This paper investigates how the type of community and personal goals affect self-presentation. In 3 online studies, we simulated the registration process in online communities, presenting either a common-bond or a common-identity community. Study 1 confirmed that members of the common-bond community presented themselves in an individualizing manner, while members of the common-identity community focused on characteristics shared among members of the community. Study 2 investigated underlying processes, showing that the goal to get in contact with other members was less salient in the common-identity than in the common-bond community. Study 3 demonstrated that community members actively manage their self-presentation in accordance with personal goals. Based on these findings, we discuss implications for research and profile design.
What to Tell About Me? Self-Presentation in Online Communities
Managing self-presentation in online communities is an integral part of private and professional life. In member profiles on social networking sites such as Facebook, people disclose intimate details that deliver a rich picture of their personality and tell stories about recent activities with colleagues, family, and friends. In communities such as Wikipedia, people selectively present personal information that is related to their effort within the community, for example, information about their interests, expertise, or professional activities. Finally, online self-presentation enables initiation of interaction, building relationships, and access to knowledge resources. The kind of personal information people present about themselves differs to a high degree between communities (Le, Beschastnikh, & McDonald, 2010; Schrammel, Köffel, & Tscheligi, 2009). Nevertheless, we do not know much about the factors that explain these differences. Are differences in self-presentation just a consequence of different opportunities to present oneself through predefined profile fields? Indeed, profiles in different communities vary in layout, content, and format. However, communities also support different kinds of interactions among their members: Social networking sites, on the one hand, support getting and staying in contact with family, friends, and colleagues. Communities focusing on content creation, on the other hand, support sharing and developing knowledge and ideas.
This paper investigates whether people adapt their self-presentation in online communities to different interaction situations, independent of the profile template the community offers. For this purpose, we build on the theoretical distinction between common-bond and common-identity groups and discuss consequences for self-presentation and communication within these groups. Thereby, we bridge social psychological theory and recent findings about self-presentation on social networking sites. We report the findings of three studies exploring members' self-presentation in different types of communities as well as the influence of personal goals interacting with community types. Finally, we discuss the implications for community design and integrate our results into the broader context of research, highlighting the importance of self-presentation for commitment and participation in communities. In sum, this paper contributes to the field by exploring self-presentation in different kinds of online communities and thus identifies new influences on online self-presentation beyond privacy concerns and personality effects.
Classification of Online Communities
Prentice, Miller, and Lightdale (1994) distinguished between common-bond groups and common-identity groups. People's attachment to common-bond groups is based on the interpersonal relations among group members, that is, on interpersonal attraction. Research on interpersonal attraction in groups indicates that the more people interact with other group members and get to know and like each other, the more they feel connected to that group (Lott & Lott, 1965; Newcomb, 1956). Importantly, common-bond groups exist only because of the interpersonal relations among their members.
In contrast, people's attachment to common-identity groups is based on the attractiveness of the group as a whole, that is, on social identification. According to Social Identity Theory and Self-Categorization Theory, getting to know other group members personally is not necessary to feel connected to a common-identity group. Instead, the perception of sharing at least one characteristic that connects all members is sufficient to foster attachment (Tajfel & Turner, 1979, 1986; Turner, 1985; Turner, Hogg, Oakes, Reicher, & Wetherell, 1987). In other words, social identity builds on shared characteristics (such as interests, attitudes, or values), on shared social categories (such as gender, nationality, or organizational affiliations), or on a common task or purpose (such as sports teams and work groups).
Exploring the effect of type of group on communication, Back (1951) showed that groups based on interpersonal attraction came up with longish and personal discussions, whereas groups with a common task discussed more efficiently and goal-oriented. Similarly, Postmes, Spears, Lee, and Novak (2005) found that members of common-bond groups contributed more words to a group discussion than members of common-identity groups. This difference in communication behavior was traced back to more off-topic communication in common-bond groups.
The differentiation of member attachment based on interpersonal attraction and group attachment based on social identity has also been applied to groups in the Internet (Ren, Kraut, & Kiesler, 2007; Sassenberg, 2002; Utz, 2003). Sassenberg (2002) showed that members of so-called off-topic chats (i.e., common-bond groups) felt interpersonally attracted to the other chat members, whereas members of so-called on-topic chats (i.e., common-identity groups) felt attached to the social identity shared with the other chat members resulting from the topic of the chat.
Accordingly, we use the term common-bond community to describe communities that are defined by the interpersonal relations among members and support their members to connect with each other. These communities are based on social networking technologies (e.g., Facebook1). Member profiles are their main content. Community-set interaction opportunities such as contact lists, asynchronous messaging, or chat are intended to support interpersonal, one-to-one communication, focusing on individual topics (i.e., off-topic communication).
In contrast, we use the term common-identity communities to describe communities that are defined by a common topic or interest and support their members in performing a common task or attaining a common goal. These communities are usually based on Wiki or forum technology (e.g., Wikipedia2 or message boards of The Internet Movie Database IMDb3). Common artifacts are their main content. Community-set interaction opportunities such as collaborative writing or commenting are intended to support communication from one member to the group as a whole, focusing on a specific topic (i.e., on-topic communication). Although member profiles are rather secondary in common-identity communities, it is usual that during the registration process a member profile is created, including, for example, username and registration data.
