Journal of the Association for Information Science

 

Bert R. Boyce
 

885
 

 RESEARCH

 

Motivations for Hyperlinking in Scholarly Electronic Articles: A Qualitative Study
Hak Joon Kim

It is easy to assume that hyperlinks in electronic journals are the equivalent of citations and may be so treated in literature analysis. Kim investigates the motivations of scholarly hyperlinking in electronic publications in order to examine this assumption. Using faculty and graduate students at Indiana University who had published at least one peer reviewed article with at least one hyperlink in it in an electronic only publication after 1995, in depth open ended semi-structured interviews were conducted. The interviewer brought a hard copy of the articles and first pages of linked articles as a basis for discussion, solicited the motivation for each link and transcribed the audiotape of the discussion. Nineteen categories of reasons were identified, and three coders achieved an agreement coefficient of 83.6%. These were grouped into scholarly, social, and technological reasons.

The provision of additional or background information was the most common purpose, and the provision of an example or illustration the second of twelve scholarly motivations. Of five social motivations, publicity for an information source, and credit to an author or institution were the most common. In the technological category, provision of an easy access mechanism was most common but a second motivation was simply the possibility of providing such a link. On the average each link has over two motivations. The social and scholarly motives are common with those for citation but more than one-third of the links were motivated at least partially by technological reasons where there is no commonality. Hyperlink counts as a measure of quality are suspect due to the complex nature of motivations for their use, however, most practices are grounded in conventional citation practices.
 

887

 

 

 

 


 

 

 

 

 

 


 

 

Narratives of New Media in Scottish Households: The Evolution of a
Framework of Inquiry
Elisabeth Davenport, Martin Higgins, and Ian Somerville

Davenport, Higgins, and Somerville attempt to investigate the way information systems (PCs and cable TV) intervene in the social group of the household in order to characterize the managing, spending, consuming and sharing of resources therein. Using questions on activities, understanding of technology, manipulation, and meaningfulness, pilot households were interviewed and narratives were stimulated that provided information that adults focused on job oriented technology and children view the devices primarily for entertainment. In phase 2, 26 households were interviewed, invited to tell stories and anecdotes as well as provide directed opinions and responses. Equipment acquisitions were joint household decisions and tended to be opportunistic rather than carefully researched. Computers tended to be purchased with some long term goal in mind while TVs were more often an instant gratification purchase. Cable subscription was a household
decision but use is male and sports dominated. TVs were in front rooms and bedrooms, computers were more likely to have their own space and often hidden from general view. TV via teletext was used to track shares and purchase holidays but not to shop. VCRs had time shifting and local scheduling advantages. Equipment users used the library and read newspapers. Interest in information was not shifting from one medium to another but expanding to find time for more. Identified narratives included: Men control viewing, children are better at new technology, men watch sports, women soaps, TV is anti-social and trivial, computers are serious. However, children view TV as a social currency and computers as entertainment machines.
 

900

 

 

 

 

 

 

 

 

 

 

 

 

 

Shifts of Focus on Various Aspects of User Information Problems During
Interactive Information Retrieval
David Robins

Robins reviews current models of the interactive retrieval process as a prelude to examining focus shifts as they occur in mediated retrieval. Using the transcribed discourse between 20 users and their intermediaries, collected earlier by Saracevic and Su, he identifies shifts (changes in conversational focus), classifies them by type and function and counts both the shifts and the utterances in each shift. In the 20 interactions 1439 shifts occurred, 1084 during the online portion of the search. Considerably more time was spent in interaction with the system than with pre-search modeling. Topic and user focus constituted only 15% of all foci indicating low attention to user problem modeling, at least as determined by discourse analysis. Search intermediaries initiated two thirds of all shifts. Shifts occur rapidly and even chaotically. The concentration is on strategy and evaluation with only moderate evidence of changing problem conception.
 

913
 

 

 

 

 

 

Users' Perception of Relevance of Spoken Documents
Tassos Tombros and Fabio Crestani

Telephone access to retrieval systems involves voice recognition as well as efficient database searching in a noisy interactive environment. Tombros and Crestani look at user perception of the relevance of document summaries presented vocally in the context of the Sonification of an Information Retrieval Environment, or SIRE project at the University of Glasgow. SIRE interactively converts telephone speech to query text for a probabilistic retrieval system, that then converts high ranked results to vocalized summaries for relevance judgment by the user. Summarizations ranked sentences by position, field, and high score based on the number of high document frequency words present, and a length normalized query score based on query terms present. The smaller of 15% or the five top sentences in original order were then presented.

