【无标题】

3Ethical, Legal, and Methodological Considerations in Cyber Research: Conducting a Cyber Ethnography of www.bullying.orgBarry E. CartwrightIntroduction

As Lee (2013) points out in Facebook Nation, we live in an age of “total information awareness.” With Google Street View, MySpace, YouTube, Facebook, Twitter, WikiLeaks, and recent revelations by Eric Snowden regarding the extent of the NSA’s surveillance activities, it may seem whimsical to talk about privacy rights, informed consent, and other ethical concerns when it comes to conducting research in cyberspace. Nevertheless, the debate concerning ethics in cyber research continues unabated, as it affects most social researchers whose activities are governed by institutional research ethics boards and by the standards of their professional associations (Graber and Graber 2013; McKee and Porter 2009).

Mixed methods in cyber ethnography

Cyber communities, like real-world communities, have shared norms and values, ongoing (sustained) communication, a sense of commonality and belonging, and an agreed-upon “netiquette” (Hine 2000). They also establish community boundaries, enforce community norms, and exclude outsiders or others who fail to conform. Thus, it is said that these cyber communities can be studied using traditional ethnographic methods – for example, observing the setting and the social actors, recording and interpreting the narratives and histories of the actors, and analyzing how norms, values, and community boundaries are established and understood (Jankowski 2006; Mcmillan and Morrison 2006).

The data set used in this cyber ethnography of www.bullying.orgconsisted of the first 475 messages posted on the website (as of 31 December 2002) and the last 475 messages posted on the site (as of 30 August 2006). This procedure captured two extended snapshots in time, providing an exact picture of what the messages on www.bullying.orglooked like at the end of 2002 and what the messages on a more recent portion of the site looked like toward the end of 2006. The 950 messages were coded and categorized in NVivo (a computer software package designed to facilitate qualitative research) and then analyzed further in the SPSS (the Statistical Package for the Social Sciences). This approach could be described as a mixed or multiple methods study (Tashakkori and Teddlie 1998) or as complementary data analysis (Sudweeks and Simoff 1999).

The messages downloaded from www.bullying.orgranged in size from a few words to over ten pages in length, with some being little more than brief commentary and others rising almost to the level of life histories. None followed a preset format – the authors of messages could say as much or as little as they wanted. In many cases, information regarding age, gender, location, and reasons for posting the message were provided by the author. In almost as many cases, this information could only be inferred from a careful reading and rereading of the messages, if it could be inferred at all. Most messages were read eight or more times from beginning to end, often side by side, for purposes of comparison, identification of significant patterns or themes, verification of the identity of author, and assessment for identity play.

The messages were subjected to content analysis because they had elements or attributes that were physically present and countable, and it was possible to summarize the entire set of messages numerically (Berger 2000; Bryman, Teevan and Bell 2009; Neuendorf 2002). The messages were subjected to discourse analysis and narrative analysis, as a number of them were fully developed stories or narratives, with beginnings, middles and ends, often recounting experiences that occurred many years ago or that took place over many years (cf. Silverman 2003; Titscher et al. 2000).

The sample included messages from bullies, victims, bully-victims, bystander-observers, school teachers and school principals, police and probation officers, parents of bullies and victims, other relatives of bullies and victims, plus teachers, parents, and other relatives who were formerly bullies, victims, or bystanders themselves. Three hundred and sixteen messages were posted by adults describing experiences from long ago (in some cases, more than 40 or 50 years ago), offering fresh insight into how little the quality or quantity of bullying has actually changed throughout the years. The messages came from all ten provinces and three territories in Canada, 22 states in the USA, ten countries in Western and Eastern Europe, seven countries in Africa, five countries in Central and South America, four countries in Asia or South East Asia, and Australia and New Zealand. Apart from illustrating the vast potential of cyber research (Kraut et al. 2004; Joinson 2005), the diversity of these messages attests to cross-cultural and transnational concern with the issue of bullying.

Copyright law and cyber ethnography

The issue of copyright was raised by the university library in response to an innocuous question about how to format and acknowledge text messages taken word for word from www.bullying.org. The library insisted that these messages could not be used “as is” without express written consent of the site owners/operators, and that in the absence of such consent, the research findings would likely be unpublishable due to copyright infringement. To buttress their position, the university librarians pointed to the copyright symbol on the home page of the site and asserted that they were relying upon a previously obtained legal opinion from a Canadian copyright lawyer.