The Importance of Community-Oriented Self-Presentation
When people become members of a community, usually the first step is to fill out a member profile, deciding to which extent they want to become visible to the group and which impression they want to convey. It is a key premise in self-presentation research that people want to be liked by their audience and want to get the audience to think favorably of themselves (Baumeister, 1982). Thereby, norm-consistent behavior is especially important for newcomers because behavior inconsistent with the norms of the group causes resistance (Hornsey, Grice, Jetten, Paulsen, & Callan, 2007). Consequently, people must select those pieces of information for their self-presentation that are relevant and appropriate in the given situation (Leary, 1993).
At present, the large and growing body of research on self-presentation in online communities (e.g., boyd & Ellision, 2007; Buffardi & Campbell, 2008; Gross & Acquisti, 2005; Stutzman, 2006; Utz & Krämer, 2009) has mainly focused on self-presentation in common-bond communities in the form of social networking sites. Extending this perspective, we not only build on research done in the context of social networking sites but also draw conclusions from previous social psychological research on self-presentation in common-identity communities.
Building on research about common-bond and common-identity groups, we inferred that the two types of online communities frame different interaction situations. Common-bond communities support interpersonal interactions among members. In contrast, common-identity communities support collective exchange on a common topic. Consequently, it might be that different goals become salient: In common-bond communities, members want to be accepted by specific others in order to get in interaction on a personal level. In common-identity communities, members want to be accepted by the community as a whole in order to be involved in topic-related exchange. These different goals might have consequences for self-presentation in common-bond and common-identity communities.
Self-presentation in Common-Bond Communities
On social networking sites, member profiles typically include information about age, gender, geographical location, interests, occupation, group affiliation, favorite music, books and movies, personal statements or an “About Me” section (boyd & Ellison, 2007; Stutzman, 2006). Additional features are a profile picture of the profile owner, a place for public messages where other members are allowed to write notes or comments, and a contact list with links to other members' profiles (boyd, 2007). The willingness to provide personal information in member profiles on social networking sites is generally high. Gross and Acquisti (2005) reported that 91% of a sample of college students using Facebook provided a personal picture, and 88% their birthday; the majority disclosed favorite music, books, or movies; 51% of profiles contained the current address of members; and even 40% of members revealed their phone number.
Exploring the reasons for the amount and kind of information people disclose in profiles, personality effects (Buffardi & Campbell, 2008; Krämer & Winter, 2008; Utz & Krämer, 2009) and privacy concerns (Acquisti & Gross, 2006; Dwyer, Hiltz, & Passerini, 2007; Lewis, Kaufman, & Christakis, 2008; Tufekci, 2008; Utz & Krämer, 2009) are the main factors discussed. Research about privacy concerns revealed sensitivity towards privacy affairs to be stronger among nonmembers than members. After the decision to join a social networking site, people even revealed information they were very concerned about being known by strangers (Acquisti & Gross, 2006). In fact, there is little or no relation between perceived Internet privacy concerns and amount of disclosure (Dwyer et al., 2007). However, people did not restrict self-disclosure itself, but adjusted the visibility of their profiles for others by changing their privacy settings to avoid undesired audiences (Tufekci, 2008; Utz & Krämer, 2009). Thereby, not only personal attitudes but also social norms play an important role. Lewis et al. (2008) showed that people were more likely to limit access to their profiles when their friends and roommates did so. To conclude, privacy concerns explain whether or not, or to whom people present themselves on social networking sites, but fail to predict how much and which kind of information they select.
Considering consequences of self-disclosure, Collins and Miller (1994) concluded in their meta-analysis that self-disclosure fosters liking. People who disclose personal information tend to be liked more than people who withhold them. This benefit of self-disclosure in common-bond communities was shown by Lampe, Ellison, and Steinfield (2007). They pointed out that especially the usage of profile fields such as hometown, school, on-campus address, and field of study predicts the number of friends in Facebook. Apparently, in common-bond communities, members gain acceptance through extensive self-presentation that facilitates establishing relationships between oneself and other members of the network. Withholding of personal information seems to be incompatible with the key motivations to join a social networking site (Debatin, Lovejoy, Horn, & Hughes, 2009; Ellison, Steinfield, & Lampe, 2007; Joinson, 2008; Lampe, Ellison, & Steinfield, 2006). In other words, it might be that the high amount of information disclosed on social networking sites is a response to the interaction situation that makes the goal of building and maintaining interpersonal relationships salient. Thus, when striving to become a member in a common-bond community, people might prefer presenting themselves in an individualizing and detailed manner that delivers a rich picture of their life and personality.
