuals using the Internet, without standard financial intermediaries”.
Mollick (2014) also discusses and gives a clear division between four different kinds of crowdfunding: the patronage, lending, reward-based and equity-based models. The
patronage model, or also called the donation-based model (Hemer, 2011), place funders in the position of philanthropists, who expect no direct return for their donations. A good example is the platform GoFundMe, where funders can help people who are in need. In the lending model, funds are offered as a loan, commonly with the expectation of some rate of return on
the capital invested. An example is the German platform Smava, which collects loan pledges from the crowd and give those to the project initiator(s). Afterwards, Smava collects
repayment instalments and give those back to the crowd-lender (Hemer, 2011). The third model is the reward-based model, where funders receive a reward for backing a project. This model treats funders as early customers, allowing them access to the products produced at an earlier date, better price, or with some other special benefit, such as being credited in a movie (Mollick, 2014; Hemer 2011). The most famous platform that uses this model is Kickstarter. Finally, equity-based crowdfunding is defined as a model in which crowdfunders receive a small equity stake in the company in return for their funding contributions (Ahlers et al. 2012) and receive financial compensation, such as revenue and profit-share arrangements
(Belleflame et al., 2013). An example of such a platform is Crowdcube.
2.2#Motives#and#motivations#
As the research is about the motivations of investors to participate in crowdfunding, it is critical to know what motivations are and what kind of motivations exist.
Motivation psychology differentiates between “motives” and “motivations”
(Bretschneider, Knaub & Wieck (2014). A motive is seen as an individually developed and content-specific, psychological disposition (Jost, 2000). Some motives are inborn while a relatively stable set of motives is developed during an individual’s socialization process (Bretschneider et al., 2014). Motivation describes the process of how an individual’s motives become activated. Ryan and Deci (2000) states that being motivated means to be moved to do something. A person who feels no impetus or inspiration to act is thus characterized as
unmotivated, whereas some who is energized or activated toward an end is considered motivated (Ryan & Deci, 2000).
Bretschneider et al. (2014) describes that an active motive will subsequently cause certain behaviour in a particular context. Certain things that an individual perceives will serve as incentives that stimulate corresponding motives in such situational contexts. The
interaction between the person’s motives and incentives from the situation results in an activated motive, which in turn causes an action or behaviour. This motivation model is illustrated in Figure 1 (Jost, 2000).
Figure 1
Ryan and Deci (2000) also states that motivation is not a unitary phenomenon, as people have not only different levels of motivation, but people also vary in the orientation of that
motivation. From that perspective, Ryan and Deci (2000) came up with the
self-determination-theory (SDT), one of the most popular motivations concepts. The concept makes a distinction between intrinsic and extrinsic motivations. When intrinsically motivated, one does something because it is inherently interesting or enjoyable. One does it for its
inherent satisfactions rather than for some separable consequence, such as external prods, pressures or rewards. Extrinsic motivation refers to doing something because it leads to a separable outcome. Extrinsic motivation thus contrasts with intrinsic motivation, which refers to doing an activity simply for the enjoyment of the activity itself, rather than its instrumental value (Ryan & Deci, 2000). The aim of extrinsically motivated behaviour is to support certain positive and avoid negative consequences. For example, one gets a bonus for doing a good job on his or her work, or avoids punishment by being on time at work. Both intrinsic and extrinsic motivations may play a role in an investor’s decision to fund a crowdfunding project.
In the next paragraph, intrinsic and extrinsic motivations will be further divided into measurable factors of motivations and hypotheses will be made.
2.3#Motivations#for#crowdfunding#
In the following paragraph, motivations to invest in crowdfunding projects are given including an explanation. These motivations are based on the research model of a working paper of Bretschneider et al. (2014).
Bretschneider et al. (2014) examined six empirical crowdsourcing studies (Füller, 2006; Walcher, 2007; Brabham, 2008; Brabham, 2010; Bretschneider, 2012; Janzik, 2012), four crowdfunding studies (Harms, 2007; Hemer, 2011; Ordanini, Miceli et al., 2011; Gerber, Hui et al., 2012) and three empirical business angel research studies (Sullivan & Miller, 1996; Brettel, Jaugey et al. 2000; Stedler & Peters, 2002) and analysed which of these motives could apply for the research.
The first motive that appears to influence people to invest in crowdfunding projects is fun. Several studies on motivation in open source participation (Hars & Ou, 2001; Hertel et al. 2003; Bonaccorsi & Rossi 2004) support that the primary motivator for participation on these projects is the pleasure found in doing hobbies (Brabham, 2010). As programming on open source projects is quite similar to crowdsourcing, Brabham (2010) predicted that fun would also be important for crowdsourcing. Colleagues such as Bretschneider (2012) and Janzik (2012) found that the fun to develop ideas and publish them was also found in crowdsourcing. In addition, fun was also named as one of the main reasons for the activities of business angels (Brettel, Jaugey et al. 2000). Although crowdfunding is a bit different than open source participation and crowdsouring, as it is about financial support rather than labor support, it is conceivable that enjoyment is important for crowdfunding too. Harms (2007) states that a review of discussions of crowdfunding participants in online newsgroups showed that many consumers emphasize that they enjoy to invest in crowdfunding projects. Therefore, it is assumed that crowdfunders have fun when investing in crowdfunding projects.
Hypothesis 1: “Fun to make investments” has a significant positive influence on the investment in a crowdfunding project.
Second, Deci and Ryan (1993) mention in their self-determination-theory that next to fun and enjoyment, interest and curiosity are possible causes of intrinsic motivation. Crowdfunders may invest in a crowdfunding project because they are curious about crowdfunding as a new investment alternative (Bretschneider et al., 2014), or because they want to escape boredom (Füller, 2006). Curiosity was found to be one of the most important motives for consumers’
willingness to engage in further virtual development activities (Füller, 2006), what can be seen as the willingness to act. Therefore, the following hypothesis is made:
Hypothesis 2: “Curiosity about crowdfunding” has a significant positive influence on the investment in a crowdfunding project.
