Old methods and new technologies: Social media and shifts in power in qualitative research

Old methods in new contexts

As qualitative researchers, we are always negotiating boundaries between ourselves and those from whom we aim to collect data. We should strive to behave ethically, from protecting the identity of the participants (ASA, 2008) to presenting ourselves as open, sympathetic, and forthright with our respondents. ¹ The success of qualitative interviewing and ethnographic access often depend on the relationships of trust a researcher is able to build, the depth of their access to participants’ social worlds, and how much they are able to ‘get’ from participants that goes deeper than surface understandings. Although researchers are respected for fully entering the lives of their participants, they are expected to fiercely protect the confidentiality of responses and identities of participants, unless it is otherwise granted (Giordano et al., 2007; Duneier, 2000b; Guenther, 2009). With increased regularity, sociologists see institutional review boards asking for detailed accounts of how researchers will protect subjects, store interview recordings or fieldnotes, and even protect potential subjects who choose not to participate, even as these procedures do not necessarily increase protection of research subjects (Librett and Perrone, 2010).

Researchers aim to present themselves to participants as sympathetic chroniclers of their lives and imagine themselves as likeable, straightforward, and maintaining a certain purity in their work (Fine, 1993: 273). Because data arise through interaction, the ethnographic encounter or in-depth interview presumably occurs in a common time and space and during interpersonal exchanges that are dialectic and consensual. Even though mutually participatory, qualitative researchers should consider power as they make choices about study design and access to participants’ worlds. As ethnography requires personal connection, it in fact almost always grants the ethnographer the final decision about how to represent subjects in published and disseminated findings (Stacey, 1988).

Online social networks allow individuals to create their own presentations of self, to design visual or textual representations of themselves. As a result, researchers specifically studying those digital productions must carve out new methodologies to approach, examine and analyze those data (Hargittai, 2007), including questioning respondents’ level of authenticity in representing offline identities (Zhao et al., 2008; Taraszow et al., 2010; Garcia et al., 2009) and the role online relationships play (Walther et al., 2008; Livingstone, 2008; Fox et al., 2005; Abril, 2007–8). Online researchers also debate whether they can or should use online surveys to collect data or analyze postings as text (Griffiths, 2010; Farrell and Petersen, 2010). They also question whether they should gain consent to participate in online forums and what that consent might look like (Reid, 1996; Kendall, 2002).

This literature does not help researchers manage online interactions that are not expressly about research questions, nor part of the research design. Claiming there is only one social world, researchers like Wellman, Boase and Chen (2002) and Garcia et al. (2009) quite rightly point out that offline and online worlds have blended, and that the debate as to whether there are two social worlds is wrong or at the very least overblown. Yet clearly, information moves differently than in the past. Despite rich discussions of the challenges of obtaining informed consent and protecting information among researchers who conduct their research in online social contexts, there is less discussion of how online technologies change offline research. This paper begins to fill this gap.

Studying across to anti-vaccine parents

First, I should say a few words about my research, which is the basis of much of the following discussion. I set out several years ago to study parents who choose not to vaccinate their children. Defining my inclusion criteria broadly, I sought parents who modify existing vaccination schedules for their children, either by opting out completely, providing consent to select vaccines on a schedule of their own choosing, or in some way negotiating alternative schedules with their health providers. I also targeted pediatricians to understand their experiences with parents who challenge nationally recommended vaccine schedules, and alternative or complementary health providers, including chiropractors and naturopaths, who work with parents who often adopt a less traditional course for their families’ healthcare. Additionally, I conducted interviews with attorneys who represent families in vaccine injury compensation program in the federal vaccine claims court system. Finally, I paired these interviews with ethnographic observations at community educational events hosted by pediatricians and alternative providers and at national conferences of organizations that oppose vaccine mandates.

Of the 30 parents who participated in this study, two are fathers. All but one parent is white and all but one identifies as heterosexual. Seven have a master’s degree or doctoral level of education, ten have a bachelor’s level degree, and 13 have completed high school or equivalent, with some having completed some college or other training. This is a relatively privileged parent population, which makes sense. Existing research shows children who lack vaccines because of parental choice are most likely to be white, live with two married parents, have a college-educated mother and a family income over $75,000 (Smith et al., 2004).

