
What Web Collaboration Research Can Learn from Social Sciences Regarding Impairments of Collective Intelligence and Influence of Social Platforms Fabian Flock¨ Institute AIFB Karlsruhe Institute of Technology Karlsruhe, Germany fabian.fl[email protected] ABSTRACT has two aspects of significance that will be discussed in the This paper gives a short overview about issues relevant for following with regard to the lessons that can be learned from online collaboration platforms that could be better under- traditional social sciences and the “scientific toolbox” they stood by applying social science research formerly conducted provide in terms of methodologies, theories and empirical ev- mostly in offline settings. Additionally, we are pointing out idence. which theories, mostly from media sociology, could merit The first facet alludes to the collective production of content to be tested in settings supposed to exhibit collective intel- and the trust that is placed in it. Because of the vast numbers ligence traits. The content we present is not intended to be of contributing users, the content generated by collaborative a complete literature review but rather tries to provide ideas platforms is plenty; while additionally, due to the “wisdom of and approaches to make use of existing knowledge in the so- the crowd” that is assumed to guide and control meaningful cial sciences for the research of online group collaboration. output of these systems, it is also trusted to be of high quality and importance, as can be well seen by the high consumption THE TWO SIDES OF ONLINE COLLABORATION COLLEC- and usage of knowledge extracted from Wikipedia, the im- TIVES portance assigned to current trends on Twitter and the use of The ability of social collectives to produce meaningful con- systems like reddit or Facebook as the main gatekeepers for tent through the online software systems they populate is the daily consumption of news by many users. one of the biggest promises of the human-machine appara- tus we call the World Wide Web. They gather and structure The collective intelligence that is the basis for the trust placed knowledge (Wikipedia), propel trends and discuss current in the output of these systems is, however, a fragile thing that topics (Twitter, reddit1, discussion boards), and add mean- hinges greatly on the social dynamics the userbase exhibits, ingful meta-data to content (Tagging-Systems), sometimes by shaped not only by the composition of the userbase itself, but incorporating a variety of opinions and viewpoints (comment also by the environment the software system provides. Cer- sections on media sites, Facebook). tain behavior patterns can lead to unwelcome results of the collective process, just like it is the case for offline scenarios, In the remainder, we refer to such online platforms on which covered by decades of research on the emergence of harm- (i) users collectively produce vast amounts of meaningful ful social interaction patterns in certain populations, leading content that can be in some way aggregated and used and to a vast body of scientific work aiming to understand, pre- (ii) content that is made available to and consumed by a large dict, and prevent the occurrence of such phenomena. These number of users on the open Web.2 Thus, this phenomenon include mass hysteria and panics, stock market bubbles and 1 disease spread; the mechanisms at work have been tried to http://www.reddit.com might not be known to every explain with the help of organizational theory, social imita- reader. It is a social news website where users submit content in the form of text, links or images (ranked 124th most visited website tion theories and psychological approaches, to name only a by alexa.com). Other users vote the submission “up” or “down”, few. In the remainder of this paper we will discuss a selected which is used to determine its position on the site’s ranking. sample of these theories and research approaches that we see 2By this, we mean output that is in some form an aggegregate prod- fit to aide in studying current online phenomena related to uct of the contributions of single users, such as an Wikipedia arcticle, collective intelligence. the ranking of “relevant” or “hot” items on a social news site (not the Secondly, one should, apart from perceiving these extremely popular platforms as digital agoras and online production places, see them in the role of the mass media they have be- Permission to make digital or hard copies of all or part of this work for come, similar to TV, radio and traditional press. This perspec- personal or classroom use is granted without fee provided that copies are not made or distributed for profit or commercial advantage and that copies tive merits being taken into account on the sole fact that gen- bear this notice and the full citation on the first page. To copy otherwise, or erally, around 50% - 90% of users of these platforms are mere republish, to post on servers or to redistribute to lists, requires prior specific permission and/or a fee. news themselves) or the folksonomy produced by an open tagging CHI’13, April 27–May 2, 2013, Paris, France. platform. Copyright 2013 ACM 978-1-XXXX-XXXX-X/XX/XX...