vol. 15 no. 3, September, 2010
Discussions of whether or not you should trust what you read in a Wikipedia article turn up in many parts of society – in the newspapers, in classrooms, and at the dinner table. But how is credibility discussed by the people who create the articles, by the wikipedians themselves? What types of criteria do they consider, as editors and readers of Wikipedia, when they assess credibility as part of their work with an article? This could concern, for instance, when they decide on which information to include in an article, when they decide what to edit, or when they use Wikipedia articles for other purposes. How do their practices compare with findings from previous research about credibility assessments on Wikipedia made by other user groups? These are the questions we approach in the paper.
The multi-language, participatory encyclopedia Wikipedia has now reached the level of renown where it hardly needs an introduction. It is one of the ten most visited Internet sites globally as well as in Sweden, where our study took place (Alexa 2010a; 2010b). The popularity of Wikipedia, as well as of similar applications, has consequences for people's information practices. Metzger and Flanagin note that
blogs, wikis, social networking sites, and other digital media applications… constitute a significant portion of today's media environment... It is crucial, therefore, to consider what new and emerging types of credibility and credibility assessment are implicated in these media tools. (2008: 10)
Particularly important, but so far not much investigated, is the issue of how the creators of participatory information environments act with regard to and view credibility issues. An example is Wikipedia, to which people turn for answers and entertainment. The Wikipedia editors are the ones shaping the articles and the tools of Wikipedia. Their work determines the quality of the resource, of which credibility is an important aspect. Previous studies have investigated the perspectives of various user groups and so-called experts when it comes to perceptions of Wikipedia's credibility. We argue that knowledge of the Wikipedia editors' views and actions is an important contribution to understanding credibility in Wikipedia.
The study has been conducted with editors of Swedish Wikipedia, which started in May 2001, approximately four months after the English language version. It is, at the time of writing, the eleventh largest language version with more than 350,000 articles ( 'Wikipedia: Svenskspråkiga Wikipedia' 2010). The number of editors is naturally much smaller than in English Wikipedia given that the number of speakers of Swedish can be estimated to be below 10 million. For instance, English Wikipedia had 4,007 very active contributors in January 2010, whereas Swedish Wikipedia had 157, which also places the version at eleventh place (Zachte 2010). Most studies of Wikipedia to date have concerned English Wikipedia, which makes it interesting to study contributors to the mid-sized language versions of Wikipedia to gain a broader view of Wikipedia editors. The terms editor and contributor are used interchangeably in this paper for people who contribute to Wikipedia in various ways, for instance through writing new articles, editing or expanding existing articles, improving language, or removing vandalism. We also sometimes use the term wikipedian to refer to editors who are particularly committed to their work on Wikipedia and who generally regard themselves as part of the Wikipedia community.
We have studied how frequent contributors to Swedish Wikipedia talk about credibility assessments as part of their participation in Wikipedia. This article is the first report from a study of contributors' everyday activities on Wikipedia and is based on interviews with contributors to Swedish Wikipedia. In our analysis, we draw on theories from library and information science, communication studies, organization studies, and education. The study shows that frequent contributors to Wikipedia use the encyclopedia in ways that are similar to user groups who are only readers/users of Wikipedia but that they sometimes make credibility assessments based on a richer understanding of an article's history and origin.
The analysis takes its point of departure in two areas of related work, of a both theoretical and empirical nature. One is a socio-cultural approach to learning and practice (Lankshear & Knobel 2008; Säljö 1999), in this case explored mainly through the concept of networks of practice (Brown & Duguid 2001). The other springs from a central concern in information literacy research, namely research on credibility. However, the scope of credibility research has been widened to include more than those researchers who would consider themselves information literacy researchers.
Practices can be understood within a socio-cultural framework as iterative tool-based and goal-directed activities (cf. Scribner and Cole 1981: 267). That is, people act by using cultural tools, tools which influence how we think about and do things, such as for instance information practices. Concurrently, practices have consequences, and these consequences contribute to shaping the very tools used (Säljö 1999). In the case of Wikipedia, the activities of hundreds of thousands of editors have very direct and practical consequences for Wikipedia itself, since the participatory encyclopedia is constantly made and remade by the contributors. Cultural tools can be both what has sometimes been identified as intellectual tools, such as language, and tools typically described as material tools, for example a computer. However, it is also important to remember that this dichotomy is not necessarily meaningful. Rather, as Säljö (1999) and others argue, both dimensions of tools are intimately integrated. For example, Wikipedia is a material tool with texts and images in a database structure, but it also embodies certain intellectual tools, such as views of how knowledge is produced and communicated in Wikipedia, and to some extent in contemporary society in general.
