“The Question Exists,but You Don’t Exist With It”: Strategic Anonymity in the Social Lives of Adolescents

Research Site: Ask.fm

Founded in 2010, Ask.fm, available in 49 languages, is a Q&A site with approximately 150 million members (About us, 2016). The original version of the platform was associated with cyberbullying in the popular press, but was purchased in 2014 by Ask.com with the goal of making the service “safer and positive” (About us, 2016).

More than 50% of registered Ask.fm users are under the age of 18 years (Van Grove, 2013). Ask.fm differs from other popular Q&A sites (such as Yahoo! Answers or Quora) in several key ways. First, rather than soliciting multiple answers by broadcasting their question to a large audience, questions are directed to individual users via their profile page. Importantly, only questions the user chooses to answer are displayed on her profile. Questions can be answered (in which case they are displayed on the user’s page and in followers’ streams), ignored (in which case they remain in the user’s question queue, but are never displayed on the user’s profile nor in followers’ streams), or deleted (in which case they are not displayed on the user’s profile nor in followers’ streams and do not remain in the user’s question queue).

Personal profiles on Ask.fm are minimal: individual users populate their personal profiles with a location, links to other personal websites or social media profiles, and a biography (with an unlimited character count, although brief biographies seem to be more common). The rest of a user’s profile consists of an “Ask me a question” box—which allows other users to ask the user a question, either anonymously or not – and a list of the user’s previously answered questions (see Figure 1). A user’s answers display chronologically, with the most recently answered question displayed first. Users may receive questions from named or anonymous users, or they may choose to answer the site-generated “Question of the Day,” added automatically to the user’s question queue each day (these are typically general questions that any user could answer, e.g., “What is your usual breakfast?”). Finally, an important social norm on Ask.fm is that users do not always submit “questions,” instead sometimes using the question-asking feature to send declarative statements that are not in the form of a query.

OPEN IN VIEWER

Figure 1.

User profile from time of data collection, showing the option to ask a question either anonymously or not.

The most notable feature of Ask.fm—and what makes it appropriate for this study—is that question-askers can choose to either associate their username with questions they ask of other users or do so anonymously. Question-receivers can select whether or not to allow anonymous questions, though the site’s default upon registration is to allow them. Another distinguishing characteristic is that Ask.fm can be deeply embedded in offline social structures: many site users (including our participants) interact primarily with individuals in their existing offline social networks. As a result, questions on Ask.fm are often more social than informational in nature. Though a full name is required at registration, users can elect to remain pseudonymous by choosing a non-identifiable username and inputting false information in lieu of a “full name.” Pseudonyms may then be shared with trusted peers. Even if a user’s specific identity is not known, aspects of her offline identity (e.g., what school he or she attends) may be discerned by signals such as whom he or she interacts with on the site.

Data Collection and Analysis

Interviews were conducted in April and May, 2015. The first author (N.B.E.) and a research assistant conducted the interviews, which took place via a university-provided secure videoconferencing system; most participants chose to connect using voice only. We received parental consent via email, as well as participant assent before the interview, and confirmed that participants had a clear understanding of the research process and how their data would be handled. We first asked participants about their general social media use and then their use of Ask.fm (e.g., what they liked and disliked about the site; their positive and negative site experiences). The protocol also included items about asking questions of other Ask.fm users as well as participants’ experiences answering questions, with an emphasis on anonymous interactions. This study was approved by the university’s Institutional Review Board (IRB).

The interviewers kept notes during each interview to identify emergent themes. Interviews were transcribed using Scribie (https://scribie.com/), checked for accuracy, and then imported into Atlas.TI for analysis. Coding occurred over several stages: first, the research team met to discuss broad themes that emerged from the data and developed a preliminary set of codes, which was used to code a small subset of interviews. Through discussions based on this initial coded sample, the list of codes was revised, such that similar codes were collapsed and new codes introduced to capture important phenomena. A research assistant used the finalized codebook to code the entire dataset; the research team met frequently during the coding and writing process to discuss interpretation of the data. Quotes have been lightly edited for clarity and readability.