To preclude the possibility of copyright infringement, advice was sought from Michael Geist, who holds the Canada Chair in Internet and E-Commerce Law at the University of Ottawa’s Law School. A written legal opinion was also obtained from Ann Carlsen, a Canadian lawyer who specializes in trademark and intellectual property law. According to the written legal opinion, the copyright symbol did not imply ownership of the messages, or for that matter, ownership of anything, unless copyright in fact existed. Site owners commonly display a copyright symbol or notice on their home page, but this usually indicates that copyright exists for the site itself. It does not imply that the site owners hold the copyright to the personal messages or stories that are posted on their site (Carlsen 2006). While some website owners might attempt to assert ownership rights to user content, the messages posted on sites such as www.bullying.orgstill belong to the individuals who posted them (Heilferty 2011; Moreno, Fost and Christakis 2008).

While the copyright still belonged to the author of each posting, it would have been impossible to seek individual permission from each author. There were literally hundreds of individuals in various countries around the world, most of whom had posted their messages anonymously. Even if the means existed to make individual contact with the authors of the messages, the nature of the Internet is such that it would have been equally impossible to verify the identity of those who claimed to be the authors of the messages and thus ensure that they had the authority to grant permission to use their messages.

The issue, then, was how to make substantial use of the messages posted on www.bullying.orgwithout engaging in copyright infringement. The Canadian Copyright Act(RSC 1985, C-42) defines copyright infringement as doing anything that only the owner has the right to do, without the owner’s consent (Government of Canada 1985). One of the sole rights that the owner has is to reproduce their work or any substantial part of their work. When it comes to what constitutes a “substantial part” of the work in question, Canadian courts have placed more emphasis on the quality than on the quantity of what was taken from the original work. Courts have also considered whether the copyright holder was adversely affected by the copying, whether the person copying the material was doing so simply to save personal time and effort, and whether the work was being used in a similar fashion to that of the copyright owner (Carlsen 2006).

In circumstances where only a part of a message or story was being quoted, it is doubtful that there would be a substantial taking of the work, as long as its use did not adversely affect the copyright owner. While the messages on www.bullying.orgwere subject to copyright, the authors voluntarily placed these postings in a public venue, often intending to share them with others (cf. Heilferty 2011; Moreno, Fost and Christakis 2008; Martin 2012). Moreover, messages harvested from the site were not copied in order to save time and effort but rather so that they could be used in a totally different fashion from that of the copyright owners. This research study was something that the owners of the copyright would be unable or unlikely to do themselves and something that researchers would be unable to do properly without copying the postings (Carlsen 2006).

Even if substantial parts of these postings were used to give “voice” to the research subjects, this would not necessarily imply that copyright infringement had taken place (Carlsen 2006; Geist 2006b; Geist 2006a). The Canadian Copyright Act(RSC 1985, C-42) states that “fair dealing for review or criticism does not infringe copyright” if the source of the work is acknowledged. The Supreme Court of Canada has stated that copying an entire work might still meet the definition of fair dealing for research purposes. If the researcher was making extensive use of text messages, but was putting them within the context of his or her own original thoughts, then this would be consistent with the principles of fair dealing, as long as the author and source of the downloaded messages were acknowledged (McKee and Porter 2009; Craig 2005).

As the researcher resided in Canada, the research took place at a Canadian university, and www.bullying.orgwas a Canadian-based website, it was decided that Canadian copyright law would take precedence if copyright infringement was claimed. Moreover, 94 percent of the messages posted on www.bullying.orgcame either from Canada, or else from the United States, the United Kingdom, Australia, or New Zealand, all of which have “fair dealing” provisions similar to those in Canada (McKee and Porter 2009; Geist 2013).

