Ethical questions related to using netnography as research method

Netnography is a relatively new research method, which adapts research techniques of ethnography to study cultures and communities through computer-mediated communications. It has become a popular research method in marketing research during the 21st century. However, the use of netnography in the field of information systems (later referred as IS) has not been studied to great extent. Thus, we have conducted a systematic literature review to investigate the ethical practices of netnographic research in the field IS. To analyse the ethical practices of netnographic research and discussion surrounding it, we have collected 52 articles which use netnography either as their sole research method or as their completing research method. These articles were selected from 77 IS journals. Our findings indicate that netnography is an emerging research method which is still moulding its ethical guidelines. Researchers, who use netnography, do not completely agree on the ethically just manner of conducting netnography. However, it is apparent that certain ways of conducting netnography are often considered to be ethically just where as some other ways might be often considered to be ethically unjust.


Introduction
Netnography is a relatively new research method, which adapts research techniques of ethnography to study of cultures and communities through computer-mediated communications. It was first defined by Kozinets (R. Kozinets, 1998) based on the developments of research methods in the field of marketing research. In the 1990s, also other researchers started to define similar type of research methods such as cyberethnography (Ward, 1999) or digital ethnography (Murthy, 2008).
Netnography has been employed as a research method in many research fields including information systems (E. Vaast & Walsham, 2013), marketing (Heinonen & Medberg, 2014), and tourism (Mkono. & Markwell, 2014). While the popularity of using netnography as a research method has grown in the resent 20 years, the discussion about the ethical guidelines of this method has started alongside its use. Scholars who give instructions for conducting netnographic research differ in their opinions of employing netnography in an ethical manner.
In 2002, Kozinets (2002) defined four principles which should be followed when conducting netnographic research. These principles are: 1. A researcher should fully disclose his or her presence to the online community members which he or she is studying. 2. The researcher should guarantee the confidentiality and anonymity of the informants. 3. The researcher should seek and incorporate feedback from the online community members. 4. The researcher should ask community member's permission to directly quote any specific posting.
While these ethical principles suit well any qualitative research which is done in certain community, all researchers do not agree that they suit well netnographic research. Langer and Beckman (2005) argue that netnography shares many characteristics with content analysis of traditional forms of media such as TV or newspapers, especially when the discussion takes place in a public forum which is accessible for all internet users without any registration, and for the most part this seems correct especially when conducting netnographic research in internet. Thus, they themselves did not follow most of the principles stated by Kozinets in 2002 when they studied consumer's onlinecommunication about cosmetic surgery. They also argue that discussion related to such sensitive topic would most probably be disturbed by informing the community members about the on-going study.
Kozinets' thoughts on ethics of netnography have also changed during the last decade. In 2009, Kozinets (2010) acknowledged that it could be ethically just to study some online communities without asking permission to study them. One criterion to decide if the online community in question could be studied in a covert fashion is whether writing or posting in the online community requires registration. He also reminds researchers that many commercial sites have proprietary rights to their content including discussions between site users. Because of this, many companies have denied the usage of online discussions for research purposes. If the material is publicly available, however, it can be used for research, and the legal question is separate from the ethical questions.
Prima facie Kozinets' rules 1, 2 and 4 seem unnecessary when doing netnography in open fora online. Rule 3, however seems always worth following, if for no other reason, then to verify the validity of the study. In closed fora however, all four rules are clearly worth following as the privacy expectation is clearly higher. If the forum under study does not have identifying information of the participants, and the researcher somehow comes by this information (especially external to the forum itself), rule 2 reapplies, as doxxing 1 members is considered bad form and can at worst be dangerous to doxed individuals.
Some researchers, such as (Xun & Reynolds, 2010), highlight that one of the benefits of using online sources for data collection is that it offers good possibilities of using direct quotes in research reports. However, other researchers, such as King (1996), question if it is ethical to directly quote online discussions, especially when the study is done in covert way. King claims that it is not always ethical to use direct quotes retrieved from online discussion forums in research reports even when the data is publicly available for anyone who is using internet. He proposes that the accessibility of the online community as well as the perceived privacy among the members of the online community should be considered before one decides, how the research findings will be reported.
In this article, we investigate the ethical issues related to using netnography as a research method. Thus, our research question is: how netnography can be conducted in an ethically justified manner? To answer this question, we conducted systematic literature review on netnographic studies which are done in the field of IS to form ethical guidelines for conducting netnography. While analysing these articles our main interest was directed to the ethical guidelines the authors had decided to follow and the ethical discussion the authors engaged in themselves.

