ALISE Conference 2010 Doctoral Poster Sessions
Title
Abstract
The conceptual framework that I develop for this study is a taxonomy of the teaching of stand-alone professional ethics courses in business, engineering, law, LIS, and medicine. In this taxonomy, I examine the literature and syllabi to review how ethics is taught in these five professional disciplines. This summary provides a perspective which shapes the questions I pose to subjects and guides my data collection, analysis, and synthesis and informs my conclusions. The findings of the taxonomy indicate that LIS stand-alone ethics courses provide instruction comparable that in other fields. They also suggest that teachers considering course revision in LIS ethics may want to expand their approach to conflict of interest and diversify their evaluation strategies using self and peer-review methods. I employ the qualitative research tradition with sequential data gathering. I select the interviewees and focus-group participants based on an exemplary study strategy – then, I categorize data using Lincoln and Guba’s modified version of the constant comparative coding method to devise, integrate and delimit the categories obtained from the transcripts. Finally, I check my findings by using criteria of trustworthiness (including member checking, peer debriefing and negative case analysis).
My findings and their implications fall into three main areas: First, there are major challenges of integrating ethics for both faculty and students . Teachers report both a dearth of pedagogical approaches and faculties’ lack of training/experience in ethics; these imply an obligation of ethics-oriented faculty to train others. LIS schools report the challenge of sequencing courses. Both teachers and administrators indicate that repetition is desirable in a complex topic such as ethics. Second, my research identifies integration strategies used in specific core courses. These focus on the usefulness of ethics-oriented learning objectives, the diversity of teaching techniques and the consistency of teachers employing the codes of ethics. Third, my research reports processes that schools have employed to initiate, implement and evaluate ethics on an institution-wide basis. These processes can help institutions enhance their emphasis on ethics.
This research is useful to LIS program administrators and teachers by providing insight into both major challenges and successful strategies for integrating ethics across the curriculum.
Title
Abstract
The Extended Case Method (Burawoy, 1991; 1998) guided my data collection and analysis to test theory against what I found in the field. I used participant observation, interviews, and document analysis to collect the data. Overcoming language and cultural barriers is important for this type of research and I therefore hired students of Library Science from the University of Namibia to help me. I used three theories to frame my research: World Society Theory (Meyer et al. 1997), New Institutional Theory (Powell and DiMaggio, 1991) and Post-Colonial Theory (Fanon, 1963). World Society Theory was developed from evidence of similarities in governmental, health and educational organizations globally that demonstrates the growth of a world culture. The theory supposes that organizations will be similar around the world despite differences in local resources and traditions. My findings support the idea that public libraries are built on a world-wide model, but show an important deviation from the theory which is local resources do have a significant impact on the services they can provide.
New Institutional Theory helps explain how organizations become similar, as well, as framing the national forces acting on organizations through the notion of an organizational field. Public libraries in Namibia have a strong history of being Whites-only and developed to cater to the pleasure reading habits of the elites. New community libraries are being established counter to this organizational form, and are attempting to follow international models of what public libraries should be. These changes may show the continuing institutionalization process of the library organizational form in Namibia.
Post-Colonial Theory helps to clarify the forces at work in a newly independent nation and frames the issues of power in this research. Namibia became independent from South Africa in 1990, but ties to that nation remain strong. The transition and the power dynamics involved in the transition (Dobell, 1998) are mirrored in the institutionalization process of the public and community libraries.
This research gives new insights into how World Society Theory works with Post-Colonial Theory to understand the power and resource dimensions in the post-colony. These power and resource dimensions have direct consequences on the quality of and accessibility to libraries in Namibia, which then has an effect on how the libraries can impact society. World models of public libraries, mostly defined by the UNESCO Public Library Manifesto, promise that libraries will help bring democracy and peace to troubled nations. My findings suggest that libraries must be well-funded and well managed before they can live up to the Manifesto and its promises.
Title
Abstract The data collection was conducted in four rounds and included the following: sampling a seed set of 11 articles, identifying a set of keywords and key phrases, performing the search via the Internet, specialized databases, and library catalogs, and following citations. Overall, 157 documents were collected and analyzed. The analysis of documents involved a combination of discourse and content analysis. The discourse analysis involved close reading of documents in relation to the context of their production (van Dijk, 1993, 1988; Fairclough, 1992, 1995). Content analysis techniques (Weber, 1985) were used to specify and quantify temporal, formal, and structural characteristics of the documents and their producers.
The theoretical framework employed in this research is a synthesis of several critically oriented theoretical concepts, namely the concepts of communicative rationality, power-knowledge, biopower, and the field (Bourdieu, 1981; Bourdieu & Wacquant, 1992; Foucault, 1980a, 1980b; Habermas, 1970, 1984). These concepts are used to conceptualize the processes of participation and knowledge production with regard to genetics and biology and address the issues of multiplicity of social institutions and actors involved in the production of science information.
The findings suggest that digital documentary forms play a significant role in the formation and organization of contemporary science communication. Web-based genres, such as blogs, forums, wikis, and so on, contribute to the stabilization and reinforcement of the existing understandings of life through excessive repetition, metaphorical language, and expert quotes. These genres are often described as participatory (Deuze, 2006), yet the analysis of documents about synthetic life shows that blogging and commenting do not contribute much to the debate about synthetic life. Despite common expectations, the information is being disseminated from the top down, i.e., from the sources of scientific information to mass media to the web. The dominance of scientific voice in the discussions about synthetic life and the relatively low participation of other social actors (e.g., educators, politicians, and citizens) create an imbalance at the very early stage of information dissemination.
