Social Media Use in Science

(1)

Social Media Use in Science

The brave new world of collaboration and altmetrics

by Fabian Kersten

Research Master Thesis in Media Studies University of Amsterdam

Supervisor: Prof. Dr. Richard Rogers

(2)

Abstract

The increasing popularity of social media across all society has also had an impact on science. Signals of change and transformation are already evident in research workflows and in the distribution and evaluation of science.

To create an account of change, four approaches will be presented in this study to explore the contemporary conditions surrounding the social media phenomenon in science. By examining its influences on the scientific reward system, scientific culture and its norms, collaboration practices and social capital theory, and then presenting empirical data on the scholarly use of social media, a case for social media’s values in science will be presented. This study argues that the use of social media meaningfully contributes to the researcher’s workflow by enhancing collaborations that in turn lead to greater productivity.

One of the agents of change is the debate around alternative metrics to measure scientific impact to enrich the landscape of conventional article-based metrics. This signals the maturation of social media adoption in science. Current assessments of altmetrics to measure scientific impact illustrate the increasing importance of social media to the scientific community. Initial positive correlations between social media and conventional measures have been proven and give scholars further impetus to foster social media, especially as the feedback mechanism is instant, increases their visibility among peers, and reaches a broader audience.

It is evident that as new modes of consumption, collaboration and measures of science emerge they bring with it a shift in scientific culture and its norms. Researcher autonomy is decreasing due to societal pressures and interactions, and the human component in the knowledge creation process is gaining more influence. Social media are accelerating this change by increasing scholarly participation. Drawing from social capital theory explains why scientists collaborate, as they expect gains in terms of productivity, funding and rewards. Social media cater to collaboration needs by not only taking away spatial and temporal limitations but also by allowing real-time interactions with dispersed networks of trusted peers that can lead to knowledge creation and scientific discovery. Building credibility and trust are essential elements in this process.

There is a positive perception of social media in science, and the increase in its usage reflects Science 2.0’s promise to increase research productivity. Current trends show the use of social media will grow and participation will evolve further around networking, sharing information and following others, which will require an online presence.

(3)

Primarily, scientists use mainstream social media platforms followed by social reference managers. The latter are unique in the scientific social media toolkit as they fulfill the need of filtering, organizing and prioritizing growing volumes of information. Even though scientists use social media for very practical reasons by pursuing some objectives (e.g. research collaboration, sharing and disseminating of information, watching the competition), there is no significant difference in the guidelines they should follow compared to its popular use etiquette. The proposed strategy for scientists is to participate in social media with the objective to link all activities to their personal brand and the image they want to create online. This mindset will allow them to take control and practice conscious sharing and make deliberate collaboration decisions. The effectiveness of their personal brand and social media engagement can be measured by alternative and conventional metrics.

(4)

Table of Figures

Figure 1: Keyword search in Google showing the popularity of the search topic...12

Figure 2: Impact measures through altmetrics (Taylor and Plume)...13

Figure 3: (2008) Professional occupation of the sample (N=1824)...59

Figure 4: (2008) Age of the sample (N=1824)...59

Figure 5: (2008) Top twelve countries of participation (N=1143)...60

Figure 6: (2008) Top ten fields of the samples’ specialty (N=1824)...60

Figure 7: (2008) Numbers of published articles (N=828)...61

Figure 8: (2008) Regular use of Blogs or Wikis (N=1579)...62

Figure 9: (2008) Ownership of blogs or wikis (N=1579)...62

Figure 10: (2008) Leisure or work related use of social media applications (N=898)...62

Figure 11: (2013) Proportion of ‘top 10’ countries* within base n=3,090...67

Figure 12: (2013) Age and number of articles published in their career n=3090...68

Figure 13: (2013) Gender distribution of the research sample (N: 3090)...68

Figure 14: (2013) Professional occupation of the sample (N=3090)...69

Figure 15: (2013) Top ten fields of the samples’ specialty (N=3090)...69

Figure 16: (2013) Percentages of survey participants using specific social media tools (N: 3090)...70

Figure 17: (2013) Most frequently used tools for professional purposes? (Users of social media)...71

Figure 18: (2013) Correlations between age and use of social media for professional purposes...72

Figure 19: (2013) Purpose for using the sites (Only users of social media) (n=2,169)...73

(7)

INTRODUCTION

To date, there has been an unclear view of the social media phenomenon and its influence on science and research. The definition of social media will be used in the broadest sense, comprising all applications that are based on user-generated content and that afford a “two-way interaction with an audience, beyond any specific recipient” (B. Hogan and Quan-Haase, 2010, p. 311).

Considering the increasing popularity of social media across all societal categories, it is highly likely that its impact on science will also increase over time, and more signals of change and transformation will be evident in research workflows and practices and in the distribution and evaluation of science. Drawing on theories and empirical studies to frame current dynamics, this thesis will add to the existing literature of social media studies by addressing the social media phenomenon in science. Whether social media use relates to scientific practices or the opportunities it creates for individual scholars, there is a need to examine social media influences from multiple dimensions to grasp the role that social media play and possibly will play in the future of science.

To create an account of change, four approaches will be presented in this study to explore the contemporary conditions surrounding the social media phenomenon in science. By examining its influences on the scientific reward system, scientific culture and its norms, collaboration practices and social capital theory and then presenting empirical data on the scholarly use of social media, a case for social media’s values in science will be presented. In a professional environment, how does social media use meaningfully contribute to research workflow by enhancing productivity?

The first approach looks at the scientific reward system from a social media angle, which provides new impulses for measuring scientific impact. This chapter will examine the existing literature on social media-based alternative metrics to answer the question of how social media influence the conventional scientific reward system, which relies on article citation measures and the journal impact factor. What is the potential impact of these new types of measurements (e.g., usage, downloads, followers and commentary) in changing the way scientific research gets rewarded? Alternative metrics are enriching the landscape of conventional article-based metrics and signal a maturing age of social media adoption in science. Although there is a growing body of literature examining various angles of the altmetrics phenomenon in relation to established metrics, this research is still in its infancy. Current assessments of altmetrics for scientific impact measurements will show social

(8)

media’s increasing importance for the scientific community. Initial positive correlations between social media and conventional measures have been proven and encourage further investigations into how altmetrics can add meaning by covering more research assets than just articles. These metrics also give scholars additional motivations to participate in social media as they provide an instant feedback mechanism, increase their visibility among their peers, and reach a broader audience.

