Exploring Shifts In Perception Of Trust In Bitcoin After The
First Time Experience With A Software Wallet
SUBMITTED IN PARTIAL FULFILMENT FOR THE DEGREE OF MASTER OF SCIENCE
ERNEST MATTA
11152028
M
ASTER
I
NFORMATION
S
TUDIES
H
UMAN-
C
ENTERED
M
ULTIMEDIA
F
ACULTY OF
S
CIENCE
U
NIVERSITY OF
A
MSTERDAM
July 1, 2016
1
stSupervisor
2
ndSupervisor
Dr. André Nusselder
Dr. Frank Nack
Exploring Shifts In Perception Of Trust In Bitcoin After The
First Time Experience With A Software Wallet
Ernest Matta
University of Amsterdam
Graduate School of Informatics
Science Park 904, Amsterdam, NL
ernest.matta@student.uva.nl
ABSTRACT
This paper provides motivation, implication and execution of a study, which explores shifts in perception of trust in Bitcoin technologies after using a software wallet for the very first time. We created a framework (TIB) in order to measure the perceived trust linking to seven distinctive areas: advantage of use, usability, perceived skills needed, security, privacy, social and institutional trust. We conducted an experiment (N=12) using a simple Bitcoin wallet (Electrum). Pre-test and post-test quantitative data were gathered and consequently evaluated using paired t-test approach in order to find out statistically significant changes in perception. The results suggest that there was no overall shift in perception of trust in Bitcoin after the first time experience across all seven areas of our TIB framework. However, there were significant changes after the experience in the areas of perceived skills, usability and security. Our additional in-depth interviews confirm these findings and clarify that the experience with the Bitcoin wallet made them trustier towards Bitcoin in regards to sending, receiving and keeping bitcoins.
Categories and Subject Descriptors
K.4.4 [Electronic Commerce]: Cybercash, digital cashGeneral Terms
Experimentation, Design, Human Factors
Keywords
Bitcoin, experience, trust, usability, blockchain
1. INTRODUCTION
Headlines are full of negative experiences with Bitcoin related payments, personalities and exchange sites. Arguably, they help to create a certain perception on cryptocurrencies surrounded by mystery of an unknown and difficult-to-grasp concepts for a payment system. It is known that when there is a lack of trust in applications that are closely related to any form of e-commerce, it causes users to stay away from using these systems [1]. As a result, the trust in Bitcoin might get more undermined by these sensational headlines. Since the existence and adoption of any currency is heavily dependent on trust given by its users [2], Bitcoin must face a lot of misunderstandings and misconceptions from the general public that might drastically slow down or completely destroy its further adoption [30]. Bitcoin wallets are no exception and they face the very same issues when it comes to their wide-spread acceptance by users. We realised that there is a trust issue regarding the adoption of bitcoins and its technologies, and we understood that deeper understanding of these perceptions on Bitcoin payments is needed.
Currently, trust concepts, models and principles either fail to address trust in decentralised systems, or they ignore the unique
perception of users on such type of payment transactions [3]. Bitcoin as an innovative technology is indeed a fully decentralised system which involves multiple stakeholders by design [3,5]. As a result, it offers unique perspective and requires novel approach in order to explore the development of trust in Bitcoin technology. In addition to having to trust the trustee, users have also to be prepared to trust the technology that mediates interaction (for example, the Internet), and their own ability to use both the underlying technology and the specific application correctly [4]. For the last two factors, usability and user experience is a key prerequisite [3,4]. However, even in case of web-based digital Bitcoin wallets as a usable and convenient solution, they require a certain level of trust when shifting the responsibilities of encryption and backups to a third party [5].
The goal of our research is to conduct a socio-technological investigation into concepts of trust fundamental to Bitcoin technologies and digital currency economy as highlighted in [3,5,6]. It further justifies the importance of understanding the perception of trust in emerging cryptocurrencies as well as explores trust in HCI regarding Bitcoin wallets. The main idea is to focus specifically on the first-time users using a widely available digital wallet software. Researching the experience of using Bitcoin technology for the very first time means mainly two things: (i) finding out how the perception of trust in a cryptocurrency differs before and after the first experience with a Bitcoin transaction, and in case of significant shift in trust, (ii) investigating the reasons why and what factors helped to change users’ perception.
2. RELATED WORK
In this section we focus on available research on user experience and trust regarding Bitcoin and blockchain based technologies. Even though the resources on the matter at this point are rather scarce, we have found a substantial support for our research motivation, implication and execution.
2.1
Algorithm
As A Currency
Bitcoin is a new form of digital cash. Instead of being released by a central bank, its distribution is controlled by a decentralised network of computers. This network relies on cryptography and network technology to regulate the supply of Bitcoins and monitor who owns them. Therefore, Bitcoin is known as a cryptocurrency [8,23]. Although Bitcoin uses a ledger, it is kept up collaboratively by the decentralised network of computers. For that reason, it is also called a distributed ledger. As new transaction entries are added to this distributed ledger, they incorporate references back to the previous groups of entries. In this way, all participants can verify for themselves the source and origin of everything on the ledger. These groups of transaction entries are called blocks, and the whole collection is a block chain. Its creator, Satoshi
Nakamoto, wrote that it was a system of “purely peer-to-peer electronic cash” which can be controlled completely by the holder. In addition, it can be sent to anybody without needing a bank’s permission or facing a risk a confiscation. Every full participant in the Bitcoin system has a copy of every transaction, arranged in these blocks, going all the way back to the start of the Bitcoin. Each block is cryptographically linked to the previous block, forming a block chain that maintains a full history of transactions. Every copy of the ledger is synchronised by algorithms which are set up to keep “consensus” about the state of the ledger. However, once a transaction is sufficiently confirmed, it can never be reversed. Users can access the ledger with different applications on a desktop or mobile phone [7,8,10,23].
2.2 A Brief Note On Bitcoin Wallets
In order for a user to be able to store and transfer bitcoins, he or she needs a special software called Bitcoin wallet. A “wallet” is basically Bitcoin’s equivalent of a bank account. There are two main types of wallets. A software wallet is one that the user installs on own computer or mobile device. One is in complete control over the security of coins, but these wallets can sometimes be tricky to install and maintain. A web wallet, or hosted wallet, is one that is hosted by a third party. They are often much easier to use, but you have to trust the provider to maintain high levels of security to protect your coins [7,8,9].
