Ren, Yuqing, & Kraut, Robert, & Kiesler, Sara (2007). Applying Common Identity and Bond Theory to
Design of Online Communities. Organizational Studies Vol. 28, No. 3. pp.377-408.
Submitted by John Blair on Tue, 02/19/2008 - 22:34.
0
points
John Blair
First, let me say I liked the reading and the work. It’s nice to have suggestions offered to resolve / test some of the points raised. However, I’m constantly amazed at how much effort is expended on attempting to direct human behavior through various social models (psychology, sociology, economic – page 402). Obviously, there’s an enormous carrot at the end of the stick for being successful in doing so, but so much of this seems so logical. Of course community design is going to determine how members interact – or if they interact. Offering functionality that members have asked for, commented on using at other sites, or that supports the purpose of the site is naturally going to please those members who inquired about it. Just like it will upset them if nothing is done or the implementation is poor.
I’m annoyed at the notion that people are like cattle and can be herded, directed, cajoled, and convinced to perform in an expected manner. The fact that this has been proven countless times is a never ending source of mystery to me. All I can say is that at some point the cattle always stampede and it’s almost always beyond the control of the cowpeople (PC term).
Community design should support community purpose. If you want your community to be identity based, then you should provide functionality that supports that (more group activities) and likewise for bond based communities (more functionality that allows interpersonal exchange). For those communities that have an integrated identity / bond mixture, this is more tricky but as the authors suggest try to “group participants into clusters with similar backgrounds or needs” page 397. Doesn’t this seem like an obvious thing to do if you trying to support both types of attachments?
The one area I can see as a problem for these dual communities is if they attempted to be everything to everybody. But that again comes down to the purpose / focus / mission of the community, not the design. That is an unrealistic goal and is doomed to failure – or severe splintering into sub groups as people found each other and their interests and went off to support these commonalties.
Maybe I’m making this too simple – or perhaps I’m just a simple social loafer whose touch to this tangent of reality is slipping.
Largely taking the form of a literature review, these authors treat online communities as social-technical systems to test theories that communities favor either identity-based or bond-based attachment (400). Simply put, identity-based attachment is a commitment to the overall group; whereas bond-based attachment to individuals within a group. Although the authors offer that users may approach sites for mixed uses (both identity- and bond-based uses) they overall seem to believe that the way sites are designed tends to lead users overwhelmingly in one direction or another. The authors posit, "we believe a group's design can increase the likelihood of mainly bond-based or identity-based attachment to the group" (381).
In an effort to allow community designers to either analyze their communities in progress, or design from the ground up, the authors provide interesting taxonomies of both types of sites:
3 main causes of group identity/commitment to group
Social categorization: collection of people belonging to same social category
Interdependence: cooperating through a joint task, etc.
Intergroup comparisions: comparing out-groups to in-groups
3 main causes of bond-based attachment to group
Social interaction: build trust through relationships
Personal information: opportunities for self-disclosure
Personal attraction through similarity: liking others with similiar preferences
The authors give an incredibly detailed list of examples and corrections to be made for sites that are self-analyzing to see which type of commitment is most beneficial. For our class this week, though, the discussion beginning on p. 390 and going on until the end of the article is of particular relevance. This is where they begin discussing newcomers and other types of users that may help/hinder community commitment. My big takeway was that in addition to the Arguello piece from this week, this reading also seems to advocate using welcoming messages, mentors, moderators, or FAQs to guide newbies into a community (395), this seems like a good idea but also an advantage to old-timers too so that these small design changes may actually encourage old-timers not to get frustrated and so stick around longer.
I really like the bond and identity concepts, but a number of the communities which I've been involved in seem to have social attractors somewhere inbetween "pure bond" or "pure identity".
PerlMonks I would classify as "identity heavy" for the following reasons:
- very little in the way of off topic discussion (known on the site as GTKY or getting to know you)
- very weak direct reciprocity (if someone answers my question, I am not likely to do the same for them)
- welcoming to newcomers (if they're polite)
- shares many functions similar to the open source usenet groups studied by lakhani and von hippel (cited as an identity community)
However, I would say that among regulars, and perhaps even more so in certain geographical and administrative subgroups there is heavy bond attraction -- Members know each other personally, build strong personal and professional relationships, and communicate with each other off the site ie the PerlMonks group on LJ.
It seems that a healthy community may exhibit both bond and identity attributes, albeit at different scales or in different dimensions. I think in some cases where an online community discourages bonding behavior, people may use external channels to form those kind of bonds once they are introduced. I think an interesting design challenge for community maintainers to encourage both kinds of group formation without getting too much "chocolate in the peanut butter" as it were.
I really wonder how the emergence of social networks relate to this distinction between bond and identity theory.
I guess it is really obvious that social networks support common bonds between individual much more than common identity. Does that have an influence on other types of communities? Are for example discussion boards more identity focused since everybody is on Facebook? We have also been talking about the groups within social networks (again Facebook as the example). Maybe the discussions in many Facebook groups don't take off because there is no need for any conversation related to individual bonds, because that is already taken care of by profiles and friend connections?
If that is true it would be another argument that a successful community should try to maintain a balanced levels between the two types.
Submitted by Matt Adamo on Thu, 02/21/2008 - 14:49.
2
points
The way community memberships interact and influence each other is an interesting question, and one the authors bring up in their discussion of theoretical issues. Do people get their "bond fix" from some communities and their "identity fix" from others? Or, do they seek out communities that have a similar balance between bond and identity focus? A related question might be how people's behavior compares across overlapping community memberships. For instance, are lurkers always lurkers? If you look at your own community memberships and behavior, do you see any patterns?
I'll admit that without significant incentive, I'm a chronic lurker. Perhaps not surprisingly then, I also tend to participate minimally in bond-based communities like Facebook (I haven't gotten around to changing my profile/picture since I joined in 2003).
Submitted by LizBlankenship on Thu, 02/21/2008 - 22:19.
0
points
I like your question about lurkers, and getting bond/identity fixes in different locations. In response, I don't think lurkers are always lurkers. Lurking can be a way for learning - as discussed in one of the community of practice readings we've had, a person might lurk and then go become the leader for a new similar community. I think we probably each do a mostly unconscious cost benefit analysis before deciding whether to be a lurker or more of a contributer. In different communities, we have differing levels of expertise, and fulfill differing levels of exepctations from our peers. We certainly consider different communities to be more social than identity (for example, the co-op community I live in is primarily social, although we do share in common that we believe in democracy and sharing), and some to be more identity-based, like the Perl Monks community John talked about.
I, too, am a chronic lurker and have left more than one bond-based extensively networked profile behind (myspace, friendster, zaadz). When I left myspace, it was admittedly because of a desire for less bonding....One too many awkward re-kindled in-an-unauthetic "I don't really want to hang out but we used to hang out in second grade and I kinda want to spy on you" kind of way......
A large handful of my friends deleted their profiles that year. Most of them had more than 200 friends on their accounts and left without warning or with a brief note about not being able to take it anymore. I can''t say what was going on for them, but for me it was a sense of over-exposure and over-obligation to always be checking it......
Friendster was the first of these to go. For me, Friendster lost me when they added the "see whose been checking your profile" feature. There is a point at which transparency and connectedness is an ever increasingly irritating over-reality at times.
I typically recover after each of these acts of profile-destruction, and succumb to the latest bond-based social networking site invite sent my way. I think the most useful notion I pulled from this article was the idea that theorists aren't necessarily good designers.
