Application Notes and Form (excludes financial section)
Context. An increasingly wide range of people need to find information on the
Internet. One would think that variations in the strategies people use to
search for information are due to differences in the particular individual's
needs and the nature of the topic on which searches are being conducted.
Whilst undoubtedly this is true, there is cumulating empirical evidence that
less logical factors are also at play. Recent studies at Sheffield University
suggest that differences in search strategies, effectiveness of searching,
satisfaction with the results of searches, choice of search tool, and choice
of keywords are significantly linked with differences in cognitive styles
(i.e. pervasive differences in the way different individuals process
information, learn and solve problems)2-6. These findings are broadly in
accord with other studies not restricted to information searching1. Cognitive
styles are likely to form a key building block in models used to guide the
development of improved education and training for Internet searching, the
development of adaptive information retrieval systems, and the development of
more effective help systems.
Aims/objectives.
Aim: To provide a detailed mapping of the effects of different cognitive
styles on information seeking strategies and effectiveness in an Internet
context, in order to facilitate the future use of this important building
block in developing models capable of improving both systems and education to
support effective information seeking on the Internet.
Objectives:
1. Develop a model of cognitive styles and associated linguistic factors
influencing the effectiveness of Internet information seeking.
2. Develop guidelines (based on the model) for the development of (i) improved
education and training for Internet searching; (ii) the development of
adaptive information retrieval systems; (iii) the development of more
effective help systems.
In relation to the second objective, there is evidence that education and
training can be made more effective by accommodating cognitive styles1, and that models of cognitive style effects can be used in the development of adaptive information systems7.
Methodology.
Approach: The research will pursue these objectives by:
1. Testing a series of hypotheses deduced from existing theory, using a quantitative approach (correlation and factor analysis). Relevant existing theory from which hypotheses have been drawn may be found in the following categories of research. An example hypothesis is given in italic for each category.
a) Research undertaken specifically in relation to the Internet on individual differences in Internet search behaviour and perceptions. Example hypothesis: Individuals with a "verbaliser" (as opposed to "imager") cognitive style may be more prone to disorientation when searching on the Internet3.
b) Research into information retrieval generally. Example hypothesis: Individuals with a "global" (as opposed to "analytic") cognitive style tend to make less use of analytic search techniques, and tend to perform broader searches5,6
c). Research into relatively fundamental influencing factors underpinning a range of cognitive activities, which may also be relevant to the new phenomenon of Internet searching (e.g. work on the effects of cognitive styles on information processing across a wide range of cognitive activities1). Example hypothesis: Imagers may be less "linguistically flexible" (i.e. less skilled at generating - as opposed to recognising - alternative verbal concept descriptions) than verbalisers.
2. Generating new theory using a Grounded Theory-based analysis of qualitative interview and "think aloud" data. Volunteers will be interviewed in order a) to illuminate the phenomena identified in quantitative studies (e.g. why do individuals with a global cognitive style tend to perform broader searches?), and b) to discover new themes of importance to people in their Internet searching - themes which, once identified qualitatively, may become available to quantitative testing.
3. Integrating the views provided from the different approaches to form a unified model.
4. Generating guidelines for harnessing the model to improve the development of education and training, adaptive information systems and help systems.
Research staff: Research staff will have appropriate Internet and searching skills. One will have skills particularly in quantitative data collection/analysis, the other with qualitative (interview/"think aloud") data collection/analysis.
Procedures: A minimum of 400 carefully designed experimental Internet searches will be conducted (four searches each by a minimum of 100 volunteers). The large number of experimental searches is made possible by the use of largely automated data gathering techniques (already under test). In addition, approximately 50 volunteers will participate in in-depth interviews and "think aloud" recording during searching.
Sample. The sample will consist of two years' intake (this 1 year project will span 2 years' intake) onto the Masters programmes in librarianship and information management in the Department of Information Studies at the University of Sheffield. This sample will provide a good mix of humanities, social science and science graduates.
