CCMC Symposia Abstracts

Bradford Demarest and Andrew Tsou
Department of Information and Library Science, School of Informatics and Computing

TED Talks is one of the leading science communication initiatives in the digital age. Although previous work has analyzed the demographics of speakers & audience reaction to TED Talks, there is a dearth of research into the actual content and discourse used in these talks. Transcripts of TED videos were downloaded from the official TED website and analyzed as to word use by different speaker classes (male academics, female academics, male non-academics, and female non-academics); in addition, we incorporated a focus on words commonly used to express speaker’s stance as well as audience and topic engagement, drawn from Hyland’s (2005) study of metadiscourse. Our preliminary results suggest that the subpopulations (males vs. females; academics vs. non-academics) exhibited marked differences in the words that they used during their talks, which may indicate different sentiments, topical preoccupations, and goals for the presentation. Moreover, gender was an important variable throughout the study, suggesting a fruitful direction for further investigation.

Jennifer Terrell
Social Informatics, School of Informatics and Computing

Scholarship that has explored the role of digital media in our social relationships has historically tended to focus the field of study on one particular medium, at times a particular platform such as World of Warcraft or Second Life. However, as scholars are beginning to focus more on social phenomena and less on particular media, there is now a call for methods that understand the ways in which people employ many different types of media in their social interaction. This paper presents empirical ethnographic research of those who participate in a phenomenon known as wizard rock, a collective of fans of Harry Potter who create, circulate, and perform fan music inspired by JK Rowling’s novel series and the music they produce. Participation in wizard rock happens across many digital media platforms such as YouTube, Facebook, Twitter, and Tumblr as well as numerous other websites. Wizard rock is also performed live in living rooms, libraries, and occasionally nightclubs around the United States. Social interaction in wizard rock can be characterized as transmediated because the participants within this group weave multiple media together in order to accomplish their social goals and there is no single medium in which the community resides. In addition to exploring the ways in which multiple forms of media are used by these groups, the paper argues for the necessity of transmediated methods to investigate such digitally mediated social phenomena.

Daphna Yeshua-Katz
Department of Telecommunications, The Media School

Media scholars often use concepts from Goffman’s dramaturgical approach to study online communities of stigmatized individuals as “backstages,” spaces where members take refuge from social disapproval. In this talk, I extend this view through an examination of in-depth interviews with bloggers from the pro-ana community, an online community for people with eating disorders. To explore how this community uses an online environment that is both anonymous and public, I fuse Goffman’s ideas about identity performance and stigma with more recent theories about boundary maintenance. In-depth interviews with pro-ana bloggers reveal that to protect this virtual group and resist stigmas associated both with their illness and with their online presence, they construct their own norms and rules in the online realm, and discipline and eject members deemed to be out-group.

Lindsay Ems
Department of Telecommunications, The Media School

Social media use by health practitioners helps articulate a subculture-centered approach to public health communication. This presentation explores public health practitioners’ communication strategies within one subculture: young (18- to 29-year-old) men who have sex with men (MSM). Interviews with staff at a public health clinic reveal the use of a variety of social media systems to engage potential MSM clients (e.g. Facebook, Twitter, Instagram, Grindr, and Google+). Public health communicators’ social media use reflect the esthetic, behavioral, and media preferences of a high-risk target group and the ethical risks associated with publicly viewable communication. Social media have become a platform for integrated multi-channel communication that comprises both top-down (i.e., institution centric) and bottom-up (i.e., subculture centric) communication flows that work together in a complementary fashion. Ultimately, subculture-centered social media use in this setting helps minimize cultural barriers between members of a public health subculture and the institutions that provide critical care.

Alisa Boguslavskaya
Kelley School of Business

Resource acquisition is an integral part of the new venture development process. Studies of established ventures highlight the importance of social ties and impression management in explaining the acquisition of resources for growth. However, unlike established ventures, new ventures often lack both social ties and an identity and therefore often find it difficult to acquire the resources needed for start-up and growth. How do new ventures build both their network and identity for successful resource acquisition? We examine this question through an in-depth content analysis of social media communication of eight new ventures attempting to raise capital through crowdfunding. The emergent model of entrepreneurial communication highlights the venture’s creation of identity and/or community as key behaviors in its attempt to gather resources. Furthermore, our findings show ventures’ both positive and negative reactions to success in that endeavor or lack thereof.

