Managing Impressions Online: Self-Presentation Processes in the Online Dating Environment


  • Nicole Ellison,

    1. Department of Telecommunication, Information Studies, and Media
      Michigan State University
    Search for more papers by this author
    • Nicole Ellison is an Assistant Professor in the Department of Telecommunications, Information Studies, and Media at Michigan State University. Her research explores issues of self-presentation, relationship development, and identity in online environments such as weblogs, online dating sites, and social networking services.

      Address: Dept. of Telecommunications, Information Studies, and Media, Michigan State University, 403 Communication Arts and Sciences, East Lansing, MI 48824 USA

  • Rebecca Heino,

    1. McDonough School of Business
      Georgetown University
    Search for more papers by this author
    • Rebecca Heino is an Assistant Professor in the McDonough School of Business at Georgetown University. She has centered her research on the use of communication technologies—such as intranets and email—in organizations, specifically focusing on organizational adoption, implementation, and privacy.

      Address: Georgetown University, McDonough School of Business, 37th & O St. NW, Old North G-04, Washington, DC 20057 USA

  • Jennifer Gibbs

    1. Communication Department
      Rutgers University
    Search for more papers by this author
    • Jennifer Gibbs is an Assistant Professor of Communication at Rutgers University. Her research interests center around how individuals connect, collaborate, and negotiate identity and differences in global, multicultural, and mediated contexts.

      Address: Department of Communication, SCILS, 4 Huntington Street, Rutgers University, New Brunswick, NJ 08901-1071 USA


This study investigates self-presentation strategies among online dating participants, exploring how participants manage their online presentation of self in order to accomplish the goal of finding a romantic partner. Thirty-four individuals active on a large online dating site participated in telephone interviews about their online dating experiences and perceptions. Qualitative data analysis suggests that participants attended to small cues online, mediated the tension between impression management pressures and the desire to present an authentic sense of self through tactics such as creating a profile that reflected their “ideal self,” and attempted to establish the veracity of their identity claims. This study provides empirical support for Social Information Processing theory in a naturalistic context while offering insight into the complicated way in which “honesty” is enacted online.


The online dating arena represents an opportunity to document changing cultural norms surrounding technology-mediated relationship formation and to gain insight into important aspects of online behavior, such as impression formation and self-presentation strategies. Mixed-mode relationships, wherein people first meet online and then move offline, challenge established theories that focus on exclusively online relationships and provide opportunities for new theory development (Walther & Parks, 2002). Although previous research has explored relationship development and self-presentation online (Bargh, McKenna, & Fitzsimons, 2002; McLaughlin, Osbourne, & Ellison, 1997; Parks & Floyd, 1996; Roberts & Parks, 1999; Utz, 2000), the online dating forum is qualitatively different from many other online settings due to the anticipation of face-to-face interaction inherent in this context (Gibbs, Ellison, & Heino, 2006) and the fact that social practices are still nascent.

In recent years, the use of online dating or online personals services has evolved from a marginal to a mainstream social practice. In 2003, at least 29 million Americans (two out of five singles) used an online dating service (Gershberg, 2004); in 2004, on average, there were 40 million unique visitors to online dating sites each month in the U.S. (CBC News, 2004). In fact, the online personals category is one of the most lucrative forms of paid content on the web in the United States (Egan, 2003) and the online dating market is expected to reach $642 million in 2008 (Greenspan, 2003). Ubiquitous access to the Internet, the diminished social stigma associated with online dating, and the affordable cost of Internet matchmaking services contribute to the increasingly common perception that online dating is a viable, efficient way to meet dating or long-term relationship partners (St. John, 2002). Mediated matchmaking is certainly not a new phenomenon: Newspaper personal advertisements have existed since the mid-19th century (Schaefer, 2003) and video dating was popular in the 1980s (Woll & Cosby, 1987; Woll & Young, 1989). Although scholars working in a variety of academic disciplines have studied these earlier forms of mediated matchmaking (e.g., Ahuvia & Adelman, 1992; Lynn & Bolig, 1985; Woll, 1986; Woll & Cosby, 1987), current Internet dating services are substantively different from these incarnations due to their larger user base and more sophisticated self-presentation options.

Contemporary theoretical perspectives allow us to advance our understanding of how the age-old process of mate-finding is transformed through online strategies and behaviors. For instance, Social Information Processing (SIP) theory and other frameworks help illuminate computer-mediated communication (CMC), interpersonal communication, and impression management processes. This article focuses on the ways in which CMC interactants manage their online self-presentation and contributes to our knowledge of these processes by examining these issues in the naturalistic context of online dating, using qualitative data gathered from in-depth interviews with online dating participants.

Literature Review

In contrast to a technologically deterministic perspective that focuses on the characteristics of the technologies themselves, or a socially deterministic approach that privileges user behavior, this article reflects a social shaping perspective. Social shaping of technology approaches (Dutton, 1996; MacKenzie & Wajcman, 1985; Woolgar, 1996) acknowledge the ways in which information and communication technologies (ICTs) both shape and are shaped by social practices. As Dutton points out, “technologies can open, close, and otherwise shape social choices, although not always in the ways expected on the basis of rationally extrapolating from the perceived properties of technology” (1996, p. 9). One specific framework that reflects this approach is Howard’s (2004) embedded media perspective, which acknowledges both the capacities and the constraints of ICTs. Capacities are those aspects of technology that enhance our ability to connect with one another, enact change, and so forth; constraints are those aspects of technology that hinder our ability to achieve these goals. An important aspect of technology use, which is mentioned but not explicitly highlighted in Howard’s framework, is the notion of circumvention, which describes the specific strategies employed by individuals to exploit the capacities and minimize the constraints associated with their use of ICTs. Although the notion of circumvention is certainly not new to CMC researchers, this article seeks to highlight the importance of circumvention practices when studying the social aspects of technology use.1

Self-Presentation and Self-Disclosure in Online and Offline Contexts

Self-presentation and self-disclosure processes are important aspects of relational development in offline settings (Taylor & Altman, 1987), especially in early stages. Goffman’s work on self-presentation explicates the ways in which an individual may engage in strategic activities “to convey an impression to others which it is in his interests to convey” (1959, p. 4). These impression-management behaviors consist of expressions given (communication in the traditional sense, e.g., spoken communication) and expressions given off (presumably unintentional communication, such as nonverbal communication cues). Self-presentation strategies are especially important during relationship initiation, as others will use this information to decide whether to pursue a relationship (Derlega, Winstead, Wong, & Greenspan, 1987). Research suggests that when individuals expect to meet a potential dating partner for the first time, they will alter their self-presentational behavior in accordance with the values desired by the prospective date (Rowatt, Cunningham, & Druen, 1998). Even when interacting with strangers, individuals tend to engage in self-enhancement (Schlenker & Pontari, 2000).