Measures:
1. Individual characteristics: a) Subject background; age and gender. b) Cognitive style (specifically, analytic/global, holist/serialist and verbaliser/visualiser styles). c) Linguistic skills. A series of measures will be used, reflecting a range of linguistic skills and flexibility relevant to information seeking. These measures are currently being developed and tested. d) Internet perceptions and attitudes. The work will develop an existing measure previously used in assessing differences in search behaviour and attitudes to the Internet3.
2. Searching activity: a) Search strategies will be recorded for analysis. Recorded elements will include linguistic features (terms, concepts, combinations and strategies used by the volunteers, and changes in these in response to evolving search results). b) "Think aloud" protocols for 50 of the volunteers during searching. c) Data from interviews conducted before and after the searches. These data will be used to gain understanding of subjects' reasoning, perceptions and reactions before, during and after searching.
3. Search effectiveness: The measures have been designed to take account of the difficulty in assessing effectiveness in what is a dynamically changing information environment. The effectiveness of a minimum of 200 searches will be assessed using the following
scheme of research, continuedmeasures for each search:a) Precision: the number of relevant items found within the first N retrieved items (an appropriate N to be determined for each search). b) Comparison with search results obtained by a trained intermediary. c) Evaluation by a trained intermediary of (i) the volunteers' search strategies; (ii) the search results produced by the volunteers.
d) Volunteers' levels of satisfaction with the search.
References
1. FORD, N. (1995) Levels and types of mediation in instructional systems: an individual differences approach. International Journal of Human-Computer Studies, 43, 241-259.
2. FORD, N. and FORD, R. (1993) Towards a cognitive theory of information accessing: an empirical study. Information Processing and Management 29, 569-585.
3. FORD, N. and MILLER, D. (1996) Gender differences in Internet perceptions and use. Aslib Proceedings, 48, 183-192.
4. FORD, N., WILSON, T.D., ELLIS, D., FOSTER, A. and SPINK, A. (in progress) Individual differences in information seeking: an empirical study. JASIS.
5. WOOD, F., FORD, N., MILLER, D., SOBCZYK, G. and DUFFIN, R. (1996) Information skills, searching behaviour and cognitive styles for student centred learning: a computer-assisted learning approach. Journal of Information Science, 22 (2), 79-92.
6. WOOD, F., FORD, N. and WALSH, C. (1992) Online Searching and Cognitive Styles. Final Report to the British Library. London: British Library.
7. FORD, N., LLOYD-WILLIAMS, M. & MORRIS, C. (1995) Intelligent adaptive systems for information delivery in the electronic library. In: Mel Collier & Kathryn Arnold (eds.) ELVIRA 2 [conference], Milton Keynes, May 1995. London: Aslib. pp. 137-143.10. OutputIs the primary product of your research intended to be (please tick one or more, as appropriate)(a) monograph, critical edition, or article(s)? ((b) creation of an electronic data, sound or image resource (see also below)(c) performance, film, audio-tape, broadcast(d) exhibition(e) other (please specify ) Electronic resource: if the primary product of the research will be an electronic resource, have you obtained guidance from the Arts and Humanities Data Service on appropriate standards and methods? yesnoPlease describe the proposed output from the research, and outline your plans for publication or public output of the research for which you are seeking an award
Deliverables
1. Model of cognitive styles and associated linguistic factors affecting the effectiveness of Internet information searching.
2. Guidelines for training and the development of adaptive information systems and help systems based on the model.
3. Recommendations for further research.
Dissemination: Findings and recommendations from the study will be widely disseminated via:
Papers in international academic and professional journals; conference presentations; a dedicated Web site; postings to relevant newsgroups, discussion lists and mailing lists; dissemination to schools of librarianship and information management; dissemination to developers of Internet search facilities; dissemination to Internet training organisations.
11. Duration of Research Project Duration of current proposal (max 5 years): start (dd/mm/yy)1 March 1999end (dd/mm/yy)31 March 2000Duration of entire research project: start (dd/mm/yy)1 March 1999end (dd/mm/yy)31 March 2000
12. Budget- EXCLUDED