Abstract

Systematically biased editing, persistently maintained, can occur on Wikipedia while nominally following guidelines. Techniques for biasing an entry include deleting positive material, adding negative material, using a one-sided selection of sources, and exaggerating the significance of particular topics. To maintain bias in an entry in the face of resistance, key techniques are reverting edits, selectively invoking Wikipedia rules, and overruling resistant editors. Options for dealing with sustained biased editing include making complaints, mobilising counter-editing and exposing the bias. To illustrate these techniques and responses, the rewriting of my own Wikipedia entry serves as a case study. It is worthwhile becoming aware of persistent bias and developing ways to counter it in order for Wikipedia to move closer to its goal of providing accurate and balanced information.

Introduction

Wikipedia is perhaps the best-known product of cooperative voluntary work (Jemielniak, 2014; Lih, 2009; Reagle, 2010). Through the efforts of thousands of contributors, it has rapidly outgrown traditional encyclopaedias in size and influence. Anyone who regularly searches the web for information is likely to see numerous links to Wikipedia entries. All sorts of people, from students to people with health problems, rely on Wikipedia for information. Wikipedia’s ascendancy was achieved within a few years of exceptional growth, based on contributions from unpaid and unheralded editors.

Yet Wikipedia has had plenty of teething problems (Lovink & Tkacz, 2011). On some controversial topics, such as abortion and George W. Bush, there have been edit wars, with committed editors seeking to impose their viewpoints (Lih, 2009, pp. 122-131; Yasseri et al., 2012). There are trolls and vandals who, for various reasons, seek to deface well-written entries. There are covert editing efforts to shape the portrayal of individuals, organisations and topics, for example when paid workers edit entries about their employer or client (Craver, 2015; Thompson, 2016). Allegations have been made about systematic bias on certain topics, for example parapsychology (Weiler, 2013: 152-183), and about a range of other problems (Wikipediocracy, 2017).

Wikipedia has instituted various measures to address problems. To interrupt editing wars, versions of some entries are locked down. To fix edits by trolls and vandals, various bots patrol entries, alerting administrators to suspicious changes (Geiger, 2011). Although nominally Wikipedia is an egalitarian enterprise in that anyone can be an editor, in practice admins have a lot of power, and some have more power than others (O’Neil, 2009).

Wikipedia has detailed sets of guidelines about various matters, for example neutral point of view, vandalism, disruptive editing, and biographies of living persons (Reagle, 2010). If these guidelines were always followed, Wikipedia would be remarkably free of problems. The real challenge is ensuring they are implemented in practice. All sorts of organisations, including governments, corporations and churches, have noble-sounding aims and rigorous rules, but these sometimes provide a façade for corruption and abuse. Ironically, Wikipedia rules are often used less to resolve disputes than as tools in waging editing struggles (Tkacz, 2015, p. 99).

Because of the possible discrepancies between rules and practice, the test of the performance of Wikipedia is through examining actual practice. There are various ways to go about this, for example looking at the processes by which vandalism is rectified (Geiger & Ribes, 2010) and at the bias in entries on political-party relevant topics as revealed by word use (Greenstein & Zhu, 2012). Because Wikipedia is so huge and in multiple languages, any attempt to look broadly at patterns of inaccuracy and bias is overwhelming in scale.

My aim here is to indicate some of the techniques of biased editing, ways of maintaining it in the face of resistance, methods for probing it and options for responding. My aim is not to estimate the prevalence or seriousness of bias but rather to show how it can be instituted and opposed.

To illustrate techniques and responses, I use a particular Wikipedia page as my primary example: the entry about myself in the first half of 2016. My central purpose is to illustrate methods for imposing and maintaining bias. Entrenched bias on some other pages is far more extensive and serious than the treatment of my page. My page is convenient for analysis because the volume of data is smaller and the trigger for rewriting is obvious.

Given that all claims can be disputed and that there are legitimate differences of perspective, a purely neutral presentation is only an aspiration or guiding principle, not an achievement. It might be said that every Wikipedia editor is biased in one way or another, usually unconsciously. However, by being challenged by other editors, with different views, the more extreme sorts of bias may be overcome, or at least that is the assumption underlying Wikipedia’s operations. The problem I address here is not the bias of individual editors, which is predictable, unavoidable and usually unconscious rather than intentional, but the systematic and sustained imposition of a particular point of view in the face of reasonable objections, using techniques or creating outcomes that are not regularly observed in other Wikipedia entries. This sort of imposed bias might also be called advocacy or partisanship, or even propaganda or disinformation, though these latter terms have connotations of government manipulation of the facts. Commercially inspired biased editing might be considered a form of covert advertising. The essence of the sort of bias addressed here is concerted and sustained promotion of a particular point of view, overruling objections.

In the next section, I list a number of methods that can be used to bias an entry, using examples from my own entry to illustrate them. In the following section, I look at methods to maintain bias in the face of resistance, concluding with a comparison between my Wikipedia entry and several others. Finally, I outline six options for responding to particular cases of persistent bias and present a general suggestion for revising Wikipedia policies.

Source: Social Science Computer Review, Vol. 36, No. 3, June 2018, pp. 379-388; doi: 10.1177/0894439317715434