Jordan Barlow
Kelley School of Business

This presentation will be an overview of two ongoing research projects on computer-mediated groupwork. I hope to achieve three goals through this symposium talk: (1) help CCMC fellows and others learn more about my research; (2) solicit meaningful feedback on the two projects to help them move forward toward publication; (3) invite potential co-authors for these two projects or future projects that will stem from them. The first project is a set of studies that uses role theory and speech act theory to explore and validate the emergence of informal roles in online decision-making and the impact these roles have on group perceptions of trust, communication, and performance. The studies use discourse analysis, text mining, and transcript and survey data from student groupwork in lab experiments. The second project I will present is a literature review showing that the relationship of several group-level factors (e.g., trust, cohesion, and goal setting) with group performance depends highly on the type of task that groups are performing. The literature review also highlights key research that may be missing. The literature review has implications for future research on computer-mediated groups and will itself be expanded into a larger empirical study examining differences in group performance.

Yanqin Lu and Jae Kook Lee
Department of Journalism, The Media School

Drawing on a national probability survey, this study explores the mechanism through which politically interested individuals avoid disagreement in political discussions on SNSs. The results reveal that SNS news use mediates the negative relationship between political interest and discussion disagreement. Nevertheless, this mediating effect only occurs on individuals who frequently discuss politics on social media. Further, politically interested individuals are more likely to avoid discussion disagreement when their online network is primarily composed of strong-tie relationships. Implications are discussed for the impacts of SNS use on deliberative democracy.

Will Allendorfer
Department of Second Language Studies, College of Arts and Sciences

The Islamic State of Iraq and Syria (ISIS) is a militant group claiming the establishment of an Islamic caliphate in areas it occupies in Iraq and Syria. ISIS relies heavily on propaganda in the form of videos distributed over social media to recruit new members, including from the U.S., to its cause, and the U.S. government has countered with anti-ISIS propaganda videos. However, the head of the FBI’s counterterrorism division recently claimed that the U.S. is losing the propaganda war. This study sheds light on that claim through a comparative multimodal content analysis of the video Flames of War, produced by ISIS, and videos on a YouTube channel produced by the U.S. Department of State (USDS). Findings reveal that the ISIS and USDS propaganda videos employ divergent rhetorical strategies; I argue that the USDS strategies are weak in several respects and could cause Muslim viewers in the U.S., the ostensible target audience, to perceive the USDS videos as ineffective. In concluding, I propose recommendations that the USDS might follow to improve its online propaganda defense against ISIS and extend the implications of my findings to other social media fronts where the ISIS vs. USDS propaganda war is being waged.

Muhammad Abdul-Mageed
Department of Information and Library Science, School of Informatics and Computing
Department of Linguistics, College of Arts and Sciences

In spite of the strategic importance of social media for self-expression and policy shaping in the Arab world, there is not much literature on Arabic computer-mediated communication (CMC). In this presentation, I report attempts to partially bridge this gap by investigating issues related to the language and sentiment of Arabic social media. More specifically, a data-driven approach is employed to identify the different language varieties, linguistic features, and functions and distribution of sentiment in four different social media genres: (1) YouTube comments, (2) Twitter tweets, (3) chat turns, and (4) Wikipedia Talk Pages. Results are situated within the wider CMC literature, and implications for computationally modeling different facets of social meaning (e.g., sentiment, users, and perspective) in the Arabic social media sphere are discussed.

Elizabeth Herring
Department of Spanish and Portuguese, College of Arts and Sciences

Guaraní code-switching and code-mixing in the South American country of Paraguay is considered a stigmatized variety of language not to be used in formal situations, but it is also a source of pride, emphasizing the impact the Guaraní culture and language has had in distinguishing Paraguay from its South American neighbors. The importance of Guaraní to Paraguay’s younger generation is exemplified by its extensive use on Twitter. Despite anecdotal evidence that Paraguayans younger than 30 are disillusioned with the state-mandated requirement to learn Guaraní in the classroom, the use of language laced with Guaraní grammar and vocabulary on Twitter indicates that its eradication is not imminent. I discuss this phenomenon by focusing on Guaraní interrogative markers. I also discuss the potential benefits of representing CMC language use in prescriptive classroom grammars. The Guaraní taught in Paraguayan classrooms has become an extinct version of the language; a language variety that has no native speakers and is not used outside of an academic atmosphere. More thorough and accurate descriptions of Guaraní usage in all modes of modern communication can impact not only language knowledge but also language attitudes and rate of use. By making Twitter and other CMC legitimate sources of language, what is taught in the classroom can appeal to a greater audience.