As mentioned in the introduction, insiders provide substantial amounts of a start-up company’s funding: the entrepreneur, other members of the team and family and friends (Moritz, 2014). In addition, Agrawal et al. (2011) found that family and friends are an important group for funding reward-based crowdfunding projects. Also, when looking at platforms such as Crowdaboutnow, which give a bit of information of the relationship from the investor with the project initiator when the investor leaves a message, shows that a lot of investors are friends or family. It makes it conceivable that crowdfunders tend to support projects to which they have an emotional relationship with the project initiator. This leads to the following hypothesis:
Hypothesis 3: “Direct identification” with the team has a significant positive influence on the investment in a crowdfunding project.
A often heard quote is: “First impressions are the most lasting” (Jonas Hanway, 1756) or “You’ll never have a second chance to create a good first impression” (Will Rogers, n.d.). The same could apply to the first impression of the project initiator(s). Research proved that the first impression to the entrepreneurs is the first step to a potential investment of business angels (Brettel, 2003; Mason & Stark, 2004). In addition, business angels place great
importance on the chemistry between themselves and the entrepreneur (Mason & Stark, 2004). Crowdfunders may invest in a crowdfunding project because they have a certain emotional relation based on emotional affection or sympathy for the project initator(s). This leads to the following hypothesis:
Hypothesis 4: “Indirect identification” with the team has a significant positive influence on the investment in a crowdfunding project.
The final identification and intrinsic motive is regional identification. It is possible to assume that somebody feels connected to a certain project, because it is from his or her village or city,
or in a certain regional range. This connection may imply that crowdfunding investors fund a project for the reason that the project is close to them and make them feel connected.
Research on this topic, as Bretschneider et al. (2014) states, found different results for regional identification. Agrawal et al. (2010) examined crowdfunding in the recording industry, and find some evidence that crowdfunding relaxes geographic constraints among funders. This could mean that crowdfunding has the potential to mitigate many of the distance effects found in traditional fundraising efforts. Mollick (2014) found similar evidence and states that crowdfunding, at least in part, reduces the importance of traditional geographic constraints. On the other hand, Lin and Viswanathan (2014) found that home bias is also present on the debt-based crowdfunding platform Prosper and not only on the traditionally funded ventures, as indicated by many researchers (Kenney and Burg, 1999; Owen-Smith & Powell, 2004). This could mean that regional identification have an important role in
crowdfunding. This leads to the following hypothesis:
Hypothesis 5: “Regional identification” with the project has a significant positive influence on the investment in a crowdfunding project.
Maslow (1943) found that people have what can be called the desire for reputation or recognition, which can be seen as an extrinsic motivation. When looking at the hierarchy of Maslow, recognition can be placed in the esteem needs and thus is a basic human need. As a lot of reward and donation-based projects provide recognition to funders (i.e. exclusive insiders updates or a record on the website), it would be logical that recognition also plays a role in crowdfunding. Crowdfunding investors may invest in crowdfunding projects to
increase their visibility and receive recognition for their investment from the community, their friends or other people. Therefore, it is hypothesized:
Hypothesis 6: “Recognition” has a significant positive influence on the investment in a crowdfunding project.
Reichwald and Piller (2009) found that customers participate in a crowdsourcing initiative because they are unhappy with current solutions that are offered to their problem (Vreeman, 2012). In addition, Ligas (2000) found that consumers choose in general those products (and services) that provide the greatest utility to them. It could be possible that the product or service under development from a crowdfunding project gives the highest utility or give a
good solution to the customers’ problem. Also, in open source software, software developers participate in open source projects because they can use and profit from the software solution or the idea itself, just like in crowdsourcing (Hars & Ou, 2001, Füller, 2006). Crowdfunding may not be about the development of a new product, service or solution, because under crowdfunding one can only fund a project, but the main motivation stays the same: The desire of the product, service or solution under development. Thus, it is hypothesized that:
Hypothesis 7: “Personal need” has a significant positive influence on the investment in a crowdfunding project.
As mentioned, there are different kinds of crowdfunding forms. It is, especially in the equity and lending-based categories, conceivable that crowdfunding investors invest because they want to get a return from their investment. In the reward-based category this could also be the case, because here the return is not a monetary unit, but a reward such as receiving the
product. Out of these arguments it is hypothesized that:
Hypothesis 8: “Return” has a significant positive influence on the investment in a crowdfunding project.
In addition to the eight hypotheses, an exploratory study will be done to look if there are any significant differences between the three different reward systems. No hypotheses are made for this part of the study.
Figure 2
3.#Methodology#
3.1#Research#design#
To answer the research question, a cross-sectional study is used. A cross-sectional study means that the study will be carried out at one time point or in a short time period. As we are interested in the motivations of the funders to invest in projects at this moment, this will be a proper study design. Motivations are measured through stated importance using a likert-scale. In addition, a vignette-study can be easily placed in a survey to find out if rewards have influence on investment decisions. The survey is Internet-based. The expectation is that everybody in this population knows how to complete a survey on the computer, because crowdfunding investors also use crowdfunding platforms on the Internet and in this way have knowledge about how to use a computer. These investors will be approached using a
sampling bias and that the sample is not representative of the entire population, but it’s easy to conduct, fast and cheap.
3.2#Procedure#and#strategy#
Before the gathering of the data, a literature check was performed to find out what other researchers had already done in this area, as can be seen in Chapter 2. The survey was self-constructed and included questions that where found in other studies, as well as self made-up questions. The survey contained information on the motivations, the likelihood to invest in a certain project and background-information of customers. The survey was online for about three weeks and can be found in Appendix 1. In Appendix 2 it is shown which questions belong to which motivation, including the sources for the questions. The link to the survey and additional information about the research was placed on different crowdfunding forums, general tech forums that have sub forums about crowdfunding as well on crowdfunding Facebook pages. Examples are crowdfundingforum.com, tweakers.net and forum.fok.nl. In addition, the link to the survey was also distributed on Blackboard, in the section of
Bachelor’s Thesis and some friends and acquaintances from the researcher, who met the requirements, have been approached to complete a survey.
