Improving the “darkness to light” ratio in user‐related information retrieval research

Abstract

Research into information retrieval (IR) cannot yet answer the basic question of how we can design IR systems to help people search for information with optimal levels of effectiveness. In relation to human‐system interaction, we have failed to develop any valid and at the same time robust user models capable of driving practical system development. If we strip away assumptions and over‐optimism relating to the generalisability of what are essentially sporadic and fragmented research efforts, the great “darkness to light” ratio characterising our knowledge of human aspects of IR becomes apparent. From a more critical and pessimistic (but by no means less realistic) perspective, we are getting nowhere fast. A range of strategies is proposed to improve the situation by supporting relatively “horizontal” as well as “vertical” knowledge integration. These consist of: greater use of pluralistic research approaches; enhanced access to research data; more multidisciplinary and multi‐perspective integrative reviews and conceptual mappings; and establishing a greater critical mass of published research findings sufficient to support the generation of a less sparse and fragmented evidence‐based knowledge map. The potential of electronic publishing and data access for helping achieve these goals is discussed.