Self-Presentation in Common-Identity Communities
Existing research does not provide a clear picture of self-presentation strategies in common-identity communities. To date, we do not know which kind of self-presentation people adopt in these communities. Instead, we know a lot about the consequences of becoming visible or staying anonymous in groups, for being identified with them as well as for information exchange. Visibility can increase but also diminish social identification and information exchange. On the one hand, research subsumed under the Social Identity Model of Deindividuation Effects (SIDE; e.g., Lea, Spears, & de Groot, 2001; Reicher, Spears, & Postmes, 1995) has demonstrated that anonymity of group members is advantageous for increasing the salience of social identity, because anonymity hides interpersonal differences between group members. Further, Cress (2005) showed that members of groups in which members were anonymous exchanged more pieces of information than members of groups in which personal photographs of members were accessible. On the other hand, the type of visibility matters. While visibility in the sense of individuation fostered differentiation between group members and reduced the salience of social identity (Lee, 2007), visibility in the sense of shared characteristics of group members (e.g., visibility of sharing the gender category, Lea, Spears, & Watt, 2007) as well as visibility through uniform pictures of group members (Wodzicki, Schwämmlein, Cress, & Kimmerle, 2011) even promoted social identification and information exchange. Consequently, it is necessary to differentiate between individualizing visibility that makes differences between the self and other group members salient, and depersonalizing visibility that highlights similarities and typical characteristics shared with other members.
Consequently, staying anonymous or highlighting characteristics shared with other community members is beneficial for social identification. However, we do not know whether people make the decision to present themselves in these ways. Research about communication in common-bond versus common-identity groups suggests that members of common-identity groups restrict their communication to on-topic information (Back, 1951; Postmes et al., 2005). Therefore, it might be that members of common-identity communities also restrict their self-presentation to on-topic information in order to refer to the shared interest of the community. For becoming a member in a common-identity community, people might prefer presenting themselves in a depersonalizing and restricted manner that introduces themselves as a typical group member.
Overview of the Current Studies
In three studies, we examined the interplay between the type of community and personal goals in affecting self-presentation. In Study 1, we examined the influence of the type of community on self-presentation. We expected that people present off-topic as well as on-topic information when presenting themselves in a common-bond community, whereas people focus on on-topic information when presenting themselves in a common-identity community (Hypothesis 1). In Study 2, we explored underlying processes measuring the influence of type of community on salience of personal goals. We expected that being confronted with a common-bond community - compared to a common-identity community - enhances the salience of the personal goal to get in contact with members of the community (Hypothesis 2). In Study 3, we investigated how personal goals brought along into the setting, namely, the goal to get in contact and the goal to contribute to the common task, affect self-presentation. We expected that people present the most off-topic information in their profiles when their goal to get in contact with other members is met with a common-bond community, whereas people present the least off-topic information when their goal to contribute to the common task is met with a common-identity community (Hypothesis 3).
To test these hypotheses, participants in all three studies were randomly assigned to one of two fictitious online communities: either the common-bond community, Cooking Friends, or the common-identity community, e-Cooking. Both communities addressed the domain of cooking to ensure comparability. To disentangle the effect of the social interaction situation (resulting from the type of community) from affordances of the profile itself, the profile template was the same in both types of communities. We ran online studies to enhance the realism of the setting.
In this study, the effect of the type of community on self-presentation was explored. We expected that the type of community affects the amount and the kind of information provided in profiles. We created prototypes of common-bond and common-identity communities to simulate the process of self-presentation in member profiles. Based on the causes of attachment in common-bond and common-identity communities discussed by Ren et al. (2007), we used the features presented in Table 1 to design the welcome pages of the common-bond community Cooking Friends and the common-identity community e-Cooking. Both communities addressed the domain of cooking to ensure comparability.
Table 1. Design features of common-bond community Cooking Friends versus common-identity community e-Cooking
Finding friends for cooking together online
Building the largest online recipe database by sharing cooking experiences
Members online and cooking events to get in contact
Newest recipes and recipe forum to share experiences
You and others
People eating together
Food and ingredients
The common-bond community, Cooking Friends, was introduced as a community for finding friends to cooking with online (see screenshot, Figure 1). The welcome page provided cues that highlighted the interpersonal nature of the community and that illustrated possibilities to get in contact with other community members. A short description of the purpose of the community used the personal pronoun “you and others.” On the left column of the welcome page, activity previews consisted of a list of members online, a list of cooking events, and pictures of people eating together.
The common-identity community e-Cooking was introduced as a community for building the largest recipe database online by sharing cooking experiences (see screenshot, Figure 2). The welcome page provided cues that highlighted shared interests and that illustrated possibilities to share experiences. In the community description, the usage of the personal pronoun “we” pointed out the cohesion of the group as a whole. Activity previews consisted of lists of new recipes and forum topics as well as pictures of food and ingredients.
A total of 29 people (20 female) took part in the study. Sixteen participants were university students, ten were employed, and three did not specify their occupation. Participants' age was between 18 and 31 years (M = 25.00; SD = 2.77). The participants were experienced in Internet use: They indicated being actively online 2.68 hours per day on average (SD = 2.04). Except for one participant, all responded to be a member of one or more existing online communities. Participants were recruited mainly via mailing lists and via links posted in StudiVZ, one of the most popular online communities in Germany. As reward for participation, respondents were given the chance to win gift certificates by lottery.
