JASIST IndexJASIST Table of Contents

Journal of the Association for Information Science and Technology

 EDITORIAL

 

In This Issue
   
Bert R. Boyce
 

255

 

 RESEARCH

 

Epistemology and the Socio-Cognitive Perspective in Information Science
    Biger Hjorland
    Published online 18 January 2002

In this issue Hjorland examines the Socio-Cognitive perspective, or as he also calls it, the domain-analytic approach, which builds on the premise that all aspects of information study are shaped in what he terms ``discourse communities,'' in which ordered and bounded communication takes place so that the focus of information science becomes social and cultural rather than individual or computational, and the relevant cognitive structures become historical in nature rather than psychological as they would be in a traditional cognitive science approach.

In domain analysis one chooses some bounded field of knowledge at a chosen level of specialization and studies the nature and structure of communication within the sociological group developing the field. Mental models of the literature of the subject are created by attempting to understand its social and cognitive organization, its language and research methods, its patterns of cognitive authority, and its patterns of secondary and tertiary document production. Since the domain analyst is interested in mental models that are held by sociological groups, the terms paradigms and epistemologies are more often used. Groups with differing epistemologies will have differing views of what is relevant on a topic. Hjorland shows by citation patterns that four paradigms in psychology use very distinct information sources in their work, indicating that information needs are to some extent determined by paradigm, and suggesting the importance of applying epistomologic theory to information systems.
 

257
 

 

 

 

 

 

 

 

 

 


 

 

 

Sampling and Concentration Values of Incomplete Bibliographies
    L. Egghe
    Published online 10 January 2002

Egghe addresses the problem of the inherent incompleteness of subject literatures as it applies to their characterization by the bibliometric distributions which depend upon the unequal distribution of items across sources. Since one will always practically be dealing with a sample of some sort, except perhaps for the unusual case of a closed dead literature, the effect of various sampling techniques on concentration of items is of interest. If one, starting with those in the least productive sources, uses systematic samples, or only the core of the bibliography, to choose items, the sampled bibliography, as measured by the relative position of their Lorenz curves, will be more concentrated than the original. If one chooses a sample of sources in the same manner the concentration remains greater in the sample. We can thus assume that the inequality present in a complete bibliography is systematically overestimated.
 

271

 

 

 

 

 

 

 

 

Research Anxiety and Students' Perceptions of Research: An Experiment. Part I. Effect of Teaching Kuhlthau's ISP Model
    Jacqueline Kracker
    Published online 18 January 2002

Using the 90 students in four sections of a class in technical writing who were assigned a 15 page minimum paper requiring research over an eight week period, Kracker tests for awareness of the cognitive aspects of the research process, awareness of the affective aspects of the research process, anxiety associated with the research process, and satisfaction with the research process in pre-tests, interim tests after a 30 minute presentation on Kuhlthau's model or a placebo presentation, and post- tests after completion of the assignment. In the final analysis pool 19 control and 20 experimental subjects remained. Mean differences for the control group between pre and post-tests were small and sometimes negative for cognitive, affective, and satisfaction variables, while the experimental group showed a less than significant gain of between a half point and over a point in these variables. The experimental group experienced a significantly larger drop in anxiety level than the control group indicating that presentation of the model will reduce anxiety but not effect other aspects studied.
 

282

 

 

 

 

 

 

 

 

 

 

Research Anxiety and Students' Perceptions of Research: An Experiment. Part II. Content Analysis of Their Writings on Two Experiences
    Jacqueline Kracker and Peiling Wang
    Published online 22 January 2002

In a second paper, Kracker and Wang, using the 90 students in four sections of a class in technical writing who were assigned a 15-page minimum paper requiring research over an eight week period, test for awareness of the cognitive aspects of the research process, awareness of the affective aspects of the research process, anxiety associated with the research process, and satisfaction with the research process in pre-tests, interim tests after a 30-minute presentation on Kuhlthau's model or a placebo presentation, and post- tests after completion of the assignment. Students were asked to write about a past memorable research experience at the beginning of the semester and the current project at the end. An analysis of words in the 76 students responses yielded 122 distinct words signifying emotions grouped into 25 categories. NUD*IST was used to assist in coding and aggregation. Three major dimensions were apparent in the categorized words: emotional states, perceptions of the process, and affinity to research. The feelings described in Kuhlthau's model are confirmed to be present, and anxiety is experienced across the research process, not just in library usage.
   

295

 

 

 

 

 

 

 

 


 

 

Methods for Measuring Search Engine Performance over Time
    Judit Bar-Ilan
    Published online 25 January 2002

Finally Bar-Ilan looks at search engine evaluation. Search engine databases change over time as material indexed is dropped and replaced by indexes of other material and perhaps later by the original material again. This, along with the changing nature of the material on the web, makes replications of past search results unlikely. Bar-Ilan believes that evaluation of such engines must include some measure of their searching stability over time. For this purpose he conducts monthly rounds of the same sets of search strategies during low use periods. How to choose a true random set of queries representing total web usage is not clear, and one must collect and likely download all search results since the reported document counts are less than reliable.

A document is defined as technically relevant if it meets the conditions posed by the query, and technical precision is the percentage of technically relevant and retrieved documents out of the total retrieval. A document is considered to be technically relevant throughout a study if it was technically relevant in the first round of searches. The number of technically relevant and retrieved documents from one engine divided by that number from all engines is relative coverage per round, while total relevant coverage considers all rounds and may be considered to estimate coverage. A count of new URLs for an engine is of those retrieved it did not retrieve in any previous round, and totally new would be a URL none of the engines retrieved in a previous round. New estimates the rate of adding new material, totally new the growth of the topic on the web. Forgotten measures the number of active URLs that disappear in a round, and can reflect all or a single search engine. Lost URLs are forgotten URLs whose content does not appear under a new URL in the round, while recovered URLs are those that return in later rounds. Well handled URLs are continuously retrieved or removed as a result of a site's disappearance. The remainder are mishandled and either recovered or disappeared. Self overlap, relevant and retrieved in a round versus in the whole period will reflect variability over time. Number of rounds in which a URL is retrieved, and number of persistent URLs, those in all rounds, also relate to stability.

Using a single word search, ``aporocactus'' Bar-Ilan searched 39 times over a year period early Sunday morning using Alta Vista, Excite, Fast, HotBot, Google, and Northern Light. All systems showed general growth over time with two major fluctuations in Google. The technical precision measure was between 76 and 99% which reflected a small number of unreachable and unfound documents. Google covered 70% of the results. Google and HotBot discover more new URLs more quickly. Even the most stable engines mishandled over 33% of URLs. Alta Vista retrieved 58% of its URLs seven or more times, while Google retrieved 78% of its URLs four times or less.
 

308

 

 

 

 

 

 

 

 

 

 


 

 

 

 

 

 

 


 

 

BOOK REVIEWS

 

The Social Life of Information, by John Seely Brown and Paul Duguid
    An-Chi Hoh Dianu
    Published online 9 January 2002
 

320
 

 

 

Intelligent Technologies in Library and Information Service Applications, by F.W. Lancaster and Amy Warner
    Ina Fourie
    Published online 9 January 2002
 

321



 

 

LETTERS TO THE EDITOR

323
 


ASIST Home Page

Association for Information Science and Technology
8555 16th Street, Suite 850, Silver Spring, Maryland 20910, USA
Tel. 301-495-0900, Fax: 301-495-0810 | E-mail:
asis@asis.org

Copyright © 2001, Association for Information Science and Technology