We present our findings in three sections. In each section, we highlight one specific example of analytic labor in response to the challenges generated by one unique property of Facebook environment (see: Appendix II).
A first example of analytic labor: Users' reflections about old friendships
One major characteristic that makes Facebook environment unique concerns the use of the website for re-initiating contact with old friends, which entails that dormant relations from the past are transformed into visible relations in the present. Although it has long been pointed out that Internet can provide opportunities for keeping in touch with friends who moved to a new location (Hampton & Wellman, 2002), Facebook's capacities in this respect seem to surpass all previous channels (Ellison et al., 2007).
We devoted one section of our interview to this topic. The overall scenario we observed was approximately as follows. Almost all our informants (35/36) sought – or themselves were contacted by – their old friends through Facebook. Many (27/36) of them ended up meeting with some of these old friends in person. The outcome of these meetings, however, was not always a success story. For the great majority, after a few gatherings, the initial excitement receded. Some informants reported that they continued to communicate with a few, but not with most, of these old friends. Nevertheless, most old friends, once contacted, remained in friend lists.
We wanted to delve into the background of this scenario and capture the experience of users by observing the dynamics in the reactivation of old friendships that go years back in time. Thus, unlike most studies on Facebook that exclusively use college students, we included many elder users.11 We believe, mainly due to this extension, our interviews reveal an ambivalent attitude that is not observed in previous studies.
On the one hand, almost all informants maintained that old friendships have a unique place in their lives. In particular, they noted the lasting effects of sharing something in common at a special period in their lives, stressing the difficulty of forming intimate friendship bonds in later ages. On the other hand, the special value attributed to old friendships was overshadowed by a kind of skepticism about effects of time. Both prior to and in the aftermath of encounters with old friends, informants were highly conscious about the changes in their own and their friends' lives.
One possible explanation for this ambivalent attitude might be that on Facebook, temporal distinctions about the past and present, which are more easily sustained in everyday life relations,12 are partly blurred. But the most important point for our concerns was that such encounters triggered in the informants a host of reflections about the nature of friendship and effects of time. For example:
Twenty five years since we have not seen each other … then thirty years later, it's suddenly my dearest pal, etc. We hug each other … I don't know the reason … I observed the same in my wife … Maybe it has something to do with age … (M, 51, Teacher).
While such general reflections were more common among elder users, informants from all age groups seemed to be concerned with a second question about what kind of a value a reestablished contact with an old friend constituted. They evaluated old friendships according to three distinct value orientations. Sometimes these relations were portrayed as “scarce commodities” in modern society because they involved a shared history and survived the test of time. At other times, especially in the aftermath of disappointing meetings, they were interpreted in instrumental terms, as potential sources of social/material benefits. Finally, sometimes informants characterized these relations as something to be enjoyed and experienced for their own sake.13
It is worth noting that these distinct types of evaluation were not mutually exclusive and could often be observed in the narrative of the same informant. Here are two examples:
… I'm thinking about the simplicity of sharing things at that time [i.e. in the past] … but when you look at it today … the relationships we know are relationships within the confines of an institution [workplace and university]. By now, we have already made our choices. … Maybe [those past relations] are more innocent, less based on expectations, but of course, … when you look at it at the end of the day, there are moments you realize that you strayed away a lot from your friends in primary or high school. There, after that point, they remain merely as people in the list. (M, 24, Undergraduate).
In this first example, one can easily note that the informant portrays old friendships in two distinct ways: first, as a “strong tie” between people without any instrumental expectations, and secondly, more as a “weak tie,” as “people in the list.” A second informant provided a more extended typology that, among others, included an esthetic/sentimental criterion, which corresponds to the third type we mentioned above:
I believe this [i.e. contacting old friends] might have sentimental reasons. ‘Ah! Let me remember the old days, let me go back to those days,’ it could be this, or you might think that people from the past know you better. … Another reason might be … what did he do, what he has been up to … let me not sever the ties, it is impolite towards that person, that may be another reason. Yet another reason might be because of work and economic reasons … (W, 27, Graduate student).
Informants often used these different types of criteria to make sense of their encounters with old friends. This analytic labor was largely facilitated by technical capabilities of Facebook. Users could see their past ties as part of the overall map of their present social network. In that map, they could identify some relationships that worked and could be reactivated as a friendship tie; then there were others that did not work so well, so they were delegated to the category of weak ties; and, finally, in a few cases, it would be best to disregard them completely: “if these were meaningful relations, they would not have ended …” (W, 21, Undergraduate). Already at this stage, then, our informants were carrying out an amateurish analysis of their social ties and networks. Below, we will encounter further examples.