Elli E. Bourlai, Muhammad Abdul-Mageed, and Susan C. Herring
Department of Information and Library Science, School of Informatics and Computing

Twitter hashtags are part of the social tagging phenomenon that emerged with the appearance of Web 2.0. Even though hashtags are originally a convention created by Twitter users for information sharing purposes (Hughes & Palen, 2009), users also employ them in other pragmatic functions. Efron (2011) identifies the use of some hashtags as “tongue-in-cheek humour,” while Page (2012) distinguishes two types: tags representing a topic and tags expressing an evaluative sentiment. Applying a grounded theory approach to a corpus of 2,000 tweets, this study identifies four pragmatic types of hashtags (topical, evaluative, ambifunctional, and extracontextual) based on structural linguistic criteria, such as the hashtags’ location, length, presence of repeated characters, number of words, grammatical function, and the presence of acronyms or abbreviations. We identify sets of structural features that have the potential to be incorporated into a machine-learning algorithm to enable the automatic identification of these four functional types. We further consider the feasibility of using this hashtag taxonomy to improve accuracy in sentiment and sarcasm detection, for example in social media opinion mining.

Susan C. Herring
Department of Information and Library Science, School of Informatics and Computing
Department of Linguistics, College of Arts and Sciences

CMC was originally text based and accessed through stand-alone clients. Increasingly, however, textual CMC has been supplemented by graphical, audio, and/or video channels of communication, and multiple modes of CMC are available on Web 2.0 platforms and smart phones. As the technological affordances of CMC systems have evolved over time, so, too, have the efforts of scholars to analyze the discourse produced using those systems. One approach is computer-mediated discourse analysis (CMDA), a specialization within the broader interdisciplinary study of CMC distinguished by its focus on language and language use and by its use of methods of discourse analysis to address that focus (Herring, 2004). I describe my efforts to develop and extend CMDA over time with respect to three historical phases of CMC: stand-alone textual clients, Web 1.0, and Web 2.0. I illustrate each phase with representative work, and conclude by proposing a theory of multimodal CMC that includes graphical phenomena such as memes, avatar-mediated communication, and robot-mediated communication involving telepresence robot avatars in physical space. Each of these phenomena mediates human-to-human communication, supports social interaction, and involves multiple modes or channels of communication, and thus constitutes fertile ground for contemporary computer-mediated discourse analysis.

Ilana Gershon
Department of Anthropology, College of Arts and Sciences

In 2007-2008, I collected breakup stories, which turned out to also be narratives about switching media. People of all ages would describe the sequence of events and conversations leading to the break-up, focusing largely on who said what in a reconstructed chronological order. As they did this, my interviewees always pointed out the medium in which each utterance took place. They demarcated any switch in media, letting me know whenever the narrated conversational exchanges switched, for example from phone to texting. It became apparent that this is a relatively typical feature of contemporary American break-up narratives. Many break-up narratives featured in movies such as /He’s Just Not That Into You/ or posted on blogs also routinely provide metadiscursive cues linking the utterance with its medium and help to stabilize diffuse conversational exchanges into an overarching type of event – the breakup. In this talk I discuss media-switching in the context of earlier sociolinguistic research on code-switching to explore the interconnections and dissimilarities.

Kathleen Bardovi-Harlig
Department of Second Language Studies, College of Arts and Sciences

Sometimes speakers find themselves in a situation in which they feel they must say something that others might find “impolite.” The decision to perform an impolite speech act is often instantaneous (or perhaps not a decision at all). However, in the case of disinvitations—the withdrawal of an expected or previously issued invitation—such decisions are often explicit as documented by online discussion posts which capture the conflict in performing disinvitations: “No one yet has invented a way to be polite while being rude. You cannot avoid being rude if you do this, terrible thing to do” [yahoo answers…] and “There's no polite or graceful way… because it's an inherently impolite and graceless thing to do” [askmetafilter.com]. I will attempt to show that there are speech community guidelines for performing disinvitations that may help interlocutors maintain their relationship, but do not transform disinvitations into polite utterances.

The petitions for advice about performing disinvitations reveals speaker intention in what amounts to speech-act slow motion, away from the tempo of conversation, as speakers solicit advice from online communities. This shows speakers and speech act performance to have an intentional and rational side, at least in this case where disinvitations are considered beforehand and are delivered as discrete speech events in writing (in spite of advice to the contrary). Community responses include negative assessments of disinvitations, admonitions not to perform them, and constraints on doing so, and much less often, the necessary components.