This study then is a study of what anthropologist Nader (1972) calls ‘studying up’ or ‘across’. Indisputably, the issues of online technologies and research look different than if this were a study of lower income or less educated people (van Deursen and van Dijk, 2011). However, the core questions remain similar: we must consider the new challenges and opportunities social media bring to research and how those should be addressed, even as they may present differentially depending on the level of online access groups may have.

Recruiting participants

When researchers talk of snowball sampling we often envision an effortless roll downhill, and yet, participant recruitment rarely runs so smoothly. My study participants were, in an old methods way, difficult to locate: parents of unvaccinated children don't have a club or street corner or regular meeting place; they are a small, albeit growing, population that is somewhat insular (Omer et al., 2009; U.S. News and World Report, 2011); these parents also distrust research. The last point is perhaps most significant. Parents who reject expert opinions and question the veracity of medical and pharmaceutical research often viewed me with a similar level of suspicion.

Because of their number, recruitment necessitated a heavy reliance upon new media and online social networks to distribute invitations for participation. I targeted listservs of groups that focus on topics like homeschooling, attachment parenting, vaccine information, alternative curricula, and natural living. At times I relied on parents in my own social networks who were members of these groups to email my invitation to their listservs. Needing multiple intermediaries and gatekeepers served as one of the first pitfalls in using electronic communication for recruitment; at times this was as much logistical or technical as intentional. Some gatekeepers needed to be persuaded to help me recruit participants, like the keepers of a home birth network who asked me many questions about my study and only circulated it in the most limited ways possible. Sometimes, this was simply technical. For example, those who generously forwarded calls to participate in my study often also received the replies from those interested. Many parents on those lists answered emails by simply hitting ‘reply’, which then necessitated the benevolent acquaintance that had sent the initial invitation for me to forward the email reply to me so I could follow up directly. It also limited my ability to protect information about who may have participated.

When I had exhausted my access to listservs and email networks, I found potential participants through online discussion boards (e.g. parents looking to expose their children to wild virus varicella, or chicken pox), Meetup.com or other online groups with focuses on home birth, extended breastfeeding, or autism support. Most often, I would email the moderator or organizer and ask if he or she would feel comfortable forwarding on my invitation and contact information. Many did not before emailing questions for clarification of my research first. In interviews, participants would frequently reference a friend with whom I should speak. Often, I would look for their contact information online, which sometimes yielded other personal information, ranging from legal proceedings to personal photos. I found that, through Facebook, I could sometimes search for those people and, without needing their email or phone number, I could send a message through Facebook and introduce myself. This seemed more respectful of their privacy, since I did not ask for personal information from their friends and acquaintances. Yet, through the social networking site, I often was able to glean profile photos and other information depending on their privacy settings, all without the conventional avenues of requesting access.

I deliberately never extended an invitation for friendship, a boundary that seemed important as even the vernacular of ‘friend’ seems to confuse my primary motive, which is research. Many parents joined the study through these email invitations, but my efforts to develop a snowball sample were unsuccessful, with few referred directly by another interviewee. Many apologized, saying they didn’t want their friends to know they participated in the research. In moments of frustration, I would joke that it was easier to recruit drug addicts (in my past study of parents in the child welfare system) than to recruit affluent parents who distrust science. Because of the limits of traditional sampling frames and techniques, I relied on virtual forms of communication to aid in recruitment. My method for the above study then represents an unsuccessful convenience sample – an ‘inconvenient sample’ where respondents don’t reliably refer others – and also one that developed through the fruits of new technologies. In the end, about half of the 30 parent participants came through emails sent through listservs or to groups who were asked to pass them on through their networks. The other half came from friends and colleagues who passed on information about the study, often through email, to individuals they knew who opposed vaccines.

To access attorneys to interview, I initially used the referral list on the website for a large national organization that defines itself as advocating for parental rights and opposing vaccine mandates, to find attorneys who specialize in vaccine injury claims. I then emailed or called attorneys to schedule appointments. Eventually, I gained access to a list from the federal court of all the attorneys certified to participate in this esoteric claims court, for which there are fewer than 100 nationally. This list included incomplete contact information, so I relied on web searches of firms and practice information to contact practitioners. New technologies were then, again, indelibly and unavoidably part of the recruitment effort.