Research design
Our systematic literature review concentrates on highly ranked IS related journals which were chosen based on the ranking done by Mylonopoulos & Theoharakis (2001) and the ranking determined by the Australian Council of Professors and Heads of Information Systems in 2013 (ACPHIS, 2013).
All together, we reviewed 77 journals (see Table 1). From these journals, we wanted to find all the articles which use netnography as their research method. In addition, we wanted to find those articles which use research methods which resembles netnography. Thus, our search terms included netnography, online ethnography, virtual ethnography, and cyberethnography. Because the search was targeted to whole articles rather than keywords or abstracts, one of these terms could have been mentioned in any section of the article, including references. Hence, our search found many articles which did not use netnography or any similar type of research method. We also were able to find articles, which appear to use netnography as their research method although they did not state it.
With this research strategy we were able to find 122 articles for further analysis.
These 122 articles were analysed independently by two researchers who are also the authors of this article. In the case of those articles, which the authors could not agree based on their initial analysis; they reread the articles and discussed their content until they were able to make a common decision to either include or to exclude the article from this review.
To be included in our review, the article had to report findings from an empirical study and the topic of the article had to be relevant for the field of IS. In addition, the reported research methods had to imply that some part of the empirical study or the whole empirical study followed the principles of netnography. These principles include the use of participant observation in online environment, conducting online content analysis, interviewing members of specified online community with deep knowledge about the online community in question, and interacting in the online community or communities.
By using these inclusion criteria, two researchers who are also the authors of this article, analysed each article independently. In the case of those articles, which the authors could not agree based on their initial analysis, they reread the articles and discussed their content until they were able to make a common decision to either include or to exclude the article from this review. After the initial analysis, we found 52 articles from 19 journals (see Figure 1.) which report results from an empirical study using netnography as a research method. However, only 12 of these articles used the term netnography to name their research method. More than half of the articles (n=31) used other terminologies such as online ethnography, virtual ethnography, cyber-ethnography, and digital ethnography. Significant part of the articles (n=9) did not define their research method as netnography, online ethnography, virtual ethnography, cyber-ethnography, or digital ethnography. Their research method resembled netnography however, since it had all the features of conducting ethnographic research through online or virtual environments. To define whether or not a research method resembles netnography, the authors had to show in-depth understanding about online communities, immersion in online environments during research period, and major data collected from online sources. We were able to find these articles because our search strategy focused on full articles, thus the search term could have been found from the list of references.

Figure : Distribution of included articles between journals 2
In the selected articles, netnography is used as a sole research method in 40 % of the articles (n=21). In the remaining articles (n=31) netnographic research is combined with other empirical research methods, such as interviews, focus groups, case studies, datamining, online surveys, discourse analysis, narrative analysis, content analysis, social network analysis, and grounded theory-inspired coding.

