Research:Value of IP Editing

From Meta, a Wikimedia project coordination wiki
Jump to navigation Jump to search

This page summarizes research on the value of IP editing. The page was written in 2019 with the goal of informing the discussion of a proposal to make changes to the way that users without accounts are tracked when they contribute to Wikimedia projects in order to hide the IP addresses currently used to identify contributions from these users. A major theme of community responses to that proposal are suggestions to disallow contributions from users without accounts. Because a majority of vandalism stems from unregistered IP editors, the argument for prohibiting unregistered editing is that requiring accounts would reduce this damage.

This page was constructed by a group of scholars and researchers who have studied Wikipedia and its contributors. A large body of evidence developed by scholars and Wikipedians–including, but not limited to, ourselves–suggests that requiring an account to contribute to Wikipedia would result in enormous collateral damage. We believe that the negative consequences of blocking edits from non-accountholders would far outweigh the benefits.

Evidence for this assessment falls into the following categories:

  1. IP-based edits are valuable
  2. Blocking IP edits may discourage newcomers from making their first edit
  3. Contributing via IP opens a pathway to deeper contributions
  4. Past trials of this idea were damaging
  5. Other considerations

We describe research that informs our position in a corresponding set of sections below.

IP-based edits are valuable[edit]

A large body of observational research into Wikipedia content quality over the last 15 years suggests that, although unregistered users are the source of most vandalism and registered users contributions are higher quality on average, an enormous amount of valuable contributions to Wikipedia flows from non-registered users.

  • IP-based edits are often high quality. Javanmardi et al. [1] use a content persistence measure called WikiTrust to just content quality. They find that although "registered users contribute higher quality content and therefore are assigned higher reputation values ... a significant number of anonymous users also contribute high-quality content."
  • IP-based edits are often kept. Anthony et al. [2] use data from French and Dutch Wikipedia and measure a "retention rate" (a different content persistence measure that similarly captures quality and productivity). They compare four groups of more and less experienced users with and without accounts. They find that first-time IP editors's contributions result in more retained content than the other four categories.
  • IP-based edits are of similar quality to the first edits made from users with accounts. Tran et al. [3] compare IP editors to several other groups of editors. They find that contributions from IP editors are of similar quality to those by users with accounts making their first edits in terms of (a) revert rates, (b) manual inspection and classification of revisions in terms of whether edits were damaging and made in good faith, and (c) assessment using ORES.
  • In indirect ways, even low-quality contributions (from IP editors and others) can be valuable. A study by Gorbatâi [4] suggests that low quality contributions by IP editors has an important effect by catalyzing additional high quality contributions by established editors.

Blocking IP edits may discourage newcomers from making their first edit[edit]

Even in ideal circumstances, account-creation takes time. If what prompted someone to click 'Edit' is a small typo or missing detail, even a very low barrier to entry can be enough to discourage participation.

  • Although it might be small, the requirement to create an account is a transaction cost—i.e., a cost over-and-above the necessary time and cost of producing the edit itself. Wikipedia's very low transaction costs have been credited by a wide range of scholars including Yochai Benkler who has frequently cited editing without the need to create an account as a contributor to Wikipedia's success (see Wealth of Networks)[5][6].
  • In Hill's study of online collaborative encyclopedia projects that preceded Wikipedia, the ability to edit without an account was put forward as one of the reason that Wikipedia succeeded where others failed.[7]

Contributing via IP opens a pathway to deeper contributions[edit]

  • Many people who go on to register do contribute via IP as described in this volume and impact study; studies of these individuals are often not possible for non WMF researchers due to privacy controls (post-registration IPs are not made generally visible).
  • The pattern of contributing via IP before creating an account has been studied in other contexts. A study by Jackson et al. [8] shows that editors on the citizen science project Zooniverse often edit for long periods of time without accounts and that they do so to "try out" participation and contribution.
  • A large body of research in social computing has documented the way that users go from more peripheral forms of participation to more engaged participation. This general pattern has been observed in many online communities, including Wikipedia. [9]
  • Bryant et al. found that being able to easily engage in contributions without first making an account is one of the important ways in which the process of transitioning from being a reader to a contributor occurs.[10]

Past trials of this idea were damaging[edit]

The strongest evidence against the proposed block of IP editors is that it has been tried before in similar contexts and has been shown to result in a decrease in quality contributions. We know of two attempts to test similar changes and in both cases the evidence suggests enormous collateral damage.

  • On Wikipedia itself, User:Halfak (WMF) conducted an experiment on asking anonymous editors to register. The experiment asked users to register an account before contributing but did not require it. The prominence of buttons allowing users to opt-out was varied in different versions of the study. The results are summarized as asking for registration pre-edit "gathered far more new registrations (+200%) and new active editors than post-edit; however, overall productivity fell by 25% due to the decreased probability of completing edits." Making it more clear that users could continue editing reduced the damage, but still led to a 15% decrease in productivity.
  • Using data from Wikia/Fandom, Hill and Shaw [11] presented work at Wikimania 2015 on using evidence from about 140 Wikia/Fandom wikis that blocked IP edits. They found that although requiring accounts was effective at reducing vandalism and low quality editing, it was also associated with a 20-40% decrease of quality contributions as measured in both non-reverted edits and persistent token revisions. Hill and Shaw also report blocking contributions from non-logged in editors causes a decrease in contributions of quality contributions from users with accounts. They suggest that the mechanism may be similar to the one described by Gorbatâi [4] whereby the cessation of IP-editing leads to less engagement by established users editing in response to these users' contributions.

Both studies suggest that some IP editors will create accounts (i.e., both show increases in the number of new accounts). That said, both also show decreases in quality contributions, suggesting that a large number simply forgo contributing at all when accounts are required.