I agree, and think its a curious predicament that accounts for much of the research-practice divide in many fields.
Submitted by Daniel Zhou on Mon, 02/25/2008 - 13:29.
1
point
I think you have made some good points Lisa. People just have different preferences. From my Facebook experience, I can see that some people prefer more privacy and maintain close relationship with a small number of friends, whereas other people prefer less privacy and have relatively loose relations with tons of friends. If a community only has a single design for all users, it just cannot meet the expectations of everyone.
I think an obvious move is to go to "personalization". It might be achieved by using the social computing or web2.0 technologyies.
Facebook groups are funny because they seem like another way of signalling a person's interests on their profile instead of actually acting as a kind of membership to something. At least for the way I use groups, it's usually to show my membership in something, but not necessarily in the group itself. For example, I'm in the "Squirrel Club" group, but I don't interact with other squirrel clubbers (hmm bad use of words?) in that group. I'm just using it to show people who look at my profile that I'm in the squirrel club.
So in this case, I think the groups are more identity based than bond based.
I would put Social Networks in Identity based communities because most of these sites focus on people interaction (and not on a certain topic that binds them). Yes, we have friends in facebook and that we might have the bond of "friendship" amongst us, but that is not the reason why we communicate (evident from the content of discussion in SN). We use SNs for keeping in touch with friends (interested in individuals) and therefore it seems that they are identity based communities.
With the authors doing such a good job listing and explaining the things that they didn't address or go into detail on, I find myself having to agree with most of what has been said. The primary take away that I got from this reading is that there is not necessarily a magic method that could easily equip an online community with the necessary characteristics that support either identity or bond-based user preferences. As the reading explains, there is just too much uncertainty when communities evolve, users come and go, and bonds form amongst regulars in a largely common-identity community. As a result, the best developers and site-leaders can do is to have a basic idea of what type of community they are interested in creating, then adjust and add/take away community features as needed. While all of the online communities that I have been involved in seem to fall into in between the two extreme ends of the identity-bond spectrum, I can definitely sense the conflict of interest between the two theories.
For example, consider IGN, a pretty popular gaming site/portal. Most games (popular ones, at least) have their own corresponding forums that act as a small community. Some of these forums have sets of regulars and several casual or new users (in many cases, there is also a lot of cross membership. My role on the IGN forums has always been a "lurker". While I personally don't interact with anyone on the forums, for the games that I am interested in, I always enjoy observing the inevitable "hype to fanboy flame wars" trends. Ideally, it would seem that on such a big public portal that consists of such a significant portion of "new users", people ought to be "identity heavy" (to borrow Nate's term). In fact, as the authors have pointed out, the site owners might even be interested in keeping the forums that way so new users aren't turned off.
However, from what I see, there is a lot of bond-based things going on. For example, regulars would post irrelevant threads titled "To XYZ-USER: Read This" or "I GOT THE GAME. WOWZ!!!1", which as one would expect, attracts a lot of complaints from new users, who question the relevance of such threads. In response, the original poster and his supporters or haters jump back in and express that they have the earned right to create such threads because they were on the forums since "day 1". Suffice to say, such interactions do indeed turn off new users and dilute the quality of the conversations. Personally, I have been a lurker for a long time, and I have no desire to mingle with some of these users.
I really like the subtle design difference Ren et al point out in terms of welcoming newcomers. The way newcomers are welcomed into a community can have a big impact on whether those newcomers will feel a bond-based or identity-based attachment to the community.
Design claim: Welcoming newcomers by helping them navigate information traffic, understand norms, engage in community conversation, and wait/observe before participating will encourage identity-based attachment in your community.
Design claim: Welcoming newcomers by helping them connect with other members, join interactions, and form relationships with a subgroup in the community will encourage bond-based attachment in your community.
Of course, these claims include the caveat that designers have limits in their abilities to shape member connections and community interactions.
I enjoyed this reading for its straight-forward presentation and interesting material, but I believe that the importance of bond-based attachment is understated.
While my understanding is that both identity- and bond-allegiances are necessary in order to create a lasting community, I place much more weight with bonds. For the average user, a specific topic is only going to be interesting for so long. Eventually you'll reach the top level of your MUD, earn your MBA and high paying job, get tired of swapping graphic design tips, forsake Perl for Python, or run out of interest in SimCity. What then? If the community has any hope of retaining you, it had better bank on some solid relationships built between you and the other members. Why would they want to retain you? Despite your waning interest, if you've been a regular for a significant amount of time you've likely developed valuable expertise. You're still an asset.
I agree with the authors when they suggest that members frequently go from identity-based to bond-based participation. In fact, I will go a step further and argue that, as suggested by the above paragraph, such a transition is inherent to any e-community. If you're building for longevity, anticipate supporting bond-based interactions.
I do particular favor this paper maybe because it gives a comprehensive literature review for building these basic senses on people's social relationships relevant to online communities. Since it is kind of general review, so there is no much to argue with. But as talked in above posts, i will say that the tension between identity or bond is pretty subtle and complicated. And of course, the two are often interwoven. People tend to be together with some others because they share something common and thus they fall in some sink as community being built; and they develop group identity and comparison with outside along their interacting process.
And I feel like that both are all needed in creating and maintaining health community: identity supplies belongingness and commonness, while bond intrigues substantial and repeated interactions over time to even enhance the community.
And I feel it is hard to manipulate a community to be more identity based or bond based, especially for online ones, where people come voluntarily. I guess it almost about the theme/topic of community, which mainly decides what the community would be like. I assume there are some attributes within the theme/topic decide this. For example, if the community is about pet or politics then it is very likely the pet community will be more of bond based and the politic one will be more of identity based. That is because the nature of topics about pet, people are intending to self-disclosure, e.g., to talk about their won pets, then all the interactions would have a strong context of the people participating in the community and bonds tend to be built; but for the political one, people would more focus on some target like Obama, good or bad, without much self-disclosure and opportunity to built interpersonal relationships.
I second that! I really enjoyed this reading I felt that it had more substance than some of the pervious papers. I hope that we continue to look at articles that have a stronger foundation in academia.
Submitted by Tracy Liu on Thu, 02/21/2008 - 19:40.
1
point
This paper study the reason of producing group identity and common bond, the effect of these two features on enhancing individuals attachment to groups and the implication of online community design.
The reasons of developing group identity are social categorization, in-group interdependence and intergroup comparison while they are social interaction, personal information and interpersonal similarity in bond theory. However, they are overlapped with each other. For instance, when people are categorized into one group, they do share similarity with each other, right?
Although these two theories are quite different in the definition, there is no systematic method to differentiate them, at least from this paper in my opinion.
I agree that individuals' attachment to certain community is dynamic and change from group identity to bond or reversal, but I also believe that these two features exist in one community together. For instance, in an online fans-club, there is certain discussion section for members to discuss their idols' new drama, which is based on topic/conversation, however, there is also chartroom for them to maintain individual relationship and lots of off-topic conversations are born here.
I don't mean to be writing all of these "me too" posts but I have to say that I agree with you. While I was writing my blog entry I felt that the bond and the identity were often inter-dependent. For instance, in my community you can form or join groups based on your interest. This could be seen as a type of identity based on intergroup comparison. But, people are likely to join groups that cater to their interests, which would imply a bond by similarity.