Second, to answer the question about what is actually changing in modern science compared with the pre-social media era, context will be provided around scientific culture, as seen through the lenses of science studies scholars. Robert Merton’s institutional norms of science, as represented through the acronym ‘‘CUDOS’’ (i.e., communism, universalism, disinterestedness and organized skepticism), conflict with social media influences. Kuhn’s, Latour’s and Popper’s approaches to science take a more human side by considering knowledge as a social construct that is built on the researcher’s sociological and psychological background. The acceptance of a researcher’s work is thus highly dependent on his or her personal status within the scientific community. This approach is more in line with the role that social media fulfill as interpersonal connectors that build meaningful relationships and an online presence that can be influential within networks. The shift from offline to online scientific practices also provides more power to social media that foster scientific sharing, transparency and accountability in relation to society. Research autonomy is decreasing due to societal pressures and interactions, and the human component in the knowledge creation process is gaining more influence. Social media are accelerating the change by increasing scholarly participation.

The third perspective on social media in science addresses collaboration as a pivotal aspect in the research workflow, comprising knowledge identification, knowledge creation, knowledge quality assurance and knowledge dissemination. Based on a multiplicity of networks, the scientific culture is a creation of many specialized communities, and collaboration is vital and encouraged to push science forward. This chapter will answer the question about what drives scientists to collaborate and how social media impact collaboration practices. Drawing from social capital theory can explain why scientists collaborate, as they expect gains in terms of productivity, funding and rewards. Social media cater to collaboration needs by not only taking away spatial and temporal limitations but also by allowing real-time interactions with dispersed networks of trusted peers that can lead to knowledge creation.

(9)

Chapters four and five cover an empirical approach to social media in science and answer the questions about what defines social media in science through the concept of Science 2.0 and how social media are used in science. Science 2.0 promotes scientific collaboration based on social media and Web 2.0 technologies. Services such as social reference tools and science-dedicated social media networks are especially unique in terms of the way social media apply in the context of science. The network effect, the long tale and the wisdom of the crowd are all concepts that fit well with scientists’ needs to increase efficiencies. By scoping the dimensions of Science 2.0, some practical implications of social media will be highlighted. Chapter five will draw on the uses and gratifications theory to add additional insights into the motivations of social media use and will examine two survey studies from 2008 and 2013 on how social media are used in science. The empirical research shows that the popularity of social media is increasing. Scientists are using academic and non-academic social media services to achieve their goals. The results fully confirm their aspirations to profile themselves, network, follow other researchers and share information as means of collaboration. Differences between countries, subject disciplines and age groups need to be considered. In general, there is a positive perception of social media in science, and the increase in usage reflects Science 2.0’s promise to increase research productivity.

As current trends show, social media use and Science 2.0 are predicted to grow. The final chapter will consider the outcomes of the four routes to explore the social media phenomenon in science and to provide guidelines for scientists on how to effectively engage with social media to ensure that they meet their objectives. Comparing guidelines for the popular use of social media with the way scientists should approach it does not yield any significant differences. Standard user etiquette (e.g., being authentic, telling the truth, being active and respecting copyrights) are applicable for all social media users. However, there are some nuances when looking at specific tools (e.g., social referencing or data sharing applications) that need to be addressed in the context of research workflow. Additional insights into scientific impact measurements through altmetrics might also influence usage recommendations in the future. The provided user guidelines help to effectively manage social media engagement and present opportunities for getting recognized and rewarded. The following elements are key when using social media in science: clarity about the user’s objective, understanding what services peers are using and the social media service’s suitability in terms of available time and enjoyment, as being active is crucial for success.

(10)

Personal branding will be introduced as a strategy to regain or take control of the uncertainty that defines social media practices in the day to day research life. By means of online self-differentiation researchers can gain visibility and determine their destiny.

(11)

CHAPTER 1 – Science Rewards and Altmetrics

Social media services are gaining popularity in science communication, and more social media services that target academia in particular are surfacing online. With social media, new and different modes of tracking and analyzing online behavior are enabled based on user-generated content. Especially in science, there is a constant quest to find a broader scope of measurements that can identify the impact and contribution of research in a specific field. Social media might provide new perspectives through for example views, followers or commentary. Advocates of these new measures of scientific impact, coined as altmetrics, welcome these developments because they broaden the scope of conventional impact measures based on bibliographic analysis. This chapter will explore how maturing social media provides alternative metrics for science impact analysis and influences the conventional scientific reward system based on citation analysis and the journal impact factor. Conventional measures have been accused of ignoring the contextual dynamic of scientific impact.

In the “Altmetrics Manifesto,” Priem, Taraborelli, Groth and Neylon (2010) lay the groundwork for widening the scope of quantifiable metrics to measure scholarly impact. The manifesto addresses the dynamics of a growing number of research outputs that are being disseminated online. New web services are capturing, reflecting and transmitting scholarly impact and are more likely to add breadth to the conventional forms of impact indicators (e.g., article citation counts, the journal impact factor and the h-index). The general explosion of research publications and information contributes to the complexity of navigating and selecting between relevant and less relevant material. With the introduction and use of social media tools, new possibilities for finding, filtering and evaluating via alternative metrics might also prove valuable. Alternative metrics, also referred to as ‘altmetrics,’1 _{are the}

creation and study of new metrics based on the social Web for analyzing and informing scholarship (Priem et al., 2010). Altmetrics is a sub-discipline of scientometrics and typically looks at individual research outputs, including journal articles or datasets and other formats that go beyond the conventional article format. In their article “Scientometrics 2.0: New metrics of scholarly impact on the social Web,” Priem and Hemminger (2010) discussed the development of new ways of measuring scientific impact and signaled a new era of scientometrics. Figure 1 shows the active and increasing interest in the field of research.

1 In the context of this paper, altmetrics will be used as a general term, which is not to be confused with the organization called Altmetrics.org. Altmetrics is also used interchangeably with alternative metrics.

