Information Needs:

Identifying Real Information Needs and Developing Digital Libraries to Meet Those Needs

(Social Context and Culture)

1. Investigating social and cultural influences on information needs and the ability of retrieval systems to satisfy them often draws attention to interface issues. For purposes of this workshop, Jonathan Grudin's distinction between a human's interface to a computer and a computer's interface to a human presents a challenge worth pursuing. The former is quite rich and varies from person to person as a function of social and cultural factors as well as training, experience, access to knowledgeable local gurus, and so on. At least in library systems, by contrast, the latter has tended to be rather rigid and uniform.

   The arguments for standardization are familiar--predictability enables people to know how to react to the technology and to develop reasonable expectations for how the technology will react to them. But given the quite different interfaces humans bring to computers, a one-size-fits-most library system can't be the best solution. It could even be argued that the same library system interface may well be both good and bad for the same human, depending on whether an information need has developed in a subject area in which the person is an expert or a novice.
   The challenge, then, is to provide a set of good standardized default options that are readily modifiable by motivated users who want to interact with the system to cope with information needs in their own preferred ways. User modifiable interfaces (including user-definable search templates) are available in other kinds of systems and should be extendable to digital library systems. Similarly, library systems could enable users to save their starting configurations, search patterns or templates, last-retrieved items, and so on--which the next use of their library card/ID could invoke.

2. In RAND research on information system users, we construe the human's interface to a computer as the means by which the user gets access to and manages its functions. We generally find that users attend chiefly to functionality--unless the interface actually interferes, functionality outweighs friendliness in their system evaluations. But where information work is being performed using a computer (say, searching, finding, re-organizing, analyzing or synthesizing information rather than operating a numerically controlled machine), interactions with the interface and with the task objects may be very closely related (and perhaps readily confused). Perhaps for this reason, interfaces to library systems seem to be structured like the task objects (e.g., in hierarchical subject-matter classifications), independently of how well this approach meets functionality criteria.

   So a second worthwhile challenge, it seems to me, is to try to classify the kinds of functions reflected in users' information needs. Getting closer to the context and culture of use is reflected in the notion of information needs; but "information needs" is so broad and generic a construct that it's difficult to handle operationally. On the other hand, if the next cut were something closer to kinds of functions served by getting library information, better ways of interacting with digital library systems could be more readily explored in empirical research. Consider, for instance, the following questions that a library search might answer:

   • What proportion of eligible voters registered, and what proportion of registered voters actually voted, in federal, state and local elections in the US in 1992?
   • Is there anything new and interesting in the literature on ethical aspects of social policy research?
   • Does KLM have a Sunday night nonstop flight from LA to Amsterdam this time of year?
   • What does the best and most recent research literature have to say that bears on issues of the feasibility and desirability of universal access to email in the US?
   The four information searches--all ones I've made recently--have little in common. They differ in scope and specificity; the time-value of the information to be retrieved varies; and iteration, filtering for relevance and quality, determination of likely completeness of the search and other such efforts are associated with some but not others. It seems reasonable to think about a functional typology to get more specifically at differences in kinds of information needs. Such a typology would respond to differences in social context and culture to the extent that these social factors condition the information functions performed.

3. Although little may be known about the social and cultural origins of information needs, a great deal is known about how finding and filtering occurs. A substantial body of research consistently shows that social networks play a critical role in linking people to information sources. Friends, colleagues and peers often provide the initial and most trusted pointers either to the needed information or to other knowledgeable individuals who may make the referral. In that process, the social roles of those involved are treated as prima facie indicators of the relevance and quality of the information retrieved.

   A third challenge, then, is to build regular options for such communicative interactions into the systems that support digital library use. Literature on computer supported cooperative work affords a number of models for providing interactive help that might be extended to the digital library. For example, studies by Wanda Orlikowski and her colleagues have shown that formal roles could be created for human "mediators" so that employees using a large corporate information system did not have to depend on the chance availability of a local guru.

   Earlier, Nathaniel Borenstein and other developers of the Andrew messaging system reported on the design and use of its help system, which triages queries and routes them to predesignated individuals for answering; individuals were assigned to question types based on their areas and levels of expertise, so that the questioner was linked to an appropriately knowledgeable respondent. And, in a recent RAND study, we noted the use of an international corporate on-line library that included facilities for users of its research documents to generate an assessment of the material used; such assessments then provided evaluative guides for other potential users of the same material, while the email address associated with assessments allowed future users to query reviewers directly for more commentary.

   In generaly, interactive or human-mediated options might comprise a viable way of helping individuals from diverse social and cultural contexts cope more effectively with their information needs in the digital library era.