However, research suggests that pressures to highlight one’s positive attributes are experienced in tandem with the need to present one’s true (or authentic) self to others, especially in significant relationships. Intimacy in relationships is linked to feeling understood by one’s partner (Reis & Shaver, 1988) and develops “through a dynamic process whereby an individual discloses personal information, thoughts, and feelings to a partner; receives a response from the partner; and interprets that response as understanding, validating, and caring” (Laurenceau, Barrett, & Pietromonaco, 1998, p. 1238). Therefore, if participants aspire to an intimate relationship, their desire to feel understood by their interaction partners will motivate self-disclosures that are open and honest as opposed to deceptive. This tension between authenticity and impression management is inherent in many aspects of self-disclosure. In making decisions about what and when to self-disclose, individuals often struggle to reconcile opposing needs such as openness and autonomy (Greene, Derlega, & Mathews, 2006).

Interactants in online environments experience these same pressures and desires, but the greater control over self-presentational behavior in CMC allows individuals to manage their online interactions more strategically. Due to the asynchronous nature of CMC, and the fact that CMC emphasizes verbal and linguistic cues over less controllable nonverbal communication cues, online self-presentation is more malleable and subject to self-censorship than face-to-face self-presentation (Walther, 1996). In Goffman’s (1959) terms, more expressions of self are “given” rather than “given off.” This greater control over self-presentation does not necessarily lead to misrepresentation online. Due to the “passing stranger” effect (Rubin, 1975) and the visual anonymity present in CMC (Joinson, 2001), under certain conditions the online medium may enable participants to express themselves more openly and honestly than in face-to-face contexts.

A commonly accepted understanding of identity presumes that there are multiple aspects of the self which are expressed or made salient in different contexts. Higgins (1987) argues there are three domains of the self: the actual self (attributes an individual possesses), the ideal self (attributes an individual would ideally possess), and the ought self (attributes an individual ought to possess); discrepancies between one’s actual and ideal self are linked to feelings of dejection. Klohnen and Mendelsohn (1998) determined that individuals’ descriptions of their “ideal self” influenced perceptions of their romantic partners in the direction of their ideal self-conceptions. Bargh et al. (2002) found that in comparison to face-to-face interactions, Internet interactions allowed individuals to better express aspects of their true selves—aspects of themselves that they wanted to express but felt unable to. The relative anonymity of online interactions and the lack of a shared social network online may allow individuals to reveal potentially negative aspects of the self online (Bargh et al., 2002).

Although self-presentation in personal web sites has been examined (Dominick, 1999; Schau & Gilly, 2003), the realm of online dating has not been studied as extensively (for exceptions, see Baker, 2002; Fiore & Donath, 2004), and this constitutes a gap in the current research on online self-presentation and disclosure. The online dating realm differs from other CMC environments in crucial ways that may affect self-presentational strategies. For instance, the anticipated future face-to-face interaction inherent in most online dating interactions may diminish participants’ sense of visual anonymity, an important variable in many online self-disclosure studies. An empirical study of online dating participants found that those who anticipated greater face-to-face interaction did feel that they were more open in their disclosures, and did not suppress negative aspects of the self (Gibbs et al., 2006). In addition, because the goal of many online dating participants is an intimate relationship, these individuals may be more motivated to engage in authentic self-disclosures.

Credibility Assessment and Demonstration in Online Self-Presentation

Misrepresentation in Online Environments

As discussed, online environments offer individuals an increased ability to control their self-presentation, and therefore greater opportunities to engage in misrepresentation (Cornwell & Lundgren, 2001). Concerns about the prospect of online deception are common (Bowker & Tuffin, 2003; Donath, 1999; Donn & Sherman, 2002), and narratives about identity deception have been reproduced in both academic and popular outlets (Joinson & Dietz-Uhler, 2002; Stone, 1996; Van Gelder, 1996). Some theorists argue that CMC gives participants more freedom to explore playful, fantastical online personae that differ from their “real life” identities (Stone, 1996; Turkle, 1995). In certain online settings, such as online role-playing games, a schism between one’s online representation and one’s offline identity are inconsequential, even expected. For instance, MacKinnon (1995) notes that among Usenet participants it is common practice to “forget” about the relationship between actual identities and online personae.

The online dating environment is different, however, because participants are typically seeking an intimate relationship and therefore desire agreement between others’ online identity claims and offline identities. Online dating participants report that deception is the “main perceived disadvantage of online dating” (Brym & Lenton, 2001, p. 3) and see it as commonplace: A survey of one online dating site’s participants found that 86% felt others misrepresented their physical appearance (Gibbs et al., 2006). A 2001 research study found that over a quarter of online dating participants reported misrepresenting some aspect of their identity, most commonly age (14%), marital status (10%), and appearance (10%) (Brym & Lenton, 2001). Perceptions that others are lying may encourage reciprocal deception, because users will exaggerate to the extent that they feel others are exaggerating or deceiving (Fiore & Donath, 2004). Concerns about deception in this setting have spawned related services that help online daters uncover inaccuracies in others’ representations and run background checks on would-be suitors (Baertlein, 2004; Fernandez, 2005). One site,, conducts background checks on their users and has worked to introduce legislation that would force other online dating sites to either conduct background checks on their users or display a disclaimer (Lee, 2004).