Second example of “analytic labor”: Systematic observation of other users
That Facebook users engage in the observation of other users' activities is mentioned in a number of studies. Raynes–Goldie (2010) gives several examples of how some of her informants “repurpose Facebook's design to violate the privacy of others.” Similarly, Muise et al. (2009, p. 443) note that “Facebook gives people access to information about their partner that would not otherwise be accessible” and argue that this leads to increased feelings of jealousy.
We should not, however, conclude on the basis of these findings that this is an exceptional practice. Data from other studies (Stutzman & Kramer-Duffield, 2010) and our own interviews indicate that observing other users' activities is a much more widespread practice than it first appears. Furthermore, there is evidence that such observations are not carried out randomly. Stutzman & Kramer-Duffield (2010, p. 1558) note that when they asked, “how many Facebook profiles … [their informants] had looked at in the past week. The mean answer, 6.4 (SD=9.3, 0|100) indicated that profile-viewing attention is focused primarily on a subset of the friend network.” Our informants' comments also indicate that, when they make observations, they are focused on a select group of users. This is quite remarkable since it implies that users carry out these activities in a conscious and systematic manner.
Interview reports indicate that the practice of observing others takes different forms and is performed with specific intentions. We were able to identify three major types of observation, which can be analyzed in terms of three basic variables: whether they are oriented towards present, future, or past relations; whether they target specific individuals or groups; and, whether they focus on weak or strong social ties. Let us look into each type closely.
Type 1 - “Checking Special Cases”: The target here is specific individuals and weak ties. It can be oriented towards both past and future relations. The most common version of the former was users' inquiries about old friends:
I mean, how many kids does he have for example, or where is he working, … last time we were together, it was at the technical school. (M, 42, Technician).
I was very curious about what my old friends, I mean, my primary school friends were doing, like, who is where, what are they doing, are they studying? Have they finished school? Are they married, where do they live? For example, I learned that a friend of mine passed away. A friend from high school. (W, 22, Journalist).
There were also future oriented observations of this kind. Some users observed potential candidates that could be included in their friend lists: “ I am making him wait in the waiting list … I am watching him closely … I might accept him in the future” (W, 37, Housewife). Others were interested in gathering more information about people whom they vaguely knew at present; to paraphrase a statement we heard in many interviews: “I do not look at everyone's page. But if I notice an interesting person, I try to find out more about him/her.” Another informant described the following future oriented use:
I think what makes Facebook to stand one step above other communication means is that in addition to communicating, two parties can follow each other's activities … It is a bit like this … take this guy, he has a conservative disposition, takes a look at the pictures [of a girl] … and, to put it bluntly, [he sees that] there is no ‘fault.’ Aha, he says, this is my kind of girl. (M, 20, Undergraduate).
Type 2 - “Keeping in Touch With the Group”: Unlike the previous type, this one is present-oriented and targets groups. Sometimes, the group observed is more or less identical with the user's social connections on Facebook: “Frankly, I check it everyday. I mean, like, regardless of whether I have work to do or not. Who's done what, where, and so on (laughing). Every evening I look at it this way and then I turn it off and do other stuff.” (W, 26, Technical drawer). Some users, however, carried out these observations to keep in touch with more specific groups, particularly with friends and relatives in other cities. “I mean, because I am away from [name of hometown], I cannot see my friends … where are they working, what are they doing, I mean, how old are their kids, how many kids they have, this is the kind of stuff I see there.” (M, 42, Technician). The best visual imagery with respect to these types of observations was provided by a female restaurant manager in her mid-50s: “I watch the children grow.” Another noteworthy comment was made by a male graduate student (25) who viewed his friend list like a microcosm of his society at large:
They are very different people in the end … Therefore, you understand what is on the agenda for a right wing person, or what is on the agenda for a left wing person … I mean it's like you follow things from one newspaper and it is very one sided … But because on Facebook everybody shares what they see as important, it is more pluralistic. … I mean they are like … like a sample to me right now … like a sample of society.