The process through which Wikipedia editors acquire shared knowledge and a shared set of cultural tools has been described in terms of legitimate peripheral participation (Bryant et al. 2005), building on Lave and Wenger's work (1991). Through participation in the practices of writing and editing articles on Wikipedia, the contributors gradually become more expert users of Wikipedia, what Bryant and colleagues refer to as 'members of the tribe' (2005: 7). However, the editors on Wikipedia at the same time make up a diverse group, not quite possible to describe in terms of a community of practice (Lave & Wenger 1991; Wenger 1998). We find it more suitable to turn to the elaboration of the concept introduced by rown and Duguid (2001), who speak of wider networks of practice, using the term network 'to suggest that relations among network members are significantly looser than those within a community of practice' (Brown and Duguid 2001: 205). Membership in a network of practice is something that is learnt over time through the acquisition of knowledge and identity through common practice. Brown and Duguid discuss in detail how knowledge can be understood in this context, turning to Polanyi and Ryle to capture knowledge as something characterised by two inextricable dimensions: a tacit dimension (know how) and an explicit dimension (know that) (2001: 201 f.). These two dimensions of knowledge have similarities with cultural tools in that they suggest both an intellectual and a material aspect to how knowledge is performed. One practice in which this takes place is when the credibility of a source is assessed.
The issue of credibility in relation to various sources, including people, organisations, and tools, has been studied within several disciplines, including 'communication, information science, psychology, marketing, and the management sciences' (Rieh and Danielson 2007: 307). In this paper, we will primarily draw on credibility research from the first two of these disciplines, library and information science and communication studies. However, when we specifically discuss research on credibility in Wikipedia, we will have a broader scope, including studies from education, philosophy, computer science, and human-computer interaction.
There are many attempts to define credibility, although none have reached general acceptance. The concept relates to such notions as quality, trust, authority, and persuasion (Rieh and Danielson 2007: 312 ff.). It is often associated with the authority provided by somebody or something that is perceived as holding trustworthiness and expertise or competence (Metzger and Flanagin 2008: 8; Wilson 1983), but as the discussion below will show, other tools may be useful in assessing credibility, such as the frequency with which a claim occurs. Wilson describes cognitive authorities as having 'influence on one's thoughts that one would consciously recognize as proper' in a particular situation (Wilson 1983: 15). In judging who or what can be regarded as a cognitive authority, what a particular society, network of practice, or individual considers to be trustworthiness and expertise has importance, even though either component may be differently valued in different situations. Wilson, along with many others, further points out that credibility is not a property of the source, but a quality attributed to the source by those making the assessment (Wilson 1983: 13), what Metzger and Flanagin call a 'receiver-based construct' (2008: 7).
A cognitive authority is not necessarily a person, it can also be an artefact, such as a document or a thermometer (Wilson 1983). Credibility, or lack thereof, can further be associated with a medium or a genre. This has been observed by those researchers who make a distinction between source credibility, message credibility and media credibility. Source credibility, according to this view, concerns the credibility of the person or organisation who is responsible for sending a message or making a claim (note that here source is used in a broader sense to also include various forms of informative artefacts), whereas message credibility captures clues that may be gained from qualities of the message, such as ' structure, content, language, and delivery' (Metzger and Flanagin 2008: 9). The trust attributed to particular media, radio, television, newspapers, Internet, has been studied as media credibility. However, there have been claims that assessing the credibility of these various objects overlap on the Internet, where more aspects of a document need to be taken into consideration (Metzger and Flanagin 2008; Rieh and Danielson 2007: 310). Our previous research has also indicated that credibility assessments are complex and often blur these types of distinctions (e.g. Sundin and Francke 2009).
