Amy Cooter is a PhD Candidate in Sociology at the University of Michigan. Her dissertation research focuses on a Michigan organization that is very internet savvy. You can read more about it on her department profile. This post is part of a series about why and how sociologists should study the internet.

I study what might generally be termed “sensitive” populations—those who, for a variety of reasons, are wary of outsiders. The internet has been an invaluable source of information in my research, but it is not one without potential peril. I use online data sources to supplement more traditional sources (interviews and ethnographic field work, primarily), and I think it’s very important to note that I do mean “supplement” here, not “replace.”

In my experience in the field, Facebook in particular has been a treasure trove of information for contextualizing or challenging information I have received in person. I have been in the field for more than 3 years now with my primary population of interest, and have good reason to believe that most of the viewpoints they share with me are truthful reflections of how they believe they see the world. However, I only see my respondents in person about twice a month. Facebook, other social networking sites, and my population’s forum all serve to alert me to interesting developments that happen in the intervening time. Additionally, I’m unlikely to see my participants in person when they are experiencing stressful events at home or at work. Internet access to my participants lets me see how they respond during these scenarios.

Importantly, this is not just information that fills in a gap in observation. People generally are more polished and, for example, more apt to express egalitarian sentiments when they are in a public space and when they are not experiencing abnormal life stressors that might activate typically-suppressed prejudices. Facebook and Twitter in particular have become records of more temporary and sometimes more poignant “statuses” that reflect how people are feeling in a given moment in response both to the occurrences of daily life and to calamities that interrupt the mundane. Instances of this later type can yield very interesting information about what may (or may not) undergird some of the more thoughtful and intentional statements in the field.

For example, a 33 year old, middle-class, white, male interviewee (let’s call him Robert) pleasantly surprised me with detailed knowledge of and empathy toward Detroit’s history of racial discrimination during our interview. Robert more generally talked at length about how he could completely understand that African Americans may not feel patriotic in the same way he does due to their long history of oppression in this country. He was conversant about Civil Rights leaders and legislation, and largely agreed with more contemporary affirmative action efforts. Several months later, Robert nonetheless posted an unbelievably racist message on Facebook after narrowly avoiding being victimized by a young, black male.

This is not to say that the anti-racist message Robert gave me in person was disingenuous. Rather, it seems that Robert genuinely tries to adhere to egalitarian norms and believes he acts in an accepting, anti-racist manner; however, a particularly stressful event primed racial stereotypes he was undoubtedly exposed to as a child and revealed that Robert’s sentiments were not as internalized as he wanted to believe. Without Facebook, I would not have been privy to this comment, or to how Robert’s defensive response after some of his friends’ chiding.

Generally, in my research, data from social networking sites has been more personal (e.g., data on relevant family dynamics), more spontaneous, and less filtered. Facebook posts have also provided clearer examples, compared with in-person interactions that occur in a group context, of individuals who disagree with their group’s ideology in some respect. These people feel uncomfortable voicing political opinions, for example, that are counter to those of the group when they are at a group meeting, but still evince their true feelings online. Having this additional information is very helpful when formulating interview questions and assessing adherence to group ideology and goals.

Because my participants’ online posts are often qualitatively different from what they express in person, I use both data sources to have a complete picture of the population. If, for example, I relied only on participants’ forum posts, I would understand them to be much more paranoid about news events than they really are. Contextualizing these posts alongside in-person information lets me gauge the extent to which these posts are, instead, largely just fodder for conversation to bridge the gap between their in-person interactions.

Using the internet as a data collection site can be tricky for other reasons. To paraphrase one of my early mentors, if you’re observing a chatroom interaction of three racists (his area of work), one is likely to be a student, one an FBI agent, and you’ll never know which one is which. You may also have difficulty sorting out which so-called groups may be the work of a lone, angsty teenager, who may be the subject of John’s research. There are strategies for improving one’s methods while researching online (such as using only closed forums where members must be approved by a group’s leaders, though access here may be another challenge), and more strategies will be developed as the number of researchers using internet data increases. There may still be special challenges if you are a qualitative researcher of sensitive populations.

I will be discussing this further at my ASA presentation this year (Monday 2:30, Regular Session: Qual Methodology II, room TBA), but there has thus far been very little work done on what it means for researchers themselves to have an online presence while in the field. Most academics minimally have some presence on a departmental website. How you describe your work there may either positively or negatively impact whether potentially research subjects will want to engage with you. (If you study CEOs of local companies, would they want to cooperate with you if your website info contained a paragraph about the evils of capitalism?) The personal information and contact information you make available may have negative consequences if you study potentially dangerous groups. Things like this seem like common sense, but I have been surprised at how many sociologists continue to say, “oh, I never thought of that,” even at this point in the conversation.

Less obvious, perhaps, is the decision to engage with your research participants on social networking sties. This was not a decision I took lightly, nor one I made immediately upon entering the field. Ultimately, it became clear that choosing to avoid Facebook interactions with my population amounted to choosing to not having a certain kind of data. I had fortunately worded my IRB application broadly enough regarding my interactions with participants that I did not have to quibble about whether social networking interactions should be considered data. Almost all projects with fieldwork or interviews should consider including online interactions in their IRB applications.

Most groups are likely to have some kind of online presence, and you should assume that potential participants will look you up on Google as well as social networking sites. Even if you choose to avoid friending participants directly, you should take care to ensure your privacy settings are adequate for your needs. Do you want participants to see your political or religious affiliation? Do you want them to see pictures that other people tag you in? Keep in mind, too, that privacy settings are not always reliable. Shortly after I friended my first participant, Facebook had a major update and some of my profile information reset to an earlier state where I had more personal information listed than was ideal.

If you choose to have an online presence while conducting fieldwork, you should regularly check your social networking information, and you should Google yourself at least every 6 months to ensure that only the information you want participants to see is readily available. Websites (like intelius.com, spokeo.com, or ussearch.com) that aggregate and list personal information are continually developed, and, as of now, must remove your information if you request that they do so. Depending on your privacy concerns, you may need to request that spouses, parents, siblings, and even long-time roommates remove their information, too, as some sites link your information to known relatives, or to people at the same address.

Issues of privacy and perception are relevant for most qualitative researchers, even those who do not research the internet or use the internet as a data source because their participants may use the internet to investigate them. Privacy is probably a good idea for everyone anyway in this era of rampant identity theft. Being attentive to these issues can add to the hassle of conducing research, but the data payoffs can also be fantastic.