The majority of online dating participants claim they are truthful (Gibbs et al., 2006; Brym & Lenton, 2001), and research suggests that some of the technical and social aspects of online dating may discourage deceptive communication. For instance, anticipation of face-to-face communication influences self-representation choices (Walther, 1994) and self-disclosures because individuals will more closely monitor their disclosures as the perceived probability of future face-to-face interaction increases (Berger, 1979) and will engage in more intentional or deliberate self-disclosure (Gibbs et al., 2006). Additionally, Hancock, Thom-Santelli, and Ritchie (2004) note that the design features of a medium may affect lying behaviors, and that the use of recorded media (in which messages are archived in some fashion, such as an online dating profile) will discourage lying. Also, online dating participants are typically seeking a romantic partner, which may lower their motivation for misrepresentation compared to other online relationships. Further, Cornwell and Lundgren (2001) found that individuals involved in online romantic relationships were more likely to engage in misrepresentation than those involved in face-to-face romantic relationships, but that this was directly related to the level of involvement. That is, respondents were less involved in their cyberspace relationships and therefore more likely to engage in misrepresentation. This lack of involvement is less likely in relationships started in an online dating forum, especially sites that promote marriage as a goal.

Public perceptions about the higher incidence of deception online are also contradicted by research that suggests that lying is a typical occurrence in everyday offline life (DePaulo, Kashy, Kirkendol, Wyer, & Epstein, 1996), including situations in which people are trying to impress prospective dates (Rowatt et al., 1998). Additionally, empirical data about the true extent of misrepresentation in this context is lacking. The current literature relies on self-reported data, and therefore offers only limited insight into the extent to which misrepresentation may be occurring. Hitsch, Hortacsu, and Ariely (2004) use creative techniques to address this issue, such as comparing participants’ self-reported characteristics to patterns found in national survey data, but no research to date has attempted to validate participants’ self-reported assessments of the honesty of their self-descriptions.

Assessing and Demonstrating Credibility in CMC

The potential for misrepresentation online, combined with the time and effort invested in face-to-face dates, make assessment strategies critical for online daters. These assessment strategies may then influence participants’ self-presentational strategies as they seek to prove their trustworthiness while simultaneously assessing the credibility of others.

Online dating participants operate in an environment in which assessing the identity of others is a complex and evolving process of reading signals and deconstructing cues, using both active and passive strategies (Berger, 1979; Ramirez, Walther, Burgoon, & Sunnafrank, 2002; Tidwell & Walther, 2002). SIP considers how Internet users develop impressions of others, even with the limited cues available online, and suggests that interactants will adapt to the remaining cues in order to make decisions about others (Walther, 1992; Walther, Anderson, & Park, 1994). Online users look to small cues in order to develop impressions of others, such as a poster’s email address (Donath, 1999), the links on a person’s homepage (Kibby, 1997), even the timing of email messages (Walther & Tidwell, 1995). In expressing affinity, CMC users are adept at using language (Walther, Loh, & Granka, 2005) and CMC-specific conventions, especially as they become more experienced online (Utz, 2000). In short, online users become cognitive misers, forming impressions of others while conserving mental energy (Wallace, 1999).

Walther and Parks (2002) propose the concept of “warranting” as a useful conceptual tool for understanding how users validate others’ online identity cues (see also Stone, 1996). The connection, or warrant, between one’s self-reported online persona and one’s offline aspects of self is less certain and more mutable than in face-to-face settings (Walther & Parks, 2002). In online settings, users will look for signals that are difficult to mimic or govern in order to assess others’ identity claims (Donath, 1999). For instance, individuals might use search engines to locate newsgroup postings by the person under scrutiny, knowing that this searching is covert and that the newsgroup postings most likely were authored without the realization that they would be archived (Ramirez et al., 2002). In the context of online dating, because of the perceptions of deception that characterize this sphere and the self-reported nature of individuals’ profiles, participants may adopt specific presentation strategies geared towards providing warrants for their identity claims.

In light of the above, our research question is thus:

RQ: How do online dating participants manage their online presentation of self in order to accomplish the goal of finding a romantic partner?


In order to gain insight into this question, we interviewed online dating participants about their experiences, thoughts, and behaviors. The qualitative data reported in this article were collected as part of a larger research project which surveyed a national random sample of users of a large online dating site (N = 349) about relational goals, honesty and self-disclosure, and perceived success in online dating. The survey findings are reported in Gibbs et al. (2006).

Research Site

Our study addresses contemporary CMC theory using naturalistic observations. Participants were members of a large online dating service, “” (a pseudonym). currently has 15 million active members in more than 200 countries around the world and shares structural characteristics with many other online dating services, offering users the ability to create profiles, search others’ profiles, and communicate via a manufactured email address. In their profiles, participants may include one or more photographs and a written (open-ended) description of themselves and their desired mate. They also answer a battery of closed-ended questions, with preset category-based answers, about descriptors such as income, body type, religion, marital status, and alcohol usage. Users can conduct database searches that generate a list of profiles that match their desired parameters (usually gender, sexual orientation, age, and location). Initial communication occurs through a double-blind email system, in which both email addresses are masked, and participants usually move from this medium to others as the relationship progresses.

Data Collection

Given the relative lack of prior research on the phenomenon of online dating, we used qualitative methods to explore the diverse ways in which participants understood and made sense of their experience (Berger & Luckman, 1980) through their own rich descriptions and explanations (Miles & Huberman, 1994). We took an inductive approach based on general research questions informed by literature on online self-presentation and relationship formation rather than preset hypotheses. In addition to asking about participants’ backgrounds, the interview protocol included open-ended questions about their online dating history and goals, profile construction, honesty and self-disclosure online, criteria used to assess others online, and relationship development. Interviews were semistructured to ensure that all participants were asked certain questions and to encourage participants to raise other issues they felt were relevant to the research. The protocol included questions such as: “How did you decide what to say about yourself in your profile? Are you trying to convey a certain impression of yourself with your profile? If you showed your profile to one of your close friends, what do you think their response would be? Are there any personal characteristics that you avoided mentioning or tried to deemphasize?” (The full protocol is available from the authors.)