Various attempts have been made to model credibility in ways that move beyond the type of distinction made between source, message, and medium above. Metzger and Flanagin suggest a model of variants of credibility in digital media (2008: 10 ff.). These include conferred credibility, which takes place when an entity that is well-regarded recommends a source, and reputed credibility, which is based on the reputation of the source or medium and which can be'perpetuated through personal and social networks' (2008: 11). Both variants of credibility also exist in older media, but the two other variants of credibility they propose, tabulated and emergent, are more typically limited to digital media. Tabulated credibility concerns peer ratings that are tabulated to give a metric for the quality or credibility of, for instance, a reviewer on Amazon or a seller at eBay. Emergent is the credibility that springs from a 'pool of resources' (2008: 12) that collaborate in an open environment, as in Wikipedia. This idea is closely related to what in Web 2.0 discourse is referred to as the wisdom of crowds or harnessing collective intelligence (O'Reilly 2005; Surowiecki 2004). (Several other researchers have similarly suggested models of types of credibility, e.g., Tseng and Fogg (1999) and Fogg (2003) list presumed credibility, surface credibility, reputed credibility, and earned credibility, and Rieh (2002) makes a distinction between predictive and evaluative credibility judgments.
Metzger and Flanagin (2008: 12) note that what they term emergent credibility has been described with a ore or less similar intent by Lankes (2008) as part of what he terms the ' reliability approach'. Lankes contrasts an authority approach to credibility with a reliability approach, where ' users determine credibility by synthesizing multiple sources of credibility judgments' (2008: 106) rather than turning to an established authority. He infers that the Web and, in particular, participatory web sites, requires that we assess reliability rather than authority. This is done primarily by comparing claims from different sources with each other, so that if a claim occurs in several sources that are independent of each other, it is likely to be credible. Wikipedia is identified by Metzger and Flanagin as a ' prime example' (2008: 12) of a source where emergent credibility is applied. We turn next to some of the aspects of credibility on Wikipedia that have attracted researchers' attention.
The issue of credibility in Wikipedia has been discussed and, to a smaller extent, investigated from a variety of perspectives. One approach is to attempt to facilitate credibility judgments by assessing or providing suggestions for vetting the credibility of an article or editor through quantitative methods using the vast Wikipedia database, something that enables, for example, tabulated credibility judgments to be made. The architecture of the MediaWiki software on which Wikipedia runs incorporates a number of features that can be used in credibility assessments of individual articles, such as the history of the edits made to the article, information about the editor, and the article talk pages where the article's content is negotiated and occasionally debated. Such features have been used in several studies to try to statistically create metrics or reputation systems that will aid users in assessing the trustworthiness of an article or editor (e.g. Dandio et al. 2006; Adler and de Alfaro 2007; Hu et al. 2007).
Much discussion has concerned how trustworthy Wikipedia articles are (e.g., Denning et al. 2005). One approach has been to compare the quality of Wikipedia articles to that of other encyclopedias (e.g., Giles 2005; Magnus 2006; 'Wikipedia schlägt Brockhaus' 2007). Duguid (2006) made a critical reading of a few biographic entries and noted that several small changes can make an article less accurate and coherent. Researchers, journalists, and others have also inserted ' fibs' or faulty information into articles and tracked the time it took for them to be corrected or erased (e.g., Magnus 2008). A particular interest in whether or not we should trust Wikipedia is displayed by scholars who apply an epistemological perspective (e.g., Magnus 2009; Fallis 2008; Tollefsen 2009). Magnus points out that Wikipedia is not quite like other encyclopedias and that it ' frustrates each and every one' (Magnus 2009: 83) of the methods he identifies as useful for evaluating the legitimacy of claims on single-author web sites. This includes the plausibility of style and content, because someone could have improved on the language of an article without correcting content errors, or have removed obviously faulty claims without removing those claims that are faulty but harder to spot. When it comes to authority, Magnus mentions that in, for instance, an online community, members will, over time, identify those members they trust and those they do not, or they will trust the recommendation when a member vouches for another member (80). McGrady (2009) associates authority in Wikipedia with adhering to and being able to correctly cite its policies and guidelines rather than with the editor's expertise.