Consistent with previous critical examinations of technoscience (Haraway, 1997; Nelkin & Lindee, 2004), this research shows that dominant understandings of life advanced by synthetic life discussions contribute to the revision of approaches to life according to the principles of control, experimentation, manipulation, and profit. It also shows that alternative approaches that offer a more complex and nuanced understanding exist in particular documentary forms, e.g., in the reviewing literature. The advancement of alternative approaches, however, may fail because the dominant discourse of life is justified by the references to progress, the advancement of knowledge, and the improvement of life.
Title
Abstract
Therefore, the purpose of the study is to examine the motivations and strategies of answerers as they provide information and social support for others. Two research questions have been proposed: 1) why do answerers participate and contribute in social Q&A? and 2) what strategies do they use to provide effective answers in social Q&A? For testing the research questions, the domain of health is chosen. Health is a critical issue in everyday life. People may not be able to consult with their doctors whenever they have questions. They may look for solutions through sources with easy access and be willing to rely on others experiences with the same problems. In some sense, this may be risky since relying on incorrect information could cause critical damage to their health. Nevertheless, health is one of the most popular topic categories in social Q&A, with a lot of traffic of both questions and answers.
An online survey and content analysis were chosen as research methods. Yahoo! Answers, one of the most popular social Q&A services in the United States, is chosen as the test bed of the study. Two types of answerers are selected as the research sample: 1) top answerers (who have contributed in generating many answers) and 2) recent answerers (who have recently posted answers for others). Under the Health category, a total of 1,000 answerers (including both top and recent answerers) will be selected. An online survey will be distributed to them via their email links embedded in the user profile pages of Yahoo! Answers. In addition to subjects responses on the survey, their answers will be collected with their permissions and used for content analysis of the types of information that they have provided facts, personal experiences, personal opinions, and encouragement as well as sources of information in answers.
In the current study, it is assumed that there are two important factors which influence the motivations and strategies: 1) level of knowledge and experiences in health, and 2) level of experience with answering questions. Thus, the relationship between the two factors and answerers motivations and strategies will also be tested. This study will be a significant endeavor in exploring a new emerging context of collaborative information seeking and sharing. This study will provide a basis for recommendations on how to promote peoples contribution on information and support generation in online environments.
Title
Abstract
The findings indicated that the participants’ leadership training facilitated the development of their self-perceived transformational leadership behaviors to a significantly higher level than the established national norms for the LPI in two areas - Modeling the Way and Enabling Others to Act. In addition, the assessment of leadership potential given during the program selection process had a positive correlation with the LPI subscale for Enabling Others to Act. Moreover, the social context of each participant’s circumstances had an impact on their self- perceived transformational leadership potential when considering the participants’ satisfaction with the support they received from their mentors, the amount of time they spent with their mentors, whether they selected or were assigned a mentor, their Graduate Record Exam scores, and the poverty level within their schools.
Title
Abstract
This research project revolves around three general research questions: (1) After more than a decade’s practice, what are the major trends in licensing work? (2) What are the biggest challenges of licensing work and how do electronic resources librarians deal with them? (3) How can LIS education better prepare future librarians for licensing-related work?
A grounded theory approach is used in this research project to explore these questions, with the aim of developing a substantive theory of licensing that describes and explains the major concern/themes in the field. Different from the positive research paradigm which emphasizes deductive logic and methods such as hypotheses and verification, the grounded theory approach applies inductive methods to generate theories (either substantive or formal) from empirical data. This approach is especially suitable for this research project because licensing is a relatively new research area that lacks theoretical and systematic analyses. Therefore, a good way to theorize licensing work is to systematically collect data from the field and use the collected data as the only source for understanding, explaining, and generalizing the major concerns in the field, instead of forcing data into preconceived theoretical frameworks that often lack of grounding in data.
Following this approach, the author collected data by interviewing electronic resources librarians from different academic institutions. Each interview was forty to sixty minutes in length, recorded on a digital recorder, and later transcribed manually. Interviews were semi-structured in nature, and the sample of participants was purposefully selected to include different types and sizes of academic libraries. The author coded the data openly and substantively, conceptualized the phenomena, constantly compared the concepts, and collapsed them into emerging categories. Eighteen interviews have been conducted and theoretically coded.
This research explains the licensing work in general, the challenges electronic resources librarians face daily, and how the librarians handle the challenges. It identifies the behaviors practitioners engage in as they cope with licensing-related work. Although licensing is still changing and expanding, the current major trends in licensing can be theorized as normalization (of licensing process) and standardization (of license terms). A major challenge of licensing work is relationship management: Licensing work is a service as well as a business full of tensions, and librarians employ multiple strategies to deal with the tensions. Licensing work is also a complicated decision-making process. Electronic resources librarians have to find the subtle balance between ideals and reality. Another major component of licensing work is self-education. Librarians often find themselves learning on the go due to the complicated legal and technological issues involved in licensing. Through data analysis, the research also discovers insufficient areas in the participants’ LIS education with regard to licensing, and provides suggestions to the design of e-resources management courses.
Title
Abstract
In the statewide chat reference consortium of this exploratory study, users can ask any information provider from any of the one-hundred and three participating information agencies and any consortium information provider can respond to questions from any user. This situation may potentially result in the assumption that for a location-based question, an information provider must determine the location or locations in the question to formulate a correct response, and because a local information provider is nearer and more related to a location or locations within their same county, a local information provider may provide a higher correct response fill rate to location-based questions than a non-local information provider.
This study’s methods utilize content analysis, quantitative analysis, focus groups, and unobtrusive testing. Data collection and data analysis of content analysis, focus groups, and unobtrusive testing provide qualitative data on the types of location-based questions and how information providers formulate responses to location-based questions. Measures of ratio and accuracy from quantitative analysis and unobtrusive testing provide quantitative data on the ratio of location-based questions, in total and by type, to total chat questions, the ratio of non-local information providers responding to location-based questions to total location-based questions responded to, and the correct response fill rate of location-based questions in a statewide chat reference consortium.