This was not, of course, a one-way street. While online technologies helped me in recruiting subjects, these same technologies allowed potential subjects to find me just as easily. For example, an outspoken chiropractor who objects to vaccines requested a meeting via email, having heard about me from a client. He then, without my knowledge or consent, referred colleagues and members of his professional network who shared his view to contact me as well. The ease with which they emailed reminded me that potential participants could find me and also recruit through their own networks those with a shared perspective to participate. This is not uncommon in offline research where gatekeepers might direct potential participants to the researcher. However, the ability to forward, edit, network, and mass distribute information about my study made the process occur faster and well outside my control, allowing potential participants to find me and to shape stories of my study. Because this is a population that clusters (May and Silverman, 2003), these stories ran the risk of also foreclosing other participants in my area.

During interviews, participants occasionally asked what my husband did for a living, a fairly pedestrian question in qualitative research. However, the answer – that he is a pediatrician at the local children’s hospital – proved to be a red flag to participants critical of mainstream medicine’s support for vaccines. Although I developed strategies to deflect concern, his professional life became part of the identity I had to manage (Goffman, 1959) and also became information participants shared with others, which on at least one occasion persuaded a local activist to change her mind about participating in my study. Positive stories also encouraged participation, like when an acquaintance forwarded my call for participants through a homeschooling network or when a handful of participants who heard about my study sought me out and emailed to ask to participate. As such, respondents and their communities held power in their ability to technologically open or close access. In illustrating these recruitment strategies, I aim to highlight how technologies potentially shift our understanding of power. My methods of recruiting participants – with all the messiness involved – also show the slippage between offline research and online technologies.

Gaining informed consent: From researcher to researched

When research is overt, ethics dictate that participants should provide informed consent; before consenting, participants should be informed of the content and goals of research, even as researchers fear participants might alter their responses to accommodate (or undermine) what they perceive to be the researcher’s goals. Informed consent allows participants the rights to weigh the risks of participation against the benefits to themselves or others and choose whether to be a part. For qualitative research, the very act of being known as a participant might carry risks, and as a result researchers are expected to protect confidentiality. Throughout this study, and often hearing parents’ unwillingness to refer others to me because they did not want to tell friends they had participated in my study, I felt protective of their identities and information they shared.

Yet, I discovered this concern was not mutual. I was surprised to learn that after sending invitations to participate in my study, potential participants were quick to research me before responding. Skepticism of research and researchers is well-described when studying elite populations (Ostrander, 1993); however, technology has made it much easier for potential subjects to scrutinize researchers. As they contacted me, they often mentioned they had already reviewed my earlier research, read my book online, or checked my university credentials to see if I was fair and could be trusted. One potential participant agreed to speak with me, but also felt compelled to add:

Your Katrina work with the interference from Feds in the Service Provider recovery looks interesting. It was sad that the Feds made some local people shut down their WIFI in the Dome. They had to control information and were not going to let civilians be able to contact relatives etc.

Parents know that they are often seen negatively for opting out of vaccines. They see frequent news coverage that labels them as selfish or ignorant. Some have already had bad experiences with journalists, who, using their real names, present them unfavorably. Thus, these efforts to understand my motives and predisposition reflect a desire to see if I am sympathetic to their views, or as Goffman (1986) suggests, among the wise who understand the stigma they face and want to increase understanding of their position. These potential participants aim to evaluate whether I will support their goals and whether I am trustworthy. These challenges and the relative difference in recruiting speaks to the varied views differently located parents had of research, knowledge of how it can be used, and the ability to control the process (Hmielowski et al., 2013; Bynum et al., 2012; Gauchat, 2012). Logically, parents in this study use the internet and new technologies as knowledge-gathering tools to gather information about me in making this determination. Groups with fewer resources might also vet me, but may be more likely to rely on other sources they might deem more credible. Illustrating this, one parent referenced information about me directly:

I may be incorrect, but given that you are a sociologist I can only assume that your position is that in a ‘society’ people should do what they are told for the ‘greater good.’ If that is the case, and again, I could be incorrect, I would probably not be interested in being interviewed by you. I cannot spend time on anything at this point that does not further my goals, which are to maintain the right of all people to make their own healthcare decisions and to make truthful information available to those seeking it regarding vaccinations and other healthcare issues.

Despite my efforts, I was reminded that not all online information is accurate. For example, entering my name yields information about my employment and publications, but also several false positives: a physician of internal medicine, an author of parenting and marital advice, and a Pilates instructor who lives in the same city as I do. While other researchers have also been misunderstood in the context of research, most are misunderstood after they have gained access to participants or field sites. New technologies raise the possibility of being misunderstood before data collection has even begun, which makes establishing rapport potentially even more challenging.