$10.00. “lurkers” [27], only consuming and not contributing actively results.5 On top of the fact that the typical users (readers and to the content, with platforms like Wikipedia and Reddit re- contributors) of a system might not be representative of the ceiving millions of unique visitors per day.3 At the same time, general online (or offline) population, for a range of online online media consumption gradually replaces offline media collaboration systems the phenomenon of a small minority of consumption. Once this view is taken, very similar questions contributors providing most of the edits and content [26, 16] arise in respect to online media that have formerly been asked has been observed. For Wikipedia, for instance, Priedhorsky towards the traditional mass media. Long standing theoretical et al. [20] show that a vast majority of the content is actually frameworks in the field of media sociology and media impact provided by a minuscule fraction of all editors. studies, like Agenda Setting Theory, the Cultivation Hypothe- sis or Selective Reception can help to better understand the ef- The specific reasons for the low representativeness that might fect of these online mass media on their recipients. Although occur in some systems are, however, not exactly clear. Gen- of course, on the Web2.0, the users are not merely passive re- der Studies can help to explain why still, a relatively small cipients, but active contributors themselves, making it neces- amount of women [10] is contributing to collaborative plat- forms, but have not yet delivered sufficient clarification on sary to modify the overcome theoretical approaches and their the issue regarding crowd collaboration. The theory of the segregation into the clearly distinct roles of journalists, media Digital Divide and associated works [2, 8] try to explain the producers and audience. Here, the above discussed aspect of collaborative production plays a crucial role, as the collective differences in media and online usage between different de- mostly takes over the roles of audience, journalist, newsroom mographics and social layers and have a long tradition under editor and gatekeeper at the same time. related terms such as the ”Knowledge Gap” and ”Knowledge Divide”. However, this research often focusses on the eco- Hence, as online collective creation and online mass media nomic variables determining the type of internet access and aspects are two sides of the same coin for these platforms could be employed more thoroughly in order to research and we will point out certain research areas in this regard in the explain why certain populations do not take part in the on- following that could profit from the application of social sci- line discourse and creation (sometimes called “second-level ence theory that was traditionally applied mostly in the offline digital divide” [22]) and if this is truly hurting the ability realm or at least not extensively used to explain collective col- of collaborative online platforms to achieve optimal results. laboration phenomena. Questions not yet satisfactorily answered in these two field are, e.g.: Do male users habitually dominate the shaping of C.I. RESEARCH ISSUES AND APPLICABILITY OF SOCIAL the collective result in a collaborative, interactive environ- ment? Do more technology-savvy users or users with spe- SCIENCE THEORY cific socio-economic premises (e.g., more spare time) have a In this section we list either issues that have (i) become salient vantage point in online discourse that leads to their eventual as problematic for the outcome of collective intelligence in prevalence in discussions and negotiations about content? the aforementioned platforms and could benefit from the ap- plication of traditional social theories to study them or (ii) theories and methods of social sciences (mainly media so- Information cascades and conformity ciology) that we think would merit a closer look and more One aspect that has been shown to impair the independent extensive application to online scenarios such as the ones de- decisions of the members of the crowd (a pillar of collective scribed by us. intelligence [24]) is certainly imitation. By abandoning their individual decision-making capabilities, single members of Representativeness the swarm don’t tap into their cognitive resources which in Whom does the userbase of a particular online collaboration turn are lost to the collective, possibly resulting in herd-like system represent? Who is effectively responsible for the con- decisions based on the choices of a few early opinion leaders. tent produced, and which portion of the general population Especially the work on the concept of social proof [3] and are these users a sample of? This pertains to questions about information cascades [1] merits a closer look upon from web the experience, knowledge and political and ethical views of scientists.
Details
-
File Typepdf
-
Upload Time-
-
Content LanguagesEnglish
-
Upload UserAnonymous/Not logged-in
-
File Pages5 Page
-
File Size-