2.3 Bitcoin-based Payment User Experience
Despite bitcoin having a number of superior user functionalities such as peer-to-peer connection, relative anonymity and affordability [10], real-world application developments confirm a rare case where practice is ahead of theory in regards of security and usability [6]. For example, in order for a digital payment system to appeal to users, it has to have a reduced complexity, be fun to use, and support money management without degrading user experience [11]. However, crypto payments such Bitcoin fall short regarding these requirements. Bitcoin does not reduce complexity, instead it increases it by asking users to manage addresses and setting up wallets [12].Moreover, introduction of additional processes increases the burden of experience for the user [12]. One example of that is the usability of Bitcoin's key management. Despite the most recent innovative attempts, it was found that its task involving the key management is made even more difficult by complex metaphors and confusing abstractions[13]. At this point, we may hypothesise that the trust of general public in bitcoin technologies is indeed connected to the usability, security and privacy concerns of using such a novel payment system [5].
2.4 Exploring Trust In Bitcoin Technology
Trust is a critical factor for user acceptance of cyber-based payment systems and their long-term success [4,6,7,8]. When designing or deploying such systems, the focus is on increasing trust perceptions rather than allowing users to make correct trust decisions [4,6]. A user-centred approach to the exploration of trust can shed light into the challenges experienced by people using Bitcoin. It could address various areas of trust such as usability, security and privacy from the user’s point of view regardless the technical details underlying the technology. It could explore feelings, opinions and attitudes of users towards this new disruptive currency. This is entirely different than the current algorithmic approach to trust in Bitcoin where it deals with keeping the technology safe from being misused or even broken and redistributes the blocks to everyone to see and approve [3,7].Research can also open up novel design opportunities to address the identified challenges and support trust. For example, one can imagine new class of interactive technologies where trust is captured, materialised, and gained or lost through exchanges. This new design space for decentralised interactive cryptocurrency technologies may not only support better adoption of Bitcoin technology but also the digital currency economy [3,8,13]. Hence, an interesting challenge has been recognised, resulting into the next section where research questions and methods were formulated in order to proceed further into the actual study.
3. RESEARCH QUESTION
We took up on this challenge and investigated the limitations of trust in a bitcoin-driven payment wallet in order to contribute to a better user adoption, or even pre-adoption [20], of Bitcoin technology in the near future. Our research question was derived from an unofficial interview, preliminary literature reviews and detailed investigation of the problem statement.
It is understood that the trust of non-users in Bitcoin technologies is connected to usability, security and privacy concerns of using such a novel payment system [5]. However, it is hypothesised that these concerns might be based on misconceptions about the usability, security and privacy as well as on general misunderstanding of the technology by the non- users [12]. Therefore, we investigate the question of:
How does the perception of trust in Bitcoin change after experiencing a Bitcoin wallet for the very first time?
We stress again that the perception of trust in our context refers to the factors of usability, security and privacy embedded in research frameworks related to the perception of trust in online payment and digital currency.
4. FRAMEWORKS
Three existing frameworks were used in order to investigate the perception of trust in Bitcoin payments as thoroughly as possible. As a result, these frameworks were implemented into one unique “super-framework” which worked for us as a guide for the measurements in the pre-test as well as post-test experiment data gathering in our randomised, controlled experiment [16].
4.1 Trust In Payment Systems
According Sasse [4], whether users are prepared to trust and engage in an currency exchange additionally depends on a number of other factors that characterise the interaction. Factors that have been identified include:
1. The number of actors involved in the exchange (ranging from dyads to potentially millions in public good dilemmas); 2. The actor type (individuals, organisations, technology such
as an e-commerce web site);
3. Whether there is synchronous or asynchronous trust exchange (asynchronous exchanges create higher strategic insecurity);
4. Whether the user can identify trust-warranting properties; 5. The type of signals employed to communicate
trustworthiness (symbols and symptoms of trustworthiness, identity and property signals);
6. The truster’s propensity to trust; 7. The truster’s knowledge of the situation; 8. The truster’s prior experience;
9. The potential benefits expected by the truster; and 10. The risk to the truster’s risks (enacted as ‘trusting action’).
4.2 Trust Framework For Bitcoin
Sas and Khairuddin [3] propose a research framework (Figure 1), integrating key aspects of trust from HCI literature with the main challenges posed by Bitcoin technology in order to ensure the exploration of trust across all the Bitcoin stakeholders.
“The framework places Bitcoin technology at its centre, and highlights how different stakeholders are involved in shaping the three different levels of trust. We define technological trust as
people’s trust in Bitcoin technology experienced before, during,
and after engaging in online transactions. This could include users’ trust that their Bitcoin account is secured and cannot be hacked, or payees’ trust that the transfer is authorised” [3, p. 340].
4.3 Implemented Framework For TIB
Our trust in Bitcoin (TIB) framework incorporates several distinctive areas of trust connected to payment technologies. We merged these two frameworks with the concepts of Bitcoin’s security and privacy elaborated in [5] in order to create seven specific domains of TIB, which were measured before and after the experiment. These seven areas consist of:
• Advantage of usage; • Perception on usability;
• Perceived skills needed (to use and interact with Bitcoin); • Perception of security;
• Perception of privacy;
• Social trust (how interaction with Bitcoin is perceived within society, friends and family) ;
• Institutional trust (how strong is the trust in the governmental bodies of cryptocurrency such as Bitcoin).
The goal was to measure and evaluate a shift in these seven areas of perception of trust in the post-experience with a software wallet. Any statistically significant difference in any of the seven areas had contributed to a change in trust towards Bitcoin.
5. METHODS
In essence, our study was an experimental research by nature involving both quantitative and qualitative approach. Our quantitative measurements from pre-test and post-test questionnaire resulted into our final dependent t-test for paired samples [14,15,18]. Our qualitative data were gathered during the in-depth, one-on-one interviews with some of our subjects after they had experienced a Bitcoin transaction and filled out the surveys. From the design science research methodology point of view [17], we were set out to investigate a knowledge question. Therefore, we used an existing Bitcoin wallet called Electrum [19] in order to find out how it influences first-time users so we could better understand the adoption phase, or even the pre-adoption phase [20], for a cryptocurrency such as bitcoin.