Potential participants’ search to intuit my motives, alongside acknowledgement that some information online is inaccurate, left me with several questions researchers conducting offline research in the era of new technologies should consider. First, what constitutes full disclosure for participation – truly informed consent – and what do potential participants deserve to know? In considering what we disclose about ourselves in research settings, we might also consider how what we deliberately reveal might be filtered through information research participants might already have gathered about us: our past research, but also a fair amount of personal information as well – marital status and information on our spouses, income (if researchers work at a public university), home phone number and location, or even family photos. Their sources may include Google searches, past scholarly publications, blogs, online petitions, social networking sites, online syllabi, public lectures, professor rating sites, campaign contributions, or professional or personal webpages.

Relatedly, these questions highlight how ethnography becomes potentially more challenging as scholars build their vitae and disseminate work. Many great ethnographic studies have in fact been conducted by young scholars, particularly doctoral students who may be advantaged by their relatively clean slates. Collectively, these questions remind researchers that they no longer control how knowledge flows, which carries far-reaching consequences. On the one hand, this may facilitate a more collaborative, reciprocal dynamic between participants and researchers. On the other, it may greatly limit our access to new populations and research sites. In all ways, new technologies and access to information require greater awareness of the multiple directions knowledge flows and the consequences that carries.

Gathering data

Part of the expectation of informed consent for participation in research is that participants will be able to choose what information to disclose, how to present themselves, and what level of access to allow the researcher. We expect participants will regulate what information they disclose, to maintain the right to refuse to answer any questions or withdraw at any time, and to withhold some information about themselves.

Yet, online technologies allowed me to research my participants in ways that tested my own comfort. Indisputably, researchers’ ability to conduct searches of participants also alters data, undermining the idealized standard of more naturalistic observation, or at least changing understandings of our subjects. I can, for example, analyze the websites of practitioners and read their self-published philosophical statements, patient testimonies, and practice guidelines. I discovered one participant remarried through an email blast marketing the services of a local health provider whose last name now matched his. In searching for the names of potential participants in Google, I have uncovered information about past political contributions or organizational affiliations. Some of this ‘presearch’ is not new. Researchers have often read newspaper articles or brochures about groups or their members in preparation for fieldwork, practices consistent with the norms of research and the official ethics statements of groups like ASA. Yet, social media allow an increasing number of discussions, social networks, online discussions, tweets, Facebook posts, or photos to be published and public in ways participants may not even recognize. The ethics of using these materials to triangulate our data are worthy of deeper consideration.

More practically, online information can make data collection easier. Many of us have struggled with the moments when we first meet an interviewee in an appointed public place; finding online photos might make it easier to simply locate them in a café or public place. Sometimes we share instant affinity; other times, the interactions feel awkward. In my efforts to manage ‘doing similarity’ with participants (Abell et al., 2006; Reich, 2013), knowing their religious or political affiliations could change how I present myself.

Not all information online is a liability, nor does it mean that respondents don’t want us to see it or include it to provide richer presentations of them as fully embodied social actors. Online searches might yield information about participants’ home businesses or consulting services or organizations for which they work hard. Individuals might be pleased with their online profile, having worked hard to craft it, or note it on their email signature, and even encourage a researcher to view it. As Wynn (2011) found in his study of walking tour guides in New York, many participants were relatively unconcerned about anonymity, claiming that they (and their professional identities) were easily found online and were, in fact, disappointed at the idea of anonymity as they valued their public persona. One participant in my study directed me to her personal site, which included photos of her nude in a tub at her recent homebirth. Yet, even as participants might suggest online information to explicate their positions or reveal information they would freely disclose, searches might also sometimes link to personal photo albums, blogs, or information they might not choose to share, even as they are publicly available. Guidelines for data collected from digital communication suggest that information published online can be utilized as text-based data for analysis (Garcia et al., 2009; Briassoulis, 2010). As researchers, we have not adequately considered the implications for how online information affects perceptions of participants in offline interviews and observations.

Other questions arise from researchers’ access to participants’ online personae. If we offer participants the opportunity to withdraw from a study at any time, for example, how might we stop or continue using online information? In a technical sense, have I begun data collection before consent has been granted? When does their withdrawal of consent end our ability to seek out other information about them? And must they consent at all? We rarely seek consent from those we observe in public settings during ethnographic observation. The ASA Code of Ethics advises, ‘Sociologists may conduct research in public places or use publicly-available information about individuals (e.g., naturalistic observations in public places, analysis of public records, or archival research) without obtaining consent’ (ASA, 2008). Yet, I have struggled with what level of post-observation online research was appropriate, even as I was confident my research met professional standards of behavior.