Findings produced from this studys methods will lead to practical recommendations that may improve information providers ability to provide correct responses to different types of location-based questions, as well as reduce barriers to participating information agencies in locating other information agencies location-based attributes, and allow chat software developers and chat consortia managers to mitigate some of the challenges associated with location-based questions by building geographic intelligence into their systems or creating features that allow users and information providers to disclose the locations that relate to formulating a correct response to location-based questions. Improving the ability of information providers from around the state to accurately respond to questions from users of all participating information agencies is vital to the success of this statewide consortium and other large chat reference consortia services.
Title
Abstract Before, during, and after the major campaigns by government, private industry, and mass media to get women into and out of the wartime economy workforce, the Whitman Authorized Editions reflected and modeled the lives of women in these stages. Women began as domestic paragons of good behavior and traditional feminine beauty, changed into women working in traditionally male jobs to maintain and defend the homefront, and then gracefully returned to the domestic sphere. Following the lead of Maureen Honey in her study, Creating Rosie the Riveter: Class, Gender, and Propaganda during World War II (1984), in which she points out the connection between American fiction written and published in magazines during World War II and propaganda story suggestions of the Office of War Information (OWI), and implementations of these suggestions by the Writers War Board (WWB), I suggest the congruence also takes place in popular children’s series literature for girls, even without formal directives from the OWI and WWB.
This is significant because story and reading matter. The power of story and the reading experience not only change peoples’ lives but also help people to make those changes. In Enchanted Hunters: The Power of Stories in Childhood (2009) Maria Tatar borrows the idea of a contact zone from Mary Louise Pratt. Both speak of the idea of colonialism, stating that when peoples of different geographic locations and cultures are forced together, there are inevitably conditions of coercion, inequality and conflict. Tatar uses the idea to think about the conditions navigated by children as they engage with stories and reading in the process of growing up and forming their own identities. Tatar treats stories as part of these conditions, as contact zones, as places of coercion and conflict that are often complicated by adult involvement with child interpretation and are, in effect, colonizing forces in child development. I extend the idea of stories as contact zones to think about what is being represented, conveyed, and communicated in stories as translated through contextual cultural forces. In other words, how do the Whitman Authorized Editions for Girls and particularly the story and character of Invisible Scarlet O’Neil represent, convey, and communicate to readers the contextual cultural forces of the imperatives of the World War II propaganda campaign in the United States? What is the popular imagination of American feminine identity and the feminine heroic represented in the popular children’s series fiction?
Title
Abstract
By setting learning goals at the course, program, and institutional level and documenting student progress toward those goals, institutions can offer stakeholders the evidence they seek. Information literacy has been widely identified as one of the essential, cross-disciplinary outcomes, and among the regional accreditation organizations, the Middle States Commission on Higher Education has been very explicit in its emphasis its importance as a learning outcome for all students. The purpose of this study is to determine whether and how institutions under the aegis of the Middle States Commission incorporate information literacy learning outcomes into their curricula. Specifically, this study investigates the extent to which:
This study proceeded in two stages. In stage one, the author contacted the 269 institutions accredited by the Middle States Commission to request copies of decennial self-study reviews, eliciting responses from 172 (63.9%) institutions, with 97 (36.1%) agreeing to participate. Using content analysis, the retrieved documents were analyzed for patterns in the usage and coverage of information literacy. Particular attention was paid to the context in which information literacy was discussed, which campus entities were involved in its planning and delivery, and at what levels planning, instruction, and assessment took place.
Once all documents had been reviewed, four institutions were selected based on the relative robustness of their programs to serve as case studies. The author visited each campus and conducted interviews with key individuals to explore the context that created and cultivated these programs.
The results of this study have implications for accreditation organizations that address information literacy learning outcomes. By demonstrating to what extent one population of institutions addresses information literacy outcomes this study offers some evidence of the level of importance institutions actually afford to those outcomes. Likewise, stakeholders concerned with accountability in higher education can see from these results how accreditors and institutions are holding themselves responsible for student learning outcomes in this area. Because information literacy is often intertwined with and supported by library instruction, library directors who are managing or implementing information literacy program will also be interested in these results, as will those professional library associations directly concerned with information literacy such as ALA. Finally, teaching faculty, who are ultimately most directly responsible for achievement of student learning outcomes will be interested to see how such outcomes are being interpreted, implemented, and assessed.
Title
Abstract
Title
Abstract
Theoretical Frameworks
Methods
I am coding threads from the three newsgroups and interview transcripts using methods outlined by Corbin and Strauss. I am using a constructivist approach to data analysis as outlined by Potter (2007).
Results
Title
Abstract
To capture, interpret, and understand the complex nature of the information behavior of online news readers - more specifically the nature of the incidental exposure of online news - the mixed method approach was used to guide this study. The mixed-method approach allows gathering reasonably complete and reliable data on the real-world information behavior of online news readers.
The present study was conducted in two phases. In Phase I, the banner advertisement to recruit participants to a web survey was posted at the local newspapervs website from March 2 to April 10, 2009. Self-administered survey data were collected using the Survey Monkey. The main goal of the survey was to collect general information about news readers and their media usage behavior. The survey also served as a screening tool to select people who are most aware of their incidental exposure to online news. In Phase II, in-depth interview with critical incident, explication interview and think aloud sessions were conducted with the selected number of respondents from the Phase I. 146 people responded to the web survey and 20 people were interviewed for this study.