In our experimental study (Figure 2), the aim was to manipulate an independent variable, which in this case is the specific Bitcoin wallet experience, and then examine the effects that this experience has on the dependent variable, which in this case is the perception of trust in Bitcoin. Since it was possible to manipulate the independent variable in terms of which software wallet and which cryptocurrency is going to be used, this experimental research had the advantage of enabling us to identify the cause and effect between variables [14,15,19].
5.1 Research Participants
12 participants (6 males, 6 females), aged between 18 - 54 were recruited through a convenience sampling approach [14,15]. Our sample consisted of 6 different nationalities of which the Dutch nationality had a predominance (7/12). Before the experiment, all of the participants had experience with some kind of popular online payment system such as credit cards, banking mobile apps or iDEAL. Only one participant (1/12) had never heard of Bitcoin before. The rest was slightly familiar with its concept due to various media coverage, news articles and word of mouth about the Bitcoin currency.
5.2 The Approach And Procedure
Our research design ended up having three distinctive parts that had to be carried out separately from each other. The research cycle consisted of two extensive survey questionnaires that on average took 10 minutes each to complete plus the time needed for downloading and setting up the Electrum wallet. Time spent on receiving and spending bitcoins depended on each participant individually. The goal was to let our participant use the Bitcoin in their natural environment, presumably at their home, and let them make any valid transfer involving bitcoins.
Figure 1. Research framework for exploring levels of trust in Bitcoin technology (left) and across stakeholders groups [3].
systems or address it from the sole perspective of users of such systems. Bitcoin is not only a decentralized system but a grassroots driven technology involving multiple stakeholders. Thus, it offers a unique perspective to explore the development of trust within and across these
stakeholders, together with its most challenging and promising issues. A deep understanding of these trust issues in Bitcoin technology may in turn challenge some of the assumptions underlying our current models of trust.
Figure 1: Research Framework for Exploring Levels of Trust in Bitcoin Technology (left) and across Stakeholders Groups (right)
RESEARCH FRAMEWORK
We now propose a research framework (Figure 1) which integrates key aspects of trust from HCI literature, with the main challenges posed by Bitcoin technology, to ensure the exploration of trust across all the Bitcoin stakeholders. The framework places Bitcoin technology at its center, and highlights how different stakeholders are involved in shaping the three different levels of trust. We define technological trust as people’s trust in Bitcoin technology experienced before, during, and after engaging in online transactions. This could include users’ trust that their Bitcoin account is secured and cannot be hacked, or payees’ trust that the transfer is authorized. Social trust is the trust that Bitcoin stakeholders develop between each other. This trust is enlisted for each type of exchange occurring across (and within) different categories of stakeholders. For example transactions involving purchase of goods enlist trust between users and merchants. Upon completion, these transactions require miners’ authorization, so both users and merchants need to trust the miners for completing their job. At the same time, selfish miners can raise issues of trust among miners (Eyal and Sirer, 2014). Social trust between users/merchants and exchangers can be also problematic1. We argue that
because of its decentralized nature, the classic definition of institutional trust does not apply to Bitcoin. However, there is a higher authoring to which Bitcoin technology is requested to be accountable, namely governmental institutions. We define institutional trust, the trust of
1 Manhattan, U. S. Attorney Announces Charges Against Bitcoin Exchangers, Including CEO of Bitcoin Exchange Company, For Scheme To Sell And Launder Over $1 Million In Bitcoins Related To Silk Road Drug Trafficking. 27.01.2014. URL: http://www. justice. gov/usao/nys/pressreleases/January14/SchremFaiellaChargesPR. php.
governmental institutions in Bitcoin technology. The main issues here relate to money laundry and deflation.
Applying the Framework to Identify Trust Challenges
We now explore how the framework can be applied to identify important trust issues which deserve stronger HCI engagement. We should note that there is limited empirical work exploring the experience of using Bitcoin and the issues of trust surrounding it. First we start by describing the Bitcoin stakeholders, grouped by Shcherbak (2014) in four categories: users, miners, exchanges and merchants. Users are people who use Bitcoin to buy goods and services from Bitcoin merchants. Merchants are businesses which accept Bitcoins as medium of exchange for goods and services and are connected to the Bitcoin network. Exchanges are the providers of online trading platforms where the registered members can exchange their Bitcoins for traditional currency and vice versa. Miners are those Bitcoin stakeholders who can record transactions (and collect reward) after they successfully solved crypto
-puzzles (Eyal and Sirer, 2014). Users’ Trust in Bitcoin
One specific challenge pertaining to users is their limited knowledge of how Bitcoin technology works and how they need to protect their bitcoins. Keeping bitcoins on one’s computer involves security risks similar to keeping large sums of cash in one’s physical wallet (Bitcoin Wiki, 2011). Although Bitcoin is decentralized and at large has no single point of failure, it is nevertheless susceptible to a form of denial of service (Quora Forum, 2011) or double-spending attack (Karame, 2012).
Merchants’ Trust in Bitcoin
Merchants’ trust is challenged by their limited knowledge about buyers, and whether their payment will be received in time or at all (Shcherbak, 2014). They also lack the
B
Mine rs Exchanges Merchants Users Governments Institutional TrustB
Technological Trust Social Trust 3405.2.1 Questionnaire Part One
We started out with the initial investigation of our participating users who had no prior experience with any cryptocurrency. We gave them a questionnaire which was carefully designed and developed according the known research frameworks for trust in information systems and bitcoin technology described earlier. The aim was to gain insights about their priori perception on trust in cryptocurrencies such as bitcoin. Except for general questions about the participants, the questionnaire consisted of quantitative five-level Likert scale items for each of seven areas of our TIB framework. The scale format was as follows: Strongly Disagree - Disagree - Neutral - Agree - Strongly Agree. A screenshot of one of the trust areas - advantage of usage - is depicted in the Figure 3.