In one example, I attended a meeting targeting women who embrace natural living for their families and congregate to learn more about nutrition, holistic healthcare, and to network with like-minded mothers. At that meeting, a baby stopped breathing and eventually died. As paramedics worked to revive the infant, I observed how women who reject much of mainstream medicine discussed the interventions in real time. After this event, I ran descriptions and symptoms by several pediatricians to hear their impressions of what might have happened; many thought lack of a newborn injection of vitamin K, which prevents a rare but lethal bleeding disorder, could be the cause. Not knowing for certain, I searched online for official documents – a death certificate, obituary, and listserv for the organization – to learn more about the tragedy. I confess I searched for two main reasons. First, I didn’t feel I could write about the event without knowing if it was really related to vaccine resistance. Second, the event haunted me for months as I ruminated over the scene and discussions that day in search of an explanation. Both the baby’s death and the information I found online were all publically available and did not require consent. Yet, I do worry about the ethics of searching for intimate details of the family’s life. The guidance that we only analyze information that has essentially been published ignores how this new information may alter our encounters, outside of the text itself, and provides little guidance on what we owe the people we observe, sometimes in the worst moments of their lives.

Researchers’ ability to glean information not provided directly to them clearly limits the participants’ power to control information and the ways they may choose to present themselves. Potential questions also arise about how to treat narrative accounts that might differ from, or contradict, online representations. Which are data? In some ways, participants may have enjoyed greater freedom in knitting together their own representations of self when other sources of information were less readily available. And while good ethnography has always triangulated findings, new technologies allow even greater opportunities to do so, albeit while potentially removing some of the power of self-representation within the context of research from the participants.

Leaving the field

No matter the project, all fieldwork must eventually end. As most ethnographers realize, extricating one’s self from a field site can be challenging and emotionally difficult (though also sometimes a relief), which often represents the interpersonal ties and intensity of interaction between the fieldworker and informants. The intensity of relationships emerges in part because much of the norm in ethnography supports the view that the fewer boundaries the researcher has, the more we see him or her as embedded in the site and getting really good data. In fact, we often assume that greater access to subjects’ social worlds leads to more authentic data, with semi-boundary-less embeddedness in the research communities treated as a sign of not just more but better data. As Lareau (1996: 197) points out, ‘we often have an overly romantic notion of field work, which emphasizes the glory of “going native” and glosses over the difficulties and problems of the endeavor.’ Yet, as social media and technologically-mediated communication have become widespread and normative, one possible new difficulty in the porous nature of field sites is that it now often comes with expectations for ongoing electronic communication, much of which provides information about researchers’ own lives they may not wish to share.

Researchers hold that what participants provide should be treated as confidential and devise ways to protect their words and identities. Should the participant garner information from the researcher, which is not prohibited, no expectation of privacy applies. There is no reciprocal requirement that participants protect information given to them by the researcher about him or herself. In one example from my research, my data collection included efforts to interview complementary health providers who oppose vaccines, including a network of chiropractors who live in different cities but communicate often and believe vaccines are unnecessary. After interviewing the first, other chiropractors contacted me through email and requested to be interviewed. In those emails, two mentioned that they hoped I was learning a lot at the anti-vaccine conference I was attending, which was only known to the first contact. This communicated to me that they knew my whereabouts without me telling them, and that I was a topic of their own digital communications. In other exchanges, participants referenced their knowledge that I was due to have a baby, even when the information did not originate from me. Admittedly, this study did not require deep membership with integration in a particular field site. These questions might be more intense for researchers in other kinds of ethnographic settings where they have become more fully a community member.

