Content uploaded by Raffaele Fabio Ciriello
Author content
All content in this area was uploaded by Raffaele Fabio Ciriello on Dec 02, 2019
Content may be subject to copyright.
Differential Effects of Cryptocurrency Incentives in Blockchain Social Networks
SIGBPS 2019 Workshop on Blockchain and Smart Contract, Munich 2019 1
The Differential Effects of Cryptocurrency
Incentives in Blockchain Social Networks
Rongen (Sophia) Zhang
Georgia State University
Atlanta, Georgia
rzhang6@gsu.edu
Junyoung Park
Georgia State University
Atlanta, Georgia
jpark136@gsu.edu
Raffaele Fabio Ciriello
IT University of Copenhagen
Copenhagen, Denmark
raci@itu.dk
Abstract
While blockchain technology entails transformative potential in many domains, little is
known about how the economic mechanisms and governance structures embedded in
blockchains affect user behavior. In this paper, we study how blockchain-based incentives
impact the user’s content creation and curation behavior on social networks. Collecting
and analyzing data from 113,209 postings on Steem, a major blockchain social network,
we examine how cryptocurrency incentives embedded in blockchain influence the
information quality of user-generated content. The preliminary results indicate that
different types of cryptocurrencies (namely liquid, vested, and stable cryptocurrency)
impact information quality in different ways. We discuss possible implications and ways
forward.
Introduction
Blockchain is a foundational technology enabling new modes of value creation for individuals,
organizations, and society (Beck et al. 2017). Through a distributed ledger, blockchain technology enables
transparency and irreversibility of records, intermediary-free transactions, pseudonymity, and
computational logic (Ilansiti and Lakhani 2017). While blockchain technology is still in its infancy, many
organizations have started to experiment with its potential in various industries.
One such application is the Blockchain Social Network (hereafter BSN), a blockchain-based social media
platform that provides economic mechanisms and governance structures for user participation and content
contribution (Ciriello et al. 2018). With growing skepticism about the security and privacy of user data on
social media platforms, BSN have gained attention as an alternative digital platform, without any single
central entity that can control user information. Another distinctive feature of BSN is that users can earn
cryptocurrency rewards by creating and ‘curating’ content via upvoting (similar to ‘liking’ on Facebook).
Such cryptocurrency incentive mechanisms enable the study of new phenomena regarding how an
economic mechanism embedded in a blockchain affects user behavior in social networks, which much to be
explored (Beck et al. 2018).
The openness of many traditional digital platforms invites broad participation with little or no evaluation,
allowing for free and democratized content creation, but also making it difficult to ensure a high level of
information quality of the content created on such platforms (Lukyanenko et al., 2014). Moreover, due to a
lack of monetary rewards for user-generated content (UGC) creation on many digital platforms, it is
challenging to design appropriate incentives that motivate qualified casual participants to contribute UGC
Differential Effects of Cryptocurrency Incentives in Blockchain Social Networks
SIGBPS 2019 Workshop on Blockchain and Smart Contract, Munich 2019 2
of high information quality (Lukyanenko et al. 2014). In this light, BSNs with cryptocurrency incentives
might be able to motivate users to contribute value.
The goal of this paper is to find out whether the cryptocurrency incentive mechanism implemented in BSN
effectively incentivizes users to create quality content. To this end, we examine the differential effects of
different types of cryptocurrency in BSN on user behavior. Initial exploratory studies have provided
preliminary insight into the need and possible ways to design BSNs with economic mechanisms and
governance structures that promote prosocial behavior (e.g. Beck et al. 2018; Ciriello et al. 2018). However,
with many blockchains in their infancy, empirical research on how these economic mechanisms affect user
behavior on a large scale is still scarce (Thelwall 2018).
Since one critical contribution in a social media context is the creation of high-quality content, we use
information quality as a focal construct for measuring the effects of BSN-provided incentives. Existing
studies have focused mostly on the behavior of Bitcoin or very few cryptocurrencies (ElBahrawy et al. 2017).
Although prior UGC literature has extensively investigated the motivation of user contribution behavior in
online communities, such as extrinsic and monetary incentives, little is still known about the differential
effects of cryptocurrencies with different characteristics on information quality of the content in
blockchain-based online community. Therefore, we raise and address the research question :
RQ: How do different types of cryptocurrency incentives provided by blockchain social networks
influence the information quality of users’ created content?