Both of these factors are going to influence each other as the member continues to interact as part of the group too. Over time the member will likely grow to identify more strongly with the group and their interests will begin to reflect the other group members’ interests as the continue to interact with them. These influence are only going to make it harder to tease the two factors apart.
I have always imagined that Paradise will be a kind of library.
One of Ren’s design claim was that to encourage bond-based communities the site administrators should phrase their mission statements to encourage members to engage in and to tolerate conversations on wide-ranging topics. The design of Newsvine seems to be encouraging this behavior and thus bond-based attachments (Ren et. al 2007). I have mentioned before that it seems that Newsvine users are more respectful in their disagreements than other communities I have seen. I wonder if bond-based attachment to a community discourages contention more than identity-based attachment. Anyone else care to speculate?
Ren et al. explained two theories of group attachment: identity and bond theory. When I look at the figure on page 393, I feel identity-based attachment seems more desirable then bond-based ones in terms of the consequences of attachment.
Instead of selecting one from these two types of group attachments, I prefer their idea on P.381. Ren et al. state that a community can have these two dimensions of members' attachment. Long-term members stay in the community because of their strong interpersonal ties, and new members join the community because of their common identity. As they highlight in the end of the reading, I would like to know how an individual's attachment transform from one type to another, and if both two change directions (identity to bond, or bond to identity) can work.
Through this week’s reading we have focused on key variances
between IBA (Identity based Attachment) groups (i.e. Sierra club, NRA (National
rifle Association)) and BBA (Bond Based Attachment) groups (i.e. bridge or
poker game group with friends) I couldn’t help but think of Facebook and Myspace.
In next month’s Fast Company magazine
there is an article on why marketers who use social networking sites may be
pitching into the void or, simply put, why marketers who use social networking
sites thinking they are targeting a specific market may be missing the point altogether. Essentially, the author makes a case for why “Tom”,
the Myspace guy, is not the age that he says he is. Actually, Tom is 5 years older than his
profile claims he is. And just like Tom,
according to the author, we all present our best “selves” within these social
networking sites. So, marketers who
think they have a lockdown on a target market have anything but authentic data
to test drive their products. The article ends by stating “ Social networkers
don’t sign on to expose their true consumer selves. We sign on to promote our own agendas”
So based on the above, then sites such as Facebook/Myspace/Friendster,etc.
are centered around identity based groups who focus on achieving a purpose, to
promote their own agenda, knowingly or unknowingly, consciously or
unconsciously? Part of me thinks this is
a bit of a stretch but perhaps this is more salient than I want to
believe. This is after all a recent
science despite the fact that there are correlations between studies done in
the 50’s (Back study in ’51, pp.389) finding that identity based groups competed
tasks efficiently and discussed matters that were only relevant to their
goals/aims, and more recent studies (Sassenberg in ’02, pp.389) finding individuals
engaging in discussions on a narrow range of topics reported high group
identity and high valuation. In
combination, these studies suggest that online bond based communities are more
likely to engage in off topic discussion and be more tolerant of off topic
discussion (pp.389)
So, if we combine the above two, what conclusion can we
make?
Submitted by Paul Resnick on Mon, 03/03/2008 - 07:09.
0
points
I don't understand where your assertion that FB/Myspace/Friendster are centered around identity based groups is coming from. I would think (and the authors assert) that these are the quintessential bonds-oriented sites.
There are some features that are identity based (signal your affiliation with a band or with a "group" that doesn't actually interact). But the basic feature is linking with individual others, not identifying with groups.
Feel free to correct me if I'm showing my ignorance of how these sites work in practice. As a very occasional user, my mental model may be wrong.
Submitted by Rozaidi Rashid on Thu, 02/21/2008 - 23:01.
1
point
The high school alumni online community I am investigating shows a high level of commitment among members (based on volume of message and number of returning and participating members) because:
1a) We belong to the same social category of school alumni, and of the same age, and ethnic group (mostly).
1b) We rely on each other to get things done, such as organizing trips, activities -- each of us is independent in our own task, but interdepend on others to get the big task done.
1c) We always pride ourselves of being associated with a special school system for the underprivileged, who does as well in the adult life as those who went to the elite schools in the country. But intra-group comparisons is limited - there doesn't seem to be comparisons among professions.
Members of this community also have a strong bond to the grouping, because:
2a) the relationship was built over 3-5 years at least, and now in the 20th year, and so, the social interaction is at a unique level compared to other groups that we are a member of.
2b) Being holed-up for 5 years in an isolated part of the country really did its job to make us known to others at a very personal level.
2c) Cliques and groups develop then and even now based on similiarities in interests or profession.
1) "If designers discourage off-topic discussion, they might lose people who would
like to talk with others like themselves." p4
What happens when designers introduce an off-topic board? Do established members forget about the original focus because they're too busy discussing their dogs? Is the problem maintenance?
2) Identity-based attachments seem to be centered on some notion of shared values/interests and indirect conversation, while bond-based attachments take this one step further by focusing on activities shared with other people and direct conversation/awareness. Pp5-13
If personal information sometimes promotes interpersonal bonds among people who haven't interacted, could this also extend to an individual's bond with another community he hasn't interacted with? I like architecture but don't have any personal connections to the community. Still I feel somehow connected when I read news and casually browse information.
3) "we suggest that people with identity-based attachment to an online community may be more likely to take over responsibilities from lurkers or slackers and compensate for their lack of contribution."
What if an individual isn't proud of his group membership? If an obese man who belongs to Jenny Craig considers himself to be a Jenny Craig member (identity) instead of classmates with people who belong to Jenny Craig (bond), would he be more or less likely to pay for a broken window?
By the way, I really like this article. If I don't force myself to find disagreements, I just don't know what to post other than "I like this article and I find it interesting", which won't be useful or interesting for anyone else.
Submitted by Paul Resnick on Mon, 03/03/2008 - 07:14.
0
points
It would be interesting to know whether the same effects (willingness to take up slack and compensate for lack of contribution; proclivity to sanction non-normative behavior) will be seen for identity-based attachment to groups that are lower status, groups where one might not want to proclaim membership with people outside the group.
My intuition is that you would still get the same effects. I don't whether there has been research testing this (my guess is that there has been).
Some communities can be extreme in being bond or identity based.
In technical support forums private message are highly discouraged. Sub communities are formed around topics related to community. Personal information is almost never revealed. Such features force only people who are attached to the identity to stay in the community. Similarly, friends network sites like Facebook, MySpace have features that are targeted towards personal and direct communication. The emphasis is clearly on establishing bonds. Users can form identity based communities within these sites, but they are rare and don't last very long.
Over a period of time identity based communities should pick-up some traits of bond based community. A veteran member will be staying in the community due to a strong attachment to the community identity, but repeated interaction with another user(s) will create a bond between them. This would lead to off line communication as private messages, IRC or real world conferences. Bonding with other users can happen even when personal information is not revealed. This happens when a user gains the respect of the community for his/her contribution but chose not to reveal any personal information.
However I'm not sure if bond based communities can drift towards getting a identity for the group.
Submitted by Satyendra on Fri, 02/22/2008 - 00:27.
2
points
Seems like based on the fundamental structure of the two types of bonds - depending on whether a community shows a greater common identity focus or a greater bond identity focus - macro or micro level level features of the group will become important.