(12)

Figure 1: Keyword search in Google showing the popularity of the search topic.

More evidence of the serious nature of altmetrics is given through the adoption by publishers (e.g., Nature, Elsevier, and the Public Library of Science (PLoS)), who started to integrate and present altmetrics in their content databases. Traditional and established formats of measuring the impact of scientific output at the article or journal level are based on article citation counts, which take time to accumulate due to lengthy publication processes. Peer-reviewed scientific articles are a slow communication medium. They are less useful as indicators of a recent work’s importance. Therefore, trends show publishers’ increased interest in considering altmetrics that cover scientific impact on the social Web, which appears more rapidly than journal article citations (Priem and Hemminger, 2010; Thelwall, Haustein, Larivière, and Sugimoto, 2013). In the context of altmetrics, Buschman and Michalek (2013) define five impact categories for scholarly research: usage (e.g., downloads, views), captures (e.g., bookmarks, readers, favorites), mentions (e.g., blog posts wikis, reviews), social media (e.g., tweets, likes, shares) and citations. Figure 2 presents Taylor and Plume’s (2014) categorization of altmetrics data sources into the following four types: social activities (e.g., Twitter, Facebook), scholarly activities (e.g., bookmarking and reference tools), scholarly commentary (e.g., blogs, F1000) and mass media coverage.

(13)

Figure 2: Impact measures through altmetrics (Taylor and Plume)

Priem, Groth, and Taraborelli (2012) narrowed down and discussed a range of social Web tools for alternative measures:

- social media (e.g., Twitter and Facebook)

- online reference managers (e.g., CiteULike, Zotero, and Mendeley) - collaborative encyclopedias (e.g., Wikipedia)

- blogs, both scholarly and for a general audience

- scholarly social networks (e.g., ResearchGate or Academia.edu) - conference organization sites (e.g., Lanyrd.com)

Altmetrics is not a new trend of finding richer indicators to measure scholarly impact. However, the accessibility and usability of data on the social Web has seen major advances since the introduction of Web 2.0 technologies, such as Application Programming Interfaces (APIs), which support the communication of various platforms’ data sources.

The following paragraphs will cover the current research trends in the development of new rich metrics that offer an alternative perspective to conventional citation counts and the journal impact factor.

1. Key areas of altmetrics research

The research literature shows that there is a need to validate these new metrics and their significance by correlating altmetrics with traditional measures (e.g., citation counts). Studies especially look at the use of social media for scholarly practices to understand the way altmetrics can add value. There are various areas of interest that fuel the altmetrics debate, which can be found in the following topics of ongoing research, as mentioned by Priem et al. in ‘The Altmetrics Collection’ (2012):

(14)

 Statistical analysis of altmetrics data sources and comparisons to established sources

o Used as early indicators to predict future citation trends

 Metric validation and identification of biases in measurements (e.g., gaming the metrics)

 Validation of models of scientific discovery/recommendation based on altmetrics (e.g., filters of information and the multiplicity of data formats at stake)

 Qualitative research that describes the scholarly use of online tools and environments

 Empirically supported theory that guides altmetrics usage

 Other research relating to scholarly impact in online tools and environments

1.1 Indicator for citation impact

Alternative metrics need to prove their added value to justify their existence and gain acceptance in the scientific community. Therefore, a series of studies investigates the insights that altmetrics provide that are beneficial for users. Some outcomes suggest that articles with higher altmetrics scores tend to produce higher citation counts in the future, and altmetrics thus could be of value in predicting future citation impact. In addition, this finding also provides evidence for a correlation between altmetrics and citations.

Eysenbach (2012) demonstrated that tweeted research articles were more likely to be highly cited than articles without social media coverage.2_{A similar study was conducted on blog} content and produced similar results (Shema, Bar-Ilan, and Thelwall, 2014).

Another approach examined the value of altmetrics by comparing the present correlation with citation counts. For example, one study looked at Mendeley readership and found a slight correlation with the number of citations received and the number of times and articles that have been read (Bar-Ilan et al., 2012).

2 Twitter mentions were positively correlated with rapid article downloads, whereas citations only appeared a long time after publication. It has not yet been determined if tweets have a long-term impact on citations or if they just highlight high-quality science that could have performed well without social media.

(15)

By examining multiple social Web services (e.g., Twitter, Facebook, blogs and forums), Thelwall et al. (2013) were able to prove the statistical significance of correlating higher altmetrics scores with citation counts. It is highly likely that there is value in altmetrics as an indicator, but to compare likes for likes, it is critical to consider timing as a factor.

Studies show that some altmetrics have value as impact indicators, but larger scale research is needed to further solidify the evidence.

1.2. Discover work that is not covered by traditional metrics.

Traditional means of measuring impact continue to focus on article citations and thus have shortcomings when other forms of research output (e.g., datasets, conference slides, blog entries and software) are considered. Measuring the online impact of scholarly publications via social media is currently also lacking. Therefore, scholars who publish online and in formats that differ from journal articles do not benefit from citation-based data metrics (Zahedi, Costas, and Wouters, 2014). Considering social media’s influence on scientific writing, a more flexible measurement system is needed to tell a more complete story. Altmetrics cover more formats than research articles and provide a richer picture by providing quantitative measures through, e.g., numbers of readers or followers, which indicates the level of awareness with scholarly work. A substantial part of altmetrics research focuses on finding patterns in the different social network services and correlations between altmetrics and conventional citation counts.

1.3. Social Media Use

The amount of information published on a daily basis is growing drastically, and the time available for research is getting scarcer. The research literature refers to an information overload that scientists, among others, have to face (Cann, Dimitriou, and Hooley, 2011; Cruz and Jamias, 2014; Gehl, 2010; Gruzd and Goertzen, 2013; Gruzd, Staves, and Wilk, 2012; Kietzmann, Silvestre, McCarthy, and Pitt, 2012; Quan-Haase, 2012). Research is all about mastering information and applying skills and techniques to identify relevant pieces of knowledge. Shirkey (2008) argues that this process sometimes entails filter failure, but if the right filters are applied, it is possible to digest the right information in less time. Approximately 90% of all of the data in the world have been generated over the past two years, and scientific data output increases at an annual rate of 30 percent (European Commission, 2014, p. 3).