Hypothesis Development
Among the many BSNs launched in recent years, we focus on Steem (steem.io), one of the largest and most
widely known ones. Launched in 2016, Steem enables creation, curation, and consumption of multimedia
content via public and permissionless blockchain database with a built-in incentive mechanism and a
reputation system. The Steem blockchain offers a decentralized rewarding system that provides
cryptocurrency incentives for content creation and curation (Ciriello et al., 2018). It rewards user
participation and contribution based on community evaluations (Larimer et al., 2018). Being essentially a
multipurpose platform enabling many different decentralized applications (or dApps), the most widely used
social media application on the Steem blockchain is Steemit (steemit.com). Steemit’s functions for content
sharing and social networking are similar to those of Facebook or Reddit (Thewall 2018).
Cryptocurrency can be one type of monetary incentives which motivates users to create better quality
content in BSN. With limited time and resources at hand, a user’s contributions to the platform can be
affected by cryptocurrency rewards for two behaviors, namely content creation and curation. Such focus
can affect the quality of posting because creating a better quality posting requires more time and effort.
Therefore, we are interested in the effects of users’ previous posting rewards and curation rewards on the
information quality of posting in BSN. A user with more previous posting rewards will have more incentive
to invest the time and effort in content creation. On the contrary, previous curation reward will drive a user
to focus more on curation behavior, which will lower the quality of posting. Therefore, we hypothesize:
H1a: Users’ previous posting cryptocurrency reward is positively associated with the information quality
of UGC.
H1b: Users’ previous curation cryptocurrency reward is positively associated with the information
quality of UGC.
Moreover, the platform runs three native cryptocurrency tokens with different characteristics, which users
can acquire as a reward for contribution and by transactions.
1) STEEM: STEEM is the fundamental unit of account on Steem. The STEEM token is a liquid
cryptocurrency that users can transact as a form of payment or trade for other external tokens,
cryptocurrencies, and fiat money, such as US Dollars.
2) STEEM Power (SP): Users can convert their STEEM into SP by committing them to a vesting
schedule. SP grants users stronger voting rights. SP is non-transferable and can only be withdrawn by
converting the SP back to liquid STEEM in 13 equal weekly payments.
3) Steem Blockchain Dollar (SBD): SBD is a stable cryptocurrency pegged to the value of the US dollar
that can be viewed as a form of debt with a relatively static value.
Differential Effects of Cryptocurrency Incentives in Blockchain Social Networks
SIGBPS 2019 Workshop on Blockchain and Smart Contract, Munich 2019 3
Based on the characteristics of the abovementioned cryptocurrencies, we categorize STEEM, SP, and SBD
as Liquid Cryptocurrency, Stable Cryptocurrency, and Vested Cryptocurrency, respectively. We hypothesize
about the impact of a user’s balances of the three types of cryptocurrency on information quality of UGC.
First, as a user hold more Liquid Cryptocurrency, the user will hope and try to increase the value of the
cryptocurrency to gain more benefit. In BSN, one major approach to increase the value of the platform is to
contribute better quality postings to the platform. Therefore, we hypothesize that:
H2a: The Liquid Cryptocurrency balance is positively associated with the information quality of UGC.
Stable Cryptocurrency was introduced to mitigate the infamous price volatility issue. A user with Stable
Cryptocurrency can rely on its relatively stable value. Therefore, the user with more Stable Cryptocurrency
will be more motivated to create better quality posting to acquire more rewards. So, we hypothesize that:
H2b: The Stable Cryptocurrency balance is positively associated with the information quality of UGC.
Finally, Vested Cryptocurrency was designed with the purpose of stabilizing the platform value. Users with
Vested Cryptocurrency can be viewed as equity-holders of the platform so that they are more involved in
the platform as stakeholders. A user with more Vested Cryptocurrency is likely to be more motivated to
create better quality content to increase the value of the platform, so that the value of their equity will
increase. Therefore, we hypothesize that:
H2c: The Vested Cryptocurrency balance is positively associated with the information quality of UGC.
Method
We designed a Python-based web crawler using the blockchain explorer API provided at steemdb.com to
extract all postings in English on the Steem blockchain spanning three periods: 08 April - 28 April, 17 May
- 2 June, 4 June - 10 June (2019). The texts from the postings went through basic pre-processing steps
including the removals of unique characters, slang, emoticons, and numbers, and stop words.