First, an informed consent form was given and the participant had to agree with it to participate in the study. Thereafter, a definition of crowdfunding was given and a question was constructed and placed in the survey so that the right population would complete the survey. The right population were people that have funded crowdfunding projects once or more themselves. Subsequently, the real survey began. First, a project was displayed with an idea of a product that the project initiator Nonda wants to produce. The project is based on a campaign from Kickstarter, from which a link can be found in in Appendix 1. In addition, certain rewards were given, based on the different rewards systems; reward-, equity- and lending-based. As there are three systems in the research, one-third were displayed the reward-, one-third the equity- and one-third the lending-based rewards. The rewards in the survey are based on similar projects found on projects of Kickstarter, Crowdcube and Fundingcircle. Afterwards, participants were asked how likely they would invest in the project and for which amount. In this way it is possible to identify differences in investment between the different rewards.
The motivations were transferred into 21 statements that represented the different kind of motivations. The motivations were measured using a five-point likert-scale, which
contained different choices: Strongly disagree, disagree, neither agree nor disagree, agree and strongly agree. Appendix 2 shows which items belong to which variables.
On the last page of the survey, demographic characteristics of the participant were asked. Age was measured by asking to chose between certain groups: 18-29, 30-49, 50-64 and 65 years and over. The gender of the respondent was measured by a question with the
categories male and female and education of the respondent was measured by asking what their highest level of education is with the following categories: did not complete high school, High school/GED, associate degree, bachelor’s degree, master’s degree. Next, the occupation of the participant was measured by using these categories: employee, entrepreneur, student and unemployed. The last question was about the annual income range, were the participants had to chose between one of these income ranges: below €20.000, €20.000-€29.999, €30.000-€39.999, €40.000-€49.999, €50.000-€59.999, €60.000-€69.999, €70.000-€79.999, €80.000-€89.999 and €90.000 or more. At the end of the survey, the participant was thanked for filling in the survey.
3.3#Analyses#
First, the data will be tested for normality and other problems, such as outliers. After that, correlations between variables will be checked. Subsequently, multiple linear ordinary least squares (OLS) is used to find the regression model that best fits the data (Field, 2013). In this way hypotheses 1 to 8 can be tested and is it possible to see if certain motivations have significant positive impacts on the likelihood to invest or on the investment amount.
Thereafter, the gathered data will be analysed by performing an independent samples t-test. The independent-samples t-test is used when there are two experimental conditions and different participants were assigned to each condition and can be used to compare the means of two independent samples (Field, 2013). For this analysis, there is an assumption that the independent variable should consist of two categorical independent groups (Field, 2013). To make the analysis work, dummies were made for the three different reward systems. In this way, the test variable was assigned to category one and the other two groups to category zero. After running the independent samples t-test, it can be derived whether there is a significant difference in the means of the groups (Field, 2013). This will be done for the likelihood to invest and the investment amount in combination with the different reward systems. All the tests are done with the use of the statistical program SPSS.
3.4#The#sample#
The sample contains N=67 crowdfunding investors, after the removal of N=18 respondents who did not meet the requirement. The requirement is that the person has invested in a crowdfunding project once or more in his life. N=5 investors did only filled in the vignette-part of the survey and therefore have not completed the motivational and demographic questions. A total of N=49 of this sample is men and N=13 are female and makes the male-female ratio 79% against 21%. 58.1% is between 18-29, 37.1% is between 30-49 and 4.8% is between 50-64. Most of the crowdfunding investors are employees (41.9%), followed by students (30.6%) and entrepreneurs (27.4%). In addition, the category “highest level of education” under the sample was the biggest for the bachelor’s degree (32.3%), followed by a master’s degree (27.4%), High School / GED (22.6%), Associate degree (22.6%) and one did not complete high school (1.6%). A graphical presentation of the sample demographics is given in Appendix 3.
3.5#Reliability#and#Correlations#
The cronbach’s Alphas have been measured for the different variables. The cronbach’s Alpha for fun is 0.758, for curiosity 0.732, for direct identification 0.906, for regional identification 0.729, for recognition 0.814 and for return 0.561. Because the cronbach’s Alpha of return was very low, the scale “One way or another, I will get back something tangible for my participation” was deleted, which turned the original cronbach’s alpha of 0.382 to 0.561. A cronbach’s alpha between 0.6 and 0.8 is considered reasonable, while a cronbach’s Alpha from 0.8 or higher is considered good. As most of the cronbach’s Alphas lay between these intervals, these Alphas should be reliable enough for further research. The variables indirect identification and personal need are a bit problematic, because they are both based on one item, due to forgotten items when creating the survey. All the descriptions of the question variables can be found in Appendix 4.
4.#Results#
4.1#Normality#check#
The dataset was tested for the underlying assumptions for the t-tests and the multiple
plots, histograms and with the help of tests, such as the Kolmogorov-Smirnov and Shapiro-Wilk test. Field (2013) states that these tests are appropriate for small samples, but not for large samples, because then outcomes become easily significant. As there are only 67 and in some parts 62 observations, these tests should be suitable for our normality check.
From the investment decisions, the likelihood to invest D(67) = 0,102, p = 0,082, did not significantly deviated from normal, but the investment amount D(67) = 0,372, p < 0,001 deviated significantly from normal. The likelihood to invest in this test is seen as normal, but when looking at the Q-Q plots the conclusion is that it is not. The problem with these
variables is that the likelihood to invest and the investment amount were centred on zero, because a lot of participants did not wanted to invest in the project. Another problem with the investment amount is that the investment is very different for each person, which makes the distribution very skewed. From the motivation variables, only fun to invest, D(62) = 0,102, p = 0,179, and recognition, D(62) = 0.093, p = 0,200, did not deviate significantly from normal. This view was also supported by the Q-Q plots. The other motivation variables significantly deviated from normal, as can be seen in Appendix 5. In an attempt to correct for bias, the data has been transformed with the help of mathematical functions, such as the log and the square root. These computations did not make the distribution any better, thus the original data has been used for further analysis.