The more indicators and filters that are available, the more efficient the work of a scientist might become. It is important to highlight that altmetrics are seen as complimentary sources

(16)

for analyzing impact and are not meant as substitutes for citation measures. The combination of multiple perspectives will provide the richest and most comprehensive picture.

Counting the article readers that are made visible in Mendeley, a social reference manager that stores and organizes academic references that interest researchers, could show similar insights to citations at an earlier stage. The reachable audience can also be much broader than scientific authors, which cannot be captured through citation counts. For example, the majority of Mendeley readers are early career researchers with fewer cited publications but high readership engagement (Mohammadi, Thelwall, Haustein, and Larivière, 2014).

Another tool that is being thoroughly investigated for altmetrics is microblogging, which has been introduced by Twitter to an even broader user spectrum. Scientists use the tool to cite, though not always in a direct way, which complicates the accurate tracking of social mentions. In addition, there is also criticism about the depth of scientific engagement with the referenced research, as the medium does not allow for extended critical discourse, though it can initiate one. The practice of utilizing the medium does not enforce professional referencing of the information source (Priem and Costello, 2010; Thelwall, Tsou, Weingart, Holmberg, and Haustein, 2013). However, according to Bik and Goldstein (2013a), online visibility helps track and improve scientific metrics.

Priem et al. (2012) additionally demonstrated that journal articles are circulated and utilized in multiple ways within the scientific community. Together, social media and article-level metrics have the potential to highlight the importance of research that cannot solely be revealed by traditional metrics. Some articles experience rather poor citation behavior but are evidently read and downloaded many times, according to social networking platforms.

1.4. Impact indicator for researchers

Another growing area of interest lies in the application of altmetrics when considering the evaluation of a scientist’s performance in the field of research. Having a social network profile and being able to connect and collaborate requires an effort to keep up to date with the online community. Building a network and getting noticed online is time consuming and should thus be encouraged and rewarded.

In “Beyond Citations: Scholars’ Visibility on the Social Web,” Bar-Ilan et al. (2012) focused on the “Web footprint”, publication and citation counts of researchers who presented at the 2010 STI conference. As such, this paper offers an example of how altmetrics research can also offer insight into academic events and their participants.Thelwall (2014) raises concerns

(17)

they should be seen as complimentary measures. Due to the lack of quality control and lack of governance in regard to social Web presence, there is a high risk of manipulation that demands caution when considering altmetrics in relation to individuals. Nevertheless, institutions show a growing interest in monitoring and possibly rewarding scientist’s social presence. Particularly in scholarly education that is facilitated through massive open online courses (MOOC), social media are pivotal for the reach and success of these initiatives and present new communication channels for participating academic institutions.

1.5. Speed of measuring impact

Traditionally, scientific impact is measured through citation analysis, which relies on a purely author-centric approach. An author cites another author’s research and thus creates an intimate link between two publications, which creates the perception of one person reading, digesting and canalizing the information in a professional manner. Speed is one of the key issues surrounding this single view on measuring research impact. It can take a minimum of two years before measurable impact can be noted, which leaves a long gap of uncertainty for the originator of the published research. Quick and valuable feedback is highly regarded and needed for any career and funding aspirations. An academic paper’s impact determines the researcher’s footprint in the scientific community and exposes him or her to increased opportunities for collaboration or funding.

Hence, it is a key concern for scientific authors to be able to understand how important and influential their research is or will be in a timely manner. As presented above, alternative metrics have the potential to fill this time gap and offer ways to get immediate indication, as shown in research conducted through various projects (Cave, 2012; Eysenbach, 2012; Thelwall, Haustein, et al., 2013).

2. Challenges with Altmetrics

Even though there are plenty of studies suggesting that altmetrics can add value to the scientific debate, there are also a number of skeptical views on the new dynamics of measuring science. Some researchers are concerned about gaming the metrics, as social media services are decentralized and uncontrolled and can be influenced because of the lack of regulation and standardization. The social Web and its services are constantly evolving, and altmetrics thus need to adapt and reflect these changes. In comparison to traditional metrics, there is far less consistency in altmetrics to build a solid foundation. The robustness of altmetrics is one of the primary concerns in the current research (Thelwall, Haustein, et al., 2013).

(18)

A researcher with a strong social presence can potentially draw more attention and coverage and thereby influence scores. Merton (1968) described similar dynamics in his famous article on the Matthew effect in science, whereby scientists with greater initial recognition tend to receive additional recognition more than scientists with less initial recognition. Gaming the metrics is also known in traditional citation and impact factor metrics and will always remain a weak spot in the art of measuring science (Priem and Hemminger, 2010). Furthermore, the majority of new metrics and traditional metrics are reporting on how research output is being used but still lack insights into the actual quality of the research.

In addition to current altmetrics, studies focus mainly on highly visible journal publications and ignore lower visibility ones, which might make a difference in the way the altmetrics research agenda is being shaped. Additional research is required to examine the differences between various subject areas and their media use to understand the value altmetrics can add to the respective research community.

As mentioned above, using altmetrics as more than complimentary measures when considering scholars is a concern (Thelwall, 2014). A lack of control on the social Web and ample means of manipulation call for caution when considering altmetrics, as they lack authority and credibility and are still a bit premature as performance measures (Cheung, 2013).

3. Summary

To summarize, altmetrics provide a new modus operandi for evaluating research: “Alternative metrics could help authors tap into the key networks and social influencers—tweeters, bloggers—that might help to maximize the immediate impact of their work—and demonstrate to funders that their research has impact within the public domain. […] For editors, altmetrics could help determine which authors, topics and article types might achieve the most impact and enable them to proactively solicit papers that are likely to generate the most social ‘buzz’. […] [S]ocial visibility is not necessarily an indicator of scholarly credibility” (Warne, 2014). There is no clear view on the type of articles that generate more social “buzz,” and it really depends on the topics that interest a wider audience, i.e., popularity aspects. Additional research and data analysis on altmetrics in correlation with traditional impact measures could lead to an understanding of patterns and further conclusions about the type of articles that succeed at both ends.