We simulated the registration process for online communities. After preliminary information about the goal of the study and the procedure, participants were randomly assigned to one of the two communities: Either the welcome page of the common-bond community Cooking Friends or the welcome page of the common-identity community e-Cooking was presented. Additionally, we illustrated the community and its features in a short video-clip to ensure that participants pay carefully attention to the characteristics. Participants were then invited to register. In a first step, they were asked to enter username and password to frame the process as realistic as possible. After logging in, we presented the same profile template in both conditions. Participants were instructed to get an overview of the profile fields before they start to create a profile. Further, it was made explicit that it is not necessary to complete all profile fields. After profile completion, the study ended with a postquestionnaire.
The profile template included 19 profile fields that addressed on-topic information strongly related to the domain of cooking (e.g., favorite food or cooking experience) and 19 profile fields that addressed off-topic information not related to cooking at all (e.g., personality or interests). Apart from topic relatedness, on-topic and off-topic profile fields were created as analogous as possible (see Table 2). We applied both free response profile fields and drop down fields. In drop down fields, participants were given the possibility to choose one out of several predefined answering options, whereas in free response fields participants were free in content creation. In total, participants had the choice to include a maximum of 38 pieces of information in their profiles, among them 13 free response fields and 6 drop down fields with on-topic as well as off-topic information, respectively. The order of profile fields was randomized, resulting in an unstructured but determined list of on-topic and off-topic fields. Self-presentation was measured by the amount (number of profile fields) and kind of information (on-topic versus off-topic fields) provided in profiles.
Table 2. Percentage of profile fields used in CB and CI
Off-topic profile fields
On-topic profile fields
Note. Description of drop down profile fields includes examples of predefined answering options in brackets.
What I like about others:
Cooking do's: (freshly ground pepper)
The optimal spaghetti: (thin)
What I dislike about others:
(bad table manners)
Cooking don'ts: (convenience food)
I like to talk about:
This equipment is essential for my kitchen: (gas stove)
My motto for a successful evening: (eating alone is boring)
My motto as a cook: (hot food is good food)
The optimal quantity of noodles per person:
Date of birth:
Best cheese ever:
My favorite place to be:
Favorite food shop in town:
How much time I spend cooking per week:
Three adjectives that characterize me:
Three adjectives that characterize me as a cook:
When it has to be fast, I cook:
What I can cook best:
These films or books influenced me:
These cookbooks or television cooks influenced me:
That differentiates me from others:
That differentiates me from other cooks:
Cooking yummy food takes (in minutes):
What else I want to tell about me:
What else I want to tell about me as a cook:
This ingredient is essential:
My own motto for a successful evening:
I as a cook have my own motto:
The postquestionnaire included questions about participants' internet experience (number of hours actively online per day, number of memberships in existing online communities) as well as demographics (sex, age, and occupation).
Results and Discussion
We predicted that participants provide both off-topic as well as on-topic information when filling out a profile for a common-bond community (CB), while participants focus on on-topic information when filling out a profile for a common-identity community (CI) (Hypothesis 1). To test this prediction, we conducted a mixed-model ANOVA. The between-groups variable was type of community (CB versus CI). The within-group variable was kind of information (number of off-topic profile fields and number of on-topic profile fields). The ANOVA yielded the predicted interaction between type of community and kind of information, F(1, 27) = 7.17, p = .012, partial η2 = .21. As Figure 3 shows, participants assigned to CB provided just as much off-topic information (M = 11.50, SD = 4.01) as on-topic information (M = 12.36, SD = 3.82), F(1, 27) = .39, p = .540, partial η2 = .01, whereas participants assigned to CI provided less off-topic information (M = 4.47, SD = 5.17) than on-topic information (M = 10.47, SD = 4.96), F(1, 27) = 20.20, p < .001, partial η2 = .43.4 Type of profile field (free response versus drop down) did not influence this interaction5.
Table 2 presents which particular profile fields were filled out by participants of CB and CI. It shows that those off-topic profile fields that are most commonly used in social networking sites, for example, relationship status, birthday, and favorite films and books, were filled out by participants assigned to the CB, while they were left out by the vast majority of participants assigned to the CI. Further, only 7% of participants of the CI wished to enhance differentiation from others by filling out the off-topic profile field “That differentiates me from others” (compared to 43% in CB).
Study 1 confirmed Hypothesis 1. The type of community affected self-presentation. In common-bond communities based on interpersonal interaction community members strategically adapted by providing a lot of personal information. They provided very private information such as contact information (e.g., e-mail address), personality descriptions (e.g., three adjectives that characterize me), or interests and favorites (e.g., topics they like to talk about) to become visible as an individual with idiosyncratic characteristics, but they also provided information related to the topic of the community (e.g., cooking do's and don'ts or kitchen equipment they find important) to facilitate getting in interaction with other members. In common-identity communities that focus on collaborative exchange community members strategically adapted by limiting their self-presentation. Instead of highly individualizing self-presentation, they provided more or less depersonalizing information that was related to the common topic. Of course, this kind of depersonalization does not mean that members left out any personal information but that they selected information that had the potential to demonstrate those personal properties which they shared with the community. In short, members of common-identity communities presented themselves more as a representative of the group than as an individual. One possible process underlying this effect might be that different goals become salient when becoming a member of CI or CB. We thus examined the influence of type of community on goal salience in Study 2.