Perceived Authenticity

Echoing the research literature on honesty of self-disclosure within non-identified contexts (e.g., Rubin, 1975), anonymous comments were perceived to be more authentic than those from identified users by our participants, who described anonymous content as a more honest reflection of the poster’s opinions or feelings (compared to those offered in identified contexts, where a speaker might incur social benefits from the exchange—for example, the social capital earned when paying a compliment). The absence of social outcomes from anonymous interactions meant that often, these interactions were perceived as more honest and more salient than comments from identified peers.

This increased salience of anonymous content was especially true in the case of anonymous compliments, which were perceived as being motivated by something other than normative expectations around friendship maintenance activities. P9 (F, 18) explained,

If your friends compliment you, it’s one thing, ‘cause it’s like, “Okay, well, they are my friends, and they are going to try to be nice to me, obviously,” but if it’s people who you don’t know, or don’t know you that well, it just—it’s something a little bit more special sometimes.

P13 (F, 18) echoed this sentiment, noting that “sometimes friends tell you things to give you something that you want to hear,” as did P22 (F, 16), who said “it’s nice that someone would just compliment me anonymously [because] they’re not getting a reward from doing it.” Anonymous compliments that could not produce social capital benefits had an altruistic dimension: as a result, those who received anonymous compliments saw them as more authentic. While the self-esteem benefits of receiving anonymous compliments were noted by almost everyone who received these kinds of messages, at least one of our participants felt otherwise, explaining that he was “not one to feel so uplifted and fantastic because of a comment online” (P1, M, 16).

For all anonymous content, the unknown nature of the contributor ignited curiosity as to the poster’s true identity, often enhancing the impact of this content. Speaking about anonymous compliments, P18 (M, 18) told us, “it’s more powerful if it was anonymous because it could be anyone. It can be someone you think is cute . . . [It] could be anyone, and so it definitely makes it more intense.” P12 (F, 18) similarly described the feeling of having an unknown person send compliments, adding,

It really makes your day when someone sends you something, telling you—complimenting you on one of your features anonymously. ‘Cause you’re kind of wondering, who could it be, who is this? And you never really know. But it’s great.

Of course, not knowing the poster’s identity amplified the salience of uncivil comments or insults shared on the platform as well. Not knowing who posted an unkind question was disconcerting to many, especially given the distressing possibility that the hurtful comments might have originated from trusted friends. P4 (F, 15) explained,

It can actually be anyone. They can be your close friends . . . And then they can have so much hate towards you, but they don’t want to tell you, they don’t want to hurt you. But they hurt you in [Ask.fm], they tell you these hurtful questions . . . because they don’t have the guts to tell you in person how they feel.

This notion that anonymity enabled the true expression of feelings otherwise inappropriate for face-to-face exchange was a common theme across participants.

Circumventing Social Expectations

For some of our participants, posting anonymously facilitated circumvention of restrictive social norms associated with offline settings (such as school), enabling information-seeking about taboo topics and interactions with weaker ties that might not otherwise materialize. For instance, many of our participants described feeling constrained in their communication with peers at school, both in topic and choice of communication partner, due to highly structured social environments and rigid codes of expected social behavior. Ask.fm provided a channel through which they could safely circumvent these perceived restrictions by asking questions anonymously and interacting with individuals with whom they would not otherwise engage.

Participants often described anonymous interactions as disembodied in that they were not associated with an individual, thus shielding their senders from the negative repercussions of sharing their opinions or questions. This kind of protection from consequences was described in both positive and negative terms. Because identity information was not associated with content when asking a question anonymously, participants’ self-presentational concerns were lowered, and they felt more comfortable sharing content which might reveal qualities about the self they would rather keep private (such as questions about sex).

For instance, P1 (M, 16) described how the anonymity of Ask.fm allowed users, especially males, to make information requests that would be awkward to articulate in offline settings. P1 (M, 16) said certain questions were harder to ask offline, and that teenage male users “want to know things, but they don’t want people to know that [they] desire to know.” P1 (M, 16) went on to say

It’s usually very weird, sexual things with girls . . . I think over 70% of the questions that are asked on there because they’re guys and they’re just trying to ask questions about [sex], and they obviously wouldn’t do it in real life, because that may be considered really weird.

