Avatar Experimentation: Human Subjects Research in Virtual Worlds

I have just posted a (rough) draft of my latest paper, entitled Avatar Experimentation: Human Snapshot_002 Subjects Research in Virtual Worlds to SSRN.  Virtual worlds make such great research testbeds precisely because people act in a lot of ways (especially economic ways) as if the virtual world were real.  But that complicates ethical research design: you can't engage in activities that threaten the subject's digital property or community, for example.  This raises human subjects research issues that a lot of Institutional Review Boards may not immediately take into consideration.  Here's the abstract -- but the important part is that this is still a work-in-progress (it's coming out in a symposium issue of the U.C. Irvine Law Review next year), and I would love comments or suggestions.  Abstract after the leap.

Abstract: Researchers love virtual worlds. They are drawn to virtual worlds because of the opportunity to study real populations and real behavior in shared simulated environments. The growing number of virtual worlds and population growth within such worlds has led to a sizeable increase in the number of human subjects experiments taking place in such worlds.

Virtual world users care deeply about their avatars, their virtual property, their privacy, their relationships, their community, and their accounts. People within virtual worlds act much as they would in the physical world, because the experience of the virtual world is "real" to them. The very characteristics that make virtual worlds attractive to researchers complicate ethical and lawful research design. The same principles govern research in virtual worlds as the physical world. However, the change in context can cause researchers to lose sight of the fact that virtual world research subjects may suffer very real harm to property, reputation, or community as the result of flawed experimental design. Virtual world research methodologies that fail to consider the validity of users’ experiences risk harm to research subjects. This article argues that researchers who put subjects’ interests in danger run the risk of violating basic human subjects research principles.

Although hundreds of articles and studies examine virtual worlds, none has addressed the interplay between the law and best practices of human subjects research in those worlds. This article fills that gap.

 Virtual worlds are valuable research environments precisely because the relationships and responses of users are measurably real. The article concludes that human subjects researchers must protect the very real interests of virtual worlds inhabitants in their property, community, privacy, and reputation.

The article proceeds in five parts. After Part I introduces the scope of the piece, Part II explains virtual worlds and discusses why the marriage of social networking with three-dimensional videogame graphics complicates experimental design. Part III explores current and developing practices in virtual worlds research, as well as the various areas of law that bear on such research. Part IV outlines solutions and best practices for human subjects research in virtual worlds, and Part V offers a conclusion.

X-posted: Concurring Opinions

Comments on Avatar Experimentation: Human Subjects Research in Virtual Worlds:

Cunzy1 1 says:

I suggest a Part VI which makes all the badness of Part V better again.

Then a couple of years later just flood the market with Parts 'VII-IX', cheap spin offs of Parts IV and V which to be honest weren't that great in the first place.

Posted Dec 2, 2010 12:37:34 PM | link

ren reynolds says:

"Although hundreds of articles and studies examine virtual worlds, none has addressed the interplay between the law and best practices of human subjects research in those worlds. This article fills that gap."

Well actually, this is true till the next edition of the International Journal of Internet Research Ethics come out which should have a paper by me and Melissa de Zwart:
The Duty To ‘Play’: Ethics, EULAs and MMOs.

Which looks exactly at the relationship between research practice and law in the context of MMOs

Abstract: In the last ten years there has been a steady increase in the attention paid to Massively Multiplayer Online Role Play Games (MMOs) as a site of academic research. A number of large and influential academic studies have been undertaken in a range of disciplines inside MMOs such as EverQuest (Taylor 2006), Star Wars Galaxies (Ducheneaut and Moore 2004), Lineage I and II (Steinkeuhler 2004) and EverQuestII (Castronova et al 2009). Generally these studies require the researcher to be embedded in-world, participating as a player in the game. In this paper we ask: what bounds on research, if any, are ostensibly set by the contractual terms of service of MMOs by which all participants are bound– in particular the duty to ‘play’; whether this has any problematic bearing on the practice of research; and, if so, whether there are solutions to these problems? This paper will consider the contractual terms used in the End User Licence Agreements (EULAs) of two MMOs, EvE Online and City of Heroes, in order to further explore the practical implications of those terms, This paper concludes that researchers need to consider whether their research is ethical, both in terms of formal compliance with Institutional Research Board (IRB) requirements and how it may impact upon or disrupt the player community. If it does have a disruptive impact, this may take the research not only outside the notion of ethical research but also outside of the concept of ‘play’ authorised by the EULA itself.

- great minds and all that :)

Posted Dec 3, 2010 3:45:22 AM | link

ren says:

Gah, OECD internet just ate my last comment.
So, again – in the context of the Association of Internet Researchers recent debate on research ethics, I’ve been arguing two things:

First - IRB’s frame actors as ‘subjects’ and often state that they should be anonymous. However, it seems to me that IP law would often frame the same actors as Authors and thus there would be a legal requirement to respect their Moral Rights. What’s more in many cases the actors want to be identified as the creators of a given work. Thus in supposedly protecting people, it seems to me, that IRB’s may be guiding researchers to breach IP law and individuals legitimate wishes.

Second – it also seems to me that IRB’s tend to adopt local ethical norms of the IRB / Researchers. However given the geographical spread of actors it’s arguable that local ethical norms should be given at lest consideration and probably primacy in many cases.

Posted Dec 3, 2010 4:05:35 AM | link

好秘书 says:

Well actually

Posted Dec 13, 2010 2:13:51 AM | link