To pick up from last week's topic for an example, suppose a group of friends joining a contract bridge club join it mostly for their friends - to interact and communicate with them. This kind of a group shows a stronger bond identity focus and a weaker group focus and it is more likely that regulation for example
will be more largely controlled by architectural decisions that impact micro interactions and interactions among small groups rather than a strong common norm.
On the other hand a group with a strong common identity like a group of Linux enthusiasts are more likely to be regulated by a strong social norm and common beliefs that bind them. This can help us make the design choice between the various regulation mechanisms based on things we may know about the expected group type.
There will also be switches from what bond one identifies with over time in a community. As a newcomer to a community a user might be drawn to it by the ideals and vision it stands for and signifies for him. Once a part of the community however, as he gets to interact with people there he may form strong relationships with them and then the site has a different or at least broader signifiance for him. This could lead us to design choices that target specific users at specific points of interactions in the community. For example for newcomers we could provide information that is more targeted towards explaining the norms and the values of the community, while for the members the focus could be to initiate discussions and enable them to form stronger bonds between themselves.
Submitted by Paul Resnick on Mon, 03/03/2008 - 07:16.
0
points
I agree that the way to resolve some of the tension is to support identity-based attachment initially and gradually move toward encouraging and supporting bond-based attachment.
The study identified factors a the group, individual and message levels that impact community responsiveness and individual commitment, which according to the author, are two key elements of success of a community. First, I should agree that conversation is the basic mechanism in online groups and the viability of a community also depends on the willingness of individuals to stick with the groups over time. To take a look at the study results, some seem interesting.
1. Group identity. It is claimed that bond-based groups have higher member commitment than topic based groups. I tend to agree with this claim as friendship is always a strong tie between members. In terms of the group I am studying, it is more like topic based. 2. Cross posting. Nowadays it seems cross posting is very common in a lot of online communities. The study in the paper shows no distinct result whether this factor is significant or not. Personally, I would think sometimes cross posting helps navigation while taking up more resources. 3. Group size and volume. I don’t actually agree that people tend to be less committed to larger groups and contribute less. The group I am studying has over million members, majority of whom demonstrate high commitment and significant contribution to the community. So size only is not a factor that would affect member’s involvement and commitment. 4. All individual factors are interesting, but I do not see much practical tips that could be drawn from it, such as newcomer status. I somewhat agree that the history of an individual affects the interactions, but it is still a question to me how site managers could apply this to their design.
Overall, the study is still interesting enough to take a look, at least, it tells that there are a number of factors affecting the individual group interaction, which is the key success factor to an online community.
Submitted by Paul Resnick on Mon, 03/03/2008 - 07:18.
0
points
I didn't see where the paper argued that bond-based groups have higher member commitment.
Also, I'm skeptical when you say your community has more than 1 million members, most of whom demonstrate high commitment and significant contribution to the community. Is it really one community?
Lots of interesting questions have been raised, and great summaries given so far, so let me just add a couple of my thoughts on top of it all.
First, this was probably my favorite reading to date - it was an academic article, but gave clear design claim related to how to design a community for identity-based or bond-based attachments.
It seems that a lot of people feel that the two are interdependent, or even both always present together in a community. I tend to agree with the authors, however, that in most communities, one will be the primary attachment type. Yes, identity-based attachment can morph into bond-based attachment in communities where members grow closer over time, but still in most cases, the design decisions the community builders make will determine what type of attachment occurs.
The author claims that if the community builders choose:
to limit the size of a community
cluster users into communities of interest
require users to register
allow off-topic disscusion
Then this will cause bond-based attachment.
On the other hand, if they choose:
unlimited growth
provide unstructured access to all content
allow to participate in the community anonymously
limit off-topic discussion
Then this will cause identity-based attachment.
So while the two types are not completely separate, the authors did acknowledge that it is very hard to find a balance between the two, since the decisions tend to favor one over the other.
On another, separate note, this reminds me of the discussion we had last week in classa about the optimal size of a community. Some were arguing vehemently that growth was always good, while some were saying it had major drawbacks. I think it would be interesting to see if people who were arguing for unlimited growth are studying identity-based communities this semester, while the others were studying bond-based ones. It seems that whatever community we're studying is what is framing our thoughts, instead of all communities in general.
Submitted by Sean Munson on Thu, 02/28/2008 - 02:13.
1
point
The authors describe the difference between bonds and identity in a way that I found very helpful. While we've talked before about why it would matter whether other members or the group are more important for user, this paper offers useful vocabulary for talking about the differences. The observation that some types of groups are more likely to have strong feelings of common bond and others are more likely to have feelings of common identity (just based on their reason for being together!) seems important, and I was surprised at how some of the groups were categorized.
One area where I had a bit of trouble was with figure 1a. It may be a minor quibble, but I felt that the framework would have benefited from making a feedback loop explicit; many of the consequences also seem to be antecedents of certain types of attachment / both types of attachment.
A second point is that while the authors talk about some communities shifting from identity-based to promoting interpersonal connections (p 401), I think some of these spaces (particularly Flickr) stand out because they are able to simultaneously provide outstanding support for both dimensions the authors discuss (common bond and common identity), not only for users who are excited about both dimensions but also for users who want mostly one or the other type of community.
The authors also note (also p 401) that they are unable to find many examples of communities moving from bonding to identity. In my view, bonds may have more of a role to play in getting people involved in identity communities. Returning to the Flickr example, I got started Flickr because of some friends, but now say that the common identity aspect of the space is also very important to me.
The authors state: "We argue that the constraints and opportunities inherent in online community design influence how people become attached to the community and whether they are willing to expend effort on its behalf."
Using social psychological studies off "common bond" (person-to-person relations) and "common identity" (person-to-group relations), they gather a literature review using snowball sampling and arrive at the following factors structure:
Causes of Common Identity
--Social Categorization
--Interdependence
--Intergroup Comparisons
Causes of Common Bond
--Social Interaction
--Personal Information
--Personal Attraction through Similarity
They note that:
"The emphasis in identity-based communities tends to be toward helping newcomers to navigate through information traffic,to understand community norms, and to engage in community conversations in meaningful ways..... In bond-based communities,the focus shifts more toward helping newcomers to connect with existing members,to join group interactions,and to form lasting relationships with a subset of community members."
The authors arrive at design implications such as "Subgroups are more compatible with bond-based communities than they are with identity-based ones."
My thoughts:
Diffuse vs. direct reciprocity is a big deal in political science, and it is interesting that users with "bond-based attachment to an online community will be more likely to engage in direct reciprocity, and those who feel identity-based attachment to the community will be more likely to engage in generalized reciprocity."
MUCH of bond based activity takes place OFF-CHANNEL, e.g. via email, and is thus difficult to understand simply by looking at the online community itself.
Given that multiple/plural identity is such a huge factor in network culture, it is unfortunate that "As far as we know, no studies have investigated the specific multiple sources of identity and bond that people have online."
Finally, Arendt specifically notes the distinction between lateral human-to-human community relations vs. vertical human-to-group community relations, noting that the latter is specifically the origin of totalitarianism (cf. Hannah Arendt, Origins of Totalitarianism).
-------------------------------------------------------- PHartzog@umich.edu
--------------------------------------------------------
The Universe is made up of stories, not atoms.
--Muriel Rukeyser
This was a very interesting article. I used to think (maybe still
do) that bond based communities are more resilient and provide more
interaction. But as per the definition and the examples (Facebook is an
identity based community), I agree that both types of communities
provide lively interaction among its members.