Each area of trust of our TIB framework was represented equally by 8 questions related to that specific “chapter” of trust. In total, the survey consisted of 60 questions. The questions were based on a thorough investigation of the problem space as elaborated in [3,5,6,7,8,9,12,13] as well as inferred from people’s opinion found on dedicated Bitcoin forums, discussion places (e.g Reddit) and unofficial interviews with members of general public. It was found that the concerns in academic literature closely overlap with opinions and attitudes of users and non-users alike. As a result, participants were directly asked about their honest judgment on those general concerns. All Likert scale questions were mandatory in order to avoid missing data. The full text of the questionnaire can be found in the appendix. As for the survey tool, we used an online questionnaire builder called SurveyGizmo [21].
5.2.2 The Bitcoin Wallet Experience
The next step was to make our subjects experience a Bitcoin transaction in a real-world. They were asked to download a specific bitcoin wallet - Electrum (Figure 4).
After submitting their public key to the researcher, they received their very own bitcoins from the researcher worth around €2. Subjects were asked to spend those bitcoins freely, e.g.on buying
any song they like from a bitcoin music marketplace such as Liberty Music Store or make a donation. As a result, they experienced a Bitcoin wallet and Bitcoin transaction directly for the very first time, using their very own Bitcoins.
The reason Electrum was chosen is twofold. Firstly, it is considered to be a simple and user-friendly software wallet in terms of usability and interface design [4,5,12]. The goal was to make the first experience as “hassle-free” as possible. As seen in the Figure 4, the Electrum window has three main parts needed to make a Bitcoin transaction right on the top: “Send”, “Receive” and “Addresses” tabs. Secondly, Electrum is a so called light-weight client [7], meaning that it does not download the whole blockchain from the network. It would make setting up and synchronising the wallet an unnecessarily annoying experience, at best, for our first-time users.
5.2.3 Questionnaire Part Two
The final step of the experiment was measuring the posteriori perception on trust. Within a day after subjects performed they first transaction, they were asked again via the questionnaire about their perception on bitcoin currency. The questionnaire was not exactly the same as the first one; however, it measured the same inquiry: the perception on trust in seven chapters of our TIB framework.
In Figure 5, a screenshot from the same chapter of our trust framework - advantage of use - is depicted. This time it is from the second, follow up questionnaire on the perception of trust. Two things had been changed in comparison with the first survey. Firstly, the questions asked are formulated in a way that corresponds with the fact that the participants had been exposed to Bitcoin and experienced an interaction with a software wallet. In other words, the tone of questions in the Likert scale were no more in a hypothetical state. For instance, the sentence “If I had an option, I would choose Bitcoin over any other online payment systems” had transformed into a more declarative “I would choose Bitcoin over other any online payment systems”. Secondly, a comment box had been added into each page in order to give participants an option to express additional thoughts on each of the seven areas of the TIB framework.
5.2.4 Additional In-depth Interviews
After the bitcoin experience, semi-structured interviews were carried out with a limited number of subjects (4 out of 12) in order Figure 3. Screenshot of the first questionnaire
Figure 4. Screenshot of the Electrum wallet
to find out deeper causes that might not had been captured via the questionnaires. The aim was to compare the outcomes with the results from the questionnaire and look for additional or confirming information [14,15].
5.3 Data Analysis
In our case of two samples that were correlated, in order to compare our two population means, we used a paired sample t-test as our statistical technique. By using the paired t-test, we could statistically conclude whether or not the experience with the Bitcoin wallet (our independent variable) had caused a change in perception of trust in Bitcoin (our depended variable). Consequently, we initially set up two hypothesis:
• H0: There is no significant change in perception of trust in Bitcoin after the Bitcoin experience.
• H1: There is a significant change in perception of trust in Bitcoin after the Bitcoin experience.
Intervals were used in order to encode the Likert scale items. Assigned quantifiable values were as follows: Strongly Disagree = 1, Disagree = 2, Neutral = 3, Agree = 4, Strongly Agree = 5. Since our alternative hypothesis essentially means “not equal”, we used a two-tailed t-test with a confidence level of 95% (α = 0.05). All the Likert scale questions were mandatory to answer; therefore, no missing values had occurred.
The interviews were analysed in a regular fashion. After transcribing and careful examination, colour coding was used for a better overview and orientation in our qualitative data. Consequently, the codes were analysed and grouped into categories. The most relevant categories were classified under distinctive concept labels in order to explore connections between them. Finally, these connections allowed us to formulate a more general theory and understanding [14, 22].
6. RESULTS
In this section we provide quantitative and qualitative results of our study. Our quantitative data stemmed from our exhaustive questionnaire while the qualitative data were mainly obtained from the in-depth interviews with our test participants. The intention was not to measure the agreement or disagreement of participants towards a particular trust issue. Rather, we aimed to explore a shift in perception regardless of their personal preferences for each case.
6.1 Paired T-test Results
The results of our experiment indicate that there has been a statistically significant shift within some of the areas of trust in Bitcoin technologies. However, these minor shifts do not provide us with the evidence that the first time experience caused an overall change in perception of trust in Bitcoin.
6.1.1 Evaluation Of T-test Results
Each single question was analysed and evaluated using dependent t-test between the pre-experiment and post-experiment data using paired inputs of each individual participant. Degree of freedom equaled 11 (12-1) at 95% confidence level. Critical value for T = +/- 2.201). The null hypothesis was rejected when the paired t-value crossed the critical t-value T in either direction (+ or -). In that case it was generally considered to be a statistically significant change between pre-test and post-test data; therefore, we could reject the H0 and accept our H1 pointing to the
statistically significant shift in perception.Detailed results for all the t-tests can be found in the appendix.
6.1.2 Observed Shifts In Perception
In fact, we observed changes only in three out of seven areas of our TIB framework: perceived usability (in 2 items out of 8), perceived skills needed (in 6 out of 8) and perceived security (only in 1 out of 8). Changes in perception of Bitcoin were observed in the following cases:
• A significant shift occurred when the participants disagreed that it takes too much time to figure out how Bitcoin works in order to make their first transaction. In this case, there was a shift in their perception towards “disagree” after using the software wallet for the very first time. We observed this significant change with t = -4.66 (M = -1.33, SD = .98). With this t-value, we even managed to cross over the critical value T for 99.9% confidence level (T = +/-4.437).
• Second statistically significant change in usability occurred in the perception of keeping an eye on exchange rates. Overall, after using the Bitcoin wallet, participants disagreed that they need to keep track of exchange rates all the time when using Bitcoin: t = -2.67 (M = -.75, SD = .97).