P1 (M, 16) described the way in which a question or comment on Ask.fm could exist outside of its association with the asker—thus, protecting the asker from reputation threats or attacks by peers:

The distinction between Ask and real life is that usually in a conversation when you ask a question . . . it has certain implications for how people view you, and the response also matters for you, because you’re tied to that conversation then. But, on a website like this, you ask a question and then you’re no longer a part of it, because it will never say who you are . . . The question exists, but you don’t exist with it . . . It just changes the dynamic of questions, because they’re no longer tied to any real thing. They’re just free-floating questions, which just changes how they function, because they’re no longer held back by people’s—I don’t know whether or not it’s their actual moral character or just how they want their moral character to be perceived or [aspects of] their own personality that they don’t want to let show in certain public spaces.

In another example, participants described using the site to confirm or disconfirm social information circulating in offline settings like schools. As one participant noted, “whenever there’s a rumor, there’s usually multiple people that ask the same person about it [on Ask.fm]” (P9, F, 18). Within a school or social group, users often knew one another’s usernames, so platform practices could be embedded in offline social networks and interactions, enabling the site to be used for social information-seeking purposes. Participants described several rumor-related Ask.fm practices, such as following people to be “in the know” (P15, F, 17) and asking questions to acquire new information and gossip, as when one participant told us, “I don’t think people really address rumors face-to-face anymore, and it’s more over social media, because you don’t want to feel stupid asking the question face-to-face if it wasn’t real” (P21, F, 18). In another example of using the platform to circumvent social expectations about appropriate behavior, P12 (F, 18) explained that she sometimes would ask herself anonymous questions in order to directly defuse rumors: “I’d ask myself something and I would explain, ‘Oh, this is what happened’, in order to show, ‘Hey, I’m innocent in this. Not my fault’.”

This disassociation between users and their content meant that a wider range of topics could be addressed than in face-to-face settings, encompassing subjects that participants felt would be awkward, embarrassing, or inappropriate to express through other channels. For instance, in talking about why she gave compliments anonymously through the platform, P12 (F, 18) explained,

I would get probably a different reaction if I said this face-to-face to them. I’d probably get an “Oh, thank you, I guess” . . . instead of on Ask.fm, it’s more, “Oh, thank you so much, this is so nice of you!” I guess people would probably just be a little bit more wary if I said that face-to-face to them, or confused.

In this description, both the anonymous and mediated nature of Ask.fm interactions was salient, in that a face-to-face interaction would be awkward (e.g., “thank you, I guess”) and would mean the asker was identified.

Our participants capitalized on the disassociation between user and content on Ask.fm to ask for advice on sensitive matters or from unfamiliar acquaintances. For example, P17 (F, 18) describes the anonymity as a “security blanket”:

If they said they had problems with their boyfriend, I don’t think they would want to reveal who they are because they wouldn’t want people to know that they have that problem . . . If their Mom was sick, if something bad had happened to them and they really don’t want to reveal who they are, they don’t need to . . . Like a security blanket, almost.

Participants also reported feeling more comfortable initiating interactions anonymously with people they didn’t know offline, as explained by P6 (M, 16): “maybe it was somebody that I have never talked to before and they didn’t want to seem weird if they just came out to me in person. So they just felt more comfortable with this online, doing it.” Similarly, P19 (F, 18) observed questions about college admission directed at high school seniors, presumably weak ties whom the asker would be hesitant to approach face-to-face.

This perceived freedom from social norms also contributed to negative social dynamics on the platform, as when participants received socially inappropriate or hurtful questions and statements. The ability to use anonymity to circumvent social consequences manifested in what some participants described as a lack of accountability, contributing to what they described as “mean” comments. As P1 (M, 16) explained,

The other aspect of it would be trying to make people feel uncomfortable or insulting them, and I guess maybe people don’t do that in public, just because it’s more work to have to actually deal with it. Because the benefit of being anonymous on [Ask.fm] is you can back out whenever you desire, and if you are shown to be the person that’s under fault for some reason, you can easily back out from it. There’s no tying you to what happened.

Similarly, P8 (M, 15) noted,

[On Ask.fm] you could ask anyone any question without any kind of worry of, “Am I going to be judged for asking this or suggesting this?” A lot of times, people would suggest really negative things like—or even straight out say, “You’re stupid,” or, “I think you’re ugly.” . . . And I think that was probably the main problem with the app, was people being able to do that and not have to face any kind of repercussions for it.