Ethical conduct and the requirement for informed consent

Informed consent is a central ethical principle in research scenarios that pose potential risk, harm, discomfort, or embarrassment to the research subjects (Thomas 1996; Comstock 2012). Internet research has drawn particular attention because of the ease with which “personal” messages can be downloaded, subjected to analysis, and quoted at length, without informing the research subjects that the study is taking place. It has been argued that the type of Internet research undertaken in this study of www.bullying.orgcould abrogate the right of the research subjects to know about the nature and duration of the research project, the potential risks and benefits, and what measures were being taken to ensure confidentiality (King 1996; Mann and Stewart 2000; Paechter 2012; Sharf 1999; Sharkey et al. 2011; Waskul and Douglass 1996).

It would have been impossible to obtain informed consent in this study because there were no means by which to verify with certainty the age or identity of the individuals giving such consent, short of conducting hundreds of offline interviews in some 29 countries around the world. According to information provided by the site participants, 284 were children when they posted their messages, while another 273 were teenagers. If informed consent was indeed required, then in many cases it would have been necessary to obtain the consent of the parents as well.

In any event, informed consent is not normally required when conducting research on public behavior in readily accessible public settings, where there is limited expectation of privacy and little if any risk of harm or embarrassment to the participants (Kitchin 2002a; Lindlof and Taylor 2002). Bullying.orgis a readily accessible website where anyone can read or post a message without having to register or obtain a password. One of the first websites to appear in an Internet search using the keyword “bullying,” the operators describe it as “the number one bullying site in the world.” It is linked to a variety of similar sites, and its operators actively seek publicity around the world. Moreover, the site operators had already anonymized messages that were not submitted anonymously.

In the final analysis, this study of www.bullying.orgwas deemed to fall under the minimal risk provisions of the guidelines for Ethical Conduct for Research Involving Humans(Canadian Institutes of Health Research, Natural Science and Engineering Research Council of Canada and Social Sciences and Humanities Research Council of Canada 2007). The Research Ethics Board (REB) at the Canadian university where the study was conducted confirmed that it was excluded from the requirement for an ethics review and thus from the requirement for informed consent. The study used readily accessible archival materials posted in a public arena, it involved no interaction with the research subjects, and it posed no greater risk than what might normally be encountered by the research subjects in their daily lives (Moreno, Fost and Christakis 2008; Kitchin 2002a).

Some observers have argued that REB approval does not necessarily mean that the research is ethically or morally acceptable (Comstock 2012; Zimmer 2012). The Internet is an ever-changing landscape, and Board members might not be familiar with the techniques that cyber researchers employ to collect their data or the type of privacy settings that the researchers might be circumventing (McKee and Porter 2009; Palys and Atchison 2012). There is no question that the www.bullying.orgcommunity would rise to the definition of a “sensitive” or “at-risk” population (cf. Paechter 2012; Sharkey et al. 2011), in that 284 of the participants in the sample were children, 228 of the messages described bullying that involved the commission of one or more recognizable criminal offenses, and 94 messages talked about suicidal thoughts, self-harm, attempted suicide, and completed suicides. If the names, contact details, and other identifying features had not already been removed by site operators, the REB disposition in this case might well have been negative (Tilley and Woodthorpe 2011).

Privacy rights in the public domain

Some ethicists argue that website participants such as those on www.bullying.orghave a “perceived” expectation of privacy (Martin 2012; King 1996; Zimmer 2012). The precise concepts of privacy and expectation of privacy can be difficult to pin down because interpretations vary significantly from individual to individual and from culture to culture – what some might regard as a highly personal secret, others might be willing to share publicly in cyberspace (Capurro 2002; Pomerance 2005). Indeed, millions of people seem quite willing to share personal information like their age, gender, photos, and contact details on social networking sites such as Facebook and MySpace (Raacke and Bonds-Raacke 2008). People can also be surprisingly forthcoming when it comes to the content of their home pages or web messages, freely sharing intimate details about their marital status, family problems, and medical concerns (Kennedy 2006; Lee 2005). Even those who do not have a home page or use social networking sites may voluntarily provide all sorts of revealing information about themselves in order to download free software or to gain access to a website that requires registration and a password (Klang 2004; Woo 2006).