Other considerations[edit]

  • IP editors are unlikely to speak for themselves in conversations like the current policy discussion, and it can be difficult for Wikipedians to imagine their perspective without deliberate effort. McDonald et al. [12] analyzed a history of conversations about "anons" on wikien-l and documented a lack of "perspective taking" by mailing list participants. When communities discuss participation from people who contribute without accounts, the community members may not be able to take on or reflect the perspective of the non-accountholding contributors. While these editors lacked sensitivity to the identity-related vulnerabilities that might make someone choose IP anonymity, most still argued in favor of allowing "anons." These editors felt IP editing to be a critical to lowering barriers for newbies who go on eventually to become "legitimate," registered users.
  • Some work has been done to understand why Wikipedians themselves sometimes seek to contribute to Wikipedia without creating an account. Forte et al. were able to interview Wikipedians about this, and found that their reasoning included such personal factors as avoiding bias created by their Wikipedia username, and preventing harassment and employment discrimination (e.g., if they are personally identifiable with their primary account, but want privacy for a given edit).[13] We think that consideration of these perspectives are in-line with the Wikimedia movement's spirit, mission, and goals.

Signatories[edit]

References[edit]

  1. Javanmardi, Sara; Ganjisaffar, Yasser; Lopes, Cristina; Baldi, Pierre (2009-12-28). "User contribution and trust in Wikipedia" (PDF). Proceedings of the 5th International ICST Conference on Collaborative Computing: Networking, Applications, Worksharing. New York, NY: ICST / IEEE. ISBN 978-963-9799-76-9. doi:10.4108/icst.collaboratecom2009.8376. Retrieved 2019-08-29. 
  2. Anthony, Denise; Smith, Sean W.; Williamson, Timothy (2009). "Reputation and Reliability in Collective Goods". Rationality and Society 21 (3): 283–306. Retrieved 2010-07-09. 
  3. Tran, Chau; Champion, Kaylea; Forte, Andrea; Hill, Benjamin Mako; Greenstadt, Rachel (2019). "Tor users contributing to Wikipedia: Just like everybody else?" (PDF). Working Paper. 
  4. a b Gorbatai, Andreea D. (2014-02-10). The Paradox of Novice Contributions to Collective Production: Evidence from Wikipedia. Rochester, NY: Social Science Research Network. Retrieved 2019-08-29. 
  5. Benkler, Yochai (2006). The wealth of networks: How social production transforms markets and freedom. New Haven, CT: Yale University Press. ISBN 0-300-12577-1. 
  6. Benkler, Yochai; Shaw, Aaron; Hill, Benjamin Mako (2015). "Peer production: a form of collective intelligence". In Thomas W. Malone, Michael S. Bernstein (eds.). Handbook of Collective Intelligence. Cambridge, MA: MIT Press. pp. 175–204. ISBN 978-0-262-02981-0. 
  7. Hill, Benjamin Mako (2013). "Almost Wikipedia: What eight early online collaborative encyclopedia projects reveal about the mechanisms of collective action.". Essays on volunteer mobilization in peer production (PDF). Cambridge, Massachusetts: Massachusetts Institute of Technology. 
  8. Jackson, Corey Brian; Crowston, Kevin; Østerlund, Carsten (2018-11-01). "Did they login?: Patterns of anonymous contributions in online communities" (PDF). Proceedings of the 2018 ACM Conference on Human-Computer Interaction. The 2018 Conference on Computer-Supported Cooperative Work. CSCW '18 2. Jersey City, NJ: ACM Press. pp. 1–16. doi:10.1145/3274346. Retrieved 2019-08-11. 
  9. Preece, Jennifer; Shneiderman, Ben (2009). "The reader-to-leader framework: motivating technology-mediated social participation" (PDF). AIS Transactions on Human-Computer Interaction 1 (1): 13–32. ISSN 1944-3900. 
  10. Bryant, Susan L.; Forte, Andrea; Bruckman, Amy (2005). "Becoming Wikipedian: transformation of participation in a collaborative online encyclopedia" (PDF). Proceedings of the 2005 International ACM SIGGROUP Conference on Supporting Group Work. GROUP '05. New York, NY: ACM. pp. 1–10. ISBN 1-59593-223-2. doi:10.1145/1099203.1099205. Retrieved 2015-05-09. 
  11. Hill, Benjamin Mako; Shaw, Aaron (2020-05-26). "The hidden costs of requiring accounts: quasi-experimental evidence from peer production". Communication Research: 0093650220910345. ISSN 0093-6502. doi:10.1177/0093650220910345. Retrieved 2020-06-25. 
  12. McDonald, Nora; Mako Hill, Benjamin; Greenstadt, Rachel; Forte, Andrea (2019). "Privacy, anonymity, and perceived risk in open collaboration: A study of service providers" (PDF). CHI Conference on Human Factors in Computing Systems Proceedings. CHI '19. New York, NY, USA: ACM. pp. 1–12. ISBN 978-1-4503-5970-2. doi:10.1145/3290605.3300901. 
  13. Forte, Andrea; Andalibi, Nazanin; Greenstadt, Rachel (2017). "Privacy, anonymity, and perceived risk in open collaboration: A study of Tor users and Wikipedians" (PDF). Proceedings of the 2017 ACM Conference on Computer Supported Cooperative Work and Social Computing. CSCW '17. New York, NY, USA: ACM Press. pp. 1800–1811. ISBN 978-1-4503-4335-0. doi:10.1145/2998181.2998273. 

Funding Disclaimer[edit]

Some of work done by Hill, Champion, and Greenstadt is based upon work supported by the National Science Foundation under grants numbers CNS-1703736 and CNS-1703049. Any opinions, findings, and conclusions or recommendations expressed in this material are those of the author and do not necessarily reflect the views of the National Science Foundation.