As recommended for qualitative research (Eisenhardt, 1989; Glaser & Strauss, 1967), we employed theoretical sampling rather than random sampling. In theoretical sampling, cases are chosen based on theoretical (developed a priori) categories to provide examples of polar types, rather than for statistical generalizability to a larger population (Eisenhardt, 1989). The Director of Market Research at initially contacted a subsample of members in the Los Angeles and San Francisco Bay areas, inviting them to participate in an interview and offering them a free one-month subscription to in return. Those members who did not respond within a week received a reminder email. Of those contacted, 76 people volunteered to participate in an interview. Out of these 76 volunteers, we selected and scheduled interviews with 36 (although two were unable to participate due to scheduling issues). We chose interview participants to ensure a good mix on each of our theoretical categories: gender, age, urban/rural, income, and ethnicity. We focused exclusively on those seeking relationships with the opposite sex, as this group constitutes the majority of users. We also confirmed that they were active participants in the site by ensuring that their last login date was within the past week and checking that each had a profile.

Fifty percent of our participants were female and 50% were male, with 76% from an urban location in Los Angeles and 24% from a more rural area surrounding the town of Modesto in the central valley of California. Participants’ ages ranged from 25 to 70, with most being in their 30s and 40s. Their online dating experience varied from 1 month to 5 years. Although our goal was to sample a mix of participants who varied on key demographic criteria rather than generalizing to a larger population, our sample is in fact reflective of the demographic characteristics of the larger population of’s subscribers. Thirty-four interviews were conducted in June and July 2003. Interviews were conducted by telephone, averaging 45 minutes and ranging from 30 to 90 minutes in length. The interview database consisted of 551 pages, including 223,001 words, with an average of 6559 words per interview.

Data Analysis

All of the phone interviews were audio-recorded, transcribed, and checked for accuracy by the researcher who conducted the interview. Atlas.ti, a software program used for qualitative content analysis, was used to analyze interview transcripts. Data analysis was conducted in an iterative process, in which data from one informant were confirmed or contradicted by data from others in order to refine theoretical categories, propositions, and conclusions as they emerged from the data (Lincoln & Guba, 1985). We used microanalysis of the text (Strauss & Corbin, 1998) to look for common themes among participants. The data analysis process consisted of systematic line-by-line coding of each transcript by the first two authors. Following grounded theory (Glaser & Strauss, 1967), we used an iterative process of coding. Coding consisted of both factual codes (e.g., “age,”“female,”“Los Angeles”) and referential codes (e.g., “filter,”“rejection,”“honesty”) and served both to simplify and reduce data as well as to complicate data by expanding, transforming, and reconceptualizing concepts (Coffey & Atkinson, 1996). New codes were added throughout the process, and then earlier transcripts were recoded to include these new conceptual categories. All of the data were coded twice to ensure thoroughness and accuracy of codes. The researchers had frequent discussions in which they compared and refined coding categories and schemes to ensure consistency. During the coding process, some codes were collapsed or removed when they appeared to be conceptually identical, while others were broken out into separate codes when further nuances among them became apparent.

A total of 98 codes were generated by the first two authors as they coded the interviews. Unitization was flexible in order to capture complete thought units. Codes were allowed to overlap (Krippendorff, 1980); this method of assigning multiple codes to the same thought unit facilitated the process of identifying relationships between codes. See Appendixes A and B for more information on codes.


These interview data offer insight into the self-presentation strategies utilized by participants in order to maximize the benefits and minimize the risks of online dating. Many of these strategies revolved around the profile, which is a crucial self-presentation tool because it is the first and primary means of expressing one’s self during the early stages of a correspondence and can therefore foreclose or create relationship opportunities. These strategies are intimately connected to the specific characteristics of the online dating context: fewer cues, an increased ability to manage self-presentation, and the need to establish credibility.

The Importance of Small Cues

When discussing their self-presentational strategies, many participants directly or indirectly referred to the fact that they carefully attended to subtle, almost minute cues in others’ presentational messages, and often seemed to take the same degree of care when crafting their own messages. As suggested by SIP (Walther, 1992), subtle cues such as misspellings in the online environment are important clues to identity for CMC interactants. For instance, one participant said she looked for profiles that were well-written, because “I just think if they can’t spell or … formulate sentences, I would imagine that they’re not that educated.” Because writing ability was perceived to be a cue that was “given off” or not as controllable, participants noticed misspelled words in profiles, interpreting them as evidence of lack of interest or education. As one female participant put it, “If I am getting email from someone that obviously can’t spell or put a full sentence together, I’m thinking what other parts of his life suffer from the same lack of attentiveness?” These individuals often created their own profiles with these concerns in mind. For instance, one participant who found spelling errors “unattractive” composed his emails in a word processing program to check spelling and grammar.

Many of the individuals we interviewed explicitly considered how others might interpret their profiles and carefully assessed the signals each small action or comment might send:

I really analyzed the way I was going to present myself. I’m not one of these [people who write] all cutesy type things, but I wanted to be cute enough, smart enough, funny enough, and not sexual at all, because I didn’t want to invite someone who thought I was going to go to bed with them [as soon as] I shook their hand. (PaliToWW, Los Angeles Female)2

In this case, the participant “really analyzed” her self-presentation cues and avoided any mention of sexuality, which she felt might indicate promiscuity in the exaggerated context of the profile. This same understanding of the signals “sexual” references would send was reflected in the profile of another participant, who purposefully included sexually explicit terminology in his profile to “weed out” poor matches based on his past experience:

The reason I put [the language] in there is because I had some experiences where I got together [with someone], we both really liked each other, and then it turned out that I was somebody who really liked sex and she was somebody that could take it or leave it. So I put that in there to sort of weed those people out. (imdannyboy, Los Angeles Male)

Participants spoke of the ways in which they incorporated feedback from others in order to shape their self-presentational messages. In some cases, they seemed genuinely surprised by the ways in which the digital medium allowed information to leak out. For instance, one male participant who typically wrote emails late at night discussed his reaction to a message that said, “Wow, it’s 1:18 in the morning, what are you doing writing me?” This email helped him realize how much of a “night owl” he was, and “how not attractive that may be for women I’m writing because it’s very clear the time I send the email.” Over time, he also realized that the length of his emails was shaping impressions of him, and he therefore regulated their length. He said:

In the course of [corresponding with others on the site] I became aware of how I had to present myself. Also, I became quite aware that I had to be very brief. … More often than not when I would write a long response, I wouldn’t get a response. … I think it implied. … that I was too desperate for conversation, [that] I was a hermit. (joet8, Los Angeles Male)

The site displayed the last time a user was active on the site, and this small cue was interpreted as a reliable indicator of availability. As one male participant said, “I’m not going to email somebody who hasn’t been on there for at least a week max. If it’s been two weeks since she’s logged on, forget her, she’s either dating or there’s a problem.”