It is still too early to say exactly what does and does not work. Given the ongoing debates, there is a lot more research ahead to validate some of the potential impacts that social media

(19)

metrics might have on measuring science. Altmetrics need to build more credibility in order to be fully considered in the scientific reward structure. Nevertheless, the increasing amount of literature on the topic suggests that altmetrics are taken seriously and have the potential to continue and expand in terms of influence. The shape and form through which altmetrics will evolve and be applied over time is still subject to change. The methods and data sources used to validate and attribute credibility to the metrics need to be widened in order to achieve greater acceptance in the scientific community. As suggested in studies by Priem and Piwowar (2012), altmetrics on an aggregated level point toward a direction whereby they add value, especially in areas where the impact can be measured through non-traditional means. Overall, despite all outstanding questions, the nature of enriching the metrics landscape by adding more color and nuances to research evaluation and embracing the dynamics of the networked scientific society can be seen as a positive trend. Being able to demonstrate that there are traces of correlation between new and old measures strengthens the argument for further investigation. Future research also needs to explore commonalities and differences in various subject areas and prove the value of different types of altmetrics. The way in which most articles get mentioned on the social Web needs to enable easier identification for altmetrics use in the future, e.g., harnessing altmetrics to gain insight into the extent to which articles from a particular field attract readers from other fields. Earlier studies show that the traditional ways of measuring scientific impact have most benefited the small portion of researchers who are already established and well-connected among their peers. Social media open new possibilities of connectivity and distribution, as new services are introduced into science on a regular basis to address these areas of improvement. As altmetrics are to a large extent fueled by social media services, one could argue that the success or failure of the new modus operandi depends on the degree of the scientist’s participation in social media. In other words, for altmetrics to mature and survive as valuable measures and true reflections of scientific engagement, a critical mass of active social media participants across all disciplines is required. What does that mean for the way science rewards its contributors? How do social media influence normal scientific practices if altmetrics become a more serious player in the art of measuring scientific impact? Altmetrics can be seen as the next level of maturity in science’s social media adoption process, but they still have to find a place in the scientific reward system. The next chapter connects the altmetrics trends by considering scientific practices from a social media perspective, with the norms of scientific culture and “good” scientific practices as postulated by Merton and Kuhn, the founding fathers of science studies. Chapter 2 explores how social media impact scientific culture and its norms.

(20)

(21)

CHAPTER 2 - Scientific Culture, Norms and Social Media

The previous chapter on altmetrics examined the trends around recognizing and rewarding scholarly work. This chapter addresses the second approach to studying the social media phenomenon in science by examining what defines “good” science in scientific culture. This chapter will draw from the ideas of theorists, such as Merton (1973), Kuhn (1962), Latour (1979) and Popper (2012), to better understand “normal” science practices in relation to current changes in modern science that are influenced by social media.

There is an undeniable impact of external forces on scientific culture, which has to do with questions around academic autonomy, economics and funding, society and social media. In the pursuit of knowledge creation and finding the truth, scientists are better connected with society and its realities than they were in the 17th century, when scientific elite clubs were self-sufficient and driven by an intrinsic motivation and a strong curiosity to understand the world. With the institutionalization of science, the previous concept of the scholarly ivory tower is no longer sustainable, as it is confronted with societal pressures and changing economic dependencies. Modern scientists operate in a marketplace with many actors (e.g., policymakers and sponsors), who influence the former autonomy in research and question the status of scientists as independent “truth tellers”. In other words, scientists might not tell the entire truth due to market demands. Nevertheless, science is likely to remain the best system for producing knowledge that is reliable and rational. A closer look at scientific culture reveals that Kuhn and Merton are considered the founding fathers of science studies (Rudolph, 2014), which treat science as an institution and practice to be interpreted within its social context, according to Restivo and Croissant (2011). Robert K. Merton (1973) gave birth to science studies with his fundamental publication on the social and cultural norms of science. Thomas Kuhn later developed the agenda of science studies in “The Structure of Scientific Revolution,” which examined the account of scientific change over time and the role of science in society. Modern society is drenched in science and technology, which drive academic progress across all disciplines. Evidently, the socio-technological impact on science evoked a gradual shift from Science 1.0, which was entirely autonomous and science-driven, to Science 2.0, when science entered into a dialogue with social stakeholders (Miedema, 2012, p. 25). For Merton, scientific autonomy is essential for good scientific practices. Science is currently moving further away from being autonomous. All interactions that science encounters through social media, along with a multiplicity of stakeholder’s influences, put pressure on the scientific community to perform according to market needs. Compared with the early days of science, there are hardly any remains of research autonomy.

(22)

Merton also studied reward and recognition systems as drivers for good scientific practices. They are essential in the networked nature of science, which is described in Siler’s (2012) article about citation choices. Citations are the foundation of invisible networks of formal and informal communication between leading, but often geographically distant, scholars. Within these virtual communities, Merton (1968) talks about the reward structures in science that are responsible for shaping scholars’ behaviors and ideas. Through citation choices, scholars make decisions about how their work will be situated in their field of research. According to Merton, recognition of scientific significance has to come from within the scientific community, and peer recognition is thus the only true value. Merton looked primarily at science as an institutional construct and stand-alone societal activity with no direct external influence on content. It is only up to peers in their respective field of study to evaluate and reward the quality of research and the potential scientific significance of good scientific works. Quality is the normal measure for evaluating science, which leads researchers profile themselves among others by receiving frequent citations in high-quality publications in very reputable journals. To ensure that scientists maintain a high level of quality and credibility, Merton defined a series of normative institutional values as ground rules for scientists to uphold. These idealized norms should not be skewed by non-academic conditions, as they would interfere with the freedom of scientific practices.

1. Merton’s scientific “Norms” in relation to social media

According to Merton, science’s responsibility to produce knowledge is achieved through the orientation of established institutional norms that define appropriate and inappropriate values, beliefs, attitudes and behaviors (N. M. Hogan and Sweeney, 2013).

In “The Social Structure of Science”, Merton (1973) formulated four sociological principles that drive scientists summarized in the acronym ‘‘CUDOS’’ (i.e., communism, universalism, disinterestedness and organized skepticism).

