Past Research Seminars (continued)
The nature of market orientation in UK libraries
Barbara Sen, Libraries and Information Society Research Group, Information School
Tuesday 14th January 2014
The research explores the nature of market orientation (MO) as a strategic option in the management of library and information services. Market orientation is the organizational culture that most effectively creates the necessary behaviours for the creation of value for users or customers and thus organizational performance. For businesses this performance means profitability, for non-profit organizations, it might mean survival. MO is considered to be relevant for libraries because of the current competitive environment, at a time when the survival of public libraries is under debate. If libraries are to continue to exist then there is a need for change, effective strategic management of library services, and marketing to the communities they service. MO could provide an effective strategic orientation for service improvement and performance. This study explores that relationship between MO and performance. The seminar outlines this three phase study. Phase One, of the study included interviews and focus groups across library sectors and was successful in defining MO in a library context. Phase Two, a survey of UK public libraries measured MO in that context, and through statistical analysis established the relationships between MO and other management concepts, innovation measured by an innovation scale, and performance measured using two scales one of performance trends, and one of performance comparisons with other libraries. Phase Three, explored the relationships further in a case study of one public library authority. This case study uncovered strategic complexity in the library context with evidence of library leaders displaying behaviours concomitant with multiple strategic orientations contributing to service success.
Implications of the Draft Data Protection Regulation for Health Care Research
Dr Mark Taylor, School of Law, Faculty of Social Sciences, University of Sheffield
Tuesday 17th December 2013
A key UK legal requirement for all health care research using personal data is compliance with the Data Protection Act 1998. There are, however, plans to change the European Data Protection Framework that the 1998 Act implements. If these changes are passed in Europe, then the implications for the lawful use of personal data for research purposes in all members states, including the UK, could be significant. Mark Taylor, School of Law, will discuss some of the proposed changes to data protection law, the potential implications for health research using identifiable person level data, and consider them in the light of some of the other current changes to information governance in the health sector in England and Wales..
Environmental voluntary groups: Towards curating data for sharing, access and preservation
Dr Mary Anne Kennan, Charles Sturt University, Sydney, Australia
Tuesday 19th November 2013
This presentation reports a recent study which investigated what data were collected by members of an environmental voluntary group (EVG) and how these data were collected, stored, managed and shared. Environmental and biological data sourced from volunteers and voluntary groups can make an invaluable contribution to formal science. An aim of the research was to understand how data management and approaches to data sharing could be improved in order to enhance such contributions while also continuing to meet individual and group needs. Interviews were conducted with members of the Australian Plants Society Victoria (ASPV) using a broadly ethnographic approach. Findings indicate that APSV members have a strong interest in conservation biodiversity, and in increasing their own, and society’s knowledge and understanding, passions often shared with professional scientists. Yet their data are often poorly managed, creating significant impediments to further use and sharing. The presentation explores these issues and outlines work in progress where options for improvement are explored, including ways to empower APSV members with skills and technology to contribute to emerging data repositories, so that their valuable data may be preserved and accessible beyond their immediate co-members.
‘Big data’ and housing market search in the United Kingdom
Dr. Alasdair Rae, Geography Department, Faculty of Social Sciences, University of Sheffield
Tuesday 22nd October 2013
This is the age of ‘big data’: a time when, we are told, many of our problems can be better understood – or even solved – if we interrogate ‘big’ datasets in the right way (e.g. Manyika et al, 2011). As with any grand narrative, however, it is incumbent upon scholars to look critically at the ways in which large datasets can actually help answer long-standing research questions, or whether they merely generate further questions and present us with computational headaches. Should we be more circumspect about claims made by proponents of big data? How do we know big data when we see it? And, importantly, what should we do with big data when we get it? This paper seeks to answer these questions by using a big data example of housing market search activity and behaviour in the United Kingdom. It also seeks to assess the extent to which big data is useful in arriving at more robust formulations about the spatiality of housing markets and geographical variations in search intensity. The paper draws upon a rich tradition in the study of housing and housing market search (e.g. Blank and Winnick, 1953; Maclennan, 1982; Maclennan and O’Sullivan, 2012) and concludes that even though big data has an important role to play in advancing understanding, we need to remain cautious about the claims we make and the conclusions we draw.
Blank, D. M., & Winnick, L. (1953). The structure of the housing market. The Quarterly Journal of Economics, 67(2), 181-208.