Overall, the mediated nature of these initial interactions meant that fewer cues were available, therefore amplifying the importance of those that remained. Participants carefully attended to small cues, such as spelling ability or last login date, in others’ profiles in order to form impressions. In a self-reflexive fashion, they applied these techniques to their own presentational messages, carefully scrutinizing both cues given (such as photograph) and, when possible, those perceived to be given off (such as grammar).

Balancing Accuracy and Desirability in Self-Presentation

Almost all of our participants reported that they attempted to represent themselves accurately in their profiles and interactions. Many expressed incomprehension as to why others with a shared goal of an offline romantic relationship would intentionally misrepresent themselves. As one participant explained, “They polish it up some, like we all probably do a little bit, but for the most part I would say people are fairly straightforward.” However, as suggested by previous research on self-disclosure and relationship development, participants reported competing desires. At times, their need to portray a truthful, accurate self-representation was in tension with their natural inclination to project a version of self that was attractive, successful, and desirable. Speaking about this tendency towards impression management, one participant noted that she could see why “people would be dishonest at some point because they are still trying to be attractive … in the sense they would want this other person to like them.”

Ideal Self

One way in which participants reconciled their conflicting needs for positive self-presentation and accuracy was to create profiles that described a potential, future version of self. In some cases, participants described how they or others created profiles that reflected an ideal as opposed to actual self: “Many people describe themselves the way they want [to be] … their ideal themselves.” For example, individuals might identify themselves as active in various activities (e.g., hiking, surfing) in which they rarely participated, prompting one participant to proclaim sarcastically, “I’ve never known so many incredibly athletic women in my life!” One participant explained,

For instance, I am also an avid hiker and [scuba diver] and sometimes I have communicated with someone that has presented themselves the same way, but then it turns out they like scuba diving but they haven’t done it for 10 years, they like hiking but they do it once every second year … I think they may not have tried to lie; they just have perceived themselves differently because they write about the person they want to be … In their profile they write about their dreams as if they are reality. (Christo1, Los Angeles Male)

In two cases, individuals admitted to representing themselves as less heavy than they actually were. This thinner persona represented a (desired) future state for these individuals: “The only thing I kind of feel bad about is that the picture I have of myself is a very good picture from maybe five years ago. I’ve gained a little bit of weight and I feel kind of bad about that. I’m going to, you know, lose it again.” In another case, a woman who misrepresented her weight online used an upcoming meeting as incentive to minimize the discrepancy between her actual self and the ideal self articulated in her profile:

I’ve lost 44 pounds since I’ve started [online dating], and I mean, that’s one of the reasons I lost the weight so I can thank online dating for that. [Because] the first guy that hit on me, I checked my profile and I had lied a little bit about the pounds, so I thought I had better start losing some weight so that it would be more honest. That was in December, and I’ve lost every week since then. (MaryMoon, Los Angeles Female)

In this case, a later physical change neutralized the initial discursive deception. For another participant, the profile served as an opportunity to envision and ideate a version of self that was future-focused and goal-oriented:

I sort of thought about what is my ideal self. Because when you date, you present your best foot forward. I thought about all the qualities that I have, you know, even if I sometimes make mistakes and stuff. … And also got together the best picture I had, and kind of came up with what I thought my goals were at the time, because I thought that was an important thing to stress. (Marty7, Los Angeles Male)

Overall, participants did not see this as engaging in deceptive communication per se, but rather as presenting an idealized self or portraying personal qualities they intended to develop or enhance.

Circumventing Constraints

In addition to impression management pressures, participants’ expressed desires for accurate representation were stymied by various constraints, including the technical interface of the website. In order to activate an online profile, participants had to complete a questionnaire with many closed-ended responses for descriptors such as age, body type, zip code, and income. These answers became very important because they were the variables that others used to construct searches in order to narrow the vast pool of profiles. In fact, the front page of includes a “quick” search on those descriptors believed to be most important: age, geographical location, inclusion of photograph, and gender/sexual orientation.

The structure of the search parameters encouraged some to alter information to fit into a wider range of search parameters, a circumvention behavior that guaranteed a wider audience for their profile. For example, participants tended to misrepresent their age for fear of being “filtered out.” It was not unusual for users who were one or two years older than a natural breakpoint (i.e., 35 or 50) to adjust their age so they would still show up in search results. This behavior, especially if one’s actual age was revealed during subsequent email or telephone exchanges, seemed to be socially acceptable. Many of our participants recounted cases in which others freely and without embarrassment admitted that they had slightly misrepresented something in their profile, typically very early in the correspondence:

They don’t seem to be embarrassed about [misrepresenting their age] … in their first reply they say, “oh by the way, I am not so many years, I am that many years.” And then if I ask them, they say, well, they tend to be attracted to a little bit younger crowd and they are afraid that guys may surf for a certain age group of women, because you use those filters. I mean, I may choose to list only those that are between X and Y years old and they don’t want to be filtered away. … They are trying to be sort of clever so that people they tend to be attracted to will actually find them. (Christo1, Los Angeles Male)

If lying about one’s age was perceived to be the norm, those who didn’t engage in this practice felt themselves to be at a disadvantage (see Fiore & Donath, 2004). For instance, one participant who misrepresented his age on his profile noted:

I’m such an honest guy, why should I have to lie about my age? On the other hand, if I put X number of years, that is unattractive to certain people. They’re never going to search that group and they’re never going to have an opportunity to meet me, because they have a number in their mind just like I do. … Everybody lies about their age or a lot of people do. … So I have to cheat too in order to be on the same page as everybody else that cheats. If I don’t cheat that makes me seem twice as old. So if I say I am 44, people think that I am 48. It blows. (RealSweetheart, Bay Area Male)

In the above cases, users engaged in misrepresentation triggered by the social norms of the environment and the structure of the search filters. The technical constraints of the site may have initiated a more subtle form of misrepresentation when participants were required to choose among a limited set of options, none of which described them sufficiently. For instance, when creating their profiles, participants had to designate their “perfect date” by selecting one from a dozen or so generic descriptions, which was frustrating for those who did not see any that were particularly appealing. In another case, one participant complained that there was not an option to check “plastic surgery” as one of his “turn-offs” and thus he felt forced to try to discern this from the photos; yet another participant expressed his desire for a “shaved” option under the description of hair type (“I resent having to check ‘bald’”).