Other studies, such as this one, investigate people's actual use of Wikipedia; to what extent and when people trust it and, in turn, how this guides their use of the database. Contrary to the philosophical approach, this latter, user-focused research, is often less concerned with whether or not Wikipedia should be trusted and more interested in whether or not people do trust it and how this influences their practices involving Wikipedia (Luyt et al. 2008). The use of Wikipedia has been investigated from a credibility perspective with a focus on how people, often from specific user groups, such as young people, speak about it and use it. However, we have not identified any previous studies specifically concerned with how contributors to Wikipedia make credibility assessments. Several studies (e.g., Sundin and Francke 2009; Luyt et al. 2008; Rieh and Hilligoss 2008; Head and Eisenberg 2010; Lim 2009) have shown that students often use Wikipedia strategically to get an overview of a topic, but that they are hesitant to use or refer to it in situations where they need to be certain of something, for instance in school assignments, where teachers are often critical of the use of Wikipedia as a source. A bit surprisingly, the study by Luyt et al. (2008) indicated that the young users mainly used Wikipedia for school purposes rather than for interests concerned with their extracurricular interests. The students were generally aware that Wikipedia could be edited 'by anyone', and so took that into account when they consulted it (Sundin and Francke 2009; Rieh andHilligoss 2008; Head and Eisenberg 2010; Lim 2009). The students also compared claims from Wikipedia articles with claims from other sources to determine if the claims were credible. In their analysis of the credibility of a Wikipedia article, few students used Wikipedia's architectural design to consult, for instance, talk pages or an article's history (Sundin & Francke 2009). Even though the students made frequent use of Wikipedia, they claimed to place more trust in print sources (Sundin and Francke 2009), which corresponds to findings on how young people relate to other types of web sites (e.g., Hilligoss and Rieh 2008: 1475 f.)
Lastly, although not directly focused on the credibility of Wikipedia, findings of great interest to how editors act with regard to credibility can be found in studies that have looked into editors' incentives to contribute to Wikipedia (e.g., Forte and Bruckman 2005; Bryant, Forte and Bruckman 2005; Luyt, Suan and Linus 2009; Johnson 2008). This research can tell us something about how frequent contributors to Wikipedia work and what role the Wikipedia community plays for them. Forte and Bruckman compared the incentives of wikipedians to those of scientists and found that ' The notion of credit exists in Wikipedia both as reward and as credibility that empowers individuals in the community' (2005: ), but that accruing credit was not a matter only of meritocracy based on contributing to articles, but also of becoming a well-known figure in the Wikipedia community through other means. Their interviews indicated that contributors to Wikipedia often recognised each other (). Bryant et al. studied contributors' move from novice to expert participants in the Wikipedia community. One of the things they noted was how contributors would come to use more of Wikipedia's architectural features, such as talk pages, article history, and watch lists, as part of their routine activities as they became more active members of the community (2005: 6). Further, they mention that one ' convention that is understood by wikipedians but not by novices is that anonymous contributions are inherently suspect' (2005: 8), indicating the importance of maintaining a user account and signing one's contributions on talk pages even if the user account is anonymous. This is also emphasised in Behavior Guidelines on Wikipedia (2005: 8).
The study is part of a larger study inspired by online ethnography which attempts to understand several aspects of Wikipedia editors' activities on Wikipedia as viewed from the editors' perspective, as well as how they express their views on credibility issues and referencing on Wikipedia as part of their everyday life. This paper builds primarily on qualitative, semi-structured interviews, but has in some cases been supplemented with studies of relevant Wikipedia policies and guidelines.
The study participants were qualitatively identified among editors active on Swedish Wikipedia. Swedish Wikipedia was selected as an environment partly because of convenience and partly because contributors to it, as well as to other mid-sized language versions of Wikipedia, have so far been in focus only for few academic studies. The participants belong to the group of active Wikipedia editors who regularly contribute on various topics. They range from high school students to people in their fifties and all but one participant were male. At the time of the interviews, the eight participants who answered this question had held a Wikipedia account for between one and five years. More than half of the study participants are also contributors to the English or, in a few cases, other language versions of Wikipedia. All but one of the participants identified themselves as wikipedians, by which they meant that they belong to a community of Wikipedia contributors. Sometimes this includes going to Wikipedia meetups, sometimes a general commitment to Wikipedia's ideals and objectives, and sometimes committed work with various Wikipedia activities. The participants were found on articles on controversial topics, as well as on the Village Pump, which is the discussion forum for general Wikipedia issues concerning policy, technical issues, and operations. They were selected qualitatively in order to get participants with different experiences and interests. Prospective participants were contacted by e-mail through their user pages.