Many participants have extended an invitation to join a virtual social network, including Facebook or LinkedIn, often after interviews have finished or I have begun exiting the field site. I often feel tempted by the possibility of gaining access to participants’ friendship networks by accepting their invitations. I also love the idea of understanding more of their social worlds through online social networking sites. I have not accepted the invitations, with the exception of LinkedIn connections, which provide little personal information. I do not want them to see me as rebuffing their efforts to connect to me. Yet, I am also aware that seeing my friendship networks, posts of left-leaning colleagues, or those of advocates in my social circles for causes with which participants may not agree might jeopardize potential respondents’ views of me and limit access or undermine whatever trust I have built. ² These dilemmas are not necessarily new. Researchers have always worked to earn trust of participants and to manage identities with them. However, the ubiquity of online communication and social media further challenges researchers’ abilities to manage it. Pascoe (2009) describes this challenge in her own work on youth and new media:

 … when I began researching teens and new media I had to make all sorts of decisions about what I did and did not reveal online (whether on personal pages or social network sites), whether I would ‘friend’ the teens I was researching, what I would do when they would write totally inappropriate things on my ‘wall’ (and they did – in front of the view of family and friends), how late at night they were allowed to send me text messages or call my cell etc. Even as an ethnographer I was struck by the clash of my personal and professional worlds. A researcher is constantly ‘reachable’ via new media when using that method for research and it can be incredibly disconcerting.

Perhaps mirroring their own curiosity (or suspicions) of me, I did wonder why participants would ‘friend’ me – whether they aim to expand their social networks, stay in contact, or simply extend friend requests to people they meet. Although I cannot know for certain, I presume they want to understand my motives, find out my political leanings, or glean information about my personal life. However, they may potentially wish to add the kind of social capital that would come with a loose tie with a university professor (Anderson, 2000: 253). For those aiming to challenge mainstream institutions and expert advice – like those who reject vaccines – my position could be a potential asset to them (even as that same connection could alienate other participants who promote vaccine uptake).

After interviews ended, participants would rely upon their research of me and my family, sometimes emailing additional questions to me, often referencing my husband’s role as a pediatrician. One participant, who had been confrontational throughout the interview, choosing the restaurant where we should meet and then arriving 30 minutes late, refusing to eat, and rebuking many of my questions, demonstrates this. Having contacted me to request to be interviewed, he saw me with suspicion. Weeks later, he emailed me to suggest ways pediatricians could take vaccine rejection and fears of autism more seriously: ‘The next time your husband has an opportunity to see an “autistic” child, would you mind showing him a list of some blood tests to choose from?’ This request accompanied a link to a site that questions whether autism diagnoses might actually be immune system dysfunctions, presumably caused by vaccines. Others asked me to ask him to locate definitive proof in pediatric journals that vaccines really work. In these ways, their knowledge of my personal life and ties to mainstream medicine gave respondents pause, but also inspired them to remain in communication in ways that continue to communicate their efforts to challenge that authority – long after data collection ended.

Knowledge of my connection to mainstream pediatrics worked in other more positive ways through communication after trying to exit. One participant, a mother who home-schooled her four children, was very distrustful of research and had several negative experiences with allopathic medicine. As she watched her infant son deteriorate from an undiagnosed digestive problem, she felt helpless. She watched her son vomit daily and fail to gain weight, and though she trusted a naturopathic doctor and acupuncturist to treat him, he was not improving. As his condition worsened she accepted a referral to a mainstream pediatric specialist. Knowing my connections to the local children’s hospital, she emailed frequently – months after our interview – with questions that ranged from background on different specialists, my thoughts on how she would be treated by doctors treating a condition that was not infectious but who might be concerned she opted out of vaccines, to recommendations for a new primary care physician in her town who would accept her parenting choices. She saw me and my familiarity with mainstream medicine not as a threat but as a resource. Though I could not attend the appointments with her when she asked, she perceived that I could be in constant proximity by email, eager to help her son, should she need help. Although this mother emailed me with a similar or greater frequency as the father in the last example, her desire for help allowed me to give something back to her that she needed, while I also gained insight into her broader fears of the healthcare system. Although many researchers are seen as potential allies, advocates, or even curious symbols of status and prestige (Bourgois, 1995), these kinds of email contact – both requests for help and inquiries intended to be confrontational – could continue indefinitely in the digital age.

There arises symmetry of interests, then. I want to understand their worlds, just as participants may wish to understand mine. Our motives, however, differ. I pore over their online utterances or transcriptions of interviews, to construct and reconstruct representations of them. I imagine they want to understand whether I can be trusted to represent them fairly, accurately, or in ways that support their political goals. Do my interests align with theirs? Making this clear, one participant advised me that if I were ‘a partner in this’ he could help me tremendously. In general, sociologists are unlikely to agree with all the policies of the organizations and individuals we study, though many choose to study groups with which they agree. These questions then are most perplexing as we study groups with which we do not socially, politically, or ideologically align.