Type 3 - “Scrutinizing Significant Others”: Although, like the first, this one also targets individuals, and like the second, it is present-oriented, this type is exclusively oriented towards strong ties. The most common form here is scrutinizing the interactions of a romantic partner (see also: Lewis & West, 2009; Muise et al., 2009). Alternatively, such scrutiny might also be carried out by parents. Here is how a female teacher in her mid-40s narrates her reasons for joining Facebook:
I have a twelve-year old son. This year, he joined Facebook because all his friends are on Facebook … So, I need to keep an eye on him because on Facebook one can get all sorts of different mail, invitations for friendship and so on … Frankly, you need to keep this under control.
There is, however, one characteristic common to all the types discussed above: The aim of these observations is not merely to gather “personal information” about other individuals but also to scrutinize the relationships between individuals. The interview material indicates that, in addition to the friend lists that give a glimpse of a user's social network, photos, and comments posted on Facebook can provide much information about a user's relation to others. Several informants noted how, from a photograph, one could learn not only where, when and with whom a user was, but also the more intimate details of what they were doing. Similarly, comments on a user's wall were read to discover the nature of the relationship between the user and writers.
Finally, it might be worth noting that, although observing other people in everyday life contexts is not an unusual activity, this rarely reaches to the same level of systematicity as in SNS. Conversely, “sampling” a specific subgroup in the population and systematically observing them is a typical activity in some professions. One notable case here is of course sociologists, whom Berger (1963) once famously characterized as “professional peeping-Toms.” Facebook users seem to fall somewhere between these two extremes, though, perhaps, they are closer to sociologists to the extent that they too take “samples” of their group and regularly observe them.
Facebook users, however, are different from sociologists in at least two respects. First, although being able to observe others was rather appealing to most users, they also felt quite uneasy, even slightly guilty, about this: “I mean, well, this is curiosity. I can't say I didn't do it. I am doing it. Everybody is doing it too (laughter)” (W, 32, Sanitation worker). Secondly, and perhaps more importantly, users were overly conscious about the fact that they themselves could be objects of observation:
I see Facebook as a kind of “Somebody is Watching Us” episode;14 … Everybody joins it out of curiosity about others' private affairs … and I too get caught up in such curiosity sometimes … After all, this is one of the biggest reasons why it is so successful … To see what other people are doing. (W, 55, Manager).
In fact, some users seemed to utilize the knowledge they gained from observing others reflexively, when they strategically displayed their own activities to others:
If there is someone I was very close to once and if nowadays things are not so good between us, then … you can do something like showing him that here I'm going on with my life fine … And, he will be watching this at this time, and I am able to do something which I cannot do in normal life, like showing/marketing myself to other people … I know what is going on between a friend of mine whom I knew when I was in the ninth grade and her present boyfriend … (W, 21, Undergraduate).
As these last two comments suggest, being observed by other users is the counterpart of being able to observe them. Indeed, users often carry out such observations with the aim of finding out more about their own observers. This is what we turn to in the next section.
Third example of analytic labor: Classifying and evaluating other users
The problem of collapsed contexts
Recent research have revealed that Facebook users are often less concerned about being observed by complete outsiders than by people in their own social circle (Raynes-Goldie, 2010; Stutzman & Kramer-Duffield, 2010; West et al., 2009). While similar concerns might also arise in offline environments, they are experienced more acutely on Facebook due to a unique property of the website: Heterogeneous segments of a user's network can be combined together in the same “list” and the user's interactions with these different segments become potentially visible to others. To use boyd's (2010) terminology, on Facebook, different contexts of interaction tend to “collapse” into a single environment.
As a result, users face several challenges that they seldom encounter in offline environments. One such challenge concerns the risks involved in blending social ties from the past and the present in the same list. As one informant put it bluntly, “there are certain things you do not want to share [about your past]” (M, 26, Graduate student). Similarly, a second challenge stems from lumping together social ties of varying strength (Donath & boyd, 2004; Gross & Acquisti, 2005; Spencer & Pahl, 2006; West et al., 2009) – e.g., when formal acquaintances from the workplace are included in the same list along with best friends. The following comment reveals the nature of these challenges in the context of generational differences:
Now, if it's a very close relative, for example, a cousin or a brother … In that case, you can treat them like a friend. But, you know, [if the person is] an uncle or an aunt, then because of generation difference, they have more traditional attitudes. They might find your frivolousness … [on Facebook] weird … Because normally you do not show them that face of yours and we have masks we put on in society. (W, 27, Graduate student).
