Computers continue to pervade the daily lives of individuals. In addition to being a tool for education and research, Internet capabilities now include commercial (e.g., Ebay.com) and personal (e.g., EHarmony.com) activities. Moreover, the development of wireless Internet connections (Internet capabilities on cellular phones and PDAs) has made it possible for individuals to stay connected wherever they travel. Schools and public libraries offer free access to online research to individuals who cannot afford the financial commitment of wireless service (Hargittai, 2003). An increasing number of Internet providers offer affordable service, free service, or service combined with other amenities (e.g., Cable Television). The pervasiveness of this technology has made the Internet accessible to even the most casual of computer users. Accordingly, the number of people venturing into cyberspace continues to grow; it is estimated that of the 79.5% of Americans with access to the Internet in 2004, 63% had accessed the Internet within the last 30 days (Mediamark Research, 2004).
While some research has explored the actions individuals perform while managing impressions on the Internet or personal web pages (Walther, 1993), little is known about how receivers perceive these intentional messages. Uncertainty Reduction Theory (Berger & Calabrese, 1975) offers insight into the strategies individuals use to assess others during initial encounters. Although traditionally used to predict face-to-face encounters, Uncertainty Reduction Theory (URT) may offer insights into the ways individuals interpret and respond to the mediated messages of others. The current study uses URT to explore the ways in which individuals utilize mediated cues in developing impressions of others. Since electronic mail is one of the most pervasive forms of computer-mediated communication (CMC), the current study focuses on email usernames or pseudonyms. Specifically, the study explores whether, in the absence of face-to-face interaction, individuals use email usernames as a basis for constructing perceptions of other individuals, and if so, in what ways.
Computer-mediated communication: impersonal vs. personal
Early research on CMC argued that the Internet and its technologies were best suited to impersonal and task-oriented interactions (Kiesler, Siegel, & McGuire, 1984). Support for these claims was based on the reduced number of nonverbal cues available to people communicating through mediated channels (Culnan & Markus, 1987; Kiesler, 1986; Kiesler, et al., 1984). “In traditional forms of communication, head nods, smiles, eye contact, distance, tone of voice, and other nonverbal behavior give speakers and listeners information they can use to regulate, modify, and control exchanges” (Kiesler, et al., 1984, p. 1125). Without these nonverbal, relational clues, CMC was seen as lacking the communicative subtext necessary to support interpersonal communication (Connolly, Jessup, & Valacich, 1990; Hiltz, Johnson, & Turoff, 1986; Hiltz, Turoff, & Johnson, 1989).
Despite these early conclusions, subsequent interest has centered on the interpersonal implications of cyberspace (Cohen & Metzger, 1998; Mitra, 1997; Parks & Floyd, 1996; Soukup, 2000; Walther, 1992). Research has identified nonverbal cues in the form of graphical icons, or emoticons, that operate in place of traditional face-to-face affiliative behaviors (Thurlow, Lengel, & Tomic, 2004). Using social identification/de-individuation (SIDE) and social information processing theories, Walther and Tidwell (1995) explored how response times (e.g., delayed versus immediate responses) to individual emails may serve as nonverbal cues in mediated contexts. In their study, long delays in responses to emails were perceived as sending a negative nonverbal message, in much the same way that avoidance of face-to-face contact is perceived in nonmediated circumstances. Moreover, length and depth of messages signal intimacy much the same way that personal disclosure operates in face-to-face interactions (Soukup, 2000), although mediated relationships seem to develop intimacy more quickly than their face-to-face counterparts (Walther, 1996; Walther & Burgoon, 1991).
An overview of CMC research suggests that communication mediated by computers can be impersonal and task-oriented as well as enabling interpersonal communication and relationships (Walther, 1996). As Cohen and Metzger (1998) argue,
[u]nderlying the motivations for both mediated and face-to-face communication is a basic need for social affiliation. The need for social affiliation is so central for communication because it stems from, and is necessary for, understanding of who we are in relation to the world around us. (p. 1)
The need for connection with others has led to a growing interest in self-presentation in cyberspace. Dominick (1999) observes that mediated behavior seems to mirror self-presentation in interpersonal settings. Walther (1996) contends that “in some cases, opportunities for selective self-presentation, idealization, and reciprocation exceed what we may accomplish face-to-face, in terms of our impression-generating and relational goals” (p. 28). Certainly the mediated nature of CMC allows for greater manipulation and more careful construction of personal information; the delayed nature of most CMC gives participants the opportunity to review, revise, and even cancel their communications before they are sent. Research on impression management confirms that much of this process is conscious; individuals rely on context to determine how they will represent themselves (Andrews & Kacmar, 2001; Berzonsky & Adams, 1999; Yoder, 2000). In CMC, there is the ability to conceal or construct identities. Gender, race, class, ethnicity, and religious affiliation can be manipulated, concealed, or accentuated in mediated contexts, according to the user’s intent.