Bond based communities may be more "useful" in social terms than
Identity based communities. For example: In cancer network, people
provide an intangible support to other members. Whereas, Facebook or
other identity based networks might be very focused on individuals
(providing similar intangible support to its members) but not to the
new comers (or other members) in the community.
Theoretical design suggestions
John Blair
First, let me say I liked the reading and the work. It’s nice to have suggestions offered to resolve / test some of the points raised. However, I’m constantly amazed at how much effort is expended on attempting to direct human behavior through various social models (psychology, sociology, economic – page 402). Obviously, there’s an enormous carrot at the end of the stick for being successful in doing so, but so much of this seems so logical. Of course community design is going to determine how members interact – or if they interact. Offering functionality that members have asked for, commented on using at other sites, or that supports the purpose of the site is naturally going to please those members who inquired about it. Just like it will upset them if nothing is done or the implementation is poor.
I’m annoyed at the notion that people are like cattle and can be herded, directed, cajoled, and convinced to perform in an expected manner. The fact that this has been proven countless times is a never ending source of mystery to me. All I can say is that at some point the cattle always stampede and it’s almost always beyond the control of the cowpeople (PC term).
Community design should support community purpose. If you want your community to be identity based, then you should provide functionality that supports that (more group activities) and likewise for bond based communities (more functionality that allows interpersonal exchange). For those communities that have an integrated identity / bond mixture, this is more tricky but as the authors suggest try to “group participants into clusters with similar backgrounds or needs” page 397. Doesn’t this seem like an obvious thing to do if you trying to support both types of attachments?
The one area I can see as a problem for these dual communities is if they attempted to be everything to everybody. But that again comes down to the purpose / focus / mission of the community, not the design. That is an unrealistic goal and is doomed to failure – or severe splintering into sub groups as people found each other and their interests and went off to support these commonalties.
Maybe I’m making this too simple – or perhaps I’m just a simple social loafer whose touch to this tangent of reality is slipping.
All for one and one for all?
Largely taking the form of a literature review, these authors treat online communities as social-technical systems to test theories that communities favor either identity-based or bond-based attachment (400). Simply put, identity-based attachment is a commitment to the overall group; whereas bond-based attachment to individuals within a group. Although the authors offer that users may approach sites for mixed uses (both identity- and bond-based uses) they overall seem to believe that the way sites are designed tends to lead users overwhelmingly in one direction or another. The authors posit, "we believe a group's design can increase the likelihood of mainly bond-based or identity-based attachment to the group" (381).
In an effort to allow community designers to either analyze their communities in progress, or design from the ground up, the authors provide interesting taxonomies of both types of sites:
3 main causes of group identity/commitment to group
3 main causes of bond-based attachment to group
The authors give an incredibly detailed list of examples and corrections to be made for sites that are self-analyzing to see which type of commitment is most beneficial. For our class this week, though, the discussion beginning on p. 390 and going on until the end of the article is of particular relevance. This is where they begin discussing newcomers and other types of users that may help/hinder community commitment. My big takeway was that in addition to the Arguello piece from this week, this reading also seems to advocate using welcoming messages, mentors, moderators, or FAQs to guide newbies into a community (395), this seems like a good idea but also an advantage to old-timers too so that these small design changes may actually encourage old-timers not to get frustrated and so stick around longer.
how does a site get sticky - bond + identity?
I really like the bond and identity concepts, but a number of the communities which I've been involved in seem to have social attractors somewhere inbetween "pure bond" or "pure identity".
PerlMonks I would classify as "identity heavy" for the following reasons:
- very little in the way of off topic discussion (known on the site as GTKY or getting to know you)
- very weak direct reciprocity (if someone answers my question, I am not likely to do the same for them)
- welcoming to newcomers (if they're polite)
- shares many functions similar to the open source usenet groups studied by lakhani and von hippel (cited as an identity community)
However, I would say that among regulars, and perhaps even more so in certain geographical and administrative subgroups there is heavy bond attraction -- Members know each other personally, build strong personal and professional relationships, and communicate with each other off the site ie the PerlMonks group on LJ.
It seems that a healthy community may exhibit both bond and identity attributes, albeit at different scales or in different dimensions. I think in some cases where an online community discourages bonding behavior, people may use external channels to form those kind of bonds once they are introduced. I think an interesting design challenge for community maintainers to encourage both kinds of group formation without getting too much "chocolate in the peanut butter" as it were.
--------------------------------------------------------------
oostendo@umich.edu
--------------------------------------------------------------
I'm also interested in how
I'm also interested in how to increase both identity attatchment and bond attatchment. I think our pivots research is one way of doing that.
Social networks & bonds vs. identity
I really wonder how the emergence of social networks relate to this distinction between bond and identity theory.
I guess it is really obvious that social networks support common bonds between individual much more than common identity. Does that have an influence on other types of communities? Are for example discussion boards more identity focused since everybody is on Facebook? We have also been talking about the groups within social networks (again Facebook as the example). Maybe the discussions in many Facebook groups don't take off because there is no need for any conversation related to individual bonds, because that is already taken care of by profiles and friend connections?
If that is true it would be another argument that a successful community should try to maintain a balanced levels between the two types.
What do people think?
Overlapping memberships & community participation
The way community memberships interact and influence each other is an interesting question, and one the authors bring up in their discussion of theoretical issues. Do people get their "bond fix" from some communities and their "identity fix" from others? Or, do they seek out communities that have a similar balance between bond and identity focus? A related question might be how people's behavior compares across overlapping community memberships. For instance, are lurkers always lurkers? If you look at your own community memberships and behavior, do you see any patterns?
I'll admit that without significant incentive, I'm a chronic lurker. Perhaps not surprisingly then, I also tend to participate minimally in bond-based communities like Facebook (I haven't gotten around to changing my profile/picture since I joined in 2003).
lurkers
I like your question about lurkers, and getting bond/identity fixes in different locations. In response, I don't think lurkers are always lurkers. Lurking can be a way for learning - as discussed in one of the community of practice readings we've had, a person might lurk and then go become the leader for a new similar community. I think we probably each do a mostly unconscious cost benefit analysis before deciding whether to be a lurker or more of a contributer. In different communities, we have differing levels of expertise, and fulfill differing levels of exepctations from our peers. We certainly consider different communities to be more social than identity (for example, the co-op community I live in is primarily social, although we do share in common that we believe in democracy and sharing), and some to be more identity-based, like the Perl Monks community John talked about.
On Lurking and Perilously Leaping Towards Design
I, too, am a chronic lurker and have left more than one bond-based extensively networked profile behind (myspace, friendster, zaadz). When I left myspace, it was admittedly because of a desire for less bonding....One too many awkward re-kindled in-an-unauthetic "I don't really want to hang out but we used to hang out in second grade and I kinda want to spy on you" kind of way......
A large handful of my friends deleted their profiles that year. Most of them had more than 200 friends on their accounts and left without warning or with a brief note about not being able to take it anymore. I can''t say what was going on for them, but for me it was a sense of over-exposure and over-obligation to always be checking it......
Friendster was the first of these to go. For me, Friendster lost me when they added the "see whose been checking your profile" feature. There is a point at which transparency and connectedness is an ever increasingly irritating over-reality at times.
I typically recover after each of these acts of profile-destruction, and succumb to the latest bond-based social networking site invite sent my way. I think the most useful notion I pulled from this article was the idea that theorists aren't necessarily good designers.