Foggy Mirror

In addition to the cases in which misrepresentation was triggered by technical constraints or the tendency to present an idealized self, participants described a third branch of unintentional misrepresentation triggered by the limits of self-knowledge. We call this phenomenon “foggy mirror” based on this participant’s explanation:

People like to write about themselves. Sometimes it’s not truthful, but it’s how they see themselves and that gives you a different slant on an individual. This is how they really see themselves. Sometimes you will see a person who weighs 900 pounds and—this is just an exaggeration—and they will have on spandex, you’ll think, “God, I wish I had their mirror, because obviously their mirror tells them they look great.” It’s the same thing with online. (KarieK, Bay Area Female)

This user acknowledges that sometimes others weren’t lying per se, but the fact that their self-image differed from others’ perceptions meant that their textual self-descriptions would diverge from a third party’s description. In explaining this phenomenon, KarieK used the metaphor of a mirror to emphasize the self-reflexive nature of the profile. She also refers to the importance of subtle cues when she notes that a user’s self-presentation choices give one a “different slant on an individual.” The term “foggy mirror” thus describes the gap between self-perceptions and the assessments made by others. The difference might be overly positive (which was typically the case) or negative, as the below example illustrates. A male participant explained:

There was one gal who said that she had an “average” body shape. … When I met her she was thin, and she said she was “average,” but I think she has a different concept of what “average” is. So I then widened my scope [in terms of search parameters] and would go off the photographs. What a woman thinks is an “average” body and what I think is an “average” body are two different things. (joet8, Los Angeles Male)

In this case, the participant acknowledged the semantic problems that accompany textual self-descriptions and adopted a strategy of relying on photographs as visual, objective evidence, instead of subjective, ambiguous terms like “average.” To counter the “foggy mirror” syndrome in their own profiles, some individuals asked friends or family members to read their profiles in order to validate them.

In regards to self-presentation, the most significant tension experienced by participants was one not unique to the online medium: mediating between the pressures to present an enhanced or desired self (Goffman, 1959) and the need to present one’s true self to a partner in order to achieve intimacy (Reis & Shaver, 1988). In their profiles and online interactions, they attempted to present a vision of self that was attractive, engaging, and worthy of pursuit, but realistic and honest enough that subsequent face-to-face meetings were not unpleasant or surprising. Constructing a profile that reflected one’s “ideal self” (Higgins, 1987) was one tactic by which participants reconciled these pressures. In general, although all of our participants claimed they attempted to be honest in their self-presentation, misrepresentations occurred when participants felt pressure to fudge in order to circumvent the search filters, felt the closed-ended options provided by the site didn’t describe them accurately, or were limited by their self-knowledge.

Establishing Credibility

The increased ability to engage in selective self-presentation, and the absence of visual cues in the online environment, meant that accuracy of self-presentation was a salient issue for our interviewees. The twin concerns that resulted from these factors—the challenge of establishing the credibility of one’s own self-descriptions while assessing the credibility of others’ identity claims—affected one another in a recursive fashion. In an environment in which there were limited outside confirmatory resources to draw upon, participants developed a set of rules for assessing others while incorporating these codes into their own self-presentational messages. For example, one participant made sure that her profile photograph showed her standing up because she felt that sitting or leaning poses were a camouflage technique used by heavier people. This illustrates the recursive way in which participants developed rules for assessing others (e.g., avoid people in sitting poses) while also applying these rubrics to their own self-presentational messages (e.g., don’t show self in sitting pose).

Participants adopted specific tactics in order to compensate for the fact that traditional methods of information seeking were limited and that self-reported descriptions were subject to intentional or unintentional misrepresentation when others took advantage of the “selective self-presentation” (Walther & Burgoon, 1992) available in CMC. As one participant noted, “You’re just kind of blind, you don’t know if what they’re saying in their profile online is true.” Acknowledging the potential for misrepresentation, participants also sought to “show” aspects of their personality in their profiles versus just “telling” others about themselves. They created their profiles with an eye towards stories or content that confirmed specific personality traits rather than including a ‘laundry list’ of attributes. As one Los Angeles male participant explained, “I attempted to have stories in my profile somewhat to attempt to demonstrate my character, as opposed to, you know, [just writing] ‘I’m trustworthy,’ and all that bit.” This emphasis on demonstration as opposed to description was a tactic designed to circumvent the lack of a shared social context that would have warranted identity claims and hedged against blatant deception.

Another aspect of “showing” included the use of photographs, which served to warrant or support claims made in textual descriptions. Profile photographs communicated not only what people looked like (or claimed to look like), but also indicated the qualities they felt were important. For instance, one man with a doctorate included one photo of himself standing against a wall displaying his diplomas and another of him shirtless. When asked about his choice of photos, he explained that he selected the shirtless photo because he was proud of being in shape and wanted to show it off. He picked the combination of the two photos because “one is sort of [my] intellectual side and one is sort of the athletic side.” In this case, the photos functioned on multiple levels: To communicate physical characteristics, but also self-concept (the aspects of self he was most proud of), and as an attempt to provide evidence for his discursive claims (his profile listed an advanced degree and an athletic physique).

To summarize, our data suggest that participants were cognizant of the online setting and its association with deceptive communication practices, and therefore worked to present themselves as credible. In doing so, they drew upon the rules they had developed for assessing others and turned these practices into guidelines for their own self-presentational messages.