Outliers#
The dataset was also checked for outliers for the variables likelihood to invest and investment amount. The variable likelihood to invest had no significant outliers, which is not surprisingly as the range is between 0-100%. The variable investment amount showed a lot more outliers, but when deleting these outliers, SPSS showed other outliers and in this way the problem couldn’t be solved. In addition, having a few high values can be seen as normal, because some investors just want and have the money to invest more than other investors. This vision was supported with graphs that adjust for the income of the investors. The dataset with 67 crowdfunding investors is also quite small, thus deleting cases is costly. Therefore, it was decided to not remove these cases.
4.2#Correlations#
Field (2013) recommends that when the data have outliers and are not normally distributed, which is the case for almost all of the variables, it is better to use versions of the correlation coefficient that works with ranked data, such as Spearman’s rho and Kendall’s tau, because
the impact of outliers will be reduced. In addition, Field (2013) states that Kendall’s tau should be used rather than Spearman’s coefficient when the dataset is small and there are a large number of tied ranks, which also applies for almost all variables, because the
motivations are ranked on a scale. Based on these assumptions, it was decided to use Kendall’s tau. The results can be seen in Table 1.
The likelihood to invest was significantly related to the investment amount, τ = 0,421, p < 0.001. This could imply that when someone is more likely to invest, they also invest more money, which makes sense. When one has a good feeling about a project, one is not only more likely to invest in the project, but it can also increase ones investment amount, as one is more positive towards the project.
From the motivation variables, only fun to invest, τ = 0,204, p = 0,025, indirect identification, τ = 0,218, p = 0,033 and recognition, τ = 0,182, p = 0,044, were significantly related to the likelihood to invest. These results are remarkable, as our hypotheses and literature study suggest that the other variables should also correlate with the likelihood to invest.
When looking at the investment amount, only the motivation variables curiosity, τ = 0,237, p = 0,015 and return on investment, τ = 0,346, p < 0,001, were significantly related. Furthermore, the reward-based group could have a negative effect on the investment amount, τ = -0,293, p = 0,005. A strong relationship between return on investment was predicted and this correlation also shows that. However, here too it is remarkable that the other variables do not correlate much with the investment amount. The negative correlation of the reward-based group with the investment amount could imply that the investment is smaller in comparison with the other groups. This makes sense, as people get the product for €79 in this group. When investing more, one will only get more products and nothing else. It is thinkable that someone doesn’t need any more products, because, for example, only one product can be used at a time. In the other groups, the return goes up proportionally with the invested amount and special rewards, such as voting decisions, which are obtained in a higher range of investment. Kendall's tau B 1 2 3 4 5 6 7 8 9 10 11 12 13 1. Likelihood 1,000 0,421** 0,093 -0,103 0,006 0,204* 0,138 0,093 0,218* 0,115 0,182* 0,174 0,132 2. Investment 0,421** 1,000 -0,293** 0,158 0,152 0,106 0,237* -0,056 0,197 -0,008 -0,016 -0,130 0,346** 3. Reward Group 0,093 -0,293** 1,000 -0,538** -0,519** 0,257* -0,018 0,109 0,288* 0,033 0,167 0,375** -0,112 4. Equity Group -0,103 0,158 -0,538** 1,000 -0,441** -0,228* 0,095 -0,012 -0,165 -0,105 -0,077 -0,357** 0,062
5. Lending Group 0,006 0,152 -0,519** -0,441** 1,000 -0,054 -0,072 -0,104 -0,149 0,066 -0,104 -0,056 0,060 6. Fun 0,204* 0,106 0,257* -0,228* -0,054 1,000 0,301** 0,139 0,265* 0,231* 0,279** 0,176 0,091 7. Curiosity 0,138 0,237* -0,018 0,095 -0,072 0,301** 1,000 -0,003 0,149 0,050 0,219* -0,124 0,195 8. Direct identification 0,093 -0,056 0,109 -0,012 -0,104 0,139 -0,003 1,000 0,020 0,594** 0,315** 0,170 -0,241* 9. Indirect identification 0,218* 0,197 0,288* -0,165 -0,149 0,265* 0,149 0,020 1,000 0,042 0,042 0,309** 0,125 10. Regional identification 0,115 -0,008 0,033 -0,105 0,066 0,231* 0,050 0,594** 0,042 1,000 0,358** 0,233* -0,119 11. Recognition 0,182* -0,016 0,167 -0,077 -0,104 0,279** 0,219* 0,315** 0,042 0,358** 1,000 0,066 -0,043 12. Personal need 0,174 -0,130 0,375** -0,357** -0,056 0,176 -0,124 0,170 0,309** 0,233* 0,066 1,000 -0,071 13. Return 0,132 0,346** -0,112 0,062 0,060 0,091 0,195 -0,241* 0,125 -0,119 -0,043 -0,071 1,000
**. Correlation is signifcant at the 0.01 level (2-tailed). *. Correlation is signifcant at the 0.05 level (2-tailed).
Table 1: Correlation matrix #
4.3#Ordinary#least#squares#regressions#
For the OLS regression, a hierarchical method is used. This means that predictors are selected based on past work and it is decided in which order to enter predictors into the model (Field, 2013). Predictors that are known from other researchers are entered into the model first in order of their importance in predicting the outcome. The backward method is used, as this method minimizes suppressor effects, which occur when a predictor has a significant effect but only when another variable is held constant (Field, 2013). The results are shown for the first and last step in the backward method in Table 2 and Table 3.