The survey data were analyzed with SPSS for descriptive analysis. Qualitative data from the second phase were analyzed inductively with grounded theory approach using the nVIVO software.
The preliminary analysis of survey confirm the trends described by Tewksbury, Weaver, & Maddex (2001) that incidental exposure to online news is becoming a typical behavior to get informed about the news events happening. The eight questions in the survey asked the respondents about their awareness of incidental exposure to news. 75 % of respondents said they often come across interesting news stories online when they browse the Internet for other purposes than news reading. 78 % of respondents reported that they find interesting news stories at times when they browse the news websites without specific goal in their mind. The poster will present the research design and preliminary findings from the present study on incidental exposure to online news in everyday life information seeking context.
The study of incidental exposure to online news would contribute to the research realm of human information behavior in incidental information acquisition, everyday life information seeking, multitasking, and future information use in information search studies.
Title
Abstract
This study addressed the use of WA for library decision-making and generalizing users’ information behavior. MU library’s web usability group was interviewed and their Google analytics Fall ‘08 data were reviewed. Data obtained from both sources –interview and Google analytics, were analyzed qualitatively. Findings indicated usefulness of analytics for decisions related to service and web-design. Nevertheless, other factors were identified that could affect decision-making indirectly –unexpected user behavioral elements as detected in the graphs and the management’s limited use of the evidence in the reports. The WA reports indicated irregular user behavior patterns. Reports indicated user behavior could differ with the choice of sources used to access the website. Reports on technology used suggest user-adoption of latest web technology. Reports also suggest use of databases as primary intention of the Library’s website visitors.
Title
To identify one of the most disadvantaged cell phone user groups in the world, the research applied a stratified purposive sampling with 6 filters: (i) citizens of India, the country with the highest number of citizens living under the poverty line defined by government, (ii) resident in a rural context which offers fewer socio-economic opportunities compared to an urban context, (iii) “backward class” population, (the term “backward class” was coined and defined by the Government of India for) a set of socio-economically and educationally disadvantaged population, (iv) women in a male dominated society, (v) individuals with daily income of less than a dollar, and (vi) cell phone owners and users.
Wilson’s (1997) Global Model of Information Behavior was tested to study the role of cell phones in shaping the information behavior of disadvantaged women from rural India. In the first phase, quantitative data was collected using group-administered surveys which were filled out by 100 women, working at Mahila Gruha Udyog, a domestic business run by women for women, in Bhor, a rural part of India. Unmarried girls (UMG) and women who were married for more than 20 years (MW) emerged as two groups with distinct information behavior. 12 UMG and 10 MW were interviewed on the phone in Marathi, their mother-tongue. Conceptual and relational content analysis of 22 interviews was guided by the global model of information behavior. Software-aided analysis revealed the role of “context of communication and information needs” as the main controller of information behavior. The analysis unfolded differences in information-seeking behavior, and processing and use of information acquired by UMG and MW. The research enriches exiting information behavior theories developed in the West. The research informed Wilson’s information behavior model, in particular with “context of communication and information needs”, a newly discovered construct. Research will definitely serve as a scholarly reference while crafting policy frameworks and designing dollar-aide strategies for building sustainable development, using effective deployment of mobile in developing nations. Research findings could benefit government policy makers, when designing mobile-Governance-related policies and developing implementation details. The private sector could apply a set of research findings for better human-centered designs and interfaces of mobile technologies as well as devising marketing strategies for sales of mobile devices in developing nations.
Title
Abstract
This research focuses on information as experienced in chronic conditions primarily because they foreground the contextual mismatch between patient experience and a healthcare perspective. Healthcare providers typically see chronically ill patients only when their condition becomes acute or for intermittent maintenance. This means that most of the ongoing management of chronic conditions, and consequently, the use of information to support this management, happens outside of the treatment setting, in patients’ everyday lives. The need to cope with the impact of illness and to integrate its management into daily life can result in a breadth of information needs and uses, many of which may not be anticipated by healthcare providers. The medical perspective on the type of information needed to manage illness can overemphasize medical expertise and downplay the realities of living with illness.
Consequently the gap between information and behavior change continues to frustrate healthcare providers and patients, and the doctor-patient disconnect has only widened. Health communications and health behavior education has focused on getting the message right for years. What LIS has to offer is the perspective that in order to truly understand what makes information useful, we need to know more about how it is perceived and actually put into use. The medical perspective tends to view information delivery as a one-way transfer of “perfect,” or complete information that enables rational decisions leading to compliance with treatment. Providers may dispense information assuming it will lead to a specific outcome, desirable from their perspective. However, people living with chronic conditions may actually use the information to address other priorities; in the context of their experience the same information could have very different meanings and uses.
The grounding orientation of my research is the articulation of the chronic illness sufferer’s perspective on information. But I also take an important translational step in turning toward the healthcare context with the aim of showing how provider work practices and perceptions act to shape the information experience of their patients. By examining provider perceptions and practices and the information exchanges that take place in provider-patient interactions, we can understand the ways this important institutional context contributes to the patient experience.
Title
Abstract
The research questions of this study are: 1) what primary patterns, if any, ELL students have in terms of cognitive, behavioral and affective dimensions, as they engage in the research project; 2) what resources and people ELL students interact with during the research project; 3) what concerns or challenges ELL students experience in each phase of the research process; and, 4) in what ways the characteristics of ELL students (gender, age, ethnicity, national origin, English language level, self-rated language proficiency in English and native language) interact with their research experience.
Participants of this study were 48 ELL students from three classes at a public high school which is located in one of the most diverse districts in New Jersey. During a four week period, 10 students from ELL class were required to write a research paper on their career and college preparation and 38 students from two biology classes were required to create a foldable on a genetic disorder disease of their choice.