While acknowledging that information posted on publicly accessible websites is in reality open to micro-observation and later inspection, a number of observers insist that researchers should disguise the source of the data in such a manner that members of a virtual community who subsequently recognized themselves or their own words in an article or website posting would feel that the location and identity of the group had been protected (King 1996; Waskul and Douglass 1996; cf. Boehlefeld 1996; Waruszynski 2002). However, many cyber ethnographers have felt comfortable in identifying the source of their data. Examples of studies where the source was identified include Kitchen’s (2002b) study of an Alcoholics Anonymous site called Alt.Recovery.AA and Sanders’ (2005) study of how sex-trade workers communicate with prospective clients through a website known as Punternet. In the case of www.bullying.org, any attempt to obfuscate the identity of the website would almost certainly have met with failure due to the site’s prominence and public visibility.

The participants in www.bullying.orgwere clearly aware that their messages were being read. Forty-eight thanked the site or the other site participants for reading their messages, 54 specifically said they wanted their story told, 126 asked for a response or advice, 28 attempted to provide their complete contact details, and 10 mentioned that their stories had been published elsewhere. Of the 950 messages in the sample, 201 of them were in fact replies to earlier messages, often expressing sympathy, providing advice, or offering friendship and support. Notwithstanding arguments in favor of respecting personal privacy and obtaining informed consent, it is now widely accepted by courts and ethics review bodies in most countries that individuals posting messages in these public venues are often seeking public visibility, or if not, are at least aware of the public nature of their behavior (Heilferty 2011; Moreno, Fost and Christakis 2008; Martin 2012).

Discussion

There has been an ongoing (sometimes heated) debate about research ethics in cyberspace, but there has yet to be a definitive conclusion regarding what is acceptable and what is not. Some researchers have insisted that Internet users have a right to privacy – that messages posted on websites should be off-limits to social researchers unless informed consent is properly obtained and all members of the website have been notified that the research is taking place. Others have argued that public postings are by definition public – that they are no different from writing a letter to the editor of a newspaper or calling into a radio talk show.

Either the deontological (rule-based) framework or the teleological (harm vs. benefit-based) framework can be applied when addressing ethical concerns of this nature, regardless of whether the research is taking place in cyberspace or in a more conventional social domain (Thomas 1996; Palys and Atchison 2012). This cyber ethnography of www.bullying.orgmet the requirements of the deontological framework – it fell under the exemptions set forth in the Canadian guidelines for Ethical Conduct for Research Involving Humansand received written confirmation from the university’s REB that it was excluded from the requirement for an ethics review and the requirement for informed consent. As noted previously, it also complied with the requirements for fair dealing for research purposes under the Canadian Copyright Act(RSC 1985, C-42).

Nevertheless, there may be a genuine perception of privacy in smaller websites with closed membership and especially those that have registration requirements and password protection (Heilferty 2011). Cyber researchers would be well advised to consider the situational context, carefully examining such issues as restricted site access, perceptions of privacy held by the site participants, and whether they are conducting research on a vulnerable or “at-risk” population (Heilferty 2011; Zimmer 2012; Nissenbaum 2011). It should be remembered that the site participants still have “moral rights” – including the right to privacy – even though these rights may not be enshrined in law in the country or countries where the website and the cyber researcher are domiciled (McKee and Porter 2009; Comstock 2012). In other words, researchers should try to avoid unnecessary intrusion into “private” spaces, and if they find such intrusion necessary, to limit the disclosure of intimate or highly sensitive information (Hull, Lipford and Latulipe 2011).

This cyber ethnography of www.bullying.orgalso met the requirements of the teleological framework. There was minimal risk of harm or embarrassment to the research subjects, and the likelihood of their being identified or contacted as a result of the study was virtually nonexistent (Heilferty 2011; Lee 1993). The study contributes to the understanding of bullying and victimization by bringing forward the voices, stories, impressions, and advice of those who have had personal, firsthand experience with bullying. In particular, it provides new insights that challenge the oft-expressed notion that bullying is on the upswing, or that current manifestations of bullying are more serious than earlier ones (cf. Haber and Glatzer 2007; Marsh et al. 2004; Olweus 1993). It also offers an explanation for why anti-bullying programs have routinely produced negligible to modest results at best (Rigby 2002; Twemlow and Sacco 2008). Moreover, it is hoped that this work will serve as a guidepost or template for other cyber researchers, bearing in mind that research decisions need to be made on a case-by-case basis (Sharkey et al. 2011) due to the complex nature of cyberspace and its many diverse cyber communities.