Communism describes the aspiration of making results accessible by sharing them with the community for the common good and scientific progress. He states that “the substantive findings of science are a product of social collaboration and are assigned to the community” (1973, p. 274). Universalism promotes the norm of scientific results that are analyzed objectively, as verifiable or repeatable and not subject to any personal or social attributes. Disinterestedness is Merton’s third norm, which requires scientists to stay away from any emotional or financial attachments to their work. The main incentive should be the reward

(23)

through citations that the community provides to recognize scientific achievement. The scientist’s claim to ‘his’ intellectual ‘property’ is limited to that of recognition and esteem (1973, p. 275). Lastly, organized skepticism requires scientists to delay his final conclusion until all of the facts are considered; only then should the theory be postulated.

These norms come with a series of difficulties, especially when the many ways that social media get integrated in the research lifecycle are considered. How resilient are these “norms” in relation to the increasing influence of society’s sociological and economic pressures, which are also articulated through social media? While elements of Merton’s norms are relevant and meaningful for keeping science as an independent producer of knowledge, they are severely challenged in modern science practices.

Vladimir Teif (2014) adds to Mertonian principles, claiming that scientists are not just intrinsically motivated but are also governed by economic principles. These principles dictate that scientists are rewarded for high productivity, competitiveness and success. They must publish as much and as often as possible and promote their work through multiple media channels to attract citations, funding and recognition. Teif (2014, p. 310) describes this process as a “job-centric” focus. Latour and Woolgar (1979) realized that great scientists were effectively selling scientific ideas by persuading their peers and public. They support the theory that the creation of scientific facts is the outcome of a social construction process. Knowledge creation is a negotiation process, whereby that credit that has accumulated from previous successes and technical feats, along with effective public relations, will contribute to the efficient selling of results. Today’s social media have the potential to support these processes throughout the entire researcher workflow, as will be demonstrated in Chapter 3 on collaboration.

Hogan and Sweeney (2013) analyzed the Mertonian norms through the lens of social media and recognized that the majority of these norms are not sufficiently compatible when applied to modern ways of scientific communication. First, Merton’s “norms” distinguish science from any other occupation within society, which interferes with the way social media are used as personal and professional media among scientists. Looking at the professional use of social media, there is growing evidence that engaging in social media is becoming mandatory for scientists to communicate their research effectively (Van Eperen and Marincola, 2011). In addition, the use of social media tools is becoming more diverse and serves multiple purposes (e.g., tracking disease activity or starting collaborations) (Maxmen, 2010). In addition, social media enable immediate dissemination, discussion and promotion of scientific output at all

(24)

research stages, which supports Merton’s concept of communism but interferes with the objective skepticism norm.

Scientific knowledge should be communally owned, as “good” science is communicated, shared and visible—something that is easily facilitated through social media tools (e.g., blogs). Furthermore, good science is seen as being peer reviewed by experts in the field, but social media are opening up the conventional way of publishing and also enable sharing with non-scientific audiences at all stages of the research process. Peers’ stamps of approval and the peer review process can be interrupted through the participation of non-scholarly audiences and the premature communication of research findings. To support Merton’s idea of good science, scientific communication should protect the merits of clarity, transparency and quality and reduce the risk of misinterpretation and misinformation when using the communal sharing power of the social Web.

Merton’s understanding of universalism is a conflicting norm through the eyes of social media, which put people and their online presence at the center. In light of Merton’s (1973) universalism claim, any coloring of the person’s identity should be avoided, and the actual science should lead rather than the personality. However, social networking sites encourage the creation of an online identity and require a personal profile to be set up and used as the connecting ID for various networks. In addition, people are more likely to connect with peers with similar profiles and shared interests. As the anonymous evaluations of experts in the field are the basis of peer review and traditional scientific practices, social networking sites and their identifiable profiles challenge these practices. In addition, post-publication peer review is open to audiences with multiple levels of expertise and varying opinions.

Merton’s notion of disinterestedness, which demands that scientists disengage from their work, is also questionable in an era when millions of articles are published each year and the publish-or-perish regiment forces authors to communicate and promote their work continuously to get noticed. Passionate and emotional storytelling is a successful element in social media outlets, and the audience is more likely to pick up on this storytelling, which leads to rewards in the form of altmetrics that potentially lead to article citations. Merton tended to neglect the social human component when postulating “disinterestedness” or value the free nature of science as a principle. In addition, scarce time and competition in research creates another complexity that challenges the ideal of exhaustive fact gathering, as organized skepticism suggests. Claiming the research space through early communication is likely to contradict Merton’s view. Furthermore, organized skepticism suggests that science should be

(25)

subject to community-wide tests and challenges. Scientists share widely before peer review, discussing early results prior to publication in networks across a multiplicity of media channels. Science is moving beyond the closed traditional examination process. The immediacy of feedback through online debate sets up new trends for how science is communicated and evaluated among peers. The audience is unpredictable, and influential commentary can be motivated by science’s entertainment value versus its scientific qualities. The entertainment factor of scholarly findings is increasingly important and finds more consideration in the research workflow, as social media are highly receptive and rewarding. Jia You (2014) gathered the top 50 scientists on Twitter with the most followers, who also use social media to gain visibility among non-scholars.

Overall, science is becoming more accessible through social technologies that partially enable and disable the ‘norms’, as proposed in CUDOS. Furthermore, external economic and social influences in modern science tend to dissolve scientific autonomy and influence the scientific agenda (Miedema, 2012). The role of academics as independent societal truth tellers is changing. Although it is science’s duty to generate knowledge that is interesting, works and proves to have an economic force, it becomes more difficult to safeguard the requirements for effective scientific production of new knowledge, as postulated by Merton. The scientific community should decide if Mertonian norms are “descriptive of aspirational behavior or prescriptive of a required academic standard” (N. M. Hogan and Sweeney, 2013, p. 646) to support “good” science practices.

Some science studies scholars, such as Thomas Kuhn and Bruno Latour, had a much stronger focus on the human component in the knowledge creation process and the interaction between science and society. As alluded in the assessment of Merton’s norms, the social aspect of research is growing. Researchers’ sociological and psychological backgrounds, focus on their jobs and drive for success are deeply ingrained in the way they conduct research to gain influence and recognition. Some of the human components in the process relate to making science available, negotiating, presenting, persuading, networking and playing politics. Researchers need to acquire all of these skills through social interaction to build a network of trusted peers, identify reliable research and make a career.