One general problem users encounter in collapsed contexts, then, is the difficulty of deciding which “face” of their self to display (Debatin et al., 2009; Donath and boyd, 2004; Lampinen et al., 2009).15 This problem is inseparable from another, which is, in a way, its flip-side: the difficulty of assessing the expectations of the audience to which the self is being presented (boyd, 2010; Hogan, 2010). In most everyday life contexts, it is possible for individuals to interact with different circles of their network in different places and at different times. In each interaction, therefore, individuals have a rather distinct sense of their audience. However, when these different circles are included in one list, the characteristics of the audience become ambiguous, making it difficult to decide which self-expression one should display.16
As we shall soon see, most of the analytic labor performed by users aim to tackle this problem. Before proceeding further, however, a brief conceptual clarification might be necessary. As is well known, Goffman's theory (1959) about “presentation of self” is central to many studies dealing with online identity formation on SNS, where information disclosed by users (about their tastes, political opinions, lifestyle, and so on) is conceptualized as socially significant identity expressions (Lewis et al., 2008a; Livingstone, 2008; Tufekci, 2008). In fact, sometimes, even information about the social ties of a user is also conceptualized similarly (Donath and boyd, 2004; Utz, 2010; Tong et al., 2008). However, it is important to note that the presentation of self in an environment made of collapsed contexts might require more complex forms of reflection – more “analytic labor” – than the cognitive processes that Goffman (1959) observes in such cases where an actor usually tries to tailor his/her self-expression in reference to a single set of expectations.17 In fact, Goffman (1951, p. 296) himself recognizes that “in situations where complex social judgments are required, the exact social position of a person is obscured and, in a sense, replaced by a margin of dissensus and doubt.” We shall return to this point.
The Facebook users we interviewed were aware of above mentioned challenges. Yet, none of them expressed a strong disappointment. Here we tend to agree with Lampinen et al. (2009, 289) that this is mainly “due to successful use of the proactive management strategies.” Since this is an amply discussed topic in the literature, we shall not reproduce a list of these strategies here (Donath and boyd, 2004; Hogan, 2010; Lampinen et al., 2009; Raynes–Goldie, 2010; Stutzman & Kramer-Duffield, 2010; Tufekci, 2008). Instead, we will focus on how, as a prerequisite of strategy formation, users perform various types of analytic labor to develop a sense of their audience.
Developing a sense of the audience: classifications of people in the friend list
The most basic form of analytic labor we observed was users' attempts to develop a map of their lists by classifying their Facebook friends. Here are a few examples:
Now, I can say it this way. Like, it might not be very possible to say it in percentages but approximately 10% or 15% are relatives. A large portion are people I studied with at primary school, high school, university … Beyond these are people I share similar viewpoints. (M, 24, Graduate student).
Actually, most friends I feel close to are not active on Facebook … I mean people who populate my Facebook page are like … if I grade the level of intimacy from 1 to 4 … they would stay at 2 and 3. Neither people I am very close to, nor people I am not close to populate my Facebook page. (M, 25, Graduate student).
I mean it's a bit like this. I have some extreme friends. But if I have to make a general distribution, let me say it this way. You know, there is a curve like this. In the middle of that curve there are mostly master's and doctoral students … people with a definite level of education … In the edges, there are friends from primary or high school. These are few in number. They too have some education. There are then a few foreigners …(W, 27, Graduate student).
Examples can be multiplied. What we need to emphasize here, however, is that in each classification above, different dimensions (kinship, intimacy, education) are used and blended in different ways. These classifications, therefore, have a different nature than those observed in some previous studies, which consist of relatively fixed categories (Lampinen et al., 2009). They are not abstract generalizations about categories of people. Rather, each user tends to tailor the categories in these classifications somewhat differently. This is how, for example, a student informant who was actively engaged in a political organization classified his list:
I mean, there are friends who are apolitical, but in terms of education, [they are] studying in the same department as me. There are friends who do not have the same education as me but who have the same political views … Beyond these there are friends who have nothing in common with me in terms of politics or education but whom I know for other reasons. (M, 20, Undergraduate).
The crucial point is that these categories reflect the user's main concerns (politics, education) while managing his interactions with the unique conglomeration of individuals in his friend list. In other words, these classifications function as a kind of analytic grid for the users to develop a sense of their audience, allowing them to become more conscious of their network and their own position in it.