Anonymous communication was described by some of our participants as “hiding behind a screen,” as illustrated by this quote from P18 (M, 18):

Yeah, it’s one of those, “Oh, I can say anything I want because I’m behind a screen.” . . . [It] gave people the power to say anything they wanted or any judgment or any idea . . . they had about someone to their face except through a screen.

Managing Identity and Self-Presentation

A third theme emerging from our data involved issues of identity management and self-presentation. Participants strategically used anonymous interaction to achieve social goals while protecting their reputation. For instance, P8 (M, 15; “Sam”) gave a hypothetical example of how he could use anonymity to chastise his friends about having forgotten his birthday by pretending to be a third party:

I would subtly suggest, “You forgot about Sam’s birthday,” . . . to guilt trip them, because I would want that attention, like them to wish me happy birthday . . . That was almost my subtle way of lashing out at them, because I wanted them to do something that they didn’t do.

By asking this question anonymously, the user was able to remind his friends to acknowledge his birthday without experiencing the social cost of doing this directly.

In another example of impression management, users strategically controlled whether certain questions showed up on their page or not by either answering or deleting them. In dealing with uncivil comments, participants appreciated the ability to control the visibility of posts on their page and described two primary strategies for dealing with them. In some cases, participants simply ignored questions: P11 (F, 14) said, “I just deleted those questions . . . I just didn’t want to make a bad impression on social media when nobody knows who’s asking the question.” In other instances, participants would answer the question—which meant it was displayed on their page—but with a nonchalant comment that conveyed a lack of engagement: P14 (M, 18) would reply “Ha ha, whatever”; P18 (M, 18) tried to create funny responses that would result in “likes” from observers; and P9 (F, 18) used “OK” to convey her indifference to mean or negative comments. She elaborated on her strategies for dealing with mean comments:

I would answer them unless they were just really vulgar and I didn’t want people reading that because it’s just gross, but I would answer most of them, and I would just answer them with a sarcastic comment, or just, “Okay,” or something that would probably make the person who asked it angry that it didn’t affect me.

In another case, P12 (F, 18) received compliments about her soccer-playing ability that she found supportive. She purposefully refrained from answering them because she did not want them displayed to her followers, which might give others the impression she was being boastful:

Like, “Hey look! I’ve gotten this saying that I’m good at this and I want to show you, everyone, that people think that I’m good at this!” I didn’t want to brag about it, so I didn’t answer it. I just kept it.

By not answering the submission, she retained the ability to revisit the compliment herself, while avoiding the perception that she wanted to broadcast it.

Initiating and Developing Relationships

Fourth, our data suggest that participants employed anonymity when they wanted to test relational compatibility and connect with individuals outside their usual social circle. In some cases, anonymous interactions on the site developed into closer relationships or led to face-to-face interactions, and many participants described using the site to initiate or develop friendships or romantic relationships. For some, this was a key motivation for using the site. When asked why he used Ask.fm, P7 (M, 18) reported, “To ask questions, to find things out about people that I don’t usually talk to, to make friendships.” The site was also used to validate existing social relationships. For instance, users would be asked anonymously for a list of their close friends or whom they found attractive, their answers serving as a form of social currency and confirmation.

Participants described conducting compatibility tests via anonymous interaction on the site, through questions such as “‘What do you do in your free time? What kind of music do you like?’ Kind of compatibility questions, feelers to see it they’re actually a cool person or not” (P3, F, 15). Similarly, P7 (M, 18) said sharing information on the site made the future interactions easier:

Basically, you’d use Ask or some other social media site to start the conversation, and then, after a few questions, or after you know that they’re interested in talking to you, or that you’re interested in talking to them, you’d ask for a phone number, and then you’d just transfer it to your phone . . . You get the weird questions out of the way . . . so you can have real, true conversations with people on your phone. [It] weeds out the people that you wouldn’t really be interested in talking to.

He went on to describe how relationships developed on the site, moving from anonymous compatibility testing to eventually “hanging out” in person:

Use it to make people see you as interesting or see you as cool, and get their phone number, and talk to them, keep talking to them and then they hang out. And you could ask a girl a question, without them judging you. You could ask them anything you wanted to, without them [thinking], “Oh, this kid’s weird for asking that question.”