For this study, I collected data from groups and individuals who position themselves in opposition around the meanings of childhood vaccines. The access this respondent was requesting, including information about my personal life and networks could potentially complicate access to data in ways that are familiar to many. Although I aim to present myself authentically in person (by disclosing, for example, to parents that opt out that I do fully vaccinate my own children, or to evangelical Christians that I am Jewish), doing so online feels more potentially damaging to my study. Managing participants’ assumptions about our common allegiances remains complicated on and offline and may affect how participants feel about the research when it is completed and as we leave the field.

Disseminating results

The questions raised in this paper represent a fundamental shift in contemporary qualitative research. Traditionally, ethnographic studies – dominated by a sort of cowboy ethnography, where the lone researcher heads out to the proverbial Wild West – have relied on spatial embeddedness where the researcher pierces a subject’s world. For example, Whyte (1993) moving into an Italian neighborhood, Bourgois moving his family into El Barrio (1995), or Kornblum (1974) working on a factory floor, Stack’s move with her son to a poor black community in the Midwest (1973), or Venkatesh’s entrance into gang life (2009) have all been valorized as deep membership and personal discomfort yielding rich data. In each case, researchers left the field and returned to their offices when they chose (Duneier, 2000a).

Inversely, new media make it possible for participants to potentially embed themselves within the researcher’s world. As lectures, vitae, publications, and professional information become increasingly publicized on websites, participants gain information about researchers’ professional lives. More generally, technology makes researchers’ lives visible in ways not common when one could leave the field and go home. They also raise more basic questions of whether a researcher can ever truly exit the field, finish data collection, or sever ties with participants when he or she is always reachable in nearly real time.

Months after completing interviews, participants who advocate against using vaccines often followed up with me via email, asking some version of, ‘How is it going? I am curious to see what your findings are so far.’ Another participant searched – presumably online – for my home phone number, which I do not give out, and left messages to let me know that he was ‘checking in’. This is hardly new, as many researchers encounter research participants later who want to know when they can read the study, whether it is published yet, or what we found. Yet, online communication makes these inquiries easier and seemingly ever-present.

Online communication also allows participants to provide feedback on research faster and more readily than researchers have traditionally experienced. Scheper-Hughes (2000) describes the anger she faced when she returned years later to the rural town of ‘Ballybran’, Ireland, the site of her celebrated ethnographic research. Yet, new computer-mediated communication would allow, for example, Ireland to come to her and share their ‘ire’ more immediately, directly, and potentially continually. Wynn’s research (2011) on tour guides, mentioned above as an example of how his informants endorsed use of their online personae, recalls one participant who posted a strongly worded response to his book on a website called ‘never trust a sociologist’. Despite signing a waiver stating he reviewed his portrayal in the manuscript prior to its publication, he nonetheless felt unhappy when the book was actually published and let Wynn and others know in an online medium.

Similarly, in reflecting on her research on ‘natural mothers’, Bobel (2002) recalls discovering a negative response to her book in the early days of the internet, as she was described on a listserv about mothering as having deceived her respondents, even as sociologists saw her work as overly sympathetic to their cause. A sense of betrayal, frustration, or disagreement with social research is also not new. But as these examples and many others show, new technologies allow participants to respond quickly, directly, and to a wider audience than was ever possible. In these ways, engagement with researchers may never have a natural end. Savvy participants committed to challenging representations or findings can often disseminate their views to a wide audience quickly with social media, perhaps quicker than scholars can. This signals both a shift in participants’ power and a transformation in how information can be disseminated and consumed.

Implications and future considerations

Qualitative research has a long tradition of examining power relationships in research and reflecting on challenges that arise in the course of data collection. As researchers who design studies that aim to collect data about the lived experiences of respondents, new technologies and their ubiquity in offline life require thoughtful examination of the meanings of consent, confidentiality, and how online communications represent and affect offline personae. Even research that intends to study offline populations must address the effects of technologically-mediated forms of communication. The discrete phases of research – from recruitment to field access, to data collection, to exit – may not capture the actual experiences of conducting research, particularly as new technologies change the meanings of these phases. As technologies become increasingly integrated in our daily lives, there is more slippage between material and virtual worlds than we readily acknowledge that demands new discussions of methodology.