Maclennan, D. (1982). Housing economics: an applied approach. Harlow: Longman.
Maclennan, D. and O’Sullivan, A. (2012). Housing markets, signals and search, Journal of Property Research, 29(4), 324-340.
Manyika, J., Chui, M., Brown, B., Bughin, J., Dobbs, R., Roxburgh, C. and Byers, A. (2011) Big data: The next frontier for innovation, competition, and productivity, McKinsey Global Institute, New York.
Using social media big data to aid national security: A presentation of the EMOTIVE (Extracting the Meaning of Terse Information in a Geo-Visualisation of Emotion) and related systems
Professor Thomas Jackson, Centre for Information Management, Loughborough University
Tuesday 12th November 2013
Abstract: Today social media streams, such as Twitter, represent vast amounts of 'real-time' daily streaming data. Topics on these streams cover every range of human communication, ranging from banal banter, to serious reactions to events and information sharing regarding any imaginable product, item or entity. It has now become the norm for publicly visible events to break news over social media streams first, and only then followed by main stream media picking up on the news. It has been suggested in literature that social-media are a valid, valuable and effective real-time tool for gauging public subjective reactions to events and entities. Due to the vast big-data that is generated on a daily basis on social media streams, monitoring and gauging public reactions has to be automated and most of all scalable - i.e. human, expert monitoring is generally unfeasible. In this paper the EMOTIVE system, a project funded jointly by the DSTL (Defence Science and Technology Laboratory) and EPSRC, which focuses on monitoring fine-grained emotional responses relating to events of national security importance, will be presented. Similar systems for monitoring national security events are also presented and the primary traits of such national security social media monitoring systems are introduced and discussed.
“Making Open Data Real”: the shaping of the UK’s Open Government Data initiative
Wednesday 10th April
Over the past four years, Open Government Data (OGD) initiatives have been developed by a number of national and local governments, the EU and International Organisations. The UK has taken a lead in these developments, and OGD has been adopted as a core policy of the UK Government. Under the UK’s Coalition Government, a broad range of policy areas, including economic growth, public service provision and transparency, are being augmented by the ‘opening’ of public data. In brief, to ‘open’ public data means to make non-personal data produced by public bodies available free of charge under an Open Licence which allows re-use for any purpose, including commercial. The OGD model challenges public data inaccessibility, as well as more restrictive forms of public data access and re-use based on charging and/or transactional licensing, and is thus often constructed discursively by advocates as a form of empowerment in relation to both state and corporate power. However, empirical data collected over the last two years through interviews, observations and online documentation suggests that a more critical interpretation of the ‘opening’ of public data is required than is generally presented. As Davies (2013) recently pointed out, it is clear that “open data shifts power”, but we don’t yet know in which direction. This paper adopts a critical political economy approach to understanding the complex ideational terrain that contributes to the shaping of the OGD initiative in the UK, and explores the direction of shifting power relations as these ideas are actualised in the developing public policy and re-use domains around OGD.
Visitors and Residents: What Motivates Engagement with the Digital Information Environment?
Dr. Lynn Silipigni Connaway, Senior Research Scientist at OCLC Research
Wednesday 20th February
Abstract: Dr. Connaway will discuss a collaborative study with JISC and the University of Oxford, in collaboration with the University of North Carolina, Charlotte. The longitudinal study tracks US and UK participants’ shifts in their motivations and forms of engagement with technology and information as they transition between four educational stages. The quantitative and qualitative methods, including ethnographic methods that devote individual attention to the subjects, yield a very rich data set enabling multiple methods of analysis. Instead of reporting general information-seeking habits and technology use, this study explores how the subjects get their information based on the context and situation of their needs during an extended period of time, identifying if and how their behaviours change.
Dr. Lynn Silipigni Connaway leads the OCLC Research User Behavior Studies & Synthesis activities theme. Prior to joining OCLC Research, Lynn was the vice president of Research and Library Systems at NetLibrary. She also was director of the Library and Information Services Department at the University of Denver, where she taught several courses in library and information science. During her tenure there, Lynn conducted research on the subjects of organization and access of electronic documents, as well as the education of information professionals. She also has served on the faculty of the University of Missouri, Columbia and as a lecturer at the University of Wisconsin, Madison.