• All of the participants disagreed with the statement that they need to be a computer/programming specialists in order to be able to use a Bitcoin software. This perception became even stronger towards “disagree” after the experience with the Bitcoin wallet: t = -2.33 (M = -.33, SD = .49).
• After the experiment, almost everyone (11/12) disagreed or strongly disagreed that they need to have a knowledge about cryptography in order to use Bitcoin. This perception changed with t = -2.28 (M = -.75, SD = 1.14).
• Participants, after the experiment, expressed their strong disagreement with the notion that Bitcoin payment systems are too technical for them to understand. There was a significant change from “agree” since we registered t = -2.59 (M = -.50, SD = .67).
• Participants, after the experiment, changed their opinion about the notion that Bitcoin payment systems are too technical for them to use: t = -3.33 (M = -.50, SD = .52).
• After using Bitcoin, almost everyone (10/12) disagreed with the notion that they might experience difficulties when making their very first transaction. We observed a shift form the “neutral” attitude with t = -2.31 (M = -.92, SD = 1.38).
• The largest shift we observed was in perceived ability to use Bitcoins. After the test, everyone realised that they are able to use Bitcoin easily: t = 4.76 (M = .92, SD = .67).
• Users’ perception shifted significantly when comparing security of Bitcoin to existing banking or payment systems. With only one participant (1/12) agreeing that Bitcoin’s system is less secure after he or she experienced the wallet while the rest stayed predominantly neutral or disagreed, we rejected the H0: t = -2.93 (M = -.66, SD = .78).
The most significant shift was observed in the areas of perceived skills needed to use Bitcoin. On the other hand, no change in perception was observed in the areas of advantages, privacy, social trust and institutional trust. Here we present descriptive summaries of the results for each of these areas. For the detailed t-test values, please look in the appendix.
6.1.3.1 Advantage Of Use
Even after using Bitcoin, the majority of first-time users did not see any advantages using a cryptocurrency such as Bitcoin. In general, participants already thought that Bitcoin makes online payments easier than other mainstream online financial services do. However, this perception remained unchanged since it moved only slightly after the bitcoin experience. Advantage of much smaller Bitcoin transaction fees versus other financial systems yield also no difference after the experiment. Interestingly, after using Bitcoin for the first time, almost half of the participants still did not consider being anonymous as an advantage. The rest stayed neutral towards this question. When participants compared the perceived advantages to existing payment systems, nothing changed in their views either.
6.1.3.2 Security And Privacy
We observed a shift in the area of security only in one particular case when users compared the security to existing banking systems. However, we did not see any change in perceived security risks when sending, receiving or keeping bitcoins. Although one forth agreed that having Bitcoins involves a high risk in terms of money management, overall the shift did not occur. After the experiment, only one participant still did not feel secure in a trustless system such as Bitcoin where there is no central entity to oversee the transactions. The rest still stayed neutral or disagreed with that statement.
Before and after the test, participants agreed that Bitcoin’s concept of sending and receiving money anonymously is an attractive idea. On the other hand, after the experiment almost half of participants still found the lack of personal information within Bitcoin transactions intimidating. Similarly, half of respondents agreed that they want to provide personal information when sending bitcoins over to someone and the rest was still mainly neutral. The difference in levels of privacy between Bitcoin and banks were perceived as being the same.
6.1.3.3 Social and Institutional Trust
Everyone disagreed that Bitcoin is only good for criminal activity and this perception did not change after the experiment. Likewise, almost everyone disagreed that people who use Bitcoin are not to be trusted. In general, majority liked the idea of Bitcoin being a totally decentralised currency system which is self-controlled by the whole community of users. Some users were still neutral about being suspicious towards Bitcoin community in terms of their potential malicious activity. Also nothing has changes when we asked participants whether they perceive Bitcoin as not being money in any kind of sense. Almost everyone disagreed.
6.2 Open Questions In The Questionnaire
The post-experiment questionnaire embodied an open question within each of seven areas of perception on trust. One third of participants (4/12) used this chance to express their thoughts throughout all the areas of trust. When it comes to use and usability of Bitcoin, users responded that it looks easy to use even for someone who has basic computer skills. They added that it is advantage that there are no costs for having an account whereas at a bank they have to pay for having an account/pin-card. When itcomes to skills that are needed, respondents thought that anyone who can use internet banking should be able to use a Bitcoin wallet. Security concerns were also discussed. They thought that they do not see any reason to trust in bank’s security more than they would trust in Bitcoin.
Thoughts on the privacy were twofold. Participants articulated their appreciation for the anonymity feature. On the other hand, this anonymity was perceived as not practical when it came to purchase verification. Regarding the social trust, participants addressed the criminal activities connected to Bitcoin payments such as a web merchant called Silk Road, but it it was remarked that “it was no reason not to use Bitcoin”.
6.3 In-depth Interviews
Before the experiment, our participants were partially informed and had a basic understanding what Bitcoin is. Although they considered the concept of a cryptocurrency rather “vague, unknown and unclear”, they grasped the idea of anonymous transactions. Participants did not “trust in news and media articles trying to picture Bitcoin as something bad”; however, they still perceived it as having a bad reputation connected to “illegal activities such as Silk Road website”. Regardless the limited knowledge about Bitcoin they had beforehand, participants were interested in ‘trying it out because of low risk involved” within our experiment. We provide our coding diagrams in the appendix.
6.3.1 Advantage Of Use And Usability
Advantage of use and usability are closely linked together. After we colour coded our interviews, we discovered several distinctive concepts our interviewees referred to: ease of use, undiscovered advantages, comparison to banking apps and practical usage. The interviews showed that Bitcoin, in practice, turned out to be easier to use than initially anticipated. It was due to easy-to-understand UI of Electrum, even though it looked different than familiar banking apps. Consequently, first-time users spent much less time figuring out how Bitcoin works in order to make their very first payment transaction. Regarding the usability of the software wallet, the experience was also positive. The set up process was “adventurous” but simple to follow, especially when provided with simple set of instructions (P1: “Well, I had to read the instructions very well, because I am not super handy with computers. But, when I read it, yeah, it was ok”). The participants did not think that they need to keep track of exchange rates when using Bitcoin, e.g. BTC versus EUR, since the prices are usually available in informative dual display. Even though our participants were aware of some of the advantages and discovered new ones, they did not think they are exclusive enough for them to start trusting Bitcoin more than before. It was also understood than they grasp the importance of these advantages; however, they do not see a practical use of them on the personal level.