The first multiple linear regression was calculated to predict the likelihood to invest based on the eight different motivations, without the use of demographic control variables. A significant regression equation was found, F (8, 53) = 2,365, p < 0,05 and explained 26,3% of the likelihood to invest in the first model. The last model, based on only three independent variables, shows different results, with a significant regression equation of F (3, 58) = 5,708, p < 0,01 and explained 22,80% of variance in the likelihood to invest. The p-values indicate that the models are significantly better at predicting the outcome than using the mean as a ‘best guess’ (Field, 2013). There were no significant effects for any of the independent variables in the first model. In the last model, significant effects were only found for the variables indirect identification (β = 0,266, p < 0,01) and recognition (β = 0,313, p < 0,05). Return on investment (β = 0,198, p = 0,097) did not have a significant effect. In this way,
support was only found for hypothesis 1 and 4 for the dependent variable likelihood to invest. These results show that predictions that were made are not right, or there is another problem, such as the dataset. This is further highlighted in Section 5.3. The models can be found in Table 2. In this table, the standardized beta shows the importance of each predictor, as a higher absolute value indicates a more important predictor (Field, 2013). The individual contribution of each variable can be found by looking at the b-value, which is the same as the coefficient (Field, 2013). If the value is positive, there is a positive relationship between the predictor and the outcome, whereas a negative coefficient represents a negative relationship with the outcome (Field, 2013). The variance inflation factor (VIF) is also included, which can identify multicollinearity. Bowerman & O’Connell (1990) indicate that an average VIF substantially greater than 1 can indicate the regression may be biased, which is the case for most of the variables and thus could be seen as a problem for the dataset and their outcomes, as it can result in untrustworthy b’s, limits the size of R and multicollinearity between predictors makes it difficult to assess the individual importance of a predictor (Field, 2013).
When controlling for age, gender, education and occupation, the variables indirect identification (β = 0,244, p < 0,05) and recognition (β = 0,319, p < 0,05) stay significant. But when the variable income was also put in the regression, indirect identification (β = 0,417, p = 0,7) and recognition (β = 0,146, p = 0,74) are far from a significant effect on the likelihood to invest. Thus, it is questionable if indirect identification and recognition have an effect on the likelihood to invest when looking at these outcomes.
R R2 R2 change B SE β t VIF Model 1 0,513 0,263 Fun 3,138 5,707 0,081 0,550 1,572 Curiosity 3,881 4,503 0,131 0,862 1,649 Direct identification 0,080 3,995 0,004 0,020 2,565 Indirect identification 6,453 4,597 0,185 1,404 1,254 Regional identification 0,512 4,995 0,019 0,102 2,453 Recognition 7,012 5,191 0,203 1,351 1,619 Personal need 3,346 3,508 0,130 0,954 1,333 Return 7,437 5,429 0,182 1,370 1,270 Model 2 0,477 0,228 -0,012 Indirect 9,261 4,091 0,266 2,264* 1,037
identification
Recognition 10,824 4,026 0,313 2,688** 1,017
Return 8,107 4,804 0,198 1,688 1,039
**. Effect is significant at the 0.01 level. *. Effect is significant at the 0.05 level.
Table 2: Regression models with dependent variable likelihood to invest
The second multiple linear regression was calculated to predict the investment amount and was also based on the eight different motivations, without any demographic control variables. The first model showed no significant regression equation, F (8, 53), p = 0,503 and explained 12,3% of the variance. The last model also had no significant regression equation, F = (1, 59), p = 0,067. This means that the models are not significantly better at predicting the outcome than using the mean. Both the first and the last model showed no significant effects for any of the independent motivation variables on the dependent variable investment amount. The results can be found in Table 3. Here also the VIF’s are substantially greater than 1, which can indicate that the outcomes are biased. The standard errors also stand out, as they are very large. Fun (b = -912,230), for example, has a standard error of 873,44. In addition, some results are strange, as this would mean that when fun is increased by one unit, the investment amount would go down with 912,23 euro. The intention was to look if variables stay
significant as we put demographic control variables in the regression, but as there are no significant values, it was decided to disregard this regression model.
R R2 R2 change B SE β t VIF Model 1 0,350 0,123 Fun -912,230 873,744 -0,168 -1,044 1,572 Curiosity 841,801 689,328 0,202 1,221 1,649 Direct identification -80,088 611,656 -0,27 -0,131 2,565 Indirect identification 413,384 703,771 0,085 0,587 1,254 Regional identification 824,780 764,735 0,217 1,079 2,453 Recognition -850,043 794,674 -0,175 -1,070 1,619 Personal need -131,281 536,986 -0,36 -0,244 1,333 Return 1171,750 831,088 0,204 0,164 1,270 Model 2 0,234 0,055 -0,008
Return 1340,629 719,485 0,234 1,863 1,000
**. Effect is significant at the 0.01 level. *. Effect is significant at the 0.05 level.
Table 3: Regression models with dependent variable investment amount
4.4#Tobit#regressions#
As indicated in Section 4.1, likelihood to invest and investment amount have a fair high skewness on the left side. The Tobit model, also called a censored regression model, can maybe control for the skewness. Tobit models are designed to estimate linear relationships between variables were there is some kind of censoring on the left or right side of the dependent variable (Amemiya, 1984). This is why Tobit regressions are done in addition to the ordinary least squares regressions. As SPSS cannot perform Tobit regressions, the analysis has been done in Stata. In the Tobit model, one has to specify the censoring limits. The
censoring limits are chosen by the researcher with the help of the frequency tables of the dependent variables.
The first Tobit regression model, based on the likelihood to invest, is based on a censor limit to the left of 15 and left N = 14 observations out of the sample. This led to N = 48
uncensored observations. The t-values indicate no significant effects, not even for indirect identification (p = 0.169) and recognition (p = 0,089), which were significant in the OLS regression model. Thus, no extra support was found in this regression model.