Interviews were conducted with the participants either through conversation (recorded Skype sessions) or by e-mail. Two participants elected to be interviewed by e-mail in order to retain the anonymity they maintain on Wikipedia. These participants were e-mailed a number of questions which they replied to quite extensively. All in all, eleven wikipedians out of twenty who had initially been contacted agreed to take part in the study. They were all informed of the purpose of the study and their right to withdraw from it, and gave informed consent to participate. Each semi-structured interview lasted between 45 and 70 minutes and took place in early autumn of 2009. A written summary of each interview was constructed and parts relevant to the topic of the study were transcribed word-by-word. The quotes used in the paper have been translated from Swedish. The participants' user names have been replaced by the letters A to K so as to make them anonymous.
As a complement to the interviews, the Wikipedia policy documents of relevance to credibility issues were included in the study to provide a point of reference in terms of the policies and guidelines the editors act in relation to. These include Verifiability ('Wikipedia:Verifierbarhet' 2010), Neutral point of view ('Wikipedia:Skriv från en neutral utgångspunkt' 2010), and No original research ('Wikipedia:Ingen originalforskning' 2010).
From the interviews, critical excerpts concerning the use of sources and credibility assessments were identified. The interview excerpts were categorised and analysed in terms of the role from which the participant is speaking at a given moment (reader, fixer, author, intermediary); how the participants read or use Wikipedia in various situations; what they consider when assessing the credibility of a Wikipedia article; what they consider when assessing the credibility of another source they are considering referring to or using facts from when editing a Wikipedia article; and what assessments they make as editors with regard to other editors' work. The findings were then considered in relation to previous research.
Many situations arise where contributors to Wikipedia need to consider trustworthiness in relation to Wikipedia articles. This takes place both in their roles as editors, when they need to make judgments as to what is a credible source to refer to or what should be kept in an article, and in their roles as readers, where they use Wikipedia articles for purposes outside Wikipedia. In the interviews, these positions merge and the participants often do not seem to distinguish them from each other. If anything, when the participants speak of credibility issues, they often do so from a position as reader or intermediator. For instance, several participants mentioned aspects of Wikipedia they think should be discussed in schools. A number of different and complementary strategies to assessing credibility can be identified in the participants' descriptions.
Wikipedia is often presented as an environment where it does not matter who you are: its tag line announces that ' anyone can edit', and can do so anonymously and collaboratively. It is therefore somewhat surprising that the author function (to speak with Foucault 1977) comes across as so important for the wikipedians. Many of the study participants considered contributions made by established contributors to be more credible than those by anonymous or new contributors.
But I guess it's mainly the history that you look at to decide who's written the latest version. If it's an IP number, well, then you might check what's been changed so that it's not some kind of vandalism or something that biases the article,...
Interviewer: So if it's an article with lots of IP numbers you get a bit more cautious?
Absolutely. And when it's user names you don't recognise or completely new users you should be a bit cautious. (I)
The participants had generally formed a sense of who the active editors were and who among them were trustworthy. Most of them trusted their fellow wikipedians:
...it's absolutely the case that I trust certain contributors fairly much or very much. So if I see that it's one of them who's written the article or the majority of the article, then I generally rely on it to be credible. Even if sources or footnotes haven't been added for all the details. So, yeah, it's very important for my own assessment. Absolutely. (D)
However, some mentioned that there may be reason to be cautious:
...some contributors seem to write mainly from their own point of view or to cause trouble. But I think that I trust most of the regular contributors to make good edits, because they're contributors that I consider to be serious ...(B).
The (mis)trust described above seems mainly to be something a contributor earns over time, through high quality contributions to the community.
The topic of authority was the subject of a debate on the Village Pump at the time of the field study (Wikipedia:Bybrunnen 2010). It concerned whether Wikipedia should be viewed as an open area for as many contributors as possible, or as a more controlled area where all contributions are checked by trusted editors. In the discussion, the German language version of Wikipedia often stood as a model for the latter view. Among the study participants, there were those who represented the former position, that there is no need in Wikipedia to fall back on traditional credentials for authorship – the task of a Wikipedia editor is to identify and summarise knowledge claims.