6.3.2 Perceived Skills Needed
Evidently, the most notable movement in users’ beliefs were observed in perceived skills they needed to posses in order to send and receive cryptocurrency using a simple software wallet. Before the experiment, the participants were convinced that they would need advanced programming, computer and cryptography skills to utilise a bitcoin transaction. They also felt that Bitcoin was just too technically advanced for them and feared that they might encounter difficulties when making their first transaction. After the experiment, however, this perception completely changed and participants realised that they are able to use Bitcoin easily (P1: “Well, I didn't really need a lot of technical skills I just had to read the instructions carefully…and just know how to scan a QR code I
guess”, P3: “Actually I did a little bit of web browsing and just read what it says…not really technological but just make sense of it all”). Although users did not feel that they need to have advanced technical or other skills, they stressed that it is important to be sceptical about Bitcoin software and do some research or educate oneself before committing his or her full trust to a wallet. Furthermore, interview evaluation revealed that most of the participants were positively surprised by how “fast and simple” it is to make Bitcoin payments with a “minimum knowledge” about it. Some participants described this post-experience as pleasant (P1: “I thought it would be more complicated…Yeah, I was pleasantly surprised how easy it works”) or even scary (P3: “It was super easy…but it is also a bit scary when you think about it, you know, you can spend money fast in this way”). In this particular area of trust, we concluded that the level of trust in Bitcoin partially increased in a positive way. This make sense since perceived user skills of an individual take into consideration his or her capabilities to use a technological system [24]. In this case, these capabilities were minimised to a simple ability to use a standard, simple desktop application with a “self-explanatory UI”. On the contrary, this over simplified UI can also be considered as the reason for distrusting the wallet (P3: “The problem I have with it is that it’s just not mature and it looks like some kind of alfa or beta version, and I have the feeling that I am working in some kind of excel sheet or something. So I believe whatever it says, but I don't have a feeling that it has, you know, valuable things like a new currency”). Hence, a too simplified UI can have negative consequences on trustworthiness. Because Bitcoin wallet can have a simple UI and can be easy to use, our participants worried that it can be easily abused by malicious people. Therefore, one should have a "confidence" about his or her trust toward wallet they use based on research or reputation of a wallet.
6.3.3 Security And Privacy
The last, very small change in perception, was observed in the area of security. Specifically, participants did not evaluate Bitcoin to be less secure when compared to established payment systems such as banks or Paypal. Interestingly, our first-time users disagreed that Bitcoin is less secure than banks, even though the bankruptcy of bitcoin’s famous Mt.Gox exchange in 2014 has raised serious doubts about cryptocurrency’s future [28]. In fact, our qualitative interview confirmed that Bitcoin puts the responsibilities for most of its security measures into the hands of the end user compared to centralised monetary systems [26, 29]. In one particular interview case a centralised system, such as bank, was trusted more for the very same reason. I addition, it also turned out that the perception of security in a Bitcoin wallet depends on its reputation.
In general, there were no security concerns regarding Bitcoin and its system and there was no difference in perceiving its security feature before and after the experiment, even though it has its perceived advantages and disadvantages. After using the bitcoin wallet, participants expressed their appreciation for the privacy feature which is central to Bitcoin. However, they pointed out that this sort of anonymity is not always “handy”, and it is even undesirable in cases when there is no confirmation about the purchase or no option for a refund.
6.3.4 Social and Institutional Trust
The main belief in a social context was that Bitcoin needs to become more “mainstream” in order to be accepted and used by our participants in the future instead of other banking applications. Even though our participants associated Bitcoin with “criminals” and “geeks”, it turned out to be not necessarily a bad thing. The
fact that it is used by criminals and geeks proves the system to be rather trustworthy in a specific context. (P3: “Because I also know that big technology breakthroughs coming from geeks and geeks often have good stuff”, P4: “If criminals use it for their needs, it (Bitcoin) must be doing something right in a way”.) After the experiment, participants did not feel an urge to share their experience with cryptocurrency with their family and friends, nor did they explicitly recommended it.
Although the participants “liked the idea of decentralised currency”, the level of trust towards decentralised and centralised money system stayed equal. They “do not trust decentralised system less nor more than centralised system such as banks”. Only in one case trust towards banks prevailed (P2: “Because it is centralised I trust it more, I guess”). This distrust was caused by recent news in media about Bitcoin that “confronted its transparency”. In general, participants considered the way it works “smart”, and they do not care how exactly Bitcoin works “behind the scenes” as far as there is someone to take care of the general consensus.
6.4 The Big Picture
At the end of each interview, we asked interviewees a very open question to reflect, in their own words, on how their trust in Bitcoin changed from before to after the experiment. They arrived to a conclusion that the first time use of the wallet did not change their overall perception of trust in the system. Whether they trusted it before or not, they did not consider the experience having an impact on their perceived beliefs about Bitcoin. However, the positive shifts in the areas of its usability and perceived skills made them more comfortable with the idea of using the bitcoin currency. These qualitative information correlated exactly with our paired t-test results from the experiment.
Furthermore, with no significant change in neutral attitude towards perceived institutional trust regarding the “back-end” of Bitcoin, we can confirm that, as typical for financial systems, people care more about how to use it than how it works [30]. The interviews also revealed that in order to trust bitcoin currency more and start using them regularly, Bitcoin must become a mainstream monetary system accepted everywhere. However, the major drawback that makes Bitcoin an illogical payment choice is that there is no existing damage control for the user when something goes wrong. In other words, no-one can guarantee money back. It was agreed that if Bitcoin should ever be considered to be utilised as a widely used payment method, something must be done about it. Otherwise, it will not gain enough trust of the general public to become a truly revolutionary monetary technology on a global scale.