The aim of Study 2 was to explore the salience of goals after community presentation. We expected that differences in self-presentation shown in Study 1 were caused by personal goals made salient through community context. Participants in the common-bond community might have provided off-topic as well as on-topic information to present themselves because the community made interaction goals in general salient; explicitly, the goal to get in contact with other members on an interpersonal level and the goal to share and develop recipes through the shared artifact. In contrast, participants of the common-identity community might have focused on on-topic self-presentation because the community made only the goal to contribute to the collaborative task salient, while the goal to get in contact on an interpersonal level remained less important.
Fifty-four respondents completed the study, 35 female. Mean age of respondents was 28.76 years of age (SD = 9.29), ranging from 20 to 64. Thirty-four participants were university students, 17 were employed, and the remaining three participants indicated to be doing an apprenticeship. On average, respondents spend 2.65 hours online per day (SD = 1.94). Forty-four participants indicated to be a member of at least one existing online community. Participants were recruited via mailing lists and were compensated by having the chance to win gift certificates by lottery.
As with Study 1, participants were randomly assigned to one of the two communities (either to CB Cooking Friends or to CI e-Cooking), and a video clip introduced the respective community to them. Afterwards, participants were asked for their personal goals by joining the respective community: We asked for their goal to get to know the other members (contact goal) and for their goal to contribute to the community by sharing recipes (contribution goal). Then, in contrast to Study 1, participants did not fill out member profiles but started immediately with the post-questionnaire. Analogous to Study 1, the post-questionnaire included questions about Internet experience as well as demographics.
We developed two scales to assess goal orientation. Each scale consisted of four 7-point Likert statements ranging from 1 - do not agree at all to 7 - strongly agree. The scale to assess the contact goal comprised four items representing the intention to get in contact with other community members on a personal level and to make acquaintances:
It is important to me to make many acquaintances in this community.
I aim to get accepted by other members.
I am looking for like-minded others.
I intend to participate in many cooking events.
Cronbach's alpha reliability for this scale was .77, indicating acceptable internal consistency.
The scale to assess the contribution goal comprised four items representing the intention to contribute valuable content:
I want my recipes to be accepted by other community members.
I aim to create recipes together with other members.
I strive to contribute really good recipes.
I intend to improve the quality of the community through good recipes.
Cronbach's alpha reliability was .86, indicating satisfactory internal consistency.
Results and Discussion
For testing whether contact goals are more salient in CB than in CI, while contribution goals do not differentiate between types of community (Hypothesis 2), we conducted a mixed-model ANOVA with type of community (CB versus CI) as between-groups variable and type of goal (contact goal and contribution goal) as within variable. Analysis yielded a marginally significant interaction effect between type of community and type of goal, F(1, 52) = 3.31, p = .075, partial η2 = .06. To break down this interaction effect, pairwise comparisons were performed showing that agreement with contact goal was stronger within CB (M = 3.83; SD = 1.31) than within CI (M = 2.72; SD = .95), F(1, 52) = 12.00, p = .001, partial η2 = .19, whereas agreement with contribution goal did not differ between CB (M = 4.76; SD = 1.12) and CI (M = 4.39; SD = 1.88), F(1, 52) = .82, p = .369, partial η2 = .02 (see Figure 4).
In line with our prediction (Hypothesis 2), participants assigned to CB revealed a stronger contact goal than participants assigned to CI. Moreover, contribution goals were strong in both types of communities. Results suggest that the difference in self-presentation between CB and CI shown in Study 1 was driven by strong or weak contact goals made salient through interaction context: The context of CB made the goal to get in contact salient and encouraged people to present a lot of information, off-topic as well as on-topic. In contrast, the context of CI made the goal to get in contact fade into the background and encouraged people to focus on on-topic information when presenting themselves.
In Study 2, personal goals were measured after community presentation. However, goals are not only affected by contextual cues; they are also a reason to seek a specific context. Therefore, in Study 3, we asked participants for their goals before we presented the community and examined the effect of personal goals on self-presentation, focusing on the interplay between goals and type of community.
People pursuing a certain goal actively explore new environments and social situations that seem to be helpful to fulfill their needs (Amiel & Sargant, 2004). Study 3 investigated the influence of personal goals, set before people join an online community, and how personal goals and community-set interaction opportunities interact in affecting self-presentation. Therefore, we randomly assigned participants to either common-bond or common-identity community after they had mentioned their personal goal. We expected that people present the most off-topic information in their profiles when their goal to get in contact with other members is met with a common-bond community, whereas people present the least off-topic information when their goal to contribute to the common task is met with a common-identity community. On-topic self-presentation was not expected to vary between communities and goals. Going beyond Study 1 and Study 2, we explored under which conditions people feel a sense of commitment to the community, namely, whether they want to maintain membership or instead do not feel attracted.