Likelihood to invest B SE t [95% confidence interval]
Fun 3,467 5,990 0,58 -8,543 15,477 Curiosity 5,405 4,900 1,10 -4,420 15,229 Direct identification 0,035 4,124 0,01 -8,232 8,304 Indirect identification 7,133 5,119 1,39 -3,13 17,395 Regional identification 0,328 5,194 0,06 -10,085 10,741 Recognition 9,548 5,513 1,73 -1,505 20,599 Personal need 3,975 3,718 1,07 -3,479 11,428 Return 7,829 5,636 1,39 -3,471 19,129
Table 4: Tobit regression model with dependent variable likelihood to invest
The second Tobit regression model, based on the investment amount, is based on a censor limit to the left of €14 and to the right of €19000, so that the largest outliers will be excluded.
In this way, 18 left-censored and 2 right-censored observations were excluded, resulting in N = 42 uncensored observations. As can be seen in Table 5, this resulted in a significant effect for return (p < 0.05) on the investment amount. This could mean that the return has a positive influence on the invested amount. This makes sense, as one tends to invest more when the return is greater. The model suggests that when return is increased with one unit, the invested amount will go up with €1931,37, which is a large result. However, the standard error of €932,47 still indicates that there is a large deviation. Thus, this model finds, in contrast to the OLS regression model, support for hypothesis 8.
Investment amount B SE t [95% confidence interval]
Fun -557,946 961,879 -0,58 -2486,397 1370,505 Curiosity 1451,191 797,3765 1,82 -147,453 3049,834 Direct identification -125,193 693,173 -0,18 -1514,921 1264,534 Indirect identification 1400,935 941,851 1,49 -487,361 3289,232 Regional identification 1218,862 864,198 1,41 -513,751 2951,475 Recognition -1269,475 883,459 -1,44 -3040,704 501,754 Personal need -482,978 583,020 -0,83 -1651,861 685,906 Return 1931,373 932,465 2,07* 61,89414 3800,852
*. Effect is significant at the 0.05 level.
Table 5: Tobit regression model with dependent variable investment amount
4.5#THtests#
Independent-samples t-tests are used, because the participants were assigned to the different reward-groups in the vignette. In this way a comparison can be made between the groups. The t-test is a parametric test based on normal distribution and the dataset in this research did not meet this requirement. Nevertheless, this test will still be used, because it can still give
valuable information. Field (2013) recommends using bootstrap if there is potential bias in the dataset, because it can estimate the properties of the sampling distribution from the sample data. In this way bootstrapping can overcome the problem of not knowing the shape of the sampling distribution (Field, 2013).
THtest#1#
On average, participants who were assigned to the reward-based group were more likely to invest in the crowdfunding project from Nonda (M = 47,15, SE = 29,80), than those
that did not belong to the reward-based group (M = 39,68, SE = 27,67). This difference, 7,47, BCa 95% CI [-6,19, 20,43], was not significant t(65) = 1,045, p = 0,300. In addition, it represents a small-sized effect, d = 0,26.
The investment amount in het project was smaller under participants who were assigned to the reward-based group ((M = 67,92, SE = 97,61), than those that did not belong to the reward-based group (M = 2087,44, SE = 4980,616). This difference, -2019,516, BCa 95% CI [3807,285, -745,802], was significant t(40) = -2,596, p < 0.05 with an medium effect size, d = 0.573. It is interesting to see that the standard deviation under the reward-based group is much smaller than under the other group. This may be caused by the reward, which is the product itself and the amount of money that participants would need to invest, which was €79 and is close to the mean.
THtest#2#
In the second test, participants who were assigned to the equity-based group were less likely to invest in the project (M = 36,76, SE = 30,49), than those that did not belong to the equity-based group (M = 45,24, SE = 27,52). However, the difference, -8,447, BCa 95% CI [-23,10, 8,92], was not significant t(65) = -1,131, p = 0,262. The effect size was also small, d = 0,219.
The investment amount in the project under the equity-based group was higher (M =1882,10, SE = 4459,42), than those that did not belong to the equity-based group (M = 1039,72, SE = 3798,92). Nonetheless, this doesn’t say much, because the standard deviations are big and the sample size N = 21 in the equity-group and N = 46 in the other group is quite small. The difference, 842,38, BCa 95% CI [-935,01, 2901,67], was not significant t(65) = 0,797, p = 0,428 and the effect size small, d = 0,203.
THtest#3#
The last test is based on the lending reward-group. Participants who were assigned to this group were not more likely to invest in the project (M = 42,75, SE = 24,79), than those that did not belong to this group (M = 42,51, SE = 30,23). The difference, 0,239, BCa 95% CI [-14,65, 15,15, was also far from significant t(65) = 0,031, p = 0,975, and had a very small effect size, d = 0,008.
The investment amount was higher under the lending-based group (M =2303,05, SE = 5585,304), than those that did not belong to this group (M =878,51,10, SE = 3079,391). The difference, 0,239, BCa 95% CI [-14,649, 15,153], was not significant t(65) = 0,031, and the
effect size is medium, d = 0,316. Here also applies that the huge standard deviation has a big impact on the significance of the test.
4.6#NonHparametric#tests
Because the assumption of homogeneity, linear relationship and normal distribution was not met for most of the variables, non-parametric tests are also included. Non-parametric tests or ‘assumption-free tests’ make fewer assumptions than the other tests (Field, 2013). Non-parametric tests overcome the problem of the shape of the distribution of scores by ranking the data (Field, 2013). As independent samples are compared, the Wilcoxon rank-sum test and the Mann-Whitney test can be used. As Field (2013) states that both tests are equivalent, the decision was made to use the Mann-Whitney test. The tests can be found in Table 6.
Mann-Whitney test Median U z p r
Likelihood to invest Model 1 Reward group 45,50 603,50 0,908 0,364 0,111 Non-reward group 40,00 Model 2 Equity group 40 408,50 -1,008 0,314 -0,123 Non-equity group 40,50 Model 3 Lending group 40 4744 0,055 0,956 0,007 Non-lending group 41 Investment amount Model 1 Reward group 79 316,00* -2,808 < 0,01 -0,341 Non-Reward group 250 Model 2 Equity group 100 594 1,509 0,131 0,184 Non-equity group 79 Model 3 Lending group 375 576 1,461 0,144 0,179 Non-lending group 79
*. Effect is significant at the 0.05 level.