You don't have to be educated to be knowledgeable and be well-read. We mustn't forget that you're supposed to refer to others on Wikipedia and not write based on what you know. That's why I don't agree at all with those who think we should try more actively to involve academics on the wiki. (J, e-mail)
One of the participants (A) highlights the difficulties that arise in determining which claims to use when writing a Wikipedia article on a topic where there are conflicting claims. When this occurs, he tries to determine which is the more authoritative source, both in terms of author and of publication history. The authority of authors and sources thus plays an important role for many of the study participants.
The main tool the participants use for identifying contributors to an article is the article's history page. Here they can see who has made what changes, and also how frequent the changes have been. Two of the participants (H and I) mention the number of changes to an article and the number of changes since a particular claim was introduced as something they may draw upon in determining the credibility of the claim in articles on controversial topics. Another help when the participants encounter an article on a controversial topic are the talk pages: 'Sometimes, if it's a controversial topic, obviously, you also check the talk page to see what's been discussed' (I). If there is much activity and a heated debate on the talk page, that is perceived as a sign to be critical in one's reading. However, not all topics invite this type of scrutiny.
Certain topics can be presumed to be more controversial than others. One participant stressed that the topic influences his assessment of an article, so that he takes into account the level of controversy and his own expertise in the area, his possibility to identify false claims or bias.
I think I place quite a lot of trust in Wikipedia within the areas where I use it... I think I also intuitively trust articles on language, geography, historical persons, etc. which in themselves are generally uncontroversial. On the other hand, I'm more sceptical when it comes to politics etc. (K, e-mail)
Another editor noted that he considers 'if it agrees with what I already know' (A). Basic subject knowledge is often required to identify a subject as controversial. For instance, some historical and political persons may be controversial whereas others are not, which was also noted by A. More in-depth subject knowledge is needed to spot problematic passages (cf. Duguid 2006).
Another strategy for assessing credibility employed by the participants is comparison. On the one hand, they compare articles on the same topic in various language versions of Wikipedia, and on the other hand, they compare Wikipedia articles with other sources. In the interviews, there are several mentions of how the Swedish Wikipedia articles are compared to those in the English Wikipedia and, to a lesser extent, to the German version. The architectural feature (cf. above) of inter-wiki links between articles in different languages make such comparison very easy.
I like to look at English and German Wikipedia to see if they include these facts as well. And if they do, you can take that as, not a guarantee, but at least an indication that it's somehow correct. So a certain part of my own credibility assessment as a writer, sometimes you look at another language version and if there are enough versions that include the same facts, I think you can probably trust it. (F)
Several participants mention that they trust English Wikipedia more than the Swedish language one, implying that the greater number of contributors results in better developed articles. Some also turn to the English version to find references, because they find more references there than in the Swedish articles.
Traditional encyclopedias also play an important role for many participants for comparison and for providing input when writing articles. One participant discusses this in relation to two Swedish encyclopedias:
On Wikpedia, you often have a look at Nationalencyklopedin [the latest comprehensive commercial Swedish encyclopedia] ...and what they say about a topic. Such as, have we missed anything, or, what do the experts generally think about the issue. Wikipedia uses lots of stuff from the old Nordisk familjebok [a comprehensive Swedish encyclopedia from the late 19th and early 20th century]... at least you're aware that it's from a source that's been around for some time, with a different perspective on certain things, but it can still be useful because it's comprehensive in many areas. (A)
How a particular subject is represented in authorised, commercial encyclopedias influences what is considered canonical knowledge. In this context, even encyclopedias that are to some degree out-of-date are sometimes considered useful, particularly since they are no longer covered by copyright.
Several of the participants mention how important they think it is to cite the original sources: ' I generally think that all claims should be supported by sources… ' (K, e-mail). Many also use the references in Wikipedia articles as a starting point to find more information on a topic. This brings us to the issue of how the participants use Wikipedia articles when they themselves are looking for information.