Seventy-seven participants took part, 41 female. Mean age of participants was 25.22 years (SD = 2.53), ranging from 21 to 34. Among participants, 47 were university students, 23 employees, two participants indicated to be doing an apprenticeship and five did not specify their occupation. On average, participants indicated being online 2.28 hours per day (SD = 1.93) and 74 participants indicated to be a member of one or more online communities. Participants were mainly recruited via mailing lists and links posted in StudiVZ. As compensation participants were given the chance to win gift certificates.
Study 3 followed the same procedure as Study 1 with one exception: At the very beginning, before the particular community description was presented, participants were asked to indicate their personal goal. They had the choice to select between the contact goal to make new acquaintances and the contribution goal to share and develop experiences. After indicating their goal preference, participants were randomly assigned to either the common-bond community Cooking Friends or to the common-identity community e-Cooking and were asked to fill out member profiles. The profile template was exactly the same as in Study 1. Again, the study ended with postquestionnaire, including an additional scale to measure the perceived commitment to community.
To measure the personal goals of participants, we asked them to select one out of two predefined goals: “On the next page, the cooking community is presented to you. What is your main goal when considering your registration?” The answering option “to make new acquaintances” represented the contact goal and the answering option “to share and develop cooking experiences” represented the contribution goal.
The postquestionnaire started with three statements to assess perceived commitment to the community:
I would like to continue membership in this cooking community.
I feel closely connected to members of this cooking community.
I identify with this cooking community.
The scale ranged from 1 do not agree at all to 7 strongly agree. Cronbach's alpha reliability was acceptable at .87. Questions concerning participants' Internet experience and demographics concluded the postquestionnaire (see Study 1).
Results and Discussion
Participants were equally distributed between conditions, χ2(1) = .06, p = .814. Among participants who chose the contact goal, 15 had been assigned to CB and 16 to CI. Among participants who chose the contribution goal, 21 had been assigned to CB and 25 to CI.
It was hypothesized that participants fill out the most off-topic profile fields in condition “contact goal/CB” and least off-topic profile fields in condition “contribution goal/CI”, while the amount of on-topic profile fields filled out remains constant between conditions (Hypothesis 3). The 2 (type of goal) × 2 (type of community) ANOVA with kind of information provided in profiles (off-topic and on-topic) as repeated measures factor did not yield this predicted three-way interaction (Type of Goal × Type of Community × Kind of Information), F(1, 73) = .65, p = .424, partial η2 = .01. However, the two-way interaction between type of goal and kind of information yielded significance, F(1, 73) = 19.91, p < .001, partial η2 = .21. Pairwise comparison revealed that participants with contact goal provided equal amounts of on-topic information (M = 15.39, SD = 4.11) and off-topic information (M = 14.77, SD = 4.29), F(1, 73) = .73, p = .396, partial η2 =.01, whereas participants with contribution goal provided less off-topic information (M = 7.67, SD = 5.40) than on-topic information (M = 12.50, SD = 4.37), F(1, 73) = 64.89, p < .001, partial η2 =.47 (see Figure 5). Concerning the influence of type of community, only the main effect reached significance, F(1, 73) = 4.75, p = .032, partial η2 = .06. Participants assigned to CB provided more information in general (M = 26.31; SD = 9.09) than participants assigned to CI (M = 22.34; SD = 9.90).
Hypothesis 3 was not supported. Self-set goals dominated over the effect of type of community. Although the descriptive pattern of means shown in Figure 5 indicated that most profile fields were filled out in condition “contact goal/CB,” type of goal and type of community did not interact in affecting off-topic versus on-topic self-presentation. Instead, there was an interaction between type of goal and type of information. Community members with contact goal presented themselves in a detailed manner including off-topic and on-topic information to address like-minded others, while people with contribution goal focused on on-topic information, referring to the shared topic of the community. Nevertheless, there is evidence that at least participants with contact goal adapted sensitively to community context; Participants with contact goal provided more information in general when assigned to CB than CI (p = .030), while participants with contribution goal made no distinction between communities (p = .465). It has to be mentioned that Levine's Test indicated heterogeneous variances for off-topic, F(3, 73) = 12.29, p < .001, as well as on-topic information, F(3, 73) = 7.01, p < .001; with smaller variances in condition “contact goal/CB” (SDoff-topic = .74; SDon-topic = 1.62) than in remaining conditions (SDsoff-topic > 5.21; SDson-topic > 4.26). However, this pattern even supports that participants in condition “contact goal/CB” agreed strongly in disclosing a lot of off-topic and on-topic information.
To explore whether commitment to community depends on fit between personal goal and type of community, we ran a 2 (type of goal) × 2 (type of community) ANOVA with commitment as dependent variable. Analysis revealed that participants with a contact goal felt more committed to the community (M = 4.22; SD = .73) than participants with a contribution goal (M = 2.62; SD = 1.49), F(1, 73) = 30.04, p < .001, partial η2 = .29, independent of the type of community that was presented to them.6 This suggests that it was not the fit between personal goal and community but goal alone that affected commitment. A possible explanation lies in the procedure. All participants were asked to fill out a user profile. Thus, it might be that filling out a profile in itself was more adequate for participants with contact goal, whereas participants with contribution goal felt uncomfortable because either they did not like predefined profile fields or they did not want to become visible at all.