As can be seen, the non-parametric Mann-Whitney test does not change significance outcomes in comparison with the t-test and in this way support the earlier findings with the parametric t-test.
5.#Discussion
5.1#General#discussion#
The aim of this research was to discover the motivations for crowdfunding investors to invest financially in a crowdfunding project. Eight motivations were identified, which were checked for the likelihood to invest and the investment amount. For two motivations, indirect
identification and recognition, support was found for the likelihood to invest in the theoretical framework and data. For the other six motivation values, no support was found. For the investment amount, only one significant effect was found, for the motivation return. However, support was only found under the Tobit regression model. The fact that only two out of eight for likelihood to invest and one out of eight for investment amount were significant, indicate that the framework of motivation on crowdfunding is not appropriate, or there are other problems, what seems more likely and will be discussed in Section 5.3. First, the results of the motivations are discussed one by one, divided by intrinsic and extrinsic motivations.
Intrinsic#motivations#
Fun to make investments, one of the main motivations for programmers in open source
projects and an important motivation in crowdsourcing and for business angels, was not found in this research for crowdfunding. If the results are correct, this would mean that investors can have fun when they look for projects, but it do not raise the likelihood to invest in a
crowdfunding project or the investment amount.
The second motivation, curiosity, was suggested as a motivational factor for
investment. Crowdfunding is a relatively new investment alternative and investors could be curious about this new source. However, this study did not found any of these effects. This could indicate that investors do not invest in crowdfunding projects because they are curious, but other reasons are important for investment decisions.
Thirdly, finding of this research suggests that there is no relationship between direct identification and investment decisions. This could imply that friends and family are not important for the investment. However, findings of Agrawal (2011), who did an extensive study on direct identification in reward-based crowdfunding, conclude the opposite.
Fourthly, indirect identification was found to be an important factor for business angels, as they want to have some kind of chemistry with the project initiator. The results indicate that the same applies for crowdfunding investors. The look and feel of the project and the person or company behind the project are, if the results are correct, important factors for the likelihood to invest, but not for the investment amount.
The fifth and final intrinsic motivation, regional identification, was found to not be a predictor for investment decisions. This result further fuels the debate about this motivation, as these results are in line with Mollick (2014), which could mean that crowdfunding has the potential to mitigate many of the distance effects found in traditional fundraising efforts, but contradict the results found by Lin and Viswanathan (2014), who found that there is some sort of home bias for projects.
Extrinsic#motivations#
For the first extrinsic motivation, recognition, support was found in this study for the
likelihood to invest. This could arise from the fact that people have the desire for reputation or recognition. In this way, crowdfunding investors may receive recognition for their investment from the community, their friends or other people.
Personal need has been found to be a motivational factor in crowdsourcing and a huge motivational factor in open source software projects. However, the same effect was not found in this study for crowdfunding investors. This effect may not be found, because the main thing that crowdfunders do is invest financially in a project. In an open source or
crowdsourcing project, ones expertise, ideas and work can further develop the new product, service or solution. This difference may cause that personal need is not found in
crowdfunding.
For the last motivation, return, it was hypothesized that it was especially important for the equity and lending-based crowdfunding categories, as the return is in monetary units and have much similarities with the stock and lending market. However, no effect was found for the likelihood to invest and an unclear effect was found for the investment amount, as the effect was only found in one of the two different models.
Differences#between#crowdfunding#models#
Three different models, reward-, equity-, and lending-based crowdfunding were checked for differences. The results suggest that the investment amount differs between the reward-based group and the other groups. This is not surprisingly, as one tends to invest in a reward-based
project because one want the product, service or solution as a reward. These amounts are often much smaller then for investment decisions based on gaining money, which is more likely to be the case in equity- and lending-based crowdfunding, where the investment amount is larger.
5.2#Managerial#implications#
If the results are right, these findings have a few managerial implications. First of all, if indirect identification found to be an important factor for the likelihood to invest,
crowdfunding platforms should encourage project initiators to try to make and maintain a certain emotional relationship with potential investors. Integrating a chat system in the platform could achieve this, so that investors can easily contact project initiators. Another way would be adding a bibliography page and make a guide for project initiators in what way they can get the most out of this page.
Second, as recognition found to be a factor for the likelihood to invest, crowdfunding platforms could implement profile sites such as Facebook and LinkedIn for investors. In this way, investors can display their personal collection of carried out investments on their profile, which makes their investments more visible and could increase the recognition-effect, as also suggested by Bretschneider et al. (2014).
Finally, as return found to be a potential factor for the investment amount, rewards must be clearly displayed at the campaign, but most of the crowdfunding platforms already do so.
5.3#Contributions#and#limitations#
Based on our predictions, the expectation was that this research could contribute to theory by offering new insights concerning motivation for participation in crowdfunding initiatives and thereby extending the body of knowledge in this field, because so far it is unclear which motivations lead to investments from the crowd in crowdfunding projects. In addition, the expectation was that this research could have a practical contribution, as knowing the motivations is a valuable insight for crowdfunding platforms and project initiators. Crowdfunding platforms could implement features that further trigger the researched
motivations in order to attract the likelihood of the crowd to invest in a project and for a larger amount. Project initiators could use the results to improve their campaign, in order to collect
more money from the crowd. Still, the research in this form does have some contributions. It is one of the first researches in the field of the motivations of investors in a crowdfunding context and shows how such a research can be done. It shows which obstacles there are and which mistakes were made, so these mistakes can be taken out in future researches.
As not all of the expected contributions were achieved, there is also something to say about the limitations of the research. Our results indicate that only indirect identification and recognition are important for the likelihood to invest and only return for the investment amount. If this is true, managerial implications as described in Section 5.2 are applicable. However, there may be some problems with the data and the data collection, which causes that a lot of motivations are not supported in this research.