Data were collected from January to May in 2009 following a pilot study: questionnaire, process surveys at three times (initiation, mid-point and completion), observation, semi-structured interviews, students’ final product, and teacher interviews. The survey answers and interview transcripts were analyzed through content analysis with an assistance of NVivo. Quantitative data were analyzed with SPSS using ANOVA, multiple regression, and logistic regression.
This study showed how ELL students experience information search and knowledge building processes through a complex research project in English while revealing what factors interact with individuals’ primary patterns in their information seeking behavior. This study contributes to Kuhlthau’s Information Search Process by adding linguistic and cultural dimensions as critical factors which influence human information behavior. At a pedagogical level, the findings will facilitate understanding of the efficient strategies and instructional interventions for ELL populations in K-12 school contexts.
Title
Abstract
This study has been proceed in three major steps. First, the Web of Science (WoS) ISI Thompson’s Citation Database was used to identify the extent of collaborative efforts between Turkish scientists and scientists located abroad, in order to see how such efforts affect patterns of changed in this post-Internet era. The seventeen years of (1990-2006) scientific research collaboration of Turkey and the G7 nations (France, Germany, Italy, Japan, UK, USA, and Canada) was examined by using WoS data. For the second part of the research study, an email questionnaire survey has been administered to Turkish scholars. The third step was to conduct semi-structured interviews with the volunteer researchers from Turkey.
Findings from this study indicate that, both internal and external factors play roles in the international collaboration behavior of Turkish scholars. These are external (uncontrollable) factors such as organizational factors and internal factors from observed reality to belief system. The results from this study provide the basis for identifying some of the crucial components for the future success in scholarly collaboration of Turkish scholars. In conclusion, I believe that the findings of this dissertation research will assist various groups of stakeholders to gain better understanding of needed changes in the context of scholarly collaboration and will provide feedback to restructure the existing framework in Turkey. I also predict that this research will produce some positive impact on the current scholarly science policy agenda in Turkey.
Title
Abstract
Since a classification must continually be revised to retain its relevance, the current research focuses on changes made to a classification over time. This focus requires historical data so that longitudinal changes to the classification can be examined. Comparing the differences between each of the classification’s revisions can illuminate how people have chosen to contribute within the context of the classificatory language. This level of detailed scrutiny to the revision process is possible in online environments that record and log changes to digital documents. For these reasons, the free encyclopedia Wikipedia serves as the source of data for the present study. Like most wikis, Wikipedia maintains a complete archive of all the edits made by contributors. These revisions can be examined unobtrusively, preserving the authentic evolution of the classification.
A dozen classifications are purposively sampled from the English-language Wikipedia, spanning the extensive range of content in the encyclopedia. Each classification is analyzed longitudinally, comparing every edit made in succession from initial creation to its most recent revision. The unit of analysis is the difference between each successive edit. A content analytic coding scheme adapted from Taylor (1986) characterizes each user’s contribution as an attempt to add value to the classification. While many different types of contributions are made over time, the most salient classificatory features are the focus of the present study. Briefly, these salient features can be characterized as changes of “position” and “inclusion.” In linguistic terms, classificatory position is a function of relational semantics: where an entity’s placement in the classificatory structure represents its relationship to the other entities. In contrast, classificatory inclusion is a function of referential semantics: where an entity’s mere presence in the structure is the basis for the classification’s indexical and navigational potential. By quantifying the fundamental ways that each contribution incrementally changes a classification, this research seeks to contribute to the field of LIS in general, and information organization in particular. It is critical that LIS develop robust, theoretically based evaluative methods for classifications; this study seeks to help contribute to that goal.
Title
Abstract
This research aims to answer whether collaborative, user-generated tags through social tagging could be used to enhance access to web resources and provide additional access points beyond professionally-generated ones, and whether we could verify the usefulness of social tagging to obtain benefit from it.
Sample and Setting
Method
Descriptive statistics estimating the distribution of tag frequency is used to explore the features of tags by subject categories which map to BUBL subject categories. The process of identifying bibliographic attributes of tags is based on the Functional Requirements for Bibliographic Records (FRBR) model.
As a second phase of analysis, this study measures three principal values of the inter-indexer consistency (1) among social tagging users, (2) between two groups of professional indexers, and (3) between social tagging users and professional indexers. In order to measure indexing consistency, this research employs the method of Inter-indexer Consistency Density (ICD) based on traditional Information Retrieval (IR) Vector Space Model, which calculates the distance between each indexer/tagger’s resulting vector and the indexing centroid and the resulting vectors explain that high density space among indexers/taggers means more similarity and higher consistency. Furthermore, inferential statistics using ANOVA is calculated to compare average Inter-Indexer Consistency Density among different subject areas.
An analysis of tagging exhaustivity and tag specificity on tagging effectiveness is subsequently conducted to ameliorate difficulties associated with limitations in the analysis of indexing consistency based on only quantitative measures.
Data Collection Technique(s)
Results
Title
Abstract
Title
Abstract
We introduce the concept of annotation dominance (AD) as one way to measure the importance of a tag. AD is a way of measuring how often the tag is used related to a resource. AD should reflect the difference in importance of tags when distribution of tags on a resource is different. Another metric, Cross Resource Annotation Usage (CRAU) is considered as a means to offset weight of general tags. CRAU is designed to remove tags that are used broadly in the document corpus. If a tag is assigned for every document in the collection, we consider it to be a weak candidate as a tag to clarify the document classification. CRAU gives a lower score for a general tag and gives a higher score to a specific tag. In addition to these metrics to assess a tag’s classification potential, we suggest special processing steps to identify and manage compound tags. Due to the limitation on inputting tags in existing social tagging systems, users are generating numerous forms of compound tags. Although compound tags are not as structured as in formal metadata, it is possible to make use of this tag information in tag data.
