Kiduk Yang

Differences between respondents and nonrespondents in a multicenter community-based study vary by gender and ethnicity☆

Abstract This study provides data on differences between respondents and nonrespondents by gender and ethnicity in a multicenter community-based study that is rarely collected and that may be useful for estimating bias in prevalence estimates in other studies. Demographic, general health, and cardiovascular risk factors were examined in black and white respondents and nonrespondents to the Atherosclerosis Risk in Communities (ARIC) Study, a prospective study investigating cardiovascular risk factors in approximately 16,000 adults that was initiated in 1986 in four U.S. communities. In one of the communities (Jackson, MS) black participants were recruited exclusively; in another (Forsyth County, NC) 12% of the eligible sample were black, whereas the samples in Washington County, MD and the northwestern suburbs of Minneapolis, MN were almost all white. Demographic and health characteristics were collected during a home interview. Subjects who subsequently agreed to complete a clinical examination were defined as respondents, while eligible participants who only took part in the home interview were considered to be nonrespondents. Approxmately 75% of age-eligible individals (45–64 years) in each community completed the home interview. In three of the communities 86–88% of those who took part in the home interview also completed the clinic examination, whereas only 63% did so in Jackson. Among white participants, response rates were similar in men and women and between communities. Among black participants, the response rates were considerably lower, particularly in men. White male respondents reported a higher socioeconomic status, better general health and a lower prevalence of cardiovascular disease and associated risk factors than white male nonrespondents. The difference between white respondents and nonrespondents were greater for men than women despite similar response rates. Among black participants, respondent/nonrespondent differences were usually of smaller magnitude or absent, particularly in women. General health status and recent hospitalization rates were almost identical in black respondents and nonrespondents. Low response rates can bias estimates of prevalence in community-based studies although differences between respondents and nonrespondents tend to exaggerate real differences beween respondents and the eligible population sampled. For example, among white males 25% of respondents and 44% of nonrespondents were current smokers, yet the estimated community prevalence of smoking was 31%. In conclusion, differences observed between respondents and nonrespondents were in the expected direction, but were greater for men than women and for whites than blacks.

Citation Analysis: A Comparison of Google Scholar, Scopus, and Web of Science

When faculty members are evaluated, they are judged in part by the impact and quality of their scholarly publications. While all academic institutions look to publication counts and venues as well as the subjective opinions of peers, many hiring, tenure, and promotion committees also rely on citation analysis to obtain a more objective assessment of an author’s work. Consequently, faculty members try to identify as many citations to their published works as possible to provide a comprehensive assessment of their publication impact on the scholarly and professional communities. The Institute for Scientific Information’s (ISI) citation databases, which are widely used as a starting point if not the only source for locating citations, have several limitations that may leave gaps in the coverage of citations to an author’s work. This paper presents a case study comparing citations found in Scopus and Google Scholar with those found in Web of Science (the portal used to search the three ISI citation databases) for items published by two Library and Information Science full-time faculty members. In addition, the paper presents a brief overview of a prototype system called CiteSearch, which analyzes combined data from multiple citation databases to produce citation-based quality evaluation measures.

Information retrieval on the web

Introduction How do we find information on the Web? Although information on the Web is distributed and decentralized, the Web can be viewed as a single, virtual document collection. In that regard, the fundamental questions and approaches of traditional information retrieval (IR) research (e.g., term weighting, query expansion) are likely to be relevant in Web document retrieva1.l Findings from traditional IR research, however, may not always be applicable in a Web setting. The Web document collection-massive in size and diverse in content, format, purpose, and quality-challenges the validity of previous research findings that are based on relatively small and homogeneous test collections. Moreover, some traditional IR approaches, although applicable in theory, may be impossible or impractical to implement in a Web setting. For instance, the size, distribution, and dynamic nature of Web information make it extremely difficult to construct a complete and up-to-date data representation of the kind required for a model IR system. To further complicate matters, information seeking on the Web is diverse in character and unpredictable in nature. Web searchers come from all walks of life and are motivated by many kinds of information needs. The wide range of experience, knowledge, motivation, and purpose means that searchers can express diverse types of information needs in a wide variety of ways with differing criteria for satisfying those needs. Conventional evaluation measures, such as precision and recall, may no longer be appropriate for Web IR, where a representative test collection is all but impossible to construct. Finding information on the Web creates many new challenges for, and exacerbates some old problems in, IR research. At the same time, the Web is rich in new types of information not present in most IR test collections. Hyperlinks, usage statistics, document markup tags, and collections of topic hierarchies such as Yahoo! (http://www.yahoo.com) present an opportunity to leverage Web-specific document characteristics in novel ways that go beyond the term-based retrieval framework of traditional IR. Consequently, researchers in Web IR have reexamined the findings from traditional IR research to discover which conventional

Design and empirical evaluation of search software for legal professionals on the WWW

