Supporting Information Search by Older Adults

Abstract

Using cognitive models of web-navigation to generate support has long been a topic of research. In this paper, we address two limitations in this area. First, these models have so far been used to generate support for navigation within a website and not for interaction with a search engine. Second, very few studies have looked at the usefulness of such model-generated support for older adults who are known to be less efficient than younger adults. An experiment with 24 younger and 24 older adults on six simple and six difficult information search tasks was conducted. Results showed that the semantic relevance of queries showed a decreasing trend across reformulations for older adults and remained constant for younger adults, indicating that as older adults reformulated, they produced queries that were further away from the target information, which could be the reason for their lower efficiency. Based on these outcomes, two types of model-generated support mechanisms for interaction with a search engine are proposed, one which visually highlights the most relevant search result given a query and the other which monitors the average semantic relevance of search results for a given query and warns the user if it falls below a threshold.