The dataset is gathered from both Open Directory Project, expert generated categories, and Delicious, a social bookmarking system. Data will be gathered based on a dataset that provides categories from Open Directory Project. Based on the category dataset, tags on each resource can be crawled from Delicious. In result, the dataset consist of URLs, categories, and tags. To make the sample dataset more reliable, the possible history of bookmarks on each bookmarks and users were collected.
The research aims to select important tags (meta-terms) and remove meaningless ones (tag noise) from the tag set. This study, therefore, suggests two main weighting methods for getting a classificatory metadata set with classification potential. Two methods of evaluating the proposed classificatory metadata terms are suggested. First method is evaluating with users’ relevant judgment comparing suggested classificatory metadata terms, high frequent tags, and category terms. The terms will be provided in a random order to the subjects. Normalized Discount Cumulative Gain at K (NDCG at K) will be used for the measurement of human judges. Basically, human judges rate on how relevant each retrieval result is on an n-point scale based on the relevance of each of the randomly proposed classification terms for the given resource. Second, to evaluate the performance of the proposed classificatory metadata terms, we will evaluate categories of web resources created based on the proposed classificatory metadata to see whether their categories are competitive with human created ones. The categorization of collected web resources based on existing categories, categories using high frequent tags, and categories created by proposed tags will be compared.
Title
Abstract
Sample and Setting: 554 collection-level metadata descriptions in 3 aggregations of digital collections: Opening History (OH), American Memory (AM), and European Library (EL); 12 weeks of web logs from Opening History aggregation; representative sample of academic historian users.
Method: Combination of quantitative and qualitative. Data Collection Techniques: Content analysis of collection metadata; web log analysis of collection-level search and browse queries and results, applying FRBR entities as the analytical framework, to assess efficacy of collection descriptions; interviews and observations with scholarly users on collection search practices and the value of collections.
Results:
The Description field plays an especially important role in collection-level subject access. Collection-level user searches in the OH aggregation can often be satisfied only through the Description field. In 93% of searches (based on the sample collected through the transaction log), at least one of the collection records retrieved had a match in Description field. In 74% of searches at least one of the retrieved collection records could be found exclusively in this field.
Across the three aggregations, an average Description field provides information about 6 collection properties, with collection subject represented most frequently, followed by object types or genres, geographic coverage, temporal coverage, collection title, size, collection development policy, navigation and functionality, participating/contributing institutions, intended audience, funding sources, and language of items in collection. Additional important properties generally not found in other metadata fields include creators of items in collection and provenance, as well as statements about the collection’s importance, uniqueness, and comprehensiveness.
User search categories are strongly associated with four collection properties most often covered in Description fields. Transaction log analysis shows that the user search on average belongs to 1.74 FRBR-based categories. The majority of collection-level searches can be categorized as FRBR Group 3 entities: concept (22%, e.g., “transportation”, “soccer”), object (35%, e.g., “steam shovel”, “cookbooks”), place (26%, e.g., “Lake Springfield”, “Central Florida”), or event (9%, e.g., “1935 meat strike”, “Civil War”) searches.
Much of the information provided in the Description field is unique, potentially enhancing subject access by concept, object, place, and event. Collection subject information beyond that found in other fields is 76% overall: 86% in AM, 76% in OH, and 70% in EL. Object types and genres are also often more fully represented in the Description field than elsewhere (33% overall: 70% in AM, 44% in EL, 30% in OH). Geographic and temporal coverage information, consistently supplied by Description field, often complements information found in other metadata fields (31% overall geographic: 39% in OH, 33% in EL, 19% in AM; 48% overall temporal: 67% in OH, 54% in AM, 15% in EL).
The next stage of analysis will determine how historians interact with collection-level metadata. Of particular interest is the value of displaying controlled-vocabulary-based metadata fields (Subjects, Objects, Geographic Coverage, Temporal Coverage, etc.) for subject searching in aggregations of digital collections.
Title
Abstract
The study employed a mixed method approach to address the complexities behind the problem, including interview, observation, questionnaire and web-based analysis. The case study was carried out at the Al-Bayan Model Girls’ Secondary School, Jeddah, Saudi Arabia. The main findings of this study were: (1) the Principal Component Analysis for student satisfaction was statistically significant, explaining 68% as total of variance, and (2) the number of components extracted by factor analysis was six: Usefulness, Relevance, Variety, Usability, Availability, and Accessibility. Reliability for each component was above .70 and the correlation matrix among all six components was statistically significant, arranging from .31 to .59. A Confirmatory Factor Analysis (CFA) to confirm construct validity for the six factors underlying student satisfaction showed perfectly fit data and statistical significance, supporting the general factor model. Structural Equation Modeling (SEM) was used to assess the effectiveness and relationship among student satisfaction, evaluation of the tool, and coursework performance. The result indicated an adequate model fit, meaning that the theoretical foundations were acceptable, supported by goodness of fit indices. The findings suggest that the system tested can facilitate more collaborative work among teachers and librarians to provide accurate, efficient, and sufficient e-resource exposure for students at Al-Bayan Model Girls School in order to empower them through knowledge and information literacy for lifelong learning in a knowledge-based society. Furthermore, these findings suggest the potential to use this approach in other developing countries where differences in languages and cultural norms have proven challenging to developing information literacy.
Title
Abstract
RQ: Does the stage of user task help predict document usefulness from the time that users spent on the documents?
There were three sub-RQs, which looks at the above RQ in both tasks, in the parallel task, and in the dependent task, respectively.