7. DISCUSSION
Even though our findings do not indicate an overall shift in the perceived trust, our qualitative data confirm that there has been a lot of misunderstanding of Bitcoin by both users and non-users alike regarding technical details [30]. From the non-user point of view, we observed that these misunderstandings were grounded in the perceived skills that users would need to operate a Bitcoin based software as well as perceived security and privacy misconceptions about owning bitcoins. Due to bad reputation connected to black market, no refund option and bitcoin exchange bankruptcies [28], we now understand why Bitcoin makes people suspicious about whether the currency can be trusted [30]. The black market usage can be considered as a justified foundation of distrust since the analysis of Google search data of Bitcoin users
from January 2015 [33] provided robust evidence that computer programming enthusiasts and illegal activity drove interest in Bitcoin. However, after using Bitcoins themselves, our research participants pointed out that connecting bitcoins to illegal activities is reasonable but completely irrelevant to their perceived trust towards using cryptocurrency in practice.
7.1 Mainstream Adoption: Trust and Usability
It is known that perception of usability is one of the cornerstones of trust in bitcoin cryptocurrency [5] and that current Bitcoin wallets do not make it easier for users to grasp the complex concept of keys and key management [26]. On the contrary, we found evidence that making bitcoin transactions was far less complicated as initially perceived. This shift in perceived usability was partly due to Electrum's design foreseeing user’s expectations of Bitcoin’s usability and anticipating user’s initial guess on what using this technology would be like [27] in order to create Bitcoin applications that take minimal effort even for non-technical users to use [3]. After users invested a little bit of time into familiarising themselves with a wallet, they were surprised how fast and easy it is to make a payment. Since they were using Electrum - a very simple Bitcoin wallet [19] - the user interface was considered to be self-explanatory and easy to use.From the user interface point of view, participants did not consider Electrum to be more or less trustworthy than well-known payment options such as iDEAL (dutch online payment system) or banking apps. This finding comes across as counter-intuitive because the issue of usable public key cryptography for user authentication has been well known [25,26], and it has been concluded that current UI of desktop and mobile apps for Bitcoin employ complex metaphors which don’t fully capture the implications of key management actions [26]. Additionally , it had all the signs of a usable software yet the lack of visual appearance did make users trust it less since it does not appeal as a “finished product” according to todays perceived standards [33].
Furthermore, our research indicates that users would hardly keep using Bitcoin in the post-research period. This is due to three factors. One being that the advantage of use over other payment systems is not perceived as superior. Nor it is perceived as inferior, nevertheless [30]. Second reason is that Bitcoin has not yet become a mainstream, globally accepted monetary system [8,23]. Third reason is that there is no option to claim money back in case something goes wrong with the transaction. This is perceived as the major drawback of Bitcoin [30]. Therefore, users prefer to use familiar online payment options which they already employ in their everyday lives.
7.2 Beyond Bitcoin
This revolutionary concept of cryptocurrency is much more than just a new way of transferring value and paying for goods and services. The “magic”, so much debated nowadays, is in the underlying technology that drives Bitcoin - the blockchain. There are dozens of other ways how this blockchain could be used in the near future: from revolutionising the ways how government transparency works [23], through effective ways of moving the ownership from person to person [7], to peer-to-peer energy trading [31]. Perhaps the most interesting and ambitions use of blockchain concepts in the future is applying its underlying philosophies into “blockchain thinking” for decentralised autonomous artificial intelligence [32]. At the dawn of this new era of blockchain paradigms, we believe that our research can shed light into understanding of its potential users and help to
prepare them in the pre-adoption phase for any of these blockchain based systems and services in the “trustless” future.
7.3 Limitations
Our experiment suffered a tedious recruitment process when it turned out that making potential users participate in our three-fold research process is more difficult than we expected. Even though the potential participants were well informed about the process involving two extensive questionnaires, setting up the software and using the Bitcoin wallet, it was expected that at least 30 candidates would join our experiment. Due to the limited time available for our whole study (90 days), we ended up with 12 participant. However, we believe that reliability would not increase significantly after 10 users being tested.
Furthermore, a small amount of data related to usability could be influenced by one negative, unforeseen difficulty when paying for a book online through purse.io with bitcoins. It turned out that it works in a very limited manner; however, this difficulty is due to a third party. This option was further not suggested for other participants and the results stemmed from this particular case did not significantly influence the paired t-test results. We also decided to use only one particular wallet in order to make the experience consistent throughout our sample. However, there are wallets with better user interfaces than Electrum has, which at the end could have a more positive effect on the perception based on visual design. Moreover, the small amount that we gave our participants to spend also limited their choices of spending the bitcoins on something more desirable which could have also influenced the overall user experience.
8. CONCLUSION
The experiment we carried out had one clear mission: to find out how perceptions of general users change after experiencing a relatively new and disrupting monetary system such as Bitcoin. Since Bitcoin itself was proposed as an electronic transaction system without trust [10], we decided to focus on the HCI part of the equation and explore changes in users’ perception of trust towards this trustless system after they experienced Bitcoin on their own for the very first time. We created a framework (TIB) out of three related and existing frameworks in order to measure the trust within seven distinctive areas: advantage of use, usability, perceived skills needed, security, privacy, social and institutional trust. On the overall scale, we did not observe a significant shift in perception throughout all of the seven areas of our TIB framework. However, we did observe a change in perception in the areas of usability, skills and security. The biggest shift occurred in the perceived skills when users realised that they do not need any advanced knowledge or technical skills to be able to use Bitcoin or make Bitcoin transactions on their own. Bitcoin was also perceived as simple and easy to use after experiencing the Bitcoin for the first time. Our participants also realised that Bitcoin is not more or less secure than banks or other online payment solutions available nowadays. On the other hand, there was no significant shift observed in the areas of advantages, privacy concerns and socio-governmental trust in Bitcoin. Moreover, it was found that even though Bitcoin is easy to use and has certain advantages regarding anonymity and speed, it is not trusted to be a fully integrated substitution for established banking apps and online payment services, just yet. Besides the reasons of not having an option to reverse a transaction and not seeing personal information included in payments, Bitcoin is not considered to be used instead of banking apps simply because there is no need to do so. Participants are fully comfortable using
monetary systems they are familiar with, although they do not necessarily trust it more or less than Bitcoin.
