This thesis reports results of a research study into the usefulness of a user-centred approach for designing information retrieval interfaces. The main objective of the research was to examine the usability of an existing Web-based IR system in order to design a user-centred prototype Web interface. This research used the Web of Science available at http://wos.mimas.ac.uk.
A series of usability experiments was carried out with the Web of Science. The first experiment was carried out using both novice and experienced users to see their performance and satisfaction with the interface. A set of search tasks was obtained from a user survey and was used in the study. The results showed that there were no significant differences in the time taken to complete the tasks, and the number of different search terms used between the two search groups. Novice users were significantly more satisfied with the interface than the experienced group. However, the experienced group was significantly more successful, and made fewer errors than the novice users.
The second experiment was conducted on novices’ learning and retention with the Web of Science using the same equipment, tasks and environment. The results of the original learning phase of the experiment showed that novices could readily pick up interface functionality when a brief training was provided. However, their retention of search skills weakened over time. Their subjective satisfaction with the interface also diminished from learning to retention. These findings suggested that the fundamental difficulties of searching IR systems still remain with the Web-based version.
A heuristic evaluation was carried out to find out the usability problems in the Web of Science interface. Three human factors experts evaluate the interface. The heuristic evaluation was very helpful in identifying some interface design issues for Web IR systems. The most fundamental of these was increasing the match between system and the real world.
The results of both the usability testing and the heuristic evaluations served as a baseline for designing a prototype Web interface. The prototype was designed based on a conceptual model of users’ information seeking. Various usability evaluation methods were used to test the usability of the prototype system. After each round of testing, the interface was modified in accordance with the test findings. A summative evaluation of the prototype interface showed that both novice and experienced users improved their search performance. Comparative analysis with the earlier usability studies also showed significant improvements in performance and satisfaction with the prototype. These results show that user-centred methods can yield better interface design for IR systems.
I would like to express my greatest appreciation to my supervisors Professor Cliff McKnight and Professor Charles Oppenheim for their guidance and support throughout this research work. Without their care and consideration, this work would likely not have matured.
I would also like to thank my Director of Research, Dr Paul Sturges for his help and encouragement during this research work.
The research used transaction logging of the Web of Science gathered through MIMAS at the University of Manchester. I like to thank Andrew Weeks for this help. The contributions of expert panel members in heuristic evaluations, namely, Martin Maguire, Anne Morris, and John Richardson, are gratefully acknowledged. The volunteers at Loughborough University who took part in various stages of this research project are also gratefully acknowledged.
This research was funded by the Commonwealth Scholarship Commission in the UK and was supported by the British Council. I would like to thank several staff members who, during the three years in which this endeavour lasted, provided me with useful and helpful assistance. I would like to thank Anna Gane and Rosalind Grimmitt, my award administrators at the ACU, and Sarah Profit and Stephen Dunnett, my award administrators at the British Council for their help and support. Thanks are also due to Ellen Elcock, British Council Regional Service Officer, for her help and assistance.
I would thank the support of staff in the Department of Information Science at Loughborough University. Special thanks must go to Linda Thornhill and Shirley Horner for their help and support.
Many thanks to all my friends and colleagues in Dhaka University who have given support during this research. Thanks are also due to Mamun Murtaza Sharif for his help with the CGI files.
My parents have always been my emotional anchors. Their support and belief in me have always inspired me in my endeavours. My wife Sarwat has always been patient and supportive. This dissertation is lovingly dedicated to them.