Ethical issues arise with research on Internet

How should IRBs handle virtual worlds?

One of the strangest new areas of research ethics involves how IRBs should handle research that involves Internet communities, including virtual communities.

IRBs have some experience and background in handling online surveys or assessing protocols that involve the collection of data available over the World Wide Web. But what do they do when an investigator proposes to visit an online world to study the community as an avatar?

"There are people behind these avatars, and a lot of times researchers are not that sure about who they need to obtain consent from," says Montana Miller, PhD, an assistant professor at the Bowling Green State University in Bowling Green, OH.

For instance, there's an online world called Second Life in which people can create a person who has a job, house, family, and friends. Some visitors will spend hours in this imaginary community, closely identifying themselves with their avatar.

"A lot of researchers feel it's too daunting to even try to get informed consent so they will avoid the whole issue and convince themselves it's just an avatar and not a person so it's not human subjects research," Miller says. "And sometimes even an IRB is confused and says it's not human subjects research."

But the ethical reality is more nuanced. Virtual worlds like Second Life advertise themselves to be places for people to connect, explore, have adventures, and find love. While a researcher might be detached when visiting these communities, chances are that other people are more personally engaged in the recreational activity.

"It may not seem like you're studying a person because you're looking at a computer screen," Miller explains. "However, research shows that when you do something to someone online, whether it's to support or bully them or some other interaction, then what happens on that computer screen has a real effect on the person behind that computer screen."

IRBs and investigators should keep in mind that these virtual avatars represent real people, real feelings and real harms, she adds.

"Strangely enough, that is something researchers seem to forget or ignore," Miller says. "Even when you're typing things out through text or even when it doesn't feel real, it may feel real to them."

Studies show that when someone's avatar is hurt, the person also feels that pain emotionally, she adds.

It's similar to the experience of "flaming" someone online, Miller says.

Internet flaming refers to deliberate acts of posting messages online with the intent of insulting and creating havoc with specific people or groups. Sometimes this act can be personal and a form of bullying, and sometimes it's meant to be provocative and directed at a general group or community.

"Flaming someone is unprovoked and nasty attacks that can make someone feel absolutely terrible," Miller explains. "You'd think, 'Why am I so upset when I don't even know this person?' But it really does have an emotional effect."

Here is how a researcher might unknowingly cause someone similar emotional distress: "Suppose you do an experiment where someone walks up to a person in Second Life as an avatar and has some interaction with the person. Then the researcher turns around and walks away," Miller says.

"That may seem innocent and benign, but to the person in Second Life that may be shocking or disturbing to have a stranger walk up to them and say something and then walk away," she adds. "Researchers have to take responsibility for how their behaviors as an avatar in a virtual world impact other people."

As IRBs review this type of research, they should consider how the researcher will represent himself or herself to the online community.

"This is a community of people who often are misunderstood by the general public, and there are issues of confidentiality and identity," Miller says.

For instance a researcher might decide to visit Second Life to do a study of romantic relationships. While conducting this research, the investigator sees avatars connecting in virtual love affairs with other avatars. These avatars are represented by people who have committed relationships in the real world, so their virtual affairs could be seen as cheating by their real-life partners, she explains.

"A researcher could do a study in which he or she unintentionally exposes identifying information, which could have an impact on their real lives," Miller says.

"Some researchers say they are just using the avatar's online handle or screen name so they're not really identifying them," she adds. "But someone could track that down through all these other places on the Internet where that screen name is used, and that's why a screen name is an identity."

The key is to change the screen name or omit it entirely from the study.

"IRBs and researchers need to figure out just how rigorous the consent needs to be," Miller suggests. "In some cases it may involve actually being able to contact the real person, give the person a real consent form."

In other cases, IRBs might decide to waive written consent, but require investigators to identify themselves as researchers and provide a public information sheet to those involved in the study.

"It depends on different factors, such as whether the online community is a public or private space, and what is the topic sensitivity," Miller says. "And how much of an intervention is the investigator doing? Is the investigator just observing or is he or she asking questions and really doing an experiment involving the avatars?"