3.3.3 Analyzing and evaluating online behavior of other users
Our informants also mentioned a number of general criteria that they used in making selections for their lists. The first was the degree of acquaintance. For some, political views also mattered, especially if they were expressed strongly – although many reported that their lists included friends with different political views. Another criterion, especially for users working in a profession, was whether there existed a formal status difference between them and the “candidate” (as in the case of a teacher who did not accept students to her list). Interview material suggests that the way in which users develop these criteria is influenced by their online activities and social background. For example, the importance attributed to degree of acquaintance seemed to vary negatively with previous Internet experience. Similarly, sensitivity to political views vary depending on informants' engagement in politics. A full exploration of these relationships, however, is beyond the scope of this study since our primary objective is to investigate widespread practices rather than differences.
In addition to the above criteria, all informants cared about what might be called the “Facebook behavior” of other users, which they mainly evaluated in relation to what kind of information a user discloses on the website and how. In this context, four distinct criteria were mentioned.
The first concerns the frequency with which a user posts information (comments, news, pictures, etc.). The main question here is how talkative or passive/quiet a user is. A second criterion concerned how private the content of information that was being posted. Informants often made reference to “some users” who posted irrelevant details about their private lives, which they often described as annoying. A third important criterion was the manner in which a user interacted on Facebook. This is essentially an ethical criterion used for evaluating to what extent a user cares about not disturbing others. Typically emphasized issues included how emotional a given user was when making comments to others and whether his/her comments involved aggression or “fanaticism.” Finally, a fourth criterion was related to the intentions of users; that is, whether a participant used Facebook for sociability or for instrumental reasons such as advertising his/her business.
Beyond its empirical significance, the utilization of such criteria for evaluating online behavior of others also has important theoretical implications. First, it means that even if the users try – as is often suggested in the literature – to form an online identity, they do this on the basis of a comparative analysis: Their self-presentation is performed according to a mental scheme consisting of multiple categories of user identities. Typically, they expressed this in statements like, “I am such and such a user, I post these kinds of comments, I do not do what some others do,” and so on:
I mean, one thing that strikes me is … for example … I had some friends from high-school whom I met on Facebook … some of them share too much of their private lives. On Facebook I post photos too but I post … photos of the city and things like that. I mean some of them are like … they are living their lives on Facebook … There he puts the pictures of where he has been last night … there he says, there I was last night … right away, the next day, they make comments, they talk about last night … I never do this. (M, 26, Government employee).
Furthermore, although to an external observer SNS might appear as sites of exhibition rather than performance – because the the stuff users post on SNS have the characteristics of an “artifact” rather than an “action” (Hogan, 2010) – such comparisons indicate that for the observing users, the manner in which an observed user displays information is interpreted as performance. In fact, when prompted further about this issue, some informants provided rather interesting, albeit improvisatory, classifications of “Facebook performances”:
Some use it like a therapy group; some use it to spread their views; some truly do it for friendship purposes and for having fun; some use it to enlarge their social circle … for some, being accepted or rejected as a friend is a matter of life and death …(W, 47, Teacher).
To summarize, all this indicates that, in trying to make sense of their audience, participants utilize various kinds of norms and criteria about user behaviors on Facebook. Moreover, through these criteria, they classify other users and try to explain or interpret their behavior.
Finally, let us note that sometimes these interpretations take the form of a general skepticism about the authenticity of online behavior:
I'm thinking that through Facebook people's capacity for creativity and humor increases because they reflect way too much to make comments and status changes and so on. … I mean, because it is something everybody sees, maybe it is done with more thought …(W, 30, Manager).
It may be that unrealistic self-presentations are more difficult on SNS due to the presence of offline acquaintances (Utz, 2010). Nevertheless, in a similar fashion to scholars who underline the constructed nature of self representations in online environments (Ellison et al., 2006; Turkle, 1997; Walther, 1996), our informants kept on stressing that Facebook was different from the “real world”:
Sometimes it appears as if you are very intimate … I mean, in reality, if I come together with him and sit in a café, I know I won't be so successful in that conversation. … There [on Facebook], you can behave towards him, how should I say, unlike what you are in reality …(M, 26, Government employee).
Thus, while, as Goffman (1959) observed, social actors might be routinely engaged in modifying their self-expressions in everyday life, the users we interviewed seemed to be all the more acutely aware of the ubiquity of this practice in Facebook environment. At this stage, however, it might be best to leave it as an open question whether this awareness might add a critical dimension to the analytic labor of users in the long run.