Twenty-four participants were recruited, each coming three times (as three experiment sessions) within a two-week period to a usability laboratory working on three sub-tasks associated with a single a general task, couched either as parallel or dependent. Participants were asked to write a three-section article on hybrid cars, with each section to be finished at the end of each session. Half the participants worked on a parallel task in which the accomplishment of one sub-task did not depend upon that of others. The other half worked on a dependent task, in which the accomplishment of some sub-tasks depended upon that of others. Data were collected by usability logging software and questionnaires. It was observed that users often visited a document multiple times, so two types of time were considered. One is total dwell time (TDT) and the other is the first dwell time. The latter is named decision time (DT) because at the end of that time period, the user usually had made a decision on the usefulness of the document. It was found that TDT only may be a good predictor for document usefulness, but this has limited implications since TCT cannot be captured until the end of a session. Meanwhile, DT can be more easily captured. Task stage can enhance the accuracy of predicting document usefulness from DT, especially in the parallel task.
Title
Abstract
This dissertation research takes up these problems and aims to 1) find out mood categories that are frequently used by real-world music listeners, through an empirical investigation of real-life social tags applied to music; 2) advance the technology in automatic music mood classification by a thorough investigation on lyric text analysis and the combination of lyrics and audio. Using linguistic resources and human expertise, 18 mood categories were identified from social tags collected from last.fm, a major Western music tagging site, and a ground truth dataset of 5,296 songs were built with mood labels given by a number of real-life users. Both commonly used text features and advanced linguistic features were investigated, as well as different feature representation models and feature combinations. The best performing lyric feature set was then compared to a leading audio-based system. In combining lyric and audio sources, both methods of feature concatenation and linear interpolation of classifiers were examined. Finally, learning curves on various training data sizes and different audio lengths were compared. The results indicate: 1) social tags can help identify mood categories suitable for real world music listening environment; 2) the most useful lyric features are linguistic features combined with text statistics; 3) lyric features outperform audio features when semantic meanings taken from lyrics tie well to the mood category; 4) multi-modal systems outperform single modal systems; 5) multi-modal systems can reduce the requirement on training data, both in size and in audio length.
Contributions of this research are three folds. On methodology, it pushes forward the state-of-the-art on text sentiment analysis and multi-modal classification in MIR. The mood categories identified from empirical social tags can complement to those in theoretical psychology models. On evaluation, the ground truth dataset built in this research is the largest one on this topic to date and has been used in the Music Information Retrieval Evaluation eXchange (MIREX) 2009, the community-based evaluation framework. The proposed method of deriving ground truth from social tags helps reduce the huge cost of human assessments on music and clear the way to large scale experiments. On applications, findings of this research help build better music mood classification and recommendation systems by optimizing the interaction of music audio and lyrics. A prototype of such systems can be accessed at http://moodydb.com.
Title
Abstract
Searching for relevant information in distributed environments transforms into a problem concerning complex networks and information retrieval (IR). Although relevant information is likely within a short radius from any location in any real network, it is only a tiny fraction of all densely packed information in the “small world.” If we allow queries to traverse the edges of a network to find relevant information, there has to be some association (clustering) between the network space and the relevance space in order to guide efficient searches.
We studied decentralized search in information networks and focused on the impact of network clustering on the retrieval of relevant information. We developed a multi-agent system to simulate peer-to-peer like networks, in which agents worked with one another to forward queries to targets that have relevant information, and evaluated the effectiveness, efficiency, and scalability of decentralized searches. Using a large benchmark IR collection from TREC, we conducted experimental simulations on various size scales of networks (of roughly 102 to 106 agents) and investigated several searching strategies including random walk (RW) and similarity-based greedy routing (SIM). By using a clustering exponent a to guide network self-organization, we examined the role of strong ties vs. weak ties and their influence on decentralized search. Interestingly, an inflection point was discovered for a, below which performance suffered from many remote connections that disoriented searches and above which performance degraded due to lack of weak ties that could move queries quickly from one segment to another. Experiments on very large networks demonstrated that clustering optimization is crucial for decentralized information retrieval. Although overclustering only moderately degraded search performance on small networks, it led to dramatic loss in search efficiency for large networks.
Clustering reduces the number of “irrelevant” links and aids in creating topical segments useful for orienting searches. With very strong clustering, however, a network tends to be fragmented into local communities with abundant strong ties but few weak ties to bridge remote parts. Although searches might be able to move gradually toward targets, necessary “hops” become unavailable. Some level of network clustering, on the balance of strong ties vs. weak ties, supports optimal semantic overlay on which decentralized searches can best function and scale. We refer to this phenomenon as the Clustering Paradox and explain the implication on scalability of decentralized IR systems.
Title
Abstract
Method: The study was guided by Activity Theory, Scenario Based Task Analysis, and survey method was used to collect data. Two groups of stakeholders (end users and curators) were identified based on hypothesized differences in their quality needs for genome annotations. The study used an earlier developed general framework of information quality assessment and a taxonomy of data quality skills to develop survey questions. In addition, to contextualize the questions and motivate subjects to provide answers, the survey’s questionnaire included two genome annotation scenarios consisted of series of the genome annotation actions. Subjects were asked to rank the importance of data quality dimensions and related data quality skills when considering these scenarios. The survey data collected from 151 subjects were further explored by factor analysis to identify the quality concepts or criteria stakeholders consider important in genome annotation process.
Result: Seventeen data quality dimensions were reduced into five factor constructs, while seventeen data quality skills were aggregated into four factor constructs respectively. Data quality dimension-skill factors were mapped into the space of different levels in data-information-knowledge. The two user groups prioritized data quality aspects differently. The database users cared much about whether annotation records came from reliable sources (‘Believability’), and the most current information (‘Up-to-date’), while the curators focused more on the displays, formats and usefulness of the annotation products (‘Consistency’, and ‘Interpretability’). Both groups believed the ‘Security’ issue was trivial since genome annotation environments were highly publicly accessible and sharable. Both groups thought that data quality error detection skills as well as data quality literacy skills were essential and necessary for improving data quality work in genome annotation.
