The retrieval of highly scattered facts and architectural images: Strategies for search and design

The development of huge sources of information in online domains like healthcare, e-commerce, and design, coupled with powerful search engines, suggests that finding comprehensive information about a topic is straightforward. However, recent studies show that while novices can easily find information for questions that have specific answers (e.g. What is a melanoma?), they have difficulty in finding answers for questions requiring a comprehensive understanding of a topic (e.g. What are the risk and prevention factors for melanoma?). This article argues that an important explanation for this difficulty is the phenomenon of information scatter: as the number of information sources about a specific topic increases, the information across the sources begins to follow a Zipf-like distribution, where a few sources have a large amount of information, and many sources have very little information. To illustrate the phenomenon of information scatter, this article presents examples from an ongoing study of how facts related to common healthcare topics are distributed across high-quality sources. These results are compared to results from a small study to explore how images of buildings designed by a well-known architect are distributed across high-quality image sources. The results from both studies suggest that the distributions of facts and images across relevant sources are Zipf-like, and pinpoint the kind of search knowledge needed to address such scatter. These results suggest the need for the development of systems and training that are "distribution conscious", to assist users in finding comprehensive information about topics across information domains.