First, I want to suggest the need for greater cognizance of these issues of confidentiality, ethics, and boundaries. As our access to information about participants increases, so potentially do our data. Yet, as we include information participants may not know we have or may not have freely offered, researchers should question whether they have honored the spirit of informed consent. This is neither condemnation for using information that helps to present a richer picture of participants nor permission to use anything available without hesitation. Rather, researchers should think hard about their choices as they collect and analyze their data. Even as I quote emails and conversations with colleagues, I struggle with when I should return for additional consent to use these examples. Calling for greater awareness and bringing these dilemmas to light does not necessarily solve the issues, but it does allow qualitative researchers to better account for their methods, and the messiness within them. I would advocate for greater accounting for the complex and messy processes of recruiting participants, gaining access, and use of technology in methodological appendices or other forums so we might understand research differently and learn from each other.

Second, researchers should prioritize greater vigilance in identifying what representations of them exist and how much they affect data. Much literature demonstrates how researchers’ identity and embodiment affect data and how qualitative data collection is intrinsically interactive (Arendell, 1997; Hutchinson et al., 2002; Lareau, 2000; Pierce, 1995; Bourgois, 2000; Stanfield II, 1994; Kong et al., 2002; Schwalbe and Wolkimir, 2001). It is worth considering that disembodied virtual representations – ones actively crafted and those that emerge from remnants of past online actions – also shape how the researcher and participant perceive each other in ways that might differ from offline impressions only. As we develop methodological rigor that situates the researcher in the social world, we should also take into account better consideration of these issues as well.

Third, ethnography unbound – without borders or boundaries – often presumes common interests with one’s research subjects. For those who define themselves as scholar-activists, the ability to be clearly identified with the cause could increase access to data. However, in studies where the researcher aims to capture multiple perspectives or while studying those that do not share a political affinity (Blee, 2003; Luker, 1985), these questions are more challenging, with a great deal at stake. In such cases, the ability to write critically, even as we lose the capacity to ever fully leave the field, raises new questions about the consequences for researchers. These risks are also quite real for those who conduct covert research. In fact, the ability to even conduct covert research may be more difficult, as researchers’ identities are easier to locate. In general, the outcomes of such research might be more serious than they were prior to the widespread use of social media. As new ethnographers are trained, they should be encouraged to not just consider how to access a field site, but also the process by which they will extricate themselves as well. This might include asking young scholars to consider their relative comfort with ongoing communication after data collection has concluded, potential negative responses of participants, or knowledge participants might seek about their personal lives. Doing so might also require communicating across generational differences that shape experience with social media and differing expectations of privacy. Nonetheless, these are important discussions to have with greater frequency.

Fourth, because of technology, participants have new avenues with which to voice their opinion of research; the stakes of them doing so are potentially high. Some examples: Piven received threats after a news program pundit vilified her during network broadcasts and on his blog for an article she wrote more than 40 years ago (Stelter, 2011), while Joffe has for more than a decade been listed on the anti-abortion website (formerly known as the Nuremberg Files) that details, chronicles, and celebrates the injury or murder of physicians who provide abortion services and their allies, including clinic owners, staff, judges, politicians, police officers, and researchers. ³ Although no research participant has claimed responsibility in these cases, they illustrate the erasures between real and virtual worlds and the ways researchers lose control over the interpretations of their work and can be more directly targeted as private information is increasingly available.

Online communities, web searches, and the digital fingerprints we leave may potentially jeopardize access in ways we have not fully considered. This newly-facilitated access to information can also potentially flatten the power differentials between the researcher and the researched and provide new resources to potentially empower participants to make decisions about what participation means to them, their communities, and the issues for which they advocate. They also reconfigure the ways researchers design and operationalize data collection that must be more honestly considered and disclosed. One example supports this latter point. One anonymous reviewer of a manuscript responded to my descriptions of methods, with mention of contacting participants through Facebook messaging, as potentially violating the Facebook terms of service, which specify ‘If you collect information from users, you will: obtain their consent, make it clear you (and not Facebook) are the one collecting their information, and post a privacy policy explaining what information you collect and how you will use it.’ ⁴ Although I did not ‘collect information from users’, I did learn information about them – just as they learned about me – and found this a complicated issue to address, both to the satisfaction of an anonymous reviewer who suggests I should treat corporate user agreements as ethical guideposts or within the policies provided by a for-profit corporation with its own distasteful research practices (Verma, 2014). Thus, this article suggests that qualitative researchers should be more honest about our methods in pursuit of best practices for ethical research that respects participants’ autonomy, while also acknowledging the new challenges and opportunities provided by social media and online communication.