The role Wikipedia plays in the participants' information seeking practices depends on the topic, as discussed above, but also on the situation in or purpose for which they need the information and on the norms and conventions in the network of practice where it will be used. Many of the participants hesitate to rely on Wikipedia for information if it is critical that the information is correct, although some compare the claims found in Wikipedia with those in other sources. They do, however, use Wikipedia extensively for gaining a general overview of a topic and for finding references to other sources they may pursue. One participant, who is a university student, explains that Wikipedia is not a source he would use at university, as it is not a respected source in that environment:
...people generally have very different opinions about Wikipedia; if it's a good source or not, so I hesitate to use it as a source... Even if I might use it to get information, I might try to find another source when I add references... If I'm just curious about something – maybe not for a paper or anything – I guess I trust Wikipedia quite a bit. (B)
The university is an environment where the students are expected to follow the conventions of a particular network of practice, and Wikipedia is not perceived to be an accepted source in this network. On the other hand, another participant mentions that he often uses Wikipedia at work to find information within his specialty and that he has never had any problems with inaccuracies (G). The situation, but also previous experiences, are thus of great importance for how seriously the participants consider an article's credibility.
The following quote sums up many of the strategies for assessing credibility on Wikipedia employed by the study participants:
The first thing is to find out if there are sources. That doesn't mean that an article without sources is no good. And see if it agrees with what I already know. Then, I don't know, I guess you look at who've done the writing and if you know whether they usually write OK articles. Check the talk page and see how much there is, and if there's been a lot of controversy... Mostly, I don't look at Wikipedia to find something if it's really important that it's correct. It's often a good place to start, but if I really need something to be correct, I either start looking there to get a general overview, or to follow links, or figure out if I've got anything else that's more credible that I can use as well. (A)
The way the participants report that they use Wikipedia, to get background and introductory information on a topic and to use as a pearl for finding references to sources they may pursue, is similar to the use previous studies have shown among students who are generally not Wikipedia editors (Sundin and Francke 2009; Rieh and Hilligoss 2008; Luyt et al. 2008). Although a few of the participants were members of both these networks of practice, students and editors, this was not the case with all participants. The situation and the purpose for which the information is to be used has importance, and both students and Wikipedia editors refrain from using Wikipedia without comparing it to other sources if it is important that the information is correct or if they think the network of practice in which it will be used will have a critical attitude towards Wikipedia (as in school reports). In what they consider less serious situations, however, several of the participants are prepared to trust Wikipedia articles. Comparison with other sources is something that is often mentioned, and can be understood in terms of assessing the reliability of a claim rather than the authority of a source (Lankes 2008).
At the same time, there are notable differences in how the editors assess credibility compared to other researched groups. The concept of network of practice is useful in bringing into focus these differences, because they spring from the editors' knowledge of the system and the community, knowledge they have developed as members of a wikipedian network of practice. This comes through both in how they relate to the author function and how they use the MediaWiki architecture and affects both their work as editors and their information seeking on Wikipedia. Even though a socio-cultural perspective leads us to assume that the purpose for which a participant considers credibility in Wikipedia is of importance, it is likely that the editors also draw on their experience as editors in their role as readers. Many of the study participants express that through the time spent on Wikipedia, trmed opinions of the other editors based on their edits and participation in discussions. They have learnt, for instance, which editors adhere to the Wikipedia policies and which editors argue a specific view. These opinions provide them with intellectual tools that help them make credibility assessments of the work put in by particular editors on an article – tools generally not available to somebody who is not a core member of the wikipedian network of practice. The fact that Swedish Wikipedia is a smaller environment than some other language versions, with fewer very active editors, may have made it easier for the participants to recognise other editors. It is, however, not unlikely that this also happens on larger Wikipedia versions, at least on particular articles (cf. Forte and Bruckman 2005: ). A contribution by an unknown contributor, on the other hand, is a sign to pay attention, something that was also found by Bryant and colleagues who noted that wikipedians were generally suspicious of anonymous contributions (Bryant et al. 2005: 8).
