Tumbleweeds

Terry Anderson a week ago commented on the "tumbleweed blowing through a ghost town" effect in community websites. Sites "have gracefully (and some not so gracefully) atrophied away - usually through lack of posting/input or contribution." My own efforts to build online community have been mixed - MuniMall and this site are successes by any measure, but some others have been less successful.

Of course, I haven't always tried to build online community per se. My most successful site - the Logical Fallacies - was simply an online document. And the largest success I had anywhere - the Referrer System, which at its peak was generating 800,000 hits a day - was simply an online service, not a community at all. Similarly, EduRSS, which is pretty popular, is nothing more than an aggregator.

That said, I have seen and taken part in thousands of communities and other web services over the years, from MUDs to mailing lists to discussion boards to websites and social networks. Some have been successful, some have struggled, and the majority simply vanished into the ether. Which leads me to ask, with Terry Anderson, what it is that promotes successful community online.

Today the aggregator picked up this paper by Guowei Jian and Leo W. Jeffres focusing on the poor cousin of online communities, shared organizational databases. Here we see the same problem surface. "Employees’ reluctance or resistance to contribute to shared electronic databases results in information undersupply and underutilization of such technologies." The empirical data presented here is of course useless but the theoretical framework offers some insight.

Jian and Jeffres, drawing on traditional explanations for participation in online community, offer a tripartite model:

"we understand employee information contribution as the outcome of three key organizational processes (see Figure 1):

- the process of costs-benefits analysis in which employees assess their personal welfare in the organization

- the process of organizational identification (OI) in which employees form their relationship with the organization, and

- the process of collaboration in which employees work together and develop positive or negative orientations toward collaboration.

From the perspective of online community, then, what Jian and Jeffres are saying is that successful communities online will be the result of three factors working in combination:

1. Utilitarian value - the community serves some practical purpose, from the point of view of the user and not (necessarily) the organization hosting the community. This means that participation in the community will in some way save time, reduce effort, increase return, or in some other way directly help the user.

Why, for example, did MySpace succeed where Friendster foundered? Because MySpace actually helped its members upload content, including digital music, which they could then share with their friends. When we look at other communities, such as YouTube and Blogger, we see this same feature. Sites like Writely and Jotspot, just purchased by Google, also exhibit this trait.

2. Identity with a group - "Research in work motivation and organizational commitment all suggests that collective action is partly a result of identification mediated by organizational values and norms." In other words, the contribution of resources supports a user's sense of belonging with the rest.

Nowhere is this more clear than in the field of open source software, where contribution of code to, say, Firefox, creates an attachment to the group. As well, when we look at discussion lists, we find that people create a niche for themselves by answering questions and otherwise adding value; nowhere is this more clear than in the work of Larry Lippman. Related to this is the user's creation of his or her own identity within the group context; a person literally becomes his or her contributions.

3. Means to collaborate - as Jian and Jeffres argue, collaboration "has become a major portion of people’s work experience in today’s work environment." Hence, insofar as a service supports this collaboration, it will be successful.

Examples of this may be found in unexpected places. Craigslist, for example, is a runaway success because it supports a collaboration between buyer and seller; eBay, Amazon, eTrade and others have managed the same feat. Sites like Yahoo Groups and Google Groups, similarly, succeed because they help people collaborate with each other.

Now, according to Jian and Jeffres, all three of these contribute to the successful fostering of participation in a community online. "We propose an additive model in which the willingness of
employees to contribute to shared electronic databases (WEC) is predicted by employees’ perceived benefits (PB) derived from the utilitarian dimension, OI from the normative dimension, and CE from the collaborative dimension."

My own view is that this is far too simple (and that is why a straight-up empirical investigation with a limited number of measurements will be useless).

For example: in an additive model each of the three elements is assumed to have the same value. But it is arguable that one element - utility, say - may be more important than the others. More to the point, it seems to me that each of the three elements described will be of different weight depending on what you're trying to do. Jian and Jeffres simply look at shared organizational databases, which locks the context into place. But even in an organizational environment, we need to look through the lens of multiple perspectives.

To illustrate this, allow me to recast the three dimensions slightly.

1. Individual - the utilitarian element becomes the individual element, that is, value to the individual (this is consistent with the account given by Jian and Jeffres).

2. Group - the identity element becomes the group element. By 'group' I am now talking about 'group' as I have described it elsewhere. Again, this accords with the Jian and Jeffres usage.

3. Network - the 'means to collaborate' element becomes what I have described as the 'network'. This introduces a bit of a change to the Jian and Jeffres discussion. They write, "A collaborative action emphasizes cooperation and coordination. It involves sharing resources and responsibility and creating shared meaning." For reasons I have described elsewhere, I would class this sort of thing as being part of 'group' behaviour.

You need not agree with my characterizations to accept the point, however. If we look at each of these three types of incentive to contribute to community online, we can see that their relative value will vary depending on the circumstances and from the point of view of the individual actor. As the authors argue, "knowing that I will be able to enjoy the benefits negotiated by the union anyway, I might not pay my dues to become a member." In this circumstance, the individual element, to this person, is most important. Forming an 'identity' or group affiliation with the union is of little or no value.

Understanding the elements in this way also leads us to understand that each of the three factors might also carry a negative weight into the success or non-success of communities online. The open source movement again offers a good example. For many people, open source is not merely a means of producing software, it is also a philosophy, which means then that in order to contribute to the community a strong sense of identification with this philosophy is required. This requirement, however, may serve to prevent, rather than encourage, involvement in open source initiatives.