Because this study examines the impact of cryptocurrency incentives on the objective quality of posted
information, rather than subjective evaluation of it, we attempt to capture the information quality from the
measures directly accessible from the contents (Agichtein et al. 2008). In doing so, we use multiple
linguistic and multimedia measures. As the linguistic measures, we extracted the number of unique words
(henceforth ‘length’), the Flesch Reading Ease Score (henceforth ‘readability’), and the Shannon entropy of
trigrams (henceforth ‘entropy’), which are used in online review literature as indicators of information
quality (e.g., Johnson et al. 2015; Singh et al. 2014). As for the multimedia measures, we extracted the
number of images, external links, and hashtags used in each post, which provide additional information
than plain texts (Janze and Risius 2017). Because the six measures use different units, the overall
information quality was constructed by standardizing each of the six measures and then calculating the
average of them.
Our explanatory variables include the author’s previous cryptocurrency reward and three types of
cryptocurrency balance. We used the values of an author’s cumulative rewards from their previous postings
and curation activities as the two types of previous cryptocurrency reward. Author’s balances of the three
types of cryptocurrency (i.e., liquid, stable, and vested) were measured as the amount of STEEM, SBD, and
SP, respectively, that the author holds at the time that the posting was created.
Finally, we control for the authors’ reputation score, number of followers, tenure in the platform (i.e., days
since registration), number of previous postings, the category of the post, and the creation date of the
posting. Among the 279,814 postings collected, we removed those with missing values and with readability
scores below 0 or above 100. Overall, our final dataset includes 113,029 postings from 12,334 authors.
The econometric model for the antecedents of posting information quality is:
where the subscript i represents posting.
Differential Effects of Cryptocurrency Incentives in Blockchain Social Networks
SIGBPS 2019 Workshop on Blockchain and Smart Contract, Munich 2019 4
Preliminary Results
Table 1 summarizes the results of the OLS regression for the antecedents of information quality. Model 1
includes only the control variables. We find that reputation and number of followers are positively
associated with information quality, whereas tenure shows negative association. Model 2 represents the full
model with the variables of interest. Both previous posting reward and previous curation reward are
negatively associated with information quality, rejecting both H1a and H1b. That is, receiving higher
rewards from both content creation and curation efforts might crowd out the user’s intrinsic motivation to
craft good-quality content.
Table 1. Results of the OLS Regression for the Antecedents of Information Quality
Variable
Model 1
(DV - IQ)
Model 2
(DV - IQ)
Model 3
(DV - Linguistic
IQ)
Model 4
(DV - Multimedia
IQ)
Intercept
-17.440**
(0.0000)
-12.105**
(2.579)
-21.073**
(5.009)
-4.0398**
(0.3327)
Previous posting
reward
-
-0.00003*
(0.0000)
-0.00007**
(0.0000)
0.00000
(0.0000)
Previous curation
reward
-
-0.00003
(0.0000)
-0.00006
(0.0001)
0.00000
(0.0000)
STEEM (liquid)
balance
-
-0.00027**
(0.0000)
-0.00052**
(0.0001)
-0.00003**
(0.0000)
SP (vested) balance
-
0.00003**
(0.0000)
0.00005**
(0.0000)
0.00000
(0.0000)
SBD (stable) balance
-
0.00032**
(0.0000)
0.00060**
(0.0001)
0.00004**
(0.0000)
Reputation
0.09874**
(0.0048)
0.10061**
(0.0048)
0.18863**
(0.0094)
0.01258**
(0.0006)
Number of followers
0.00042**
(0.0000)
0.00032**
(0.0000)
0.00057**
(0.0001)
0.00008**
(0.0000)
Tenure
-0.00120**
(0.0003)
-0.00106**
(0.0003)
-0.00224**
(0.0005)
0.00012**
(0.0000)
Number of previous
postings
0.00006**
(0.0000)
0.00006**
(0.0000)
0.00011**
(0.0000)
0.00002**
(0.0000)
Category dummy
YES
YES
YES
YES
Date dummy
YES
YES
YES
YES
R-squared
0..4977
0.4984
0.4981
0.46
Adjusted R-squared
0.467
0.4678
0.4674
0.427
n=113,209, *p<0.05, **p<0.005
Regarding the role of cryptocurrency balance, one major surprising finding is that SP (vested) and SBD
(stable) balances are significantly positively associated with information quality, but STEEM (liquid)
balance is negatively associated with information quality. Therefore, H2b and H2c are supported, while
H2a is not supported. This result suggests that different types of cryptocurrency play different roles as
incentives for creating better quality postings. It has been acknowledged that the volatility of a
Differential Effects of Cryptocurrency Incentives in Blockchain Social Networks
SIGBPS 2019 Workshop on Blockchain and Smart Contract, Munich 2019 5
cryptocurrency generates risk, which discourages consumers from holding the cryptocurrency for a long
time (Devries 2016). In that sense, users with more Liquid Cryptocurrency (i.e., STEEM) that is volatile
might be less motivated to contribute with higher quality postings.