Another aspect that is learnt as one becomes a member of the wikipedian network of practice is the use of the material tools, such as Wikipedia's architecture (Bryant et al. 2005: 6 f.). Several of the participants, who are familiar with the structure and architecture of Wikipedia, use it to identify paratexts (Genette 1997) of various types, such as an article's history, talk page, and the same entry in other language versions of Wikipedia, that will aid them in making their credibility assessments. Talk pages are valuable for articles on potentially controversial topics where the discussion around the article's content and form may indicate if one should exercise care. In this case, some notion of what is potentially controversial is required beforehand. The talk page is also a forum for discussing the credibility of references and other sources used to verify claims made in the article. The history of the article is used to find out who has edited the article and when this has taken place. Such use of an article's architecture was found quite rarely among students (Sundin and Francke 2009) but was frequently mentioned by the editors.
The participants belong to several different networks of practice, which guides their interaction with Wikipedia in different situations. In some cases, the network of practice of wikipedians may not be the main one influencing their actions. However, the participants' roles as editors on Wikipedia guide their credibility practices so that they consider cultural tools that are less commonly applied by other users. Furthermore, their participation in the wikipedian network of practice influences both their credibility activities as editors and administrators on the site and their assessments of credibility when they use articles on Wikipedia for information seeking or entertainment in practices outside Wikipedia editing.
When wikipedians make credibility assessments, both in their roles as editors and as readers of Wikipedia, they use tools that to some degree concern what Metzger and Flanagin (2008) term conferred and reputed credibility. These are strategies that are also common in assessments of credibility in traditional media and that are used by other groups in relation to web sites and participatory information environments. Much research has been concerned with trying to design tools that facilitate tabulated credibility, but this is not something mentioned by the study participants. They do sometimes take the number of edits of an article into consideration and in that case consider the power of the collective, but this is often pushed to the background compared to the trust they place in authors and references to authoritative sources. The emergent credibility that is often associated with participatory information environments and Web 2.0 is thus downplayed compared to more traditional tools for assessing credibility. The view of knowledge creation expressed by the participants is not as radically postmodern as has sometimes been suggested about contributors to participatory media (cf. Haider and Sundin 2010). Whether or not this is particular to Wikipedia compared to other participatory information environments, to this particular time period, or to Swedish Wikipedia compared to other language versions, are interesting issues to continue to investigate.
That a reliance on authoritative and published sources is not particular to the editors who took part in this study is shown by the fact that authoritative and published sources are emphasised in several Wikipedia policies and guidelines. Verifying claims and avoiding original research are part of an ongoing work to increase article quality that has taken place on several Wikipedia language versions over the past few years. The policies and guidelines are created by wikipedians. Learning about them and adapting to them is part of becoming a member of the wikipedian network of practice, of becoming a respected contributor to Wikipedia. We have not, as part of this study, looked at the potential role the study participants have had in influencing policy decisions, but it is clear that the policies form an important part of the identity that is embraced by the members of the network of practice. The practices concerning credibility assessments that are associated with knowledge about the policies and guidelines can be expected to take different forms among more peripheral members of the wikipedian network.
Based on our study, it can be concluded that both cognitive authorities and a reliability approach are important cultural tools among wikipedians for making credibility assessments; tools that are also common among other groups. Because of their membership in the wikipedian network of practice, however, authority is sometimes attributed based on an insider's knowledge of who is a credible and competent author, a knowledge not appropriated by non-members of the wikipedian network of practice.
We wish to warmly thank the study participants for sharing their experiences with us. This study was conducted within the project Expertise, Authority and Control on the Internet (EXACT) which is funded by the Swedish Research Council. The project is part of the Linnaeus Centre for Research on Learning, Interaction and Mediated Communication in Contemporary Society (LinCS) at the University of Gothenburg and the University of Borås.
Helena Francke is a Senior Lecturer at the Swedish School of Library and Information Science, University of Borås, Sweden. She received her PhD from the University of Gothenburg, Sweden. Her research interests concern the credibility and authority of sources, information literacy, document studies, electronic publishing, scholarly communication and knowledge organization. She can be contacted at email@example.com.
Olof Sundin is a Professor at The Swedish School of Library and Information Science, University of Borås, Sweden and a Senior Lecturer and Docent at the Department of Cultural Sciences, Lund University, Sweden. He received his PhD from the University of Gothenburg, Sweden. His research interests concern information literacy, the credibility and authority of sources and information practices. He can be contacted at firstname.lastname@example.org.
|Find other papers on this subject|
© the authors, 2010.
Last updated: 5 September, 2010