People actively manage their self-presentation in member profiles, adapting to different interaction settings. They modify their self-presentation depending on type of community and personal goals, even if profiles do not push them to do so. Study 1 demonstrated that the type of community itself guides self-presentation. Consistent with effects of group type on communication mentioned in the introduction, members of a common-identity community provided less off-topic information than members of a common-bond community. Concerning underlying processes, Study 2 supported the prediction that members of a common-identity community did not reveal the goal to get in contact. Therewith, results of Study 1 and Study 2 complement each other, suggesting that differences in self-presentation (Study 1) correspond with differences in goal salience (Study 2). By demonstrating the impact of personal goals on managing self-presentation, Study 3 underpinned the findings of Study 1 and 2. In Study 3, members with the goal of contributing to the common task provided less off-topic-information than member with the goal of getting in contact. Although the influence of community did not disappear, at least community members with contact goals adapted sensitively to the type of community showing higher self-disclosure in the common-bond than in the common-identity community, personal goals dominated the influence of type of community. That means that, on the one hand, if people do not bring along strong personal goals, the type of community guides self-presentation by prompting interaction goals; on the other hand, if people bring along strong self-set goals before they encounter an interaction situation, they shape their self-presentation primarily in accordance with their goals.
Finally, personal goals, whether brought along to the setting or prompted by community context, affect self-presentation. In situations in which people strive to get in contact, they show high self-disclosure, presenting themselves in a quite extensive and individualizing way; while, in situations in which they do not strive for interpersonal contact, they prefer a more focused, depersonalizing self-presentation, highlighting those aspects of the self that define themselves as group members.
Furthermore, Study 3 revealed that members with contribution goal were attracted to neither community, possibly because of the extensive profiles we provided. Consequently, extensive and individualizing profiles we know from social networking sites appear not to be suitable in situations in which people strive for information exchange.
Our studies investigated initial self-presentation when people entered a community for the first time. However, self-presentation is also a dynamic process that is highly influenced by audience responses and ongoing interactions. It is certainly possible that, even when people become member of a community aiming at sharing cooking experience, they might find friends and make acquaintances in the course of their interaction within the community. Or, vice versa, people who initially seek interpersonal contact might become engaged in collective exchange with subgroups within a network. This dynamic process of group formation and change is addressed by Postmes, Haslam, and Swaab's (2005) interactive model of identity formation, which describes the interplay of interpersonal attraction and social identification. Future research on online self-presentation should take these dynamics into account, for example, by examining differences between initial self-presentation and self-presentation in established interpersonal relationships and group memberships.
Implications for Research on Consequences of Self-Presentation
This research is in line with research about self-disclosure on social networking sites discussed in the beginning of this paper. Once again it shows how much information people are willing to disclose if they strive for interpersonal contact and the community supports interpersonal interactions. Through individualizing self-presentation people do exactly what is helpful to make acquaintances.
If people strive to contribute to a common task, in contrast, they present themselves as group members by highlighting the shared property of the community. That might have positive consequences for the individual as well as for the community as a whole. The individual might consolidate his or her social identification and enhance the chance to be accepted by the community (Klein, Spears, & Reicher, 2007). The community benefits because homogeneity fosters commitment and attraction (Lieberman, Wizlenberg, Golant, & Di Minno, 2005). Generally, people invest a lot of effort and engagement in support of the group, if they strongly identify with it (Chiu, Hsu, & Wang, 2006; De Cremer & Van Vugt, 1999; Karau & Williams, 1993).
However, findings also reveal the risks profiles have for commitment in common-identity communities. The effect that community members with the goal to contribute to the common task showed (independent of the kind of community) less commitment to the community than members with the goal to get in contact (Study 3) might be due to the fact that we offered a quite extensive profile template. Possibly, community members with contribution goal perceived the profiles to be inappropriate and, therefore, perceived the community as less suitable for reaching their goal. Thus, it appears to be that profiles themselves affect community perception, and that people use not only the community description but also the profile to evaluate whether the community is attractive for them or not.
Consequently, the perfect match of personal goals, community-set interaction opportunities and profile design might have the potential to enhance members' satisfaction and engagement, while the mismatch involves the risks of dissatisfaction followed by shifting to another community, or even worse, of disturbing the interaction of members. To exploit the full potential of profiles, future research is necessary that investigates the appropriateness of more or less individualizing profiles in different types of communities - in experimental as well as real-world settings. Thereby, the effect of different profile templates on satisfaction with self-presentation itself and on commitment, as well as participation, can be considered. Further, it is an open question whether community members who strive for collaborative exchange and social identification with the group as a whole want to become visible at all. Although the profile templates we offered did not push a specific kind of self-presentation, they still suggested disclosing at least a minimum of information.