In a similar manner, suppose management says that contributing to a corporate database supports workers' identification with the goals and values of the corporation. This may be interpreted as a means of motivating contributions, however, in some workforces the suggestion may be counterproductive, as contributions of this sort may be seen as 'sucking up' to the management. This is especially the case when the workers' primary loyalty is to their own group or cluster, and not to the organization as a whole.

Failing to appreciate these factors may result in the expectation that a community online will be successful when in fact significant disincentives have been built in from the outset. And it seems to me tht this has especially been the case in education, where the weights of various factors have not been properly considered.

Consider, from Jian and Jeffres again, what constitutes a common belief among educators: "Wenger states that spreading information in a community of practice is a form of engagement and declaration of allegiance to a community. Information contribution is not only an act of sharing or exchanging resources but also a symbolic one that forges and maintains a relationship between an individual and a collective." Or Siemens: "Virtual and physical communities share many similar traits... These aspects of community address our social needs as learners. Much of our learning comes through informal, social means... Small communities, loosely joined, are the future of effective life-long learning (connected specialization)."

What's happening here is that the mechanism of group affiliation is being used in order to provide a motivation for participation, and it is this motivation that in turn is intended to support learning. "Collaborative learning, particularly when engaging learners in complex authentic tasks or projects, results in significantly higher outcomes," write Moisseeva, Steinbeck and Seufert, due in part to "high motivational level of learning and awareness of individual responsibility for the success in leaning." We can see the same point made elsewhere, here, for example, or here or here.

There seems to me no doubt that identification with a group increases motivation, and that motivation increases participation (and output). But where I have a concern - which I have expressed elsewhere - is that this same element may carry a negative weight in some circumstances for some individuals. The comments of Konrad Glogowski weigh heavily here:

I could, in a totalitarian fashion, put my students in groups (keeping in mind, of course, their homeroom teacher’s suggestion that Jeremy and Sean should never be in the same group because of the bullying episode last year). I could even let them choose their own groups. Of course, if I let them choose groups, Phil and Vanessa would have to go through yet another humiliating moment of looking left and right or staring at their desk only to discover that no one wants them in their groups, that no one asked, that no one really seems to care that they even exist.

I have no doubt that an empirical study of such an environment would, indeed, show an increase in overall educational attainment. However, would such an attainment be worth the cost - is it worth, once again, causing Sean to face the prospects of working with the bully, of Vanessa, once again, feeling the humiliation of being left out?

This problem is, in my mind, exaggerated by the unwillingness or inability on the part of so many to acknowledge that there can be more than one type of community. Morten Flate Paulsen, in his contribution at the recent (2006) EDEN conference, tried to capture this by distinguishing between 'collaboration', which is required, and 'cooperation', which is voluntary. Even so, some commentators continued to push for (involuntary) collaboration, on the ground that some students might benefit.

Yet we still see - and will probably continue to see - remarks such as were offered by Scott Wilson: "Oh, and bridges aren't built by networks of uncoordinated individuals with a shared need to get across a river (that looks something like this). They are built by teams of engineers, crane operators, welders, and the like coordinated by a project manager and a chief architect.." The presumption is that anything that is not a forced collaboration is some sort of unorganized chaos. This off-the-cuff comment echoes his more seriously intended questions about the need for coordination in community activities. "A good balance of individual and group activity keeps things interesting. Don't kill group activities just because we like blogging and social networks and this particular technology is better at supporting this type of individually-focussed work (a concern Catherine has with the PLE generally, echoed by Juliette)," he writes.

My own response, not so off-the-cuff as it seems, is that a prison riot 'keeps things interesting', but that this isn't thereby a good reason to keep encouraging them. Sure, a prison riot may well create a sense of shared community and common interest among the prison population, but the costs outweigh the benefits. We need, I emphasize again, to consider both the harm as well as the benefit in assessing how to encourage participation in learning and community.

But more to the point, my response is that the opposite of 'group' is not 'individual'. That there are ways of contemplating the organization of community in ways that do not require the identification of the self to the vision or the values represented by the whole. That even 'coordination' can be understood in these two ways, one voluntary, the other involuntary, which is the difference between (say) a tour and a kidnapping.

So where does this leave us?

I am in rough agreement with Jian and Jeffres that utility, identity and collaboration each contribute to the success of community online, but that their additive formulation is too simple, and that these factors need to be examined and weighed not only according to their positive contributions in certain circumstances but also to their negative impact as well.

In building my own communities I have not sought to elicit identification with the group - indeed, in many cases I have explicitly discouraged it - and have tended to prefer individual (utilitarian) and network (collaborative, or perhaps I should say, cooperative) values.

The question of whether community can succeed online without explicit self-identification with the group is an empirical question (but not one that will be measured in ridiculous n=80 surveys). But that said, one cannot help but look at the successful communities online to observe that it is exactly those that do not require identification with the group that are most successful - communities, in other words, with loose, cognitive and unemotional affiliation.

In practical terms, what this means, I think is that successful community online will be the result of:

- individual benefit - the community will in some way save time, reduce effort, increase return, or in some other way directly help the user, and

- cooperation - the community will help people connect in ways that allow them to accomplish (work, hobby or school related) tasks.