Abstract Our research focuses on designing effective search aids for legal researchers interested in law-related information on the world wide web. In this paper we report on the design and evaluation of two software systems developed to explore models for browsing and searching across a user-selected set of WWW sites. A directory services tool, LIBClient, provides a hierarchical index of legal information resources in an interface emphasizing ease-of-use by Internet novices and management of multiple-site searching. To study the relative effectiveness of LIBClient in the hands of legal professionals, nineteen law students were observed using LIBClient and, in separate trials, the popular general-purpose search services to perform known-item searches within a fixed time limit. The experiment indicates the value of LIBClient for focused searching, most properly as a supplement to general-purpose search engines. Motivated by observations from the LIBClient study, a second retrieval experiment explores the effectiveness of a radically different LIBClient design in which the LIBClient interface is combined with a crawler-enhanced search engine, IRISWeb. The LIBClient–IRISWeb system enables full-text searching using natural language queries across a set of WWW pages collected by the IRISWeb crawler. The page harvesting process relies on a cascading set of filters to define the final set of WWW pages to be collected, including user selections in LIBClient, search results from site-specific search engines, and the hyperlink structure at target sites. To evaluate the LIBClient–IRISWeb method, the queries used in the user study are submitted to the system, with excellent retrieval results. In conclusion, our research points to the promise of WWW search tool designs that tightly couple directed browsing with query-based search capabilities using new forms of search automation.

Analysis of publication patterns in Korean library and information science research

This study assessed research patterns and trends of library and information science (LIS) in Korea by applying bibliometric analysis to 159 Korean LIS professors’ 2,401 peer-reviewed publications published between 2001 and 2010. Bibliometric analysis of publication data found an increasing trend for collaboration, robust publication patterns, increasing number of international publications, and internationalization of LIS in Korea. The maturation and internalization of LIS research was evidenced in increased number of publications in high impact journals (e.g., SSI, SSCI), growing participation in leading international conferences (e.g., ASIST, TREC), increasing proportion of Korean LIS faculty with international degrees, and high publication rates by professors with international degrees. Though limited in its evaluative power without citation data, publication data can be a rich source for bibliometric analysis as this study has shown. The analysis of publication patterns conducted by the study, which is a first step in our aim to establish a multi-faceted approach for assessing the impact of scholarly work, will be followed up in a future study, where the question of quantity versus quality will be examined by comparing publication counts with citation counts.

An interactive WWW search engine for user-defined collections

Given the dynamic nature and the quantity of information on the WWW, many individual users and organizations compile and use focused WWW resource lists related to a particular topic or subject domain. The IRISWeb system extends this concept such that any user-defined set of WWW pages (a virtual collection) can be retrieved, indexed, and searched using a powerful full-text search engine with a relevance-feedback interface. This capability adds full-text searching to highly customized subsets of the WWW. Here we describe the IRISWeb software and an experiment that highlights its potential.

Passage feedback with IRIS

We compare a user-defined passage feedback (pf) system to a document feedback (df) system. Df employed the adaptive linear model for retrieval, while pf used weighted query expansion based on positive and negative feedback. Twenty-four searchers performed the same six tasks in varying search and system-order per TREC-8 guidelines. We hypothesized that pf, which featured interactive query expansion, would outperform df, which relied on automatic query expansion. Initial analysis appeared to reject this hypothesis, as df showed slightly higher overall performance than pf. However, analysis by system-order groups indicates only the first pf use had lower performance. These data suggest that pf was more difficult to learn than df, though the second pf use yielded competitive performance. If performance of pf is indeed affected by learning, an improved pf system with usability enhancements may prove to be an effective mechanism for interactive information retrieval.

WIDIT: fusion-based approach to web search optimization

To facilitate both the understanding and the discovery of information, we need to utilize multiple sources of evidence, integrate a variety of methodologies, and combine human capabilities with those of the machine. The Web Information Discovery Integrated Tool (WIDIT) Laboratory at the School of Library and Information Science, Indiana University-Bloomington, houses several projects that employ this idea of multi-level fusion in the areas of information retrieval and knowledge discovery. This paper describes a Web search optimization study by the TREC research group of WIDIT, who explores a fusion-based approach to enhancing retrieval performance on the Web. In the study, we employed both static and dynamic tuning methods to optimize the fusion formula that combines multiple sources of evidence. By static tuning, we refer to the typical stepwise tuning of system parameters based on training data. “Dynamic tuning”, the key idea of which is to combine the human intelligence, especially pattern recognition ability, with the computational power of the machine, involves an interactive system tuning process that facilitates fine-tuning of the system parameters based on the cognitive analysis of immediate system feedback. The rest of the paper is organized as follows. The next section discusses related work in Web information retrieval (IR). Section 3 details the WIDIT approach to Web IR, followed by the description of our experiment using the TREC .gov data in section 4 and the discussion of results in section 5.

A Bibliometric Analysis of Faculty Research Performance Assessment Methods

Effective assessment of faculty research performance should involve considerations of both quality and quantity of faculty research. This study analyzed methods for evaluating faculty research output by comparing the rankings of Library and Information Science(LIS) faculty by publication counts, citation counts, and research performance assessment guidelines employed by Korean universities. The study results indicated that faculty rankings based on publication counts to be significantly different from those based on citation counts. Additionally, faculty rankings measured by university guidelines showed bigger correlations with rankings based on publication counts than rankings by citation counts, while differences in universities guidelines did not significantly affect the faculty rankings. The study findings suggest the need for bibliometric indicators that reflect the quality as well as the quantity of research output.

CiteSearch: next-generation citation analysis

The coverage of citations in citation databases of today is disjoint and incomplete, which can result in conflicting quality assessment outcomes across different data sources. Fusion approach to quality assessment that employs a range of citation-based methods to analyze data from multiple sources is one way to address this limitation. The paper discusses a citation analysis pilot study that measured the impact of scholarly publications based on the data mined from Web of Science, Scopus, and Google Scholar.