I agree, and think its a curious predicament that accounts for much of the research-practice divide in many fields.
Lisa McLaughlin
I think you have made some
I think you have made some good points Lisa. People just have different preferences. From my Facebook experience, I can see that some people prefer more privacy and maintain close relationship with a small number of friends, whereas other people prefer less privacy and have relatively loose relations with tons of friends. If a community only has a single design for all users, it just cannot meet the expectations of everyone.
I think an obvious move is to go to "personalization". It might be achieved by using the social computing or web2.0 technologyies.
Group = identity bonds
Facebook groups are funny because they seem like another way of signalling a person's interests on their profile instead of actually acting as a kind of membership to something. At least for the way I use groups, it's usually to show my membership in something, but not necessarily in the group itself. For example, I'm in the "Squirrel Club" group, but I don't interact with other squirrel clubbers (hmm bad use of words?) in that group. I'm just using it to show people who look at my profile that I'm in the squirrel club.
So in this case, I think the groups are more identity based than bond based.
Social Networks
I would put Social Networks in Identity based communities because most of these sites focus on people interaction (and not on a certain topic that binds them). Yes, we have friends in facebook and that we might have the bond of "friendship" amongst us, but that is not the reason why we communicate (evident from the content of discussion in SN). We use SNs for keeping in touch with friends (interested in individuals) and therefore it seems that they are identity based communities.
There is indeed a conflict (example: forums on IGN)
With the authors doing such a good job listing and explaining the things that they didn't address or go into detail on, I find myself having to agree with most of what has been said. The primary take away that I got from this reading is that there is not necessarily a magic method that could easily equip an online community with the necessary characteristics that support either identity or bond-based user preferences. As the reading explains, there is just too much uncertainty when communities evolve, users come and go, and bonds form amongst regulars in a largely common-identity community. As a result, the best developers and site-leaders can do is to have a basic idea of what type of community they are interested in creating, then adjust and add/take away community features as needed. While all of the online communities that I have been involved in seem to fall into in between the two extreme ends of the identity-bond spectrum, I can definitely sense the conflict of interest between the two theories.
For example, consider IGN, a pretty popular gaming site/portal. Most games (popular ones, at least) have their own corresponding forums that act as a small community. Some of these forums have sets of regulars and several casual or new users (in many cases, there is also a lot of cross membership. My role on the IGN forums has always been a "lurker". While I personally don't interact with anyone on the forums, for the games that I am interested in, I always enjoy observing the inevitable "hype to fanboy flame wars" trends. Ideally, it would seem that on such a big public portal that consists of such a significant portion of "new users", people ought to be "identity heavy" (to borrow Nate's term). In fact, as the authors have pointed out, the site owners might even be interested in keeping the forums that way so new users aren't turned off.
However, from what I see, there is a lot of bond-based things going on. For example, regulars would post irrelevant threads titled "To XYZ-USER: Read This" or "I GOT THE GAME. WOWZ!!!1", which as one would expect, attracts a lot of complaints from new users, who question the relevance of such threads. In response, the original poster and his supporters or haters jump back in and express that they have the earned right to create such threads because they were on the forums since "day 1". Suffice to say, such interactions do indeed turn off new users and dilute the quality of the conversations. Personally, I have been a lurker for a long time, and I have no desire to mingle with some of these users.
Welcoming newcomers - 2 ways
I really like the subtle design difference Ren et al point out in terms of welcoming newcomers. The way newcomers are welcomed into a community can have a big impact on whether those newcomers will feel a bond-based or identity-based attachment to the community.
Design claim: Welcoming newcomers by helping them navigate information traffic, understand norms, engage in community conversation, and wait/observe before participating will encourage identity-based attachment in your community.
Design claim: Welcoming newcomers by helping them connect with other members, join interactions, and form relationships with a subgroup in the community will encourage bond-based attachment in your community.
Of course, these claims include the caveat that designers have limits in their abilities to shape member connections and community interactions.
The Argument for Bonds
I enjoyed this reading for its straight-forward presentation and interesting material, but I believe that the importance of bond-based attachment is understated.
While my understanding is that both identity- and bond-allegiances are necessary in order to create a lasting community, I place much more weight with bonds. For the average user, a specific topic is only going to be interesting for so long. Eventually you'll reach the top level of your MUD, earn your MBA and high paying job, get tired of swapping graphic design tips, forsake Perl for Python, or run out of interest in SimCity. What then? If the community has any hope of retaining you, it had better bank on some solid relationships built between you and the other members. Why would they want to retain you? Despite your waning interest, if you've been a regular for a significant amount of time you've likely developed valuable expertise. You're still an asset.
I agree with the authors when they suggest that members frequently go from identity-based to bond-based participation. In fact, I will go a step further and argue that, as suggested by the above paragraph, such a transition is inherent to any e-community. If you're building for longevity, anticipate supporting bond-based interactions.
Maybe not easy to design
I do particular favor this paper maybe because it gives a comprehensive literature review for building these basic senses on people's social relationships relevant to online communities. Since it is kind of general review, so there is no much to argue with. But as talked in above posts, i will say that the tension between identity or bond is pretty subtle and complicated. And of course, the two are often interwoven. People tend to be together with some others because they share something common and thus they fall in some sink as community being built; and they develop group identity and comparison with outside along their interacting process.
And I feel like that both are all needed in creating and maintaining health community: identity supplies belongingness and commonness, while bond intrigues substantial and repeated interactions over time to even enhance the community.
And I feel it is hard to manipulate a community to be more identity based or bond based, especially for online ones, where people come voluntarily. I guess it almost about the theme/topic of community, which mainly decides what the community would be like. I assume there are some attributes within the theme/topic decide this. For example, if the community is about pet or politics then it is very likely the pet community will be more of bond based and the politic one will be more of identity based. That is because the nature of topics about pet, people are intending to self-disclosure, e.g., to talk about their won pets, then all the interactions would have a strong context of the people participating in the community and bonds tend to be built; but for the political one, people would more focus on some target like Obama, good or bad, without much self-disclosure and opportunity to built interpersonal relationships.
Yup
I second that! I really enjoyed this reading I felt that it had more substance than some of the pervious papers. I hope that we continue to look at articles that have a stronger foundation in academia.
Social Identity, interpersonal bond? How to differentiate?
This paper study the reason of producing group identity and common bond, the effect of these two features on enhancing individuals attachment to groups and the implication of online community design.
The reasons of developing group identity are social categorization, in-group interdependence and intergroup comparison while they are social interaction, personal information and interpersonal similarity in bond theory. However, they are overlapped with each other. For instance, when people are categorized into one group, they do share similarity with each other, right?
Although these two theories are quite different in the definition, there is no systematic method to differentiate them, at least from this paper in my opinion.
I agree that individuals' attachment to certain community is dynamic and change from group identity to bond or reversal, but I also believe that these two features exist in one community together. For instance, in an online fans-club, there is certain discussion section for members to discuss their idols' new drama, which is based on topic/conversation, however, there is also chartroom for them to maintain individual relationship and lots of off-topic conversations are born here.
Me too
I don't mean to be writing all of these "me too" posts but I have to say that I agree with you. While I was writing my blog entry I felt that the bond and the identity were often inter-dependent. For instance, in my community you can form or join groups based on your interest. This could be seen as a type of identity based on intergroup comparison. But, people are likely to join groups that cater to their interests, which would imply a bond by similarity.