Ten subjects judged relevance from direct human read descriptions, descriptions read by a human over a telephone, and descriptions read by a speech application while situational variables are controlled. Answer sheets as to relevance of the top 50 documents, evaluated in a five-minute period for each query, and a questionnaire on utility and clarity of the treatment were utilized. Elapsed time, precision (relevant & identified / total indicated as relevant), and recall (relevant & identified / number examined & relevant) are used as measures. Time increases, and recall decreases as one moves down the continuum from screen views to artificial voice. Natural voice on the telephone brings higher precision than direct voice or synthesized voice on the telephone. The difference between times for the various treatments is statistically significant, and time spent on judgment declines with the later queries.
 

929

 

 

 

 

 

 

 

 

 

 

 


 

 

Impact of Prior Electronic Publication on Manuscript Consideration
Policies of Scholarly Journals
Stephen P. Harter and Taemin Kim Park

Using the source journals of the Institute for Scientific Information, Harter and Park selected four sub-populations, mathematics, physics, psychology, and education, based on their early activity in web publishing. The journals were sorted by an impact factor limited to the number of citations received in a year to papers published in the previous two years. The top and bottom 20 journals in each field were selected plus 42 random journals from Arts and Humanities Citation Index. The editors of these 202 journals were identified and surveyed as to policies on acceptance of items existing on the Internet and on factors that would influence a decision in this area achieving a response rate of 57%. Most editors had no formal policy. Most would consider papers already electronically available in something like preprint form, but only one-fourth would consider a paper previously published in an electronic journal. Fewer than half would consider a paper previously published in print. Impact factor and location have little effect on these decisions. Editors and fields of study disagree on the factors that control such decisions. Arts and humanities editors are less likely to consider publication of material previously available in electronic form.
 

940
 

 


 

 

 

 

 

 

 

 

End User Searching on the Internet: An Analysis of Term Pair Topics
Submitted to the Excite Search Engine
Nancy C. M. Ross and Dietmar Wolfram

Ross and Wolfram study a database containing 1,025,910 Excite queries from a single day each containing a searcher identifier, times, and the full query statement. Identical queries from the same machine in succession were included only once resulting in 363,282 unique queries. Terms per query, term frequency distribution, co-occurrence frequencies, queries per identifier, and pages visited per query were computed. The 1,054 most frequently occurring term pairs (from 6,116 to 45) were analyzed using NUD-IST software for indexing and grouping, resulting in categories to apply to the term pairs. Using the number of queries in which pairs occurred to apply weights to category overlaps, a matrix was created for cluster analysis and multidimensional scaling. Thirty categories were produced with 85 pairs unclassified. Using Ward's amalgamation rule with either Pearson's r or the City block method produced clear and meaningful clusters. The largest cluster (about 26%) concerns adult topics, the second, computer oriented pursuits, while the others breaking out at a lower linkage distance represent education, products and services, formal
reference, and popular culture. For MDS Pearson's coefficients were calculated for each category pair, and a three-dimensional solution adopted for queries from personal to institutional, general to specific, and leisure to formal.
 

949
 

 

 

 

 

 

 

 

 


 

 

In Memory of Belver C. Griffith
Howard D. White and Katherine W. McCain
 

959

 

 BOOK REVIEWS

 

U.S. Government on the Web: Getting the Information You Need, by Peter Hernon, John A. Shuler, and Robert E. Dugan
Mike Steckel
 

963


 

 

Information Seeking in the Online Age: Principles and Practice, by Andrew Large, Lucy A. Tedd, and R.J. Hartley
Ethelene Whitmire
 

964


 

 

Web Style Guide: Basic Design Principles for Creating Web Sites, by
Patrick J. Lynch and Sarah Horton
Terrence A. Brooks
 

964


 

 

Developer's Guide to the Java Web Server: Building Effective and
Scalable Server-Side Applications, by Dan Woods, Larne Pekowsky, and Tom Snee
Pascal V. Calarco
 

965

 

 

 

The Clock of the Long Now: Time and Responsibility, by Stewart Brand
Jo Ann Oravec
 

966

 

 LETTERS TO THE EDITOR

 

Incremental Benefit of Human Indexing
Jian Qin

968
 


2000 , Association for Information Science