Engaging in anonymous uncertainty reduction activities enabled users to avoid uncomfortable feelings if they discovered they were not compatible with someone with whom they were interacting. P13 (F, 18) said she would “maybe first ask them anonymous questions, so they won’t know who you are in case they react negatively to your question. You’d rather just not have that awkward tension between y’all afterwards.”

Relationship development on Ask.fm, where relationships often started online and moved offline, was one domain in which we saw clear evidence of how site use was shaped by and shaped offline social structures. The socially embedded nature of Ask.fm interactions was also evident when participants admitted to trying to guess the identity of their anonymous commenters. Some of our participants mentioned attempt to discern others’ identities, either through direct questioning or by assessing the sparse cues that did exist, such as linguistic idiosyncrasies. For instance, one participant knew who sent her a compliment because it included a specific word (“woo”) used by one of her acquaintances; another thought she could identify someone that sent her a rude comment because the commenter used “4” instead of “for.”

Learning about the Self

Finally, other affordances of the platform enabled some participants to learn about themselves through the process of answering questions in textual form: the anonymous context shaped the answers participants provided and the kinds of questions they received. Our data suggest that the freedom to engage in different kinds of expression supported identity exploration and articulation processes. For instance, P4 (F, 15) enjoyed answering the “question of the day”—a question posed to each user by the site itself—because this gave her an opportunity to write about her feelings and opinions in ways not possible in other environments. She tried to write in detail on Ask.fm because it was “the only time that I can say my opinion, and I can make it long.” She explained that writing on Ask.fm was different than offline conversations:

If a person asks you, you just have to make it short because they wouldn’t want to listen to your whole story of how your day was. And Facebook, no one would want to ask me . . . On Ask, I can actually say so much of how my day was. I can actually say it in detail, without people saying, “You’re annoying me, just stop saying so much.” . . . At school they ask you for your opinion. But it’s the opinion that the teacher asks you to take . . . I love writing so much, and so [on Ask.fm], I find the questions that are really interesting and that I know that I can write a lot about. (P4, F, 15)

Participants described learning about aspects of the self through interactions on the site, whether through comments made by other people, their own reactions to questions they received, or retrospective engagement with their own content. These instances of self-knowledge ranged from mundane information about the self (e.g., P9 [F, 18] was asked why she did not blink when she listened to people, which she then confirmed through observation) to more meaningful insights (e.g., P9 [F, 18] also learned she could be very defensive; P6 [M, 16] was told he was good at giving advice). P11 (F, 14) described the way in which the questions on the site prompted her to think about her relationship with her sister and to try to improve it:

Some of the questions, they actually make me think about things that I did or things that I have done in the past . . . I have a twin sister, and oftentimes people just ask me about that . . . So oftentimes, I have to think about what I do with my sister . . . That actually makes me learn a lot. Like, how can I improve myself, and how can I improve my interactions with my sister, because we fight a lot. (P11, F, 14)

In response to a question about what she learned from questions about her sister, she explained,

[Before] I answer a question, I think back about what my sister and I have done and what we do . . . and I think, “Oh, how have I treated her?” or “What have I done bad to her?” “What have I done good to her?” And I kind of learn from myself, because without that question, I just probably wouldn’t have thought of “What have I done like in the past and stuff?” And from that question . . . I learned how to treat her better and tolerate her. (P11, F, 14)

Others described perusing old Ask.fm content to look for changes in how they answered questions over time. P17 (F, 18) said she “learned a lot about the person I used to be . . . going back and looking at how I answered things.” This act of viewing their historical written traces allowed users to reflect on how their opinions, writing style, and other qualities evolved over time.

Circumventing Social Expectations and the Development of Social Relationships

This work suggests that anonymity is an important mechanism for information gathering about new or little-known acquaintances in known networks, enabling adolescents to reduce uncertainty about their peers. Social relationships develop over time and through a process of uncertainty reduction. Uncertainty Reduction Theory (URT), first introduced by Berger and Calabrese (1975), argues that social actors are motivated to reduce uncertainty about strangers or acquaintances through interpersonal communication processes—for example, the initial sharing of simple biographical or demographic information between strangers. Over time, this allows interlocutors to reduce their uncertainty and better predict others’ beliefs or behaviors, increasing the likelihood that a relationship will form (Berger & Calabrese, 1975).