Conclusion: Analysis of annotation user survey results revealed that users have specific sets of “virtues” or criteria in the genome annotation context. ‘Accessibility’ and ‘Accuracy’ are the most important data quality dimensions in genome annotation work. Users with different roles value diverse data quality dimensions and skills differently. The end users, as data consumers, indirectly assess the data quality of annotation records by relying on source credibility. The curators, as data intermediaries and/or produces, directly assess the data quality and virtue in genome annotation. Identify data quality dimensions and correspondent skills in the levels of data-info-knowledge help us propose effective approaches to provide accessible, interpretable, and useful signal transmissions among these levels. Identifying data quality requirements of different stakeholders and defining models for quality can help systematize quality assurance activities.
Title
Abstract
Title
Abstract We designed an experimental study that investigated emotional patterns expressed during the online search, and examined the role of mood in the online information retrieval session. During the experiment, participants were asked to conduct two online searches of various complexities. Participants’ moods during the search were measured using the Positive Affect Negative Affect Scale (questionnaire). Searchers’ emotions were inferred from their facial expressions (based on Paul Ekman’s theory of emotions). Participants’ were video recorded; their search behaviors were logged using the TechSmith Morae software.
We conducted Analysis of Variance statistical tests to identify emotions that varied significantly around the search actions. We found that every analyzed search action (such as a mouse click or a scroll) was characterized by a unique emotional pattern. In most instances, search actions did not lead to the immediate improvement of emotional state. The only click which lead to the consistent improvement of emotional state was wheel scroll up, suggesting that re-examination of the information was one of the few emotionally rewarding actions during the search.
We analyzed relationships between mood and search performance using a multivariate Regression and Canonical Correlation Analysis statistical techniques. Positive mood was found to be associated with fewer search activities (such as fewer visited sites and reformulated queries), while negative mood was associated with increased search activities. We found that mood recorded before searching did not vary significantly during the course of the search and did not have a significant impact on the quality of the search results.
The study produced findings that have significant implications for understanding the role of emotions and affect in searching and design of affective information systems. The fact that we identified and classified emotional expressions around certain search actions can inform development of the systems capable of anticipating (and possibly reacting to) searchers’ actions and emotions (e.g., if the system recognizes a certain sequence of emotions it can predict the action that will follow; and by knowing the action, the system will be able to anticipate a certain emotional reaction). Our findings indicate that searching is a learning process that does not lead to instant gratification, but a thorough searcher is rewarded with the more frequent positive emotions during the search. We gained understanding of the reciprocal relationships between mood and the search process. While we found that searching does not affect participants’ mood, we found that moods affected search efficiency.
The study findings can benefit library education programs (e.g., in setting up the right expectations for the search) as well as the affective computing agenda. Additional strengths of the study include the use of the operational definitions of the affective concepts that are rarely defined in the LIS research, and the use of the interdisciplinary methodology that can be incorporated in the future studies.
Title
Abstract
Sample and Setting: An email survey was randomly sent to 1,010 directors of various degree-granting colleges and universities within the United States; 596 (59%) responded. The survey was based on a review of library literature and on Bolman and Deal’s change management model. Multiple choice questions tracked the directors experiences with change management, the approaches used, and the factors that may have influenced these approaches. When applicable, directors were also encouraged to write their own views and experiences. This allowed for any “other” possible categories outside of the Bolman and Deal model.
Method (qualitative, quantitative): The collected quantitative and qualitative data were analyzed using descriptive (frequencies, percentages, means, standard deviations) and inferential (bivariate crosstabulations, chi-square tests, correlations, binary and multinomial logistic regressions) statistics. Multinominal logistic regression was used to determine the relationships between a dependent variable with multiple categories and more than two predicator variables. The qualitative data of the open-ended questions were analyzed using content analysis.
Data Collection Techniques (email survey, pilot study, large-scale study): Initially 18 directors, chosen by stratified random sampling, participated in a pilot study of the email survey via surveymonkey.com. Following their suggestions and comments, revisions were made to the survey before it was applied to the large-scale study in a similar manner.
Results: Most directors managed both planned and unplanned change and used multiple approaches. The structural and human resource approaches were the most frequently used single approaches, although dual approaches were also common. Correlation and regression analysis confirms that demographics, human capital data, and library variables play significant roles in managing change. Regression results show that older directors and those who worked for a higher academic degree institution were more likely to use multiple approaches when managing change. The latter also used multiple approaches to handle conflicts.
Title
Abstract
Title
Abstract
Sample and Setting
Method
Data Collection Technique(s)
Results
Browsing by resource (29.27%), i.e. exploring the relevant resources of a particular resource, is the most prevalent approach to finding resources, followed by searching (27.35%), encountering (17.78%), browsing by tag (12.95%), browsing by user (11.54%), and monitoring (1.11%) successively. The found resources are not necessarily what the users need. In terms of the find-to-collect rate, browsing by tag (37.66%) is the highest, followed by encountering (32.12%), searching (29.99%), browsing by user (29.99%), browsing by resource (24.75%), and monitoring (18.12%) successively. Hence, social cataloging is very effective in helping users find resources of interest. Users differ greatly in the length of their tracks and in the frequency they collect resources. But there is a positive correlation (r = .48) between these two parameters. Most users (74.37%) employ only one accustomed mode to proceed along their tracks. So it is possible to characterize them with their favorite modes, although other users (25.63%) alternate multiple modes. |