(26)

2. Kuhn’s concept of “Normal” Science

In “The structure of scientific revolution,” Thomas Kuhn (1962) pushed for transforming Merton’s view into a reality. He formulated the concept of “normal science,” based on collective thoughts and its orientation alongside the theoretical paradigms in which scientists operate. This concept was the result of examining the process of discovery and describing the emergence of facts while considering the context of discovery. According to Kuhn, science is happening within a broad framework with established theories that are not questioned or challenged. Research thinks within paradigms that are built on a scientist’s previous research experience and theories; by adhering to these paradigms, research also limits itself with a narrow view of reality by observing and naming facts.

Kuhn (1962) states that the majority of science conducted by “normal” scientists is articulated within existing paradigms that endure quite a few challenges over time, during which scientists resolve conflicts with experiments by adjusting the theory or data. Therefore, objective knowledge cannot be achieved due to the psychological and sociological aspects of being human. Evidently, sciences go through various phases of consensus and disruption via normal science in the event of anomalies. Eventually, a point is reached when there are simply too many phenomena that the theory cannot explain, and through the recognition of the scientific community, a shift to a new paradigm occurs. However, Kuhn was able to demonstrate that observations and new findings that opposed old theories did not necessarily translate into their rejection. If something does not fit within an existing framework, peers will usually ignore it. The researcher’s status, recognition and credibility within the scientific world generally determines whether new ideas will be accepted, as a paradigm shift does not happen automatically (Miedema, 2012, p. 30). The phenomenon of falling in love with a theory and personal status challenges Merton’s norm of disinterestedness.

In addition, Karl Popper argues that accepted scientific results are merely a social construct that relies on trusting a huge number of assumptions that colleagues have previously made (Miedema, 2012, p. 66). That trust is derived from the researcher’s knowledge and expertise, as acquired over many years of practice and rational considerations. Therefore, new knowledge can only be “truly” achieved if it is accepted as such by peers. Building one’s trusted network through social media presents an ideal opportunity to broaden the scope of influence and power within the community, as authority plays a key role in the dynamics of development and acceptance of a theory. Postmodern cultural constructivism argues that all knowledge is powerfully influenced by both social and personal circumstances, and various

(27)

insights into the world are all subject to change (Miedema, 2012, p. 64). Hence, scientists will continue to utilize their trusted personal networks as the basic means of shifting and using new knowledge to make an impact.

3. Summary

The increasing use of social media tools for networking and collaboration is a condition that influences research practices in modern science. According to Miedema (2012, p. 111), “As an interest group, scientists are prepared to lobby for government funds and to compete with each other for a share of the finances available to sponsor their activities and institutions”. Researcher autonomy has become an elusive phenomenon. The more scientists follow the public agenda, the easier it gets to effectively sell their ideas. Research is increasingly conducted within a framework of wider external interest, and collaboration is an effective method for achieving common goals and earning recognition that is supported by social media. A researcher’s credibility remains vital in gaining influence and acceptance, and quality research is still highly regarded in this context. However, social media open up more opportunities to build a personal brand online via networking, which caters to the Matthew effect.

Even though Kuhn follows a rather dogmatic approach to define normative science, the changing impact that social media has on scientific culture still finds support in the way Kuhn and Popper understood knowledge as human creation. Therefore, as discussed, the personal status is of major importance and carries much weight in terms of getting recognized by the community. Social media can support profiling and provide access to desired networks. On the other hand, Merton’s norms and his institutional perspective conflict with social media practices, as they ignore the human component. Social media will further strengthen the ties between science and society and speed up the process of making science more receptive to external influences. How can scientists embrace these changing dynamics and keep scientific culture without losing control? Productivity and effectiveness through collaboration with trusted networks via social media might be the answer. Normal science is characterized by a balance between open-mindedness and dogmatism, and the benefits and links between the traditional scientific practices and social media practices require a better understanding of how researchers collaborate and use social media for disseminating, discussing and promoting their research. Recognizing and heeding the impact that social media have on the sociology of science will help preserve the quality levels to which scientists aspire.

(28)

In the next chapter, scientific collaboration, as the pivotal point in science, will be presented in relation to social media, while some of the prominent characteristics of scientific research that define the scientific community will be discussed.

(29)

CHAPTER 3 – Science Collaboration and Social Media

Due to the increasing volume of knowledge, conducting research without collaboration is becoming increasingly difficult, and collaboration has become a more common practice on a global scale. To master the available information, researchers tend to participate simultaneously in several collaborative networks to exchange and harness other collaborators’ expertise. This chapter will explore the social media phenomenon in relation to scientific collaboration practices, as the linking element in the academic research workflow, which is comprised of knowledge identification, knowledge creation, knowledge quality assurance and knowledge dissemination.

“Social media platforms play an active role in the research lifecycle. They assist researchers in staying abreast of updates in their fields, discovering related work, sharing and discussing research data and results, connecting with other researchers and citizen scientists, collaborating online and getting early feedback on their own work” (Alhoori and Furuta, 2014).

Although collaboration was previously locally bound, social media are opening up opportunities that are not tied to time, place, and funding. Being able to connect with global researchers in the same field can create new opportunities. In addition, scaling up the conversation from one-on-one to many-on-many can have a positive effect on information management in terms of crowd-sourced wisdom.