First of all, the dataset is relatively small. Only on Kickstarter alone, there are roughly nine million investors, so the total amount is even larger (Kickstarter, 2015). As the dataset of this research only contains 67 participants, it could be that the different participants have too much influence on the outcomes in the research and do not give the right reflection of the total population. A larger dataset would also help to overcome problems such as the non-normal distribution in the data. In this way, the models can better predict the outcomes, as fewer or no assumptions would be violated.
Second, questions for the survey were selected and fabricated with the greatest care, but it is possible that the questions did not measure the motivations in the right way. The number of questions for the measurement of some motivations, such as personal need, was also limited.
Thirdly, it could be that the project in the survey was badly chosen. A lot of
participants from the 67 participants decided to not invest in the project, or for a very small amount. When excluding these participants, the dataset become even smaller. If the research would be done again, the selection of an appropriate project should focus on a product, service or solution in which more people would be interested in.
5.4#Recommendations#for#future#research#
Future research could focus on the same research, but with an improved survey and data collection. In this way, more reliable results can be obtained and compared with this research. As the theoretical framework model is extensible, it is also possible to add more motivations or to add moderating effects.
It is also thinkable that different types of investors value the motivations differently. In this way it is possible to examine and cluster different types of crowdfunding investors. This could lead to valuable insights for crowdfunding platforms and could increase their targeting.
Finally, it would also be interesting to make different surveys with different crowdfunding projects. It is possible that crowdfunding projects in a different context or category would show different results. In this way it can be checked to which extent the results of a similar research are transferable to other projects in another context.
6.#Conclusion#
The research question that was investigated in this research was: ‘What motivates crowdfunding investors to invest financially in crowdfunding projects and how do the different reward systems influence investment decisions?’ To answer this question, a
literature scan was performed in order to find motivations that were already distinguished for crowdfunders to invest financially in crowdfunding projects. This resulted in five intrinsic motivations: fun, curiosity, direct, indirect and regional identification and in three extrinsic motivations: recognition, personal need and return.
It was hypothesized that the motivations would have a positive influence on the likelihood to invest and the investment amount. In order to measure this, a survey was created, which was completed by 67 crowdfunding investors. Participants were randomly assigned to one of the three vignettes, which contained a project with particular rewards. These rewards were based on the three different reward systems: reward, equity and lending-based systems. Participants were asked how likely they would invest in the project and for which amount. In addition, participants were asked 21 questions about why they invested in a crowdfunding project, which were based on the eight motivations. Finally, five demographic questions were asked.
After running different analysis, it turned out that only two out eight motivations were positively related to the likelihood to invest. These were indirect identification and
recognition. In addition, only one out of eight motivations was positively related to the investment amount, which was return. For the other motivations fun, curiosity, direct and regional identification and personal need, no positive relationships were found. When also taking the demographic factors into account, no values were positively related towards the
likelihood to invest and the investment amount. These findings contradict the predictions of this research.
If these results were correct and when demographic factors are taken out of the equation, crowdfunding platforms and project initiators would be best off to only focus on indirect identification and recognition to raise investors likelihood to invest and on return, which can increase the investment amount of investors.
The different vignettes indicate that there is one big difference between the different reward systems. Participants invested less in the reward-based vignette in comparison with the equity- and lending-based vignettes. This could mean that reward-based crowdfunding has a negative effect on the investment amount in comparison with the other crowdfunding forms. This is something where project initiators should think about, before choosing the right platform for their project.
In conclusion, further research is necessary to give definite answers about the
motivations to invest, as there could be a problem with the dataset. Furthermore, it should be stated that the research field on crowdsourcing is relatively young and was only first
mentioned in 2006 (Gobble, 2012). Therefore, researches on this topic is limited and mainly in the explorative phase.
References#
Ahlers, G. K., Cumming, D. J., Günther, C., & Schweizer, D. (2012). Signaling in equity crowdfunding. Available at SSRN 2161587,
Allison, T. H., Davis, B. C., Short, J. C., & Webb, J. W. (2015). Crowdfunding in a prosocial microlending environment: Examining the role of intrinsic versus extrinsic cues.
Entrepreneurship Theory and Practice, 39(1), 53-73.
Amemiya, T. (1984). Tobit models: A survey. Journal of Econometrics, 24(1), 3-61. Belleflamme, P., Lambert, T., & Schwienbacher, A. (2013). Individual crowdfunding
practices. Venture Capital, 15(4), 313-333.
Belleflamme, P., Lambert, T., & Schwienbacher, A. (2014). Crowdfunding: Tapping the right crowd. Journal of Business Venturing, 29(5), 585-609.
Berger, A. N., & Udell, G. F. (1998). The economics of small business finance: The roles of private equity and debt markets in the financial growth cycle. Journal of Banking & Finance, 22(6), 613-673.
Bonaccorsi, A., & Rossi Lamastra, C. (2003). Altruistic individuals, selfish firms? the structure of motivation in open source software. The Structure of Motivation in Open Source Software,
Bonaccorsi, A., & Rossi, C. (2003). Why open source software can succeed. Research Policy, 32(7), 1243-1258.
Bowerman, B. L., & O'Connell, R. T. (1990). Linear statistical models: An applied approach PWS-Kent Boston.
Brabham, D. C. (2008). Moving the crowd at iStockphoto: The composition of the crowd and motivations for participation in a crowdsourcing application. First Monday, 13(6) Brabham, D. C. (2010). Moving the crowd at threadless: Motivations for participation in a
crowdsourcing application. Information, Communication & Society, 13(8), 1122-1145. Bretschneider, U. (2012). Die ideen community zur integration von kunden in die frühen
phasen des innovationsprozesses Springer-Verlag.
Bretschneider, U., Knaub, K., & Wieck, E. (2014). Motivations for crowdfunding: What drives the crowd to invest in start-ups?
Brettel, M. (2003). Business angels in germany: A research note.
Brettel, M., Jaugey, C., & Rost, C. (2013). Business angels: Der informelle beteiligungskapitalmarkt in deutschland Springer-Verlag.