Electronic mail (Email)
Electronic mail is particularly susceptible to impression management. One of the most pervasive forms of computer-mediated communication (Newby, 1993; Thurlow, Lengel, & Tomic, 2004), email is useful for achieving both task-oriented and personal goals. The use of email requires system administrators to establish an “email account” which in turn requires individuals to set a specific username and password to ensure privacy and security. Email is popular worldwide; it is easily accessible and can be used by even the most basic computer user. Moreover, individuals often utilize a single email account across a variety of interactions, including work and personal communications. Therefore, a single email username must function within a wide array of contexts.
It follows that one’s email username would be prone to impression management, much as one’s physical self-presentation is used to manage impressions during face-to-face encounters. Previous research on impression management has suggested that such “management” is conscious and context dependent. However, little is known about the ways in which impressions are received by others, particularly in mediated contexts. The previously cited research suggests that individuals are likely to construct “context driven” impressions, even in computer-mediated settings. Additional research suggests that CMC contains many “signals” that provide recipients clues to the emotional state and development of relationship with senders of messages (Tidwell & Walther, 2002; Walther & Burgoon, 1992). Therefore, it seems that an individual’s email username could be viewed and interpreted by others as a potential source of information about its user. However, once the information is received, it is not clear what happens next. Uncertainty Reduction Theory offers potential insight into this process, including the role of information gathering and decision-making when uncertainty clouds receivers’ perceptions.
Uncertainty reduction theory and computer-mediated communication
Uncertainty Reduction Theory (URT) was initially proposed by Berger and Calabrese (1975) as an explanation for information-seeking behaviors of individuals during initial interactions. This theory of seven axioms and 21 theorems suggests that individuals are distressed when denied the ability to predict or explain others’ behaviors. While its authors acknowledge that much of social interaction is rule-governed—as in, for example, a typical “hi-how-are-you-fine-how-are-you” greeting script—URT contends that our predictive/explanatory powers are dependent on context, relationships with others, and the likelihood of future interaction with this person (e.g., Parks & Adelman, 1983; Planap & Honeycutt, 1985). According to URT, in moments when prediction/explanation are weakest, individuals will experience discomfort, anxiety, or stress—an experience the authors characterize as “uncertainty.” Because uncertainty is an unpleasant state, individuals tend to avoid situations likely to provoke uncertainty (or situations they cannot adequately predict or explain).
URT has been used extensively in research addressing the interpersonal components of CMC (e.g., Parks & Floyd, 1996; Pratt, Wiseman, Cody, & Wendt, 1999; Walther & Burgoon, 1991). For example, when comparing face-to-face groups with similarly matched online groups, Walther and Burgoon (1991) discovered that the face-to-face groups were more task-oriented than the CMC groups, and suggested that “[i]t is plausible that uncertainty reduction needs combined with [CMC’s] convenient time and channel for expression allow selective self-presentation and relational behavior to emerge more readily than in traditional group settings” (p. 79).
It should be noted that URT and other relational theories predate the explosion of computer-mediated relationships. Hence, the scope of such theories is limited in their application to CMC. However, this does not mean that URT cannot be of use in understanding new phenomena that arise in CMC. Given the nature of email pseudonyms, incorporating theoretical perspectives that address nonverbal and informational-focused aspects of relational dynamics is appropriate. URT acknowledges these valuable components of communication and does not require physical proximity or frequent interaction as conditions for relational development. The following paragraphs highlight the theory’s assumptions, main tenets, and explanatory capability for recipients of mediated messages.
URT and Email communication
As previously mentioned, URT assumes that humans have an innate desire to predict and explain the world around them. Prediction can be defined as the ability to forecast one’s own and others’ behavioral choices, while explanation is the ability to interpret the meaning of those behavioral choices. In email communication, these predictive and explanatory desires surface when deciding whether or not to open an incoming message (e.g., “Does it have a virus?;”“Is this likely to be a student complaint about class?”), constructing our own messages (“Does this message sound ok?”), and finally when attempting to make sense of others messages (“Why did she type in ALL CAPS?”).
According to the theory, when these predictive and explanatory abilities are weakened, individuals experience uncertainty. This uncertainty, an aversive state that absorbs emotional and psychological energy, is an experience most individuals would prefer to avoid and/or reconcile. During face-to-face encounters, individuals take steps to increase their predictive abilities by paying attention to the surrounding contextual and nonverbal clues. For instance, meeting someone at a bar and noticing his/her smile and eye contact might give insight into how this person would respond to a dinner invitation. In mediated contexts, the same initial uncomfortable uncertainty exists. Similarly, mediated individuals are driven to reduce uncertainty by attending to available clues. However, in the absence of contextual and facial nonverbal clues, individuals must rely on the mediated information to reduce their uncertainty. In email contexts, this mediated information includes the email username or pseudonym attached to the incoming message. Therefore, we hypothesize that:
H1: Study participants will utilize email usernames as sources of information about the sender of the message.