César Félix-Brasdefer
Department of Spanish and Portuguese, College of Arts and Sciences

Pragmatic routines are generally defined as formulaic expressions used by members of a wide community of practice to express an interpersonal or relational function. Research on formulaic sequences has advanced our understanding of how learners process and produce formulaic language (so-called ‘chunk learning’ or prefabricated speech) that serves to promote the learners’ development of second language (L2) grammar in natural contexts, under experimental conditions, and in different modes of communication. Different criteria have been used to identify formulaic patterns in learner corpora including, but not limited to, frequency, association, native norms, and context (Bardovi-Harlig 2012; Ellis 2012).

Using an email corpus of natural student-initiated requests sent to faculty, in this talk I examine the formulaic patterns of pragmatic routines (recurrence, association, and context) used by L2 learners who address their instructors in their native or target language to achieve a particular perlocutionary effect. The L2 email users belong to the same academic community of practice in which a variety of low and high-imposition requests occur. Specifically, I look at pragmalinguistic variation in greetings (including address form selection) and closings of email messages sent by learners to their instructors. Issues of frequency and context are reviewed for the purposes of pragmatic formula research.

Xinping Jiao
Visiting Scholar, Department of Information and Library Science, School of Informatics and Computing

This research analyzes a 1,416,844-word corpus automatically collected in 2012 from the Sina Weibo microblogging service in China with respect to the sociolinguistic concepts of power and solidarity (Brown, 1961) and appraisal (Martin, 2003). We chose 30 institutions (government departments, businesses, TV stations, newspapers), 30 celebrities (pop stars, hosts/hostesses, writers, scholars), and 30 grassroot microbloggers (anonymous unidentified) as the sample. The analysis combines quantitative and qualitative methods. With regard to overall linguistic features, the corpus has a higher Standardized Type Token Ratio (STTR) compared with the Lancaster Corpus of Mandarin Chinese (LCMC). The Weibo Corpus integrates neologisms, multi-modal elements, and both archaic and fashionable expressions. With regard to power and solidarity, institutions use the most words for power differences, whereas celebrities use the most address terms for solidarity. Institutions also use neutral/aloof expression most. With regard to appraisal, the celebrity and grassroots microbloggers express more negative attitudes and “anger, sadness, fright and disgust” than the institutions. Celebrities and institutions express “happiness” more than the grassroots, and celebrities use the most polarized emphatic words. The research shows that power and solidarity and appraisal are powerful constructs in Weibo discourse. However, some modifications are needed to apply these constructs to Chinese CMC. We incorporated humble and contemptuous expressions into the power pole and intimacy into the solidary pole. Also, we applied appraisal to multimodal discourse elements like emoticons. The research is the first attempt in China to study Weibo discourse from the above two perspectives.

Amy Gonzales
Department of Telecommunications, The Media School

Nationally representative survey data suggests widespread Internet (77%) and cell phone use (84%) among the nation's poor, but there is little data on the stability of that access. This talk describes findings from two qualitative studies on the consequences of unstable cellphone access for low-income US populations using a new theoretical framework of technology maintence. Technology maintenance states that as the poor widely adopt digital technologies, disparities will manifest in the ability to maintain that access. In the first study, in-depth interviews with public housing residents in New York City reveal that phones are essential for safety, but residents often have broken and disconnected access. These data are especially troubling given that more than half of low-income residents in the US no longer have landline telephones. In the second study, interviews with local HIV patients find that temporary cellphone disconnection disrupt access to healthcare, employment, and social support. Findings underscore the importance of cell phone access for health and well being, and at the same time indicate that disparities in cell phone access still persist and may be worsening. Directions for future research will be explored with the audience.

Jessica Myrick
Department of Telecommunications, The Media School

Social network sites (SNSs) like Twitter continue to attract users, many of whom turn to these spaces for social support for serious illnesses like cancer. Building on literature that explored the functionality of online spaces for health-related social support, we propose a typology that situates this type of support in an open SNS-centered cancer community based on the type (informational or emotional) and the direction (expression or reception) of support. A content analysis applied the typology to a two-year span of Twitter messages using the popular hashtag “#stupidcancer.” Given that emotions form the basis for much of human communication and behavior, including aspects of social support, this content analysis also examined the relationship between emotional expression and online social support in tweets about cancer. Furthermore, this study looked at the various ways in which Twitter allows for message sharing across a user’s entire network (not just among the cancer community). This work thus begins to lay the conceptual and empirical groundwork for future research testing the effects of various types of social support in open, interactive online cancer communities.