As a post-hoc analysis, we present the results for the analysis where we run separate models for information
quality measured using only linguistic (Model 3) and multimedia measures (Model 4). The results for
linguistic information quality are consistent with the main result. The results for Model 4 shows that
STEEM (liquid) balance affect multimedia information quality negatively while SBD (stable) affects
multimedia information quality positively. Also, the coefficients for SP (vested) balance and previous
posting and curation rewards are insignificant. Overall, the results for Model 3 and Model 4 indicates that
linguistic information quality is the driving force of the overall information quality.
Intended Contributions
In further developing this work, this study will make the following theoretical contributions. First, it will
illustrate blockchain-enabled incentives provided in blockchain economies and their effects on user
behavior. While the existing blockchain discourse is still nascent and often focuses on the proof of concepts
and potential use cases, this study will provide empirical insights. Second, this study will categorize
different cryptocurrencies based on their characteristics of liquidity and volatility, which offers a novel
perspective to future research on blockchain and cryptocurrency. Third, this study will unpack the
differential effects of cryptocurrency incentives with different characteristics on information quality of
content in blockchain-based online communities.
Moreover, our study will inform the designers of BSN systems about the interrelated effects of different
incentives on user behavior. Practitioners and design-oriented researchers can build on our study and
engage in the design of effective incentive schemes for BSNs. For example, digital platforms that wish to
adopt cryptocurrency incentive mechanisms should be cautious about implementing rewards for curation
and creation to avoid the here-identified negative effect of curation rewards. Also, introducing stable and
vested cryptocurrency can be effective to facilitate the improvement of content quality in blockchain-based
online communities.
References
Agichtein, E., Castillo, C., Donato, D., Gionis, A., and Mishne, G. 2008. “Finding High-Quality Content in Social
Media,” in Proceedings of the 2008 International Conference on Web Search and Data Mining, Palo Alto, CA, pp.
183-194.
Beck, R., Avital, M., Rossi, M., and Thatcher, J. B. 2017. “Blockchain Technology in Business and Information
Systems Research,” Business & Information Systems Engineering (59:6), Springer Fachmedien Wiesbaden, pp.
381–384.
Beck, R., Müller-Bloch, C., King, J. L., Beck, R., Müller-Bloch, C., King, J. L., and Thatcher, J. 2018. “Governance in
the Blockchain Economy: A Framework and Research Agenda.,” Journal of the Association for Information
Systems (19: March), pp. 2–36.
Ciriello, R. F., Beck, R., and Thatcher, J. B., 2018. “The Paradoxical Effects of Blockchain Technology on Social
Networking Practices,” in Proceedings of the 39th International Conference on Information Systems, San
Francisco, CA, pp. 1–17.
ElBahrawy, A., Alessandretti, L., Kandler, A., Pastor-Satorras, R., and Baronchelli, A. 2017. “Evolutionary Dynamics
of the Cryptocurrency Market,” Royal Society Open Science (4:11), p. 170623.
Ilansiti, M., and Lakhani, K. R. 2017. “The Truth About Blockchain,” Harvard Business Review (February).
Janze, C., Risius, M., 2017. “Automatic Detection of Fake News on Social Media Platforms,” in Proceedings of the
Pacific Asia Conference on Information Systems, p. 261.
Johnson, S. L., Safadi, H., Faraj, S. 2015. “The Emergence of Online Community Leadership,” Information Systems
Research (26:1), pp. 165-187.
Larimer, D., Scott, N., Zavgorodnev, V., Johnson, B., Calfee, J., Vandeberg, M. 2018. “Steem: An incentivized,
blockchain-based social media platform,” Self-published, March. (https://steem.com/steem-whitepaper.pdf,
accessed Novermber 8, 2019).
Lukyanenko, R., Parsons, J., and Wiersma, Y. F. 2014. “The IQ of the Crowd: Understanding and Improving
Information Quality in Structured User-Generated Content,” Information Systems Research (25:4), pp. 669–689.
Singh, P. V., Sahoo, N., Mukhopadhyay, T. 2014. “How to Attract and Retain Readers in Enterprise Blogging?,”
Information Systems Research (25:1), pp. 35-52.
Thelwall, M., 2018. “Can Social News Websites Pay for Content and Curation? The SteemIt Cryptocurrency Model,”
Journal of Information Science (44:6), pp. 736–751.