Both of these factors are going to influence each other as the member continues to interact as part of the group too. Over time the member will likely grow to identify more strongly with the group and their interests will begin to reflect the other group members’ interests as the continue to interact with them. These influence are only going to make it harder to tease the two factors apart.
I have always imagined that Paradise will be a kind of library.
-Jorge Luis Borges
Does bond-based attachment discourage arguments online?
One of Ren’s design claim was that to encourage bond-based communities the site administrators should phrase their mission statements to encourage members to engage in and to tolerate conversations on wide-ranging topics. The design of Newsvine seems to be encouraging this behavior and thus bond-based attachments (Ren et. al 2007). I have mentioned before that it seems that Newsvine users are more respectful in their disagreements than other communities I have seen. I wonder if bond-based attachment to a community discourages contention more than identity-based attachment. Anyone else care to speculate?
Bond or identity based commmunity
Ren et al. explained two theories of group attachment: identity and bond theory. When I look at the figure on page 393, I feel identity-based attachment seems more desirable then bond-based ones in terms of the consequences of attachment.
Instead of selecting one from these two types of group attachments, I prefer their idea on P.381. Ren et al. state that a community can have these two dimensions of members' attachment. Long-term members stay in the community because of their strong interpersonal ties, and new members join the community because of their common identity. As they highlight in the end of the reading, I would like to know how an individual's attachment transform from one type to another, and if both two change directions (identity to bond, or bond to identity) can work.
Behavioral Outcomes
Through this week’s reading we have focused on key variances
between IBA (Identity based Attachment) groups (i.e. Sierra club, NRA (National
rifle Association)) and BBA (Bond Based Attachment) groups (i.e. bridge or
poker game group with friends) I couldn’t help but think of Facebook and Myspace.
In next month’s Fast Company magazine
there is an article on why marketers who use social networking sites may be
pitching into the void or, simply put, why marketers who use social networking
sites thinking they are targeting a specific market may be missing the point altogether. Essentially, the author makes a case for why “Tom”,
the Myspace guy, is not the age that he says he is. Actually, Tom is 5 years older than his
profile claims he is. And just like Tom,
according to the author, we all present our best “selves” within these social
networking sites. So, marketers who
think they have a lockdown on a target market have anything but authentic data
to test drive their products. The article ends by stating “ Social networkers
don’t sign on to expose their true consumer selves. We sign on to promote our own agendas”
So based on the above, then sites such as Facebook/Myspace/Friendster,etc.
are centered around identity based groups who focus on achieving a purpose, to
promote their own agenda, knowingly or unknowingly, consciously or
unconsciously? Part of me thinks this is
a bit of a stretch but perhaps this is more salient than I want to
believe. This is after all a recent
science despite the fact that there are correlations between studies done in
the 50’s (Back study in ’51, pp.389) finding that identity based groups competed
tasks efficiently and discussed matters that were only relevant to their
goals/aims, and more recent studies (Sassenberg in ’02, pp.389) finding individuals
engaging in discussions on a narrow range of topics reported high group
identity and high valuation. In
combination, these studies suggest that online bond based communities are more
likely to engage in off topic discussion and be more tolerant of off topic
discussion (pp.389)
So, if we combine the above two, what conclusion can we
make?
networking sites are more bonds focused, not identity
I don't understand where your assertion that FB/Myspace/Friendster are centered around identity based groups is coming from. I would think (and the authors assert) that these are the quintessential bonds-oriented sites.
There are some features that are identity based (signal your affiliation with a band or with a "group" that doesn't actually interact). But the basic feature is linking with individual others, not identifying with groups.
Feel free to correct me if I'm showing my ignorance of how these sites work in practice. As a very occasional user, my mental model may be wrong.
Boarding School Bonding
The high school alumni online community I am investigating shows a high level of commitment among members (based on volume of message and number of returning and participating members) because:
1a) We belong to the same social category of school alumni, and of the same age, and ethnic group (mostly).
1b) We rely on each other to get things done, such as organizing trips, activities -- each of us is independent in our own task, but interdepend on others to get the big task done.
1c) We always pride ourselves of being associated with a special school system for the underprivileged, who does as well in the adult life as those who went to the elite schools in the country. But intra-group comparisons is limited - there doesn't seem to be comparisons among professions.
Members of this community also have a strong bond to the grouping, because:
2a) the relationship was built over 3-5 years at least, and now in the 20th year, and so, the social interaction is at a unique level compared to other groups that we are a member of.
2b) Being holed-up for 5 years in an isolated part of the country really did its job to make us known to others at a very personal level.
2c) Cliques and groups develop then and even now based on similiarities in interests or profession.
Disagreements, Confusion
1) "If designers discourage off-topic discussion, they might lose people who would
like to talk with others like themselves." p4
What happens when designers introduce an off-topic board? Do established members forget about the original focus because they're too busy discussing their dogs? Is the problem maintenance?
2) Identity-based attachments seem to be centered on some notion of shared values/interests and indirect conversation, while bond-based attachments take this one step further by focusing on activities shared with other people and direct conversation/awareness. Pp5-13
If personal information sometimes promotes interpersonal bonds among people who haven't interacted, could this also extend to an individual's bond with another community he hasn't interacted with? I like architecture but don't have any personal connections to the community. Still I feel somehow connected when I read news and casually browse information.
3) "we suggest that people with identity-based attachment to an online community may be more likely to take over responsibilities from lurkers or slackers and compensate for their lack of contribution."
What if an individual isn't proud of his group membership? If an obese man who belongs to Jenny Craig considers himself to be a Jenny Craig member (identity) instead of classmates with people who belong to Jenny Craig (bond), would he be more or less likely to pay for a broken window?
By the way, I really like this article. If I don't force myself to find disagreements, I just don't know what to post other than "I like this article and I find it interesting", which won't be useful or interesting for anyone else.
Good Q: effects of identity-based attachment to low-status group
It would be interesting to know whether the same effects (willingness to take up slack and compensate for lack of contribution; proclivity to sanction non-normative behavior) will be seen for identity-based attachment to groups that are lower status, groups where one might not want to proclaim membership with people outside the group.
My intuition is that you would still get the same effects. I don't whether there has been research testing this (my guess is that there has been).
There can be extreme cases
Some communities can be extreme in being bond or identity based.
In technical support forums private message are highly discouraged. Sub communities are formed around topics related to community. Personal information is almost never revealed. Such features force only people who are attached to the identity to stay in the community. Similarly, friends network sites like Facebook, MySpace have features that are targeted towards personal and direct communication. The emphasis is clearly on establishing bonds. Users can form identity based communities within these sites, but they are rare and don't last very long.
Over a period of time identity based communities should pick-up some traits of bond based community. A veteran member will be staying in the community due to a strong attachment to the community identity, but repeated interaction with another user(s) will create a bond between them. This would lead to off line communication as private messages, IRC or real world conferences. Bonding with other users can happen even when personal information is not revealed. This happens when a user gains the respect of the community for his/her contribution but chose not to reveal any personal information.
However I'm not sure if bond based communities can drift towards getting a identity for the group.
Design Choices and Communities
Seems like based on the fundamental structure of the two types of bonds - depending on whether a community shows a greater common identity focus or a greater bond identity focus - macro or micro level level features of the group will become important.
To pick up from last week's topic for an example, suppose a group of friends joining a contract bridge club join it mostly for their friends - to interact and communicate with them. This kind of a group shows a stronger bond identity focus and a weaker group focus and it is more likely that regulation for example
will be more largely controlled by architectural decisions that impact micro interactions and interactions among small groups rather than a strong common norm.