The scientific culture is changing in the sense that science occurring in isolation no longer exists due to the increased demand for more openness and transparency from funding bodies, institutional administrators, public opinion and policymakers. Observing and learning from the environment is essentially as important as interacting and communication with peers to stay up-to-date. Social media and collaboration practices go hand in hand, facilitating the benefit of working together, as ‘collaborations spawn fresh ideas and boost productivity’ (Ledford, 2008). Collaboration has existed since the early days of science and gains more attention as knowledge and the amount of data increase. In particular, international collaboration has experienced a strong upward trend among scientific authors over the past few decades. Wagner and Leydesdorff (2005) investigated the increase in collaboration and provided evidence that international collaborations doubled between 1990 and 2005. Although not at the same pace, national collaboration has also grown, but it is not as heavily cited as international collaboration. These trends are still ongoing, and they are highly likely to continue, given the opportunities that technologies have to offer. In addition, the number of

(30)

authors has increased, contributing to the growing amount of co-authorship, which represents the main indicator for measuring collaboration (Jian, Ziming, Jia, and Giles, 2008; J. S. Katz and Martin, 1997). However, multiple-author publication as the standard for measuring collaboration does not provide the full picture because collaboration does not always result in an academic article (Frame and Carpenter, 1979; J. S. Katz and Martin, 1997). These instances are not included in the metrics and thus remain unrecorded in statistical figures. In addition, light interactions and minor research support confront the same lack of attention. Through increased engagement with social media in the researcher lifecycle and the resulting measurable exposure of researchers’ online activities, new ways of detecting collaboration through altmetrics can be considered. The possibilities for measuring connections, commentary, readership and more social media-related activities provide more new insights into the way scholars interact and collaborate compared with merely focusing on co-authorship.

Collaboration is essential for complex research problems that require ample information, knowledge and resources to accommodate all possible variants of a project. Because of the increasing knowledge base and advancing technologies, more expertise is needed to master specific knowledge areas. Ultimately, it is too time-intense and complex for one person to be an expert in all of the areas that some of the research problems touch upon. The complexity of scientific information naturally encourages people to become more specialized and focused on a particular method or subject. Hence, collaboration proves to be very beneficial for researchers, as it enables them to share resources and knowledge and thus save time (Hara, Solomon, Kim, and Sonnenwald, 2003, p. 953). Time can be saved if experts filter information and divide it into relevant and non-relevant research. Applying filters can also be supported through technical solutions and measures, as indicated in the literature review on altmetrics, by using the “wisdom of the crowd” (Surowiecki, 2004). Questions remain about the metrics’ reliability. The combination of human and mechanical intervention might provide the most reliable combination.

Even though collaboration is considered essential for researchers, there are a few nuances in terms of what defines and constitutes collaboration. The collaboration phenomenon is one of the most studied and least understood aspects of science. Researchers compete for funding, tenure, publications, awards, discoveries and authority. The reward is usually given to the most competitive researcher. On the other hand, researchers also collaborate to combine the expertise that is necessary to conquer complex problems, share limited or exclusive

(31)

equipment, and exchange data and other information resources. All of the studies referred to in this paper provide evidence for these collaborative practices.

This chapter will first provide an overview of the current understanding of collaboration and then examine motivational factors and theory to answer the question about collaboration in relation to social media.

1. Collaboration in science

Collaboration is defined as working jointly with others or together, especially in an intellectual endeavor (“Webster,” 2014). It is also understood as a joint venture whereby intellectual exchange seeks to achieve a common goal. Collaboration is often used interchangeably with the notion of cooperation or coordination, and it is thus difficult to make a clear distinction between them. Defining research collaboration and its constituencies is not a straightforward task, as the notion strongly depends on social conventions within the scientific community.

A lack of agreement about where loose ties end and collaboration starts also complicates the defining process. Some consider a group of linked people to not immediately fall into the collaboration category, whereas others interpret such a group differently. What defines collaboration is in constant flux and can vary within academic institutes and on a transnational scale.

Katz and Martin approached the subject by focusing on the collaborators, who determine to what extent collaboration takes place (1997, pp. 7–8). They define a collaborator as someone who provides input on a particular piece of research and meets a few requirements that correspond with the concept of collaboration. They narrow down to what extent the contribution is still considered valuable and counts as true commitment. Taking responsibility, being part of the original research proposal and contributing throughout the duration of the project are just a few examples of what defines a collaborator.

Through the introduction of social media services, collaboration has reached a new dimension and changed the possibilities of how collaborators interact with each other. Yang defines online collaborators as any virtual users who interact with others to accomplish the goals of resource discovery, access, knowledge sharing, group communication and discussion (Yang J. H., Chen Y. L., Huang, and Fan, 2007). Social networks help researchers to overcome spatial and temporal limitations when working together and provide new opportunities to reach peers across geographical distances and disciplines.

(32)

There are two essential aspects that recur in every approach to define collaboration (Beaver and Rosen, 1978; Hara et al., 2003; Katz and Martin, 1997, p. 7). The first one is the act of working together to achieve a common goal. Second, collaboration requires the act of knowledge sharing, which benefits everyone involved and the research itself. It provides participants with enough resources to achieve the goal and to manage different individual expectations and perspectives that deploy the research in current theoretical models or establish which hypothesis to test next. As everyone benefits, there is also a notion of reciprocity that keeps researchers engaged and promotes a sense of community, no matter its offline or online presence. As the means of collaboration are continually evolving, a broader and generic concept will be applied in this paper, including interactions in social media. Although science requires collaboration in practice, it can also create various challenges. One of the most studied aspects of research collaboration is the question of why people want to share and collaborate, which is also very relevant for professional social networking sites, as their DNA is based on conversations and the act of sharing and collaborating.

1.2. Why people share and collaborate

Motivational drivers have been studied in several research projects that provide answers and help explain the incentives behind collaboration. Beaver and Rosen (1978) presented an extensive list of 18 motivations that encourage researchers to collaborate with their peer, including the benefits of access to expertise, equipment, resources, or other things that one does not have. Furthermore, collaboration can improve access to funds, speed up processes, provide new skills or techniques, reduce errors and mistakes, advance knowledge and learning, stimulate intellectual interest, reduce isolation and involve pleasure, which is a key reason that people use social media.

Other research splits incentives, or motivations, to collaborate into external (e.g., reputation, funding, and publications) and internal (e.g., personal motivation) categories (Hara et al., 2003). As mentioned previously, engaging in collaboration and social media can have a positive impact on scholarly output by opening up distribution channels and inviting more citations (Figg et al., 2006; Priem and Costello, 2010). Impacting the researcher’s scientific footprint can lead to new career opportunities.

However, collaboration is not always successful, and group dynamics are a significant element that decides how collaborators pursue their goals. The following section will be devoted to the underlying factors of research collaboration and provide additional insights into why scholars engage or do not engage in professional social networking activities.

Social Media Use in Science