The primary goal of the online dating participants interviewed for this study was to find someone with whom they could establish a dating relationship (although desired commitment level and type of relationship varied across participants). Given this, they attempted to achieve their goals while contending with the unique characteristics of the online environment, engaging in strategies designed to circumvent the constraints of the online dating environment while exploiting its capacities. One constraint—the lack of nonverbal cues—meant that the task of interpreting the remaining cues became paramount in regards to both assessment of others and presentation of self. Since the goal of most online dating participants was to identify and interact with potential romantic partners, individuals strove to highlight their positive attributes and capitalize on the greater perceived control over self-presentation inherent in the medium. However, the future face-to-face interaction they anticipated meant that individuals had to balance their desire for self-promotion with their need for accurate self-presentation. In response to the risk of misrepresentation online, made possible by the selective self-presentation affordances of CMC, participants adopted various strategies to demonstrate the credibility of their identity claims, recursively applying the same techniques they employed to uncover representational ruses in others. Our findings suggest that participants consistently engaged in creative workarounds (circumvention strategies) as they went through the process of posting a profile, selecting individuals to contact, and communicating with potential romantic partners. Our data also highlight the recursive process by which some participants constructed rules of thumb for assessing others (e.g., an inactive account indicates a lack of availability or interest) while simultaneously incorporating these rules in their own messages (e.g., frequently making slight adjustments to the profile).

Theoretical Implications

As individuals make initial decisions about potential partners, they form impressions that help reduce uncertainty about the other (Berger & Calabrese, 1975). For this to happen in the context of CMC, SIP argues, individuals will adapt their behaviors to the cues that are available (Tidwell & Walther, 2002; Walther, 1992) to convey information to one another. While empirical support for SIP has been demonstrated (see Walther & Parks, 2002 for a review), this article is among the few to provide evidence for SIP in a naturalistic setting. Our data show that in the initial interactions of online dating participants, stylistic aspects of messages such as timing, length, and grammar appear equally as important as the content of the message itself; this is consistent with SIP’s formulation that when nonverbal cues are decreased, the remaining cues become more salient to users. Previous laboratory studies of SIP have tended to focus on the manipulation of a subset of cues. A unique contribution of this study’s extension of SIP is its demonstration of the organic interplay of these alternative sources of social information online.

Although much of the public debate about online dating has centered on the medium’s inability to ensure participants’ truthful self-descriptions, our interview data suggest that the notion that people frequently, explicitly, and intentionally “lie” online is simplistic and inaccurate. Exploring the question of whether participants created a playful or fantastical identity online (Stone, 1996; Turkle, 1995) or were more open and honest (Rubin, 1975), we found that the online dating participants we spoke with claimed that they attempted to present an accurate self-representation online, a finding echoed in our survey data (Gibbs et al., 2006). This study highlights the fact that creating an accurate online representation of self in this context is a complex and evolving process in which participants attempt to attract desirable partners while contending with constraints such as those posed by technological design and the limits of self-knowledge.

In some cases, the technical constraints of the site may have unintentionally enabled acts of misrepresentation, for instance when participants slightly altered information in situations in which they felt an arbitrary data point (in age, for example) would significantly harm their chances of being discovered by a potential mate. Additionally, self-reported descriptions that use subjective terms (e.g., “pretty” or “average”) could also result in unintentional misrepresentation due to different interpretations of these terms. Additionally, as Shah and Kesan point out, “Defaults have a legitimating effect, because they carry information about what most people are expected to do” (2003, p. 7). In the case of online dating, it may be that the default settings in the search field (i.e., an age range, whether searches are limited to profiles with photographs) influence user perceptions of the desirability or appropriateness of certain responses.

Additionally, our interview data suggest that online representations of one’s ideal self—when combined with the increased accountability engendered by an anticipated face-to-face interaction—may serve as a tool to enable individuals to minimize the discrepancy between their actual and their ideal selves. The ideal self refers to qualities or achievements one strives to possess in the future (Bargh et al., 2002). In the realm of online dating, it is interesting that participants reported using the profile to ideate a version of self they desired to experience in the future. For some, the act of constructing an online profile may begin a process of self-growth as they strive to close the gap between actual and ideal self, such as the woman who misrepresented her weight but then was able to achieve her goal of weight loss over time. Future research is needed to assess the extent to which this phenomenon exists and its long-term consequences for processes of self-growth.

More research is also needed to understand fully whether strategies designed to circumvent constraints (technical or other) are perceived to be deceptive by users and, if so, which norms govern their use. The literature on deception explores a wide range of deceptive acts, ranging from the more mundane “diversionary responses” to outright “lies” (Buller & Burgoon, 1994). Future research could work to develop a taxonomy of online deception and acceptability, which takes into account the nuances of social norms and the fact that some misrepresentation may be unintentional or socially accepted. For instance, if a profile includes incorrect information that is rectified immediately over email, is it a “lie?” More importantly, is it acceptable? Also, more research is needed to understand more clearly the extent and substance of participants’ actual concerns regarding online dating (i.e., misrepresentation, effectiveness, safety) and how they overlap with the often sensationalized discourse about online deception as represented in media accounts and social narratives.

Practical Implications

Given that deceptive practices are a concern for online dating participants, future research should explore the ways in which online dating sites could implement design features aimed at addressing these issues. For instance, they could acknowledge and incorporate aspects of a shared social context, similar to social networking sites like Friendster (Donath & boyd, 2004), through the use of testimonials or social network visualizations. Online dating sites could adopt some of the design features used in e-commerce sites, such as testimonials, user rating systems, or social network visualizations, where participants also must operate in an uncertain environment in which warranting is difficult and deception can be costly (Resnick & Zeckhauser, 2002).

A second design consideration is the possibility that the technical characteristics of some online dating sites may privilege objective characteristics (such as demographic features) and de-emphasize the process of seeing others as individuals rather than as amalgams of various traits. The benefit, or capacity, of online dating is that participants can use specific search parameters to cull a subset of profiles from a larger database. Participants acknowledged that the online dating environment placed more emphasis on certain kinds of information—information that might not be very important in a face-to-face setting when chemistry was already established. To compensate for or to circumvent these constraints, participants tried to create profiles that stood out or evidenced aspects of self that they were particularly proud of rather than a laundry list of features. They struggled to present themselves as unique individuals within the constraints of a technical system that encouraged homogeneity, negotiating a desire to stand out with the need to blend in. Future research might examine the potential for developing self-presentation tools that allow individuals more nuanced ways of expressing themselves in the online environment, such as video presentations, more sophisticated communication tools, or triangulated information from others on the site. Online dating sites may need to reconsider the ways in which profiles are structured and the characteristics they include; as Fiore and Donath argue, “the features of a person that presents as salient to romance will begin to have some psychological and cultural influences if 40 million Americans view them every month” (2004, p. 1395). If we accept this claim, then it stands to reason that participants’ visions of self may be impacted by their online self-presentations, especially if these presentations are constrained.