On the other hand a group with a strong common identity like a group of Linux enthusiasts are more likely to be regulated by a strong social norm and common beliefs that bind them. This can help us make the design choice between the various regulation mechanisms based on things we may know about the expected group type.
There will also be switches from what bond one identifies with over time in a community. As a newcomer to a community a user might be drawn to it by the ideals and vision it stands for and signifies for him. Once a part of the community however, as he gets to interact with people there he may form strong relationships with them and then the site has a different or at least broader signifiance for him. This could lead us to design choices that target specific users at specific points of interactions in the community. For example for newcomers we could provide information that is more targeted towards explaining the norms and the values of the community, while for the members the focus could be to initiate discussions and enable them to form stronger bonds between themselves.
targeting/personalization
I agree that the way to resolve some of the tension is to support identity-based attachment initially and gradually move toward encouraging and supporting bond-based attachment.
The study identified factors
The study identified factors a the group, individual and message levels that impact community responsiveness and individual commitment, which according to the author, are two key elements of success of a community. First, I should agree that conversation is the basic mechanism in online groups and the viability of a community also depends on the willingness of individuals to stick with the groups over time. To take a look at the study results, some seem interesting.
1. Group identity. It is claimed that bond-based groups have higher member commitment than topic based groups. I tend to agree with this claim as friendship is always a strong tie between members. In terms of the group I am studying, it is more like topic based. 2. Cross posting. Nowadays it seems cross posting is very common in a lot of online communities. The study in the paper shows no distinct result whether this factor is significant or not. Personally, I would think sometimes cross posting helps navigation while taking up more resources. 3. Group size and volume. I don’t actually agree that people tend to be less committed to larger groups and contribute less. The group I am studying has over million members, majority of whom demonstrate high commitment and significant contribution to the community. So size only is not a factor that would affect member’s involvement and commitment. 4. All individual factors are interesting, but I do not see much practical tips that could be drawn from it, such as newcomer status. I somewhat agree that the history of an individual affects the interactions, but it is still a question to me how site managers could apply this to their design.
Overall, the study is still interesting enough to take a look, at least, it tells that there are a number of factors affecting the individual group interaction, which is the key success factor to an online community.
really?
I didn't see where the paper argued that bond-based groups have higher member commitment.
Also, I'm skeptical when you say your community has more than 1 million members, most of whom demonstrate high commitment and significant contribution to the community. Is it really one community?
Thoughts
Lots of interesting questions have been raised, and great summaries given so far, so let me just add a couple of my thoughts on top of it all.
First, this was probably my favorite reading to date - it was an academic article, but gave clear design claim related to how to design a community for identity-based or bond-based attachments.
It seems that a lot of people feel that the two are interdependent, or even both always present together in a community. I tend to agree with the authors, however, that in most communities, one will be the primary attachment type. Yes, identity-based attachment can morph into bond-based attachment in communities where members grow closer over time, but still in most cases, the design decisions the community builders make will determine what type of attachment occurs.
The author claims that if the community builders choose:
Then this will cause bond-based attachment.
On the other hand, if they choose:
Then this will cause identity-based attachment.
So while the two types are not completely separate, the authors did acknowledge that it is very hard to find a balance between the two, since the decisions tend to favor one over the other.
On another, separate note, this reminds me of the discussion we had last week in classa about the optimal size of a community. Some were arguing vehemently that growth was always good, while some were saying it had major drawbacks. I think it would be interesting to see if people who were arguing for unlimited growth are studying identity-based communities this semester, while the others were studying bond-based ones. It seems that whatever community we're studying is what is framing our thoughts, instead of all communities in general.
Thanks for framing the design claims
I also like your final point is well-taken. Let's find out in class whether there's a correlation.
The authors describe the
The authors describe the difference between bonds and identity in a way that I found very helpful. While we've talked before about why it would matter whether other members or the group are more important for user, this paper offers useful vocabulary for talking about the differences. The observation that some types of groups are more likely to have strong feelings of common bond and others are more likely to have feelings of common identity (just based on their reason for being together!) seems important, and I was surprised at how some of the groups were categorized.
One area where I had a bit of trouble was with figure 1a. It may be a minor quibble, but I felt that the framework would have benefited from making a feedback loop explicit; many of the consequences also seem to be antecedents of certain types of attachment / both types of attachment.
A second point is that while the authors talk about some communities shifting from identity-based to promoting interpersonal connections (p 401), I think some of these spaces (particularly Flickr) stand out because they are able to simultaneously provide outstanding support for both dimensions the authors discuss (common bond and common identity), not only for users who are excited about both dimensions but also for users who want mostly one or the other type of community.
The authors also note (also p 401) that they are unable to find many examples of communities moving from bonding to identity. In my view, bonds may have more of a role to play in getting people involved in identity communities. Returning to the Flickr example, I got started Flickr because of some friends, but now say that the common identity aspect of the space is also very important to me.
Common Identity and Bond Theory
The authors state: "We argue that the constraints and opportunities inherent in online community design influence how people become attached to the community and whether they are willing to expend effort on its behalf."
Using social psychological studies off "common bond" (person-to-person relations) and "common identity" (person-to-group relations), they gather a literature review using snowball sampling and arrive at the following factors structure:
Causes of Common Identity
--Social Categorization
--Interdependence
--Intergroup Comparisons
Causes of Common Bond
--Social Interaction
--Personal Information
--Personal Attraction through Similarity
They note that:
"The emphasis in identity-based communities tends to be toward helping newcomers to navigate through information traffic,to understand community norms, and to engage in community conversations in meaningful ways..... In bond-based communities,the focus shifts more toward helping newcomers to connect with existing members,to join group interactions,and to form lasting relationships with a subset of community members."
The authors arrive at design implications such as "Subgroups are more compatible with bond-based communities than they are with identity-based ones."
My thoughts:
Diffuse vs. direct reciprocity is a big deal in political science, and it is interesting that users with "bond-based attachment to an online community will be more likely to engage in direct reciprocity, and those who feel identity-based attachment to the community will be more likely to engage in generalized reciprocity."
MUCH of bond based activity takes place OFF-CHANNEL, e.g. via email, and is thus difficult to understand simply by looking at the online community itself.
Given that multiple/plural identity is such a huge factor in network culture, it is unfortunate that "As far as we know, no studies have investigated the specific multiple sources of identity and bond that people have online."
Finally, Arendt specifically notes the distinction between lateral human-to-human community relations vs. vertical human-to-group community relations, noting that the latter is specifically the origin of totalitarianism (cf. Hannah Arendt, Origins of Totalitarianism).
--------------------------------------------------------
PHartzog@umich.edu
--------------------------------------------------------
The Universe is made up of stories, not atoms.
--Muriel Rukeyser
Bond vs Identity: [Reposting from blog]
This was a very interesting article. I used to think (maybe still
do) that bond based communities are more resilient and provide more
interaction. But as per the definition and the examples (Facebook is an
identity based community), I agree that both types of communities
provide lively interaction among its members.
Bond based communities may be more "useful" in social terms than
Identity based communities. For example: In cancer network, people
provide an intangible support to other members. Whereas, Facebook or
other identity based networks might be very focused on individuals
(providing similar intangible support to its members) but not to the
new comers (or other members) in the community.