Moreover, given that increased predictability should lower uncertainty, it is possible that individuals will seek as much information about other individuals as possible. Therefore:
H2: Study participants will provide more descriptive information for creative email usernames (sober4alilbit) than for plain usernames (ZH7624).
An underlying assumption exists regarding the nonverbal cues exchanged and gathered in face-to-face settings. Research suggests that individuals treat most nonverbal behavior as intentional and designed to communicate information to others (e.g., Allen & Atkinson, 1981; Buck & VanLear, 2002; Manusov & Rodriguez, 1989; Richmond & McCroskey, 2004). For instance, yawning during a lecture is often perceived as an intentional message to the “boring” professor; the graphic t-shirt worn to a concert denotes one’s socio-economic status and cultural affiliation, and so forth. Consistent with the impression management literature, individuals in mediated contexts may assume that the usernames employed in cyberspace contain purposeful clues as to the personality or identity of the sender, and that they moreover serve as unique identifiers, since email addresses cannot be duplicated. Therefore, an individual’s perceptions of volition should influence the impact that a given username has on his or her perceptions. In other words, usernames perceived as selected and designed by their users should offer more information (and greater predictability) to others. Therefore:
H3: Study participants will provide greater amounts of information for those email usernames identified as chosen by the sender than for usernames considered to have been assigned to the sender.
H4: Study participants will view creative email usernames as having been chosen by senders more often than plain usernames. Conversely, plain usernames will be identified as having been assigned to the senders more often than creative usernames.
Berger and Calabrese (1975) also believed that levels of uncertainty could have relational implications. Assuming that an inability to predict/explain another’s behavioral choices creates uncertainty, and that this uncertainty is aversive and creates cognitive stress, it follows that highly unpredictable relationships could be difficult to maintain. URT relates this argument to “liking” for another: Increases in uncertainty produce decreases in liking, while decreases in uncertainty produce increases in liking. For email users, liking could be tied to the amount of information contained in the username. For instance, email pseudonyms that offer insight into a sender’s personality (e.g., Happygrl@any.com) or work ethic (e.g., firstname.lastname@example.org) might increase predictability, such that they increase others’ desire to work or hang out with them. Recent research has explored the impact of computer mediation on perceived personality (Rouse & Haas, 2003), and research in group communication and social perception supports the premise that an individual’s preference to work with others is guided by the social perceptions he or she has of them (Ambady, Hallahan, & Rosenthal, 1995; Cesario, Plaks, & Higgins, 2006). Therefore, we advance the following hypotheses:
H5a: Owners of creative usernames will be perceived by study participants as more productive and desirable to work with in group settings than owners of plain usernames.
H5b: Owners of creative usernames will be perceived as having more positive personality traits (e.g., “more fun”) than owners of plain usernames.
A final axiom of URT identifies information seeking as a common and successful strategy for the reduction of uncertainty in initial interactions. High levels of uncertainty cause increases in information seeking behavior, and as uncertainty levels decline, information seeking behavior decreases.
In face-to-face interactions, information-seeking behaviors typically manifest as questions. In order to reduce uncertainty, individuals engage in question and answer exchanges. As answers accumulate, individuals gain a greater ability to predict the others’ behaviors, and uncertainty is reduced. In the context of email pseudonyms, however, the information seeking changes form. In the absence of a traditional, face-to-face, “back and forth” question session, recipients may treat username information as an implicit first question/answer session. For instance, suppose that an unknown individual sends an email message, providing subject line and email username. This message information acts as an implicit answer to an unasked question. It provides the recipient the chance to further reduce uncertainty by opening the message (thus continuing the “back and forth” session), or to end the interaction by deleting. In order to reduce uncertainty, the recipient needs to continue the session (by opening the email message), yet the current influx of unsolicited commercial email (spam) leads many users to delete messages based on the perception of the origins of those messages. The resulting uncertainty may compel individuals to open incoming messages even from those usernames they do not recognize. Email usernames that provide more information about the sender may be more likely to be opened, as the pseudonyms provide recipients with information about the senders’ intentions and goals. Therefore, we hypothesize that:
H6: Study participants will be more likely to open a message from a person with a creative username than a message from a person with a plain username.
RQ1: What reasons to participants give for opening as opposed to deleting an incoming message?