Results
We first analysed whether the selected 52 articles referred to any ethical guidelines while presenting their research approach. Then we continued by analysing the ethical discussion these articles engaged themselves with.
Only a minority of the articles included in our review referred to any ethical guidelines regarding their netnographic choice of methods. Ethical guidelines were mentioned in 16 articles either directly or indirectly. In the majority of these articles (n=10), the authors disclosed their presence to the online community. In four articles, authors have conducted covert netnographic research. In the case of two articles, it was not clearly stated whether the researcher revealed their presence and their motives to the online community.
The oldest study in this review is done by Ward 2 . She refers to ethical guidelines of the British Sociological Association when she explains why she revealed her presence in the online community which she was studying. She also felt uncomfortable not disclosing her purpose to those she studied, as there were no visual or audible cues (which still rely mostly on virtually created cues), misleading the research subjects would have been too easy in her opinion.
Although, disclosing ones presence and research intentions seems to be the right way to study the online communities, it might not be the only way. Reilly and Trevisan (2016) argue that covert observation of the online community might sometimes be the ethically just way to conduct research online. They give example from their own research in which focused on politically active group. This group organized flag protests in Northern Ireland. The members of the community were vary of sharing information on the protests in the Facebook. Thus, Reilly and Trevisan suspected that the members of the community would reduce or change their communication in Facebook, if they would have been aware about the study. De Koster and Houtman (2008) also decided that they do not need to ask informed consent from the users of the online community which they were studying. They justified their decision by two characteristics of the online community in question. Firstly, discussion within the online community were easily accessible because one could read them without being a member of the community. Secondly, the members of the online community were aware that "non-members with diverse backgrounds read the postings on the forum".
In some cases, it was difficult to interpret whether the authors had revealed their presence to the online group which they were studying. Vast and Levina (2015) observed online community for bankers and conducted some interviews among its members. Their article does not clarify whether any permission was asked to observe the online community or if any existing ethical guideline was followed during the research proses. However, they clearly state that the anonymity of the site and the respondents was maintained by using pseudonyms.
One of the important ethical issues for authors was the anonymity of the research participants. Skågeby (2009) explains that he wants to guarantee full anonymity to all the participants of his study. Thus, he decided to avoid using direct quotes from online discussion because it might be possible, at least in the future, to find the original posting with the direct quote. Some other researchers, such as Vaast and Levina (2015) followed less strict reporting style and aimed to maintain the anonymity of the site and the participants by using pseudonyms. In rare occasions researchers have decided that there has been no need to use pseudonyms for the participants because they knew that their comments were publicly available. De Kostera and Houtman (2008) justify their choice of using original user names alongside direct quotes by explaining that the users of the site were aware of the possibility of outsiders reading it. They also argue that each quote could be found through a search engine, thus using pseudonyms for the user names was unnecessary.
If one decides to avoid using direct quotes from online discussion, it might raise a question how to report the comments and opinions of the participants while reporting the research results. Markham (2012) has suggested that fabrication might be an ethical way to present the participant's viewpoints when there is a need to protect the anonymity of the participants and a risk to reveal them by using direct quotes from the collected data. She defines fabrication to include practices such as creating composite accounts of persons and building fictional narratives. However, Trevisan and Reilly (2014) make a clear stance against fabrication when the research concentrates on studying social groups which are typically marginalized or which experience a condition of oppression and dependence. They argue that fabrication represents a high risk of distorting the online voices of the members of these social groups because "it works on the assumption that the researcher should elaborate 'proxy' accounts to prove rhetorical points". Balancing authenticity with suitable amount of anonymity is clearly not an easy decision, and seems to be more case dependant rather than there being a one-size-fits-all choice. Beaulieu and Estalella (2012) that ethics of netnographic (or internet ethnographic, as they call it) research extends beyond the question of guaranteeing anonymity to the research participants and the acceptance of covert research. This is due to engaging in disclosed netnographic research creates new ethical dilemmas which differ from those ethical dilemmas which researchers face while conducting disclosed ethnographic research. They give an example from Estalella's research in which he creates a blog related to his research which studied blogs and bloggers. One reason to create the blog, was to make the researcher visible to the field and thus avoid doing covert research, while also making Estalella's research more interesting to possible informants. However, maintaining this blog created new ethical dilemmas. Through the blog Estalella's research became visible to anyone interested in the research or in the Estalella himself. Thus, colleagues, friends and family members could start a discussion with Estalella related to his research while he was conducting the research. At times, such discussion might have been beneficial for him and his research, but sometimes they distracted him. In addition, Estalella's blog made it possible to trace some of the bloggers he was studying, because blogs of those bloggers might refer to Estalella's blog or they have left comments to each other's blogs. This kind of situation could be especially problematic if the researcher, (unlike Estalella) was studying a highly volatile environment such as Silk Road (Maddox, Barratt, Allen, & Lenton, 2016), a site used to buy and sell various items from guns through drugs to government secrets with BitCoins.

Conclusions
The vast majority of articles included in our analysis did not refer to any ethical guidelines while explaining their choice of research methods. Only 16 out of 52 articles referred to ethical guidelines. In the case of three articles, ethics of netnographic research was the topic of the article itself.
While analysing the articles we found that there are many different ways to conduct netnographic research in ethically justified manner. There appears to be at least three relevant questions that might help the researcher to choose an ethically justified way to conduct netnographic research in a certain online community. These questions are: -Do you need to ask the informed consent of the members of the online community in question? -Do you need to protect the anonymity of the members of the online community in question? -How important is the accountability of your research?
In addition, we wish to draw attention to those situations, where the information shared between the members of the online community might be harmful for them or it is clear to the researcher that the members of the online community have given information to each other under the impression of confidentiality, and the researcher is but an accidental bystander viewing it. We suggest that information that has these characteristics should not be included in the form of direct quotes in the research reports because it could be easy to trace it back to the person who posted it. This suggestion applies even to those situations, where the researcher has informed consent of the person revealing the confidential information during direct interaction with the researcher, because the person in question might not be aware that direct quotes could be traced back to them. It is also possible that a person who is interacting with the researcher forgets that the discussion is recorded for research purposes and thus never intended the confidential information to be included in the research. Because of this, it might sometimes be ethically justified to remove some parts of the discussions with informants from the research data itself. This review has led us to find out that not many researchers seem to be all that interested in -at least disclosing their -ethical practices relating to netnography. Most (38 papers) do not seem to concern themselves with the topic at all, and we are left with only 16 papers in which the ethicality of the research is deemed of such importance that it is worth mentioning. The field of IS research is clearly in need of ethical guidelines which do not rely only on traditional ethnography inspired ethical guidelines, as the environment, especially when conducting the research in internet typically do not seem to coincide with the kinds of closed environments often studied in traditional ethnography. This remains a topic for future study, as in this paper we only look at the current situation within the field.