Computer-mediated communication affects these information acquisition processes in novel ways: most notably, individuals can engage in a wider range of passive information-seeking strategies—such as visiting the Facebook profile of a new acquaintance—rather than relying on interactive strategies such as asking directly or questioning mutual friends (Tong, 2013). As Tong (2013) argues, SNSs “obviate the need for direct, interactive strategies by providing inconspicuous routes for information gathering” (p. 789). Ramirez, Walther, Burgoon, and Sunnafrank (2002) introduce a four-part typology of information-seeking strategies, which vary in directness: (1) interactive (involving direct interaction with the target), (2) active (acquiring information from other individuals without interacting directly with the target), (3) extractive (non-interactive strategies such as using a search engine), and (4) passive (acquiring information through observations, such as viewing a user profile or reading previously posted messages). Our findings highlight the ways in which social media platforms that enable multiple forms of identity expression complicate and destabilize these traditional information gathering categories: Our participants blended several strategies at once, in order to reduce uncertainty about their peers, while protecting their own social position (e.g., by interacting directly—but anonymously—with a target). Because the technical affordances of Ask.fm allow users to remain anonymous during the early stages of the information gathering process, adolescents may covertly and safely engage in uncertainty reduction strategies that may then increase the likelihood that a future relationship will form (Berger & Calabrese, 1975).

Our findings suggest that the ways in which adolescents engage in selectively anonymous interaction with known peers constitutes a shift from themes seen in previous work on adolescence and anonymity online, which has focused on interactions with strangers (Valkenburg & Peter, 2008). Adolescents, however, are particularly sensitive to the importance of offline social structures and thus may strategically employ anonymous interaction in different ways than adults. This work suggests some useful extensions of Social Information Processing (SIP) theory (Walther, 1996), in that users attempted to identify anonymous askers based on sparse linguistic cues or other identity detritus. Future work could explore how SIP might apply to situations in which impression formation processes unfold in an environment that is nominally anonymous but which, in practice, often supports interactions between individuals who know something about one another or may know one another offline. Similarly, platforms that enable selective anonymity offer scholars a new context for exploring how users express identity in these spaces—and whether relationships that begin in an anonymous context are bound by the same interpersonal processes identified in other work on modality switching (Ramirez, Sumner, Fleuriet & Cole, 2015).

Ask.fm enabled our participants to circumvent social expectations in other ways as well. Existing literature on the benefits of anonymous interaction focuses primarily on its implications for more honest self-disclosure. Our data echo these themes: participants described anonymous content as performing a social task, such as information confirmation, without being linked to an identified user—or in the words of one participant, “The question exists, but you don’t exist with it.” However, our findings extend this work by surfacing other benefits specific to selective anonymous interaction among a network of known peers. Participants reported using anonymous questions to diffuse rumors, for instance, or to ask questions that would reinforce friendship bonds or possibly initiate a dating relationship (e.g., “Who do you think is cute?”). This ability to circumvent expectations for normative social behavior allowed adolescents to execute necessary social functions without risking impression management concerns or social norm violations.

Textual, Anonymous Interaction and Learning about the Self

Ask.fm has several affordances that allowed adolescent users to learn about themselves. The focus on writing itself—especially longer narratives that would be less normative on sites such as Twitter or Facebook—permitted some participants to engage in writing practices that they would not otherwise, or to compare their responses over time in order to chart the evolution of their ideas and emotions. Two participants described the Q&A format on Ask.fm as more conducive to longer and more well-thought-out responses than Facebook or casual face-to-face conversations. Questions from peers may also contribute to identity-forming processes: for instance, questions from peers asking for advice about sports may reinforce an adolescent’s identity as an athlete. Importantly, the platform gave users control over which questions (and therefore answers) would be associated with their pseudonymous profile, thus providing another channel for identity expression as well as a way to deflect “mean” comments. These benefits to identity expression are especially pertinent to this age group, given the developmental advantages of successful identity formation during adolescence (Collins & Steinberg, 2006).

Identity work was also facilitated by the salience of anonymous content (such as insults or compliments), which were perceived to be more authentic than comments from identified peers. Adolescents are highly sensitized to social norms around appropriateness, as indicated by their assertion that friends would feel obligated to say kind things in an identified context. These anonymous comments, however, were shared without expectation of reciprocity or acknowledgement, and thus bolstered self-esteem in ways that compliments from identified peers did not.