We chose to conduct interviews with online dating participants in order to gain insight into how they perceived their experiences and the processes through which they learned to avoid the pitfalls and exploit the possibilities of online dating. However, there are several limitations that should be acknowledged in our method and sample. Limitations of this study include the sampling of only participants located on the West Coast. While members are worldwide, we cannot assess if regional or national differences affect the online dating experience. A major limitation is the potential for self-selection bias, as participants volunteered for the study. While demographically diverse, those that chose to volunteer might be biased toward a more positive outlook on online dating or potentially more honest in their online dating practices.

In addition, the self-reported nature of the data may have resulted in a social desirability bias, making participants less likely to admit to intentional misrepresentation. Finally, many of our findings may be specific to’s model of online dating, in which participants post profiles and select with whom they want to communicate. Other online dating sites, such as eHarmony, utilize a very different model, acting as online matchmakers where individuals who are found to be compatible are paired based on personality tests developed by “expert” psychologists. Future research could assess whether variables like self-efficacy predict which model users choose to utilize. Although our observations in this article were based on the sample as a whole, we acknowledge that there may be differences (for instance, along gender lines) which are beyond the scope of this article but which could be explored in future research.

Although self-presentation and relationship formation have been studied in other online contexts, tracing how these processes take place in the online dating realm offers researchers unique insights into the crucial role of circumvention techniques, the complicated nature of “honesty” in online environments, and the social and psychological implications of the design and structure of these sites. From a historical perspective, the goals of online dating participants are not that different from those described by poets throughout the ages. What is different is the tools in their repertoire and the constraints and opportunities they present. As O’Sullivan writes, “From a functional perspective, it appears new technologies may be providing nothing terribly new— just new ways of doing things that people have been doing throughout the history of social interaction” (2000, p. 428). This study has attempted to elucidate and explain some of these social practices as a window into the ways in which new communication technologies are shaping us—and we are shaping them—in the ongoing pursuit of romantic relationships.


The authors thank Karen Aroian, Ulla Bunz, Annika Hylmo, Edythe Hough, Patrick O’Sullivan, Charles Steinfield, Joe Walther, and two anonymous reviewers for their helpful comments and suggestions regarding this manuscript.


  1. Similar to the concept of “workaround” employed by designers and software engineers, users engage in circumvention strategies to neutralize constraints—or turn them into capabilities. Prior CMC research has identified similar processes in interpersonal contexts. For instance, O’Sullivan (2000) found that users chose mediated channels over face-to-face communication in situations where a preferred impression was expected to be violated in order to capitalize on the face-saving capabilities of mediated interaction. Similarly, CMC researchers working in other contexts have noted the process by which individuals adapt their behavior to compensate for the limitations imposed by the medium in order to pursue their communication goals (Walther, Loh, & Granka, 2005).

  2. All identifying information about our participants has been changed to protect their confidentiality, although we have attempted to use pseudonyms that reflect the tone and spirit of their chosen screen names. Additionally, at the request of our research site, we have used a pseudonym in place of the site’s actual name.

About the Authors

  1. Nicole Ellison is an Assistant Professor in the Department of Telecommunications, Information Studies, and Media at Michigan State University. Her research explores issues of self-presentation, relationship development, and identity in online environments such as weblogs, online dating sites, and social networking services.Address: Dept. of Telecommunications, Information Studies, and Media, Michigan State University, 403 Communication Arts and Sciences, East Lansing, MI 48824 USA

  2. Rebecca Heino is an Assistant Professor in the McDonough School of Business at Georgetown University. She has centered her research on the use of communication technologies—such as intranets and email—in organizations, specifically focusing on organizational adoption, implementation, and privacy.Address: Georgetown University, McDonough School of Business, 37th & O St. NW, Old North G-04, Washington, DC 20057 USA

  3. Jennifer Gibbs is an Assistant Professor of Communication at Rutgers University. Her research interests center around how individuals connect, collaborate, and negotiate identity and differences in global, multicultural, and mediated contexts.Address: Department of Communication, SCILS, 4 Huntington Street, Rutgers University, New Brunswick, NJ 08901-1071 USA


Table Appendix A:.  Sample Codes and Quotes
CodeSample Quotes
Honesty“So I think some people intentionally lie, and then I think some people truly try to be honest, but they still don’t really know what they want. And then I think there’s a third group of people, that truly are honest and truly know what they want and are just picky as hell, and they either meet someone or they just give up.”
“I don’t know why anyone would want to do that [be dishonest]. I mean unless you just like have nothing else to do and you just want to play with people I guess. I mean, sure I could make myself a multimillionaire (laugh) and a Playboy model I guess. If I was going to do it I might as well go all out.”
Cues Masked/Amplified“Another guy was really, like, seemed so awesome on the phone, had a really fun personality, worked in TV and we had the greatest conversation. I met him in Santa Monica, where he lived, and first of all he had no personality. And I was like, ‘oh my God, either it just radically changed overnight or he is only good on the phone.’”
“I think if someone can write really well, that would come out more easily than in person. What is good about the Internet for a shy person is that you can show what you have to offer more than you can in person. What is difficult for a shy person is that it takes a little more time to open up and show what you have to offer, and you can do that on the Internet more, I think.”
Table Appendix B:.  Most Important Codes with Frequencies*
  • *

    All codes with more than 100 occurrences.

Supply and demand180
Desired qualities in partner180
Cues/characteristics masked or amplified by medium175
Strategies—online (things learned about online dating)171
Online vs. traditional dating163
Physical attractiveness/appearance150
Effectiveness/efficiency of online dating128
Context/weak, strong ties119