9. FUTURE WORK
In the future we would like to see a similar experiment which measures and evaluates trust after a long term use of Bitcoin. In this study we focused particularly on a single experience where we limited our focus on the adoption phase. It could go also in a direction of deploying different software wallets such as mobile apps or online web wallets. From the design point of view it would be vital to investigate what is the fine line between a simplified user interface and a simplified yet trustworthy interface when involving cryptocurrency transactions. Furthermore, we recommend to provide the participants with a substantial amount of bitcoins in order to broader their buying options during the experiment.
10. ACKNOWLEDGEMENTS
The authors would like to thank André Nusselder for his valuable feedback as supervisor to this study. Additionally, we would like to thank our participants who gave up their precious time in order to participate in our research.
REFERENCES
1. Princeton Survey Research Associates. 2002. A matter of trust: what users want from web sites. Consumer WebWatch. Retrieved from http://consumersunion.org/wp-content/ uploads/2013/05/a-matter-of-trust.pdf
2. Richey, S. 2007. Manufacturing trust: community currencies and the creation of social capital. Political Behaviour 29, 1, 69-88.
3. Sas, C., and Khairuddin, I. E. 2015. Exploring trust in bitcoin technology: A framework for HCI research. Proceedings of the Annual Meeting of the Australian Special Interest Group for Computer Human Interaction, 338-342.
4. Sasse, M. A. 2005. Usability and trust in information systems. Trust and Crime in Information Societies. Edward Elgar Publishing, Cheltenham, UK.
5. Krombholz, K., Judmayer, A., Gusenbauer, M., and Weippl, E. 2016. The other side of the coin: User experiences with bitcoin security and privacy. Financial Cryptography and Data Security 2016.
6. Bonneau, J., Miller, A., Clark, J., Narayanan, A., Kroll, J. A., and Felten, E. W. 2015. SoK: Research perspectives and challenges for bitcoin and cryptocurrencies. In Proceedings of the 2015 IEEE Symposium on Security and Privacy (SP), 104-121.
7. Antonopoulos, A.M. 2014. Mastering Bitcoin: Unlocking Digital Cryptocurrencies. O’Reilly Media, Sebastopol, CA. 8. Swan, M. 2015. Blockchain: Blueprint for a New Economy.
O’Reilly Media, Sebastopol, CA.
9. Bitcoin Simplified. How to set up a wallet. 2016. http://bitcoinsimplified.org/get-started/how-to-set-up-a-wallet/
10. Nakamoto, S. 2008. Bitcoin: A peer-to-peer electronic cash system. Retrieved from https://bitcoin.org/bitcoin.pdf 11. Mainwaring, S., March, W., and Maurer, B. 2008. From
Meiwaku to Tokushita!: Lessons for digital money design
from Japan. In Proceedings of the SIGCHI Conference on Human Factors in Computing Systems (CHI ’08), 21-24. 12. Gao, X., Clark, G. D., and Lindqvist, J. 2016. Of two minds,
multiple addresses, and one ledger: characterizing opinions, knowledge, and perceptions of Bitcoin across users and non-Users. Rutger University & ACM CHI’16.
13. Eskandari, S., Barrera, D., and Clark, J. 2015. First look at the usability of Bitcoin Key Management. USEC ’15. 14. Bryman, A. 2008. Social Research Methods. Oxford
University Press, Oxford, UK.
15. Lazar, J., Feng, J. H., and Hochheiser, H. 2009. Research Methods In Human-Computer Interaction. Wiley, Hoboken, NJ.
16. Dallal, G. E. 2005. The analysis of pre-test/post-test experiments. Retrieved from http://www.jerrydallal.com/ lhsp/prepost.htm
17. Wieringa, R. J. 2014. Design Science Methodology for Information Systems and Software Engineering. Springer, New York City, NY.
18. Laerd Statistics. 2016. Types of variable: experimental and non-experimental research. Retrieved from https:// statistics.laerd.com/statistical-guides/types-of-variable.php 19. Electrum Bitcoin Wallet. 2016. https://electrum.org/#home 20. Rogers, E. M. 2003. Diffusion of Innovations (5 ed.). The
Free Press, New York City, NY.
21. SurveyGizmo. About SurveyGizmo. 2016. https:// www.surveygizmo.com/company/about/
22. Bernard, H. R. 2006. Research Methods In Anthropology: Qualitative And Quantitative Approaches. AltaMira Press, Walnut Creek, CA.
23. UK Government Chief Scientific Adviser. 2016. Distributed ledger technology: beyond block chain. UK Government Office For Science.
24. Nielsen, J. 2000. Designing Web Usability. New Riders, Indianapolis, IN.
25. Bonneau, J., Miller, A., Clark, J., Narayanan, A., Kroll, J. A., and Felten, E. W. 2015. SoK: Research perspectives and challenges for Bitcoin and cryptocurrencies. 2015 IEEE Symposium on Security and Privacy, pp. 104-121.
26. Eskandari, S., Barrera, D., Stobert, E., and Clark, J. 2015. A first look at the usability of bitcoin key management. Workshop On Usable Security (USEC).
27. Davis, F. D. 1986. A technology acceptance model for empirically testing new end-user information systems: Theory and results. Doctoral Dissertation. Sloan School Of Management, MIT.
28. Evans-Pughe, C., Novikov, A., and Vitaliev, V. 2014. To bit or not to bit. Engineering and Technology Magazine, May 2014, pp 82-85.
29. Krombholz, K., Judmayer, A., Gusenbauer, M., and Weippl, E. 2015. The other side of the coin: User experiences with bitcoin security and privacy. SBA Research.
30. Gao, X., Clark, G. D., Lindqvist, J. 2016. Of two minds, multiple addresses, and One Ledger: characterizing opinions, knowledge, and perceptions of Bitcoin across users and non-users. Rutgers University. ACM CHI ’16.
31. Mihaylov, M., Jurado, S., Avellana, N., Moffaert, K. V., Abril, de I. M, and Nowé, A. 2014. NRGcoin: Virtual currency for trading of renewable energy in smart grids. 11th International Conference on the European Energy Market (EEM14), pp. 1-6.
32. Swan, M. 2015. Blockchain Thinking : The brain as a decentralized autonomous corporation [Commentary]. IEEE Technology and Society Magazine, vol. 34, no. 4, pp. 41-52 33. Reiss, E. 2012. Usable Usability: Simple Steps for Making
