« You think you bought the music? | Main | Can incentives make the skies friendlier? »

July 27, 2008

"Web science": recognizing integral role of human behavior

Sir Tim Berners-Lee and his colleagues have been advocating for a field of "web science" for several years. They describe it as a multidisciplinary, systems science needed to understand and engineer the future Web.

I've not fully grokked what they are proposing: the research questions they suggest to define seem a bit vague, and I don't quite see what defines this "science" other than a set of topics (the semantic web prominent among them, natch) in which this group of people is interested. I'm not saying I think it's not science, I'm just not sure what field definition they are proposing (so that, for example, many universities could start offering courses, or even creating departments of web science) -- it feels like the space of current interest to a particular research center (and indeed, there is a Web Science Research Institute).

They have published another manifesto (also available from the WSRI site) (there have been several in the past few years), this time in the Communications of the ACM. Maybe I'm just paying more attention, but I think I'm starting to get parts of it. In any case, one thing seems clear to me: incentive-centered design (ICD) fits comfortably in their framework. It's a subset of their very ambitious agenda, but I think it's clearly a central piece of. Put another way, they are saying some of the same things our ICD group has been saying independently over the last several years.

For example,

We show there is significant interplay among the social interactions enabled by the Web's design....However, the study of the relationships among these levels is often hampered by the disciplinary boundaries that tend to separate the study of the underlying networking from the study of the social applications.

I agree, and this has been part of the ICD manifesto from the beginning. It is rather vague, but here's a clearer statement of the common starting point: "It is the interaction of human beings creating, linking, and consuming information that generates the Web's behavior."

They suggest, as do we, that this inquiry should rely (among other things) on the sciences of motivated behavior, such as economics and psychology. However, I think there is one way in which we diverge: for the most part, I have seen these authors talking about computer scientists and web engineers needing to understand how people behave so they will understand the consequences of web design decisions. But, I have not really seen much evidence that they recognize the role of incorporating motivated human behavior directly into the design loop, which is the essence of ICD: design incentives or motivations for the humans who will be interacting with and over the web in order to obtain desired consequences. What we propose is more a more central recognition of the malleability of human behavior, and the social (or commercial) value to be gained from designing for that malleability.

On the other hand, one large area of research (and not the only one) that the "web science" promoters claim that is not inside the boundaries of what we call ICD, is a micro-behavioral science of understanding and explaining observed macro web phenomena. For example, they point out that social network analysts (like my SI colleague Lada Adamic, U Mich's Mark Newman, or HP Labs's Bernardo Huberman) rarely explore or test the underlying human behavior that generates the macro phenomena they observe and characterize.

Posted by jmm at July 27, 2008 11:38 PM

Comments

Login to leave a comment. Create a new account.