Alan Dennis
Kelley School of Business

Understanding and improving team creativity has been a major focus of CMC research for almost three decades. Most research has focused on social psychology-based factors such as anonymity, group size, and the communication process itself. Few studies have used the lens of cognitive psychology to examine how individual cognition affects team performance. Cognitive psychology shows that much of human behavior is influenced by subconscious cognition, rather than deliberate rational conscious cognition, and that it is possible to influence subconscious cognition and thereby change behavior.

In this presentation, I report on a series of studies showing that we can design CMC technology to deliberately influence subconscious cognition and thereby improve team creativity. We use a technique called priming, in which an external stimulus is presented to team members, either before or during the creativity task. We examine the impacts of three of different forms of priming stimuli (words, pictures, and the work environment itself). In one study, we use techniques from cognitive neuroscience (i.e., EEG) to examine how the priming stimuli influence cognition in different regions of the brain and link those changes to changes in creative performance. The effect sizes due to priming were large (most Cohen’s d above .80), which are as large as the effect sizes typically observed when comparing CMC to verbal brainstorming. In other words, the impact of designing CMC environments to trigger subconscious cognition is as strong as the choice to use CMC itself.

Noriko Hara and Pnina Fichman
Department of Information and Library Science, School of Informatics and Computing

Websites such as Facebook, Twitter, and Wikipedia facilitate knowledge sharing in online communities that extend beyond the constraints of existing organizational boundaries (Qualman 2011). Vakkayil (2012) contends that the concept of boundaries, such as physical, cognitive, social, and political (e.g., Carlile, 2004), is informative for understanding knowledge sharing and creation among heterogeneous users with diverse cultural, organizational, socio-economic and spatial backgrounds. However, these boundaries may function quite differently in Web 2.0 environments outside organizational settings. Although it is easier and faster to share knowledge and collaborate online, few studies have investigated how boundaries and boundary-crossing in Web 2.0 environments affect online collaboration, knowledge sharing, and participation. This talk explores the usefulness of employing existing boundary frameworks to facilitate knowledge sharing and collaboration.

To test and modify existing theoretical frameworks of boundaries, we analyzed Wikipedia entries about the Japanese nuclear power plant accident triggered by the Tōhoku earthquake in March 2011. This event attracted attention from around the world and is useful for our research because the case involves multiple boundaries and various processes of boundary-crossing. Based on the data and the literature we propose a refined framework for boundaries and boundary-crossing. In so doing, this paper advances understanding of boundaries for knowledge sharing in online communities and identifies how these boundaries facilitate or hinder equal participation.

Xiaozhong Liu
Department of Information and Library Science, School of Informatics and Computing

All the popular social media systems, e.g., Facebook, Twitter, and Weibo, can only reach a limited group of users. For instance, Twitter is forbidden in China for political reasons, and Weibo (Chinese) is not being used in the US or Europe due to language barriers. In this project, we design innovative algorithms, models, and system to efficiently compare Twitter and Weibo communities from different perspectives, which can be used to better understand the similarities and differences of very large user communities. For instance, given the same topic, Iraq War or iPhone 6, Twitter and Weibo users may access, mention, and discuss it in very different ways. In this study, we use sophisticated text mining and graph mining algorithms along with big data indexation methods to compare Twitter and Weibo users. In addition, we have proposed a new method to efficiently integrate Twitter and Weibo data, and to generate a Pseudo Global Social Media Network (PGSMN); to the best of our knowledge, this is the first global social media network. By using PGSMN, we can implement global information recommendation, e.g., to recommend Twitter information/topics to Weibo (Chinese) users and vice versa.

Sung-Yeon Kim
Visiting Scholar, Department of Literacy, Culture, & Language, School of Education

This study aims to explore potentials of online chatting as a pre-writing task, focusing on the words shared between text-based chat scripts and written essays. Data for the study comprise chat scripts and essays produced by 100 college students. The participants were asked to read a passage and participate in electronic chatting in groups of three to four for 40 to 60 minutes. After a week interval, they were asked to write argumentative essays in response to the reading text. A total of 26 chat scripts from group discussion and 100 individual essays were collected and analyzed for vocabulary overlap between the two different types of texts. The findings of the study indicate that over 90% of the tokens and 64% of word types were shared between the two texts. In other words, a substantial number of words used in chatting were recycled in essay writing. It can be inferred from the findings that online chatting was as useful as essay writing in facilitating the use of productive vocabulary. These findings suggest that online chatting seems to be a valid option as a pre-writing task. More detailed findings of the study will be reported, along with their pedagogical implications.