Function of Profiles and Profile Design
In common-bond communities, the function of profiles is to facilitate establishing and maintaining contacts, whether searching for acquaintances known offline or looking for new contacts. Consequently, self-presentation through member profiles aims at fostering interpersonal attraction and interaction. For this reason, it is important to make it possible to provide not only contact information and information about members' location but also information related to specific topics and members' interests which might enhance the chance of bringing people together.
In contrast, self-presentation in common-identity communities aims at connecting members to the group and fostering collaborative exchange. If profiles represent group members instead of individuals, self-presentation supports the perception of homogeneity of the community and thereby commitment. Consequently, we recommend profiles that provide opportunities to demonstrate similarities among community members and guide members to address the group as a whole instead of particular individuals. Further, to avoid individualizing self-presentation through personal information that is irrelevant for the community, it might be useful to stimulate reflection about what is important and interesting to be known for the community. Providing information about the importance of one's information for other members was shown to enhance the quality of information exchange in laboratory setting (Cress, Kimmerle, & Hesse, 2006). Likewise, field studies in the context of online communities demonstrated the positive consequences of, on the one hand, homogeneity of the group, and on the other hand, feedback concerning uniqueness of contribution for participation (Cosley, Ludford, & Terveen, 2003; Ling et al., 2005; Ludford, Cosley, Frankowski, & Terveen, 2004; Rashid et al., 2006). It might appear to be difficult to balance members' similarity as well as uniqueness in member profiles, but in the end, profiles should enable to transport those characteristics members share with the group (e.g., shared demographic characteristics, interest, attitude, value, or taste). Differentiation should only be supported in the sense of complementarity. Disclosure of special task- or topic-based skills, for example, demonstrates in what way members are beneficial for the group - what might foster perceived instrumentality of members' contributions. In short, common-identity communities require profiles that allow members to present themselves as valuable group members. If profiles deliver this message, self-presentation might foster commitment and information exchange at the same time.
In our studies, we offered the opportunity to select those profile fields that focused on the shared interest of the community. Another possibility to turn members' attention to the group is a so-called group profile page that represents a group of community members and show the shared taste and activity of the group. Ren et al. (Working Paper) investigated the effect of group profile pages versus individual profiles and concluded that group profile pages, combined with repeated information about activities within the group, increased commitment and participation. Similar to topic-based profiles, group profile pages make the group and its goal salient. Fostering the idea of member profiles instead of user profiles, there is room to develop new profile features that support collaborative exchange in common-identity groups.
Additionally, visualizations that capture communication and activity within a community (so called data portraits, Donath, Dragulescu, Zinman, Viégas, & Xiong, 2010) might be useful to support identification and participation. PeopleGarden (Xiong & Donath, 1999), for example, depicts members of a discussion group as flowers. The height of a flower and number of its petals represent how long and often a member contributes. Thereby, the activity of a particular group member and his or her role within the community becomes visible. The accumulated content of member profiles is also suitable to show what constitutes the group, for example, through visualizations of frequently used words describing members' interests and expertise. Through accumulation of the content of member profiles, self-presentation is able to support group formation.
This paper provides insight into reasons for self-presentation beyond inter-individual differences. Our findings illustrate why privacy concerns fail to explain self-disclosure on social networking sites. Relying on high self-disclosure, members of common-bond groups do what is functional to reach their goal in the given interaction situation. That is also the case in common-identity communities. There, members prefer to highlight shared characteristics, a useful strategy for becoming accepted and reaching shared goals. Consequently, differences in self-presentation between communities are not simply an artificial effect of different opportunities to present oneself through predefined profile fields. Instead, people actively select those pieces of information that help to fulfill their goals.
Both authors contributed equally in conducting the presented research. We thank two anonymous reviewers for their helpful comments on the first version of the manuscript and we thank Klea Simon and the participants in experimental research courses at the University of Tübingen for their assistance in material development and data collection.
To control for gender effects, in each study, we run an ANOVA with gender as independent factor and kind of information provided in profiles (Study 1 and 3) or type of goal (Study 2) as repeated measures factor. Across all three studies, we did not found an effect of gender (all p-values > .423, all partial η2 < .02). We think that the missing effect on information provided might be due to the fact that the used profile fields did not allow for very intimate self-disclosure as it has been usually considered in self-disclosure research (e.g., the self-disclosure questionnaire of Jourard & Lasakow, 1958), in which gender effects have been found.
To control for the effect of type of profile field, a repeated-measures ANOVA was computed with type of community as between-groups variable and kind of information as well as type of profile field as within-group variables. Three-way interaction was not significant, F(1, 27) = .08, p = .785, partial η2 = .00.
Levine's Test indicated heterogeneous variances with smaller variances for conditions with contact goal (SD < .85) than contribution goal (SD > 1.42), F(3, 73) = 5.99, p = .001. Because the biggest variances are in conditions with most participants, the analysis is more conservative in testing the hypothesis (see Field, 2005).
Eva Schwämmlein is Ph.D. student at the Leibniz Graduate School for Knowledge Media Research. Her work focuses social aspects of self-presentation in online communities.
Katrin Wodzicki, Ph.D., is researcher at the Knowledge Media Research Center. Her work focuses on social psychological and motivational aspects of knowledge exchange, especially on social networking sites.