Examining the Limits of Crowdsourcing for Relevance Assessment
Dr Paul Clough, Information School
Thursday 1st November at 4pm in the Information School lecture room (Regent Court, RC-204) with refreshments afterwards
Evaluation is instrumental in the development and management of effective information retrieval systems and ensuring high levels of user satisfaction. Using crowdsourcing as part of this process has been shown to be viable. What is less well understood are the limits of crowdsourcing for evaluation, particularly for domain specific search. I will present results comparing relevance assessments gathered using crowdsourcing with those gathered from a domain expert for evaluating different search engines in a large government archive. While crowdsourced judgments rank the tested search engines in the same order as expert judgments, crowdsourced workers appear unable to distinguish different levels of highly accurate search results in a way that expert assessors can. The nature of this limitation in crowd sourced workers for this experiment is examined and the viability of crowdsourcing for evaluating search in specialist settings is discussed.
Designing a decision model for an e-procurement decision support system (DSS) in public sector using multi-criteria decision analysis (MCDA)
Mohamed Adil, Information School
Wednesday 14th November at 1pm in the Information School Lecture Room (Regent Court, RC-204)
Public sector procurement has a rigid structure enforced by law and regulations. This results in an organised step-by-step procedure for public sector procurement. However, this research focuses only on decision making based on evaluation of the performances of the suppliers against a preset list of criteria. This situation creates a context in which MCDA techniques can be applied to the evaluation. There are multiple MCDA methods. However, the applicability of these methods to the problem context is not known due to the multiple constrains and expectations of the public sector. This research will study the applicability of a set of published MCDA methods which involves linear weighting methods, single synthesising criterion or utility theory, outranking methods, fuzzy methods and mixed methods to the problem context and identify the best acceptable decision model. The presentation will highlight the research project and its progress to date.
The culturally competent public librarian
Mostafa Syed, Information School
Tuesday 20th November at 11am in the Information School Meetings Room (Regent Court, RC-231)
Public libraries in Britain serve a wide range of people in today’s multicultural Britain, including those from Black and Minority Ethnic (BME) communities. In order to be able to do so in an effective manner, many library staff participate in training that comes under terms such as ‘Diversity Training’ or ‘Cultural Awareness Courses.’Observational data coupled with findings in the literature show that such training is too short, rarely evaluated and is focused too strongly on anti-discrimination in the workplace as opposed to cultural understanding.What is needed is a training model that caters to the needs of library staff and helps them to be confident and culturally empathic in interacting with library users from a BME background. This is a skill set known as 'cultural competency' and this presentation will look into what this actually is, and the issues that surround it.
Twitpic-ing the riots: analysing images shared on Twitter during the 2011 UK riots
Dr Farida Vis, University of Sheffield, Information School | email@example.com
Thursday 29th November at 4pm in the Information School lecture room (Regent Court, RC-204) with refreshments afterwards
Crisis events, such as natural disasters and civil disobedience can be intensely visual in nature and this is often how we come to know and remember them. With the development of social media, such events have allowed for the production of vast amounts of user-generated content, including still images and videos circulated within a wider media ecology. Although earlier events such as the 2004 Tsunami, Hurricane Katrina and the London bombings, both from 2005, have all received attention through the lens of ‘citizen journalism’ and have to some extent framed research on subsequent events, the role of images in these as a subject worthy of study in their own right remains under explored. Recent interest in the use of amateur images in the news as well as moves to rethink citizen journalism as ‘witnessing’ may signal important changes. This paper presents research from a collaborative interdisciplinary project and highlights that no model as yet exists to understand the specificities of the relationship between digital imaging technologies and Twitter, but at the same time it introduces theoretical frameworks and critically assesses methods and data collection strategies that might work towards establishing such a model. This will also need to take into account understandings of digitally native image practices such as the creation and circulation of ‘memes’. Advances in software for capturing and researching Twitter data open up important and exciting possibilities for future research to focus more centrally on image production and sharing practices on Twitter and beyond, which will help to enrich our understanding of the platform. The paper examines image production and sharing practices on Twitter during the four days of the 2011 UK riots. In particular, a set of images that powerfully signaled the start of the riots: the setting on fire of a Double Decker bus during the first night (August 6) of rioting in Tottenham.
Farida Vis is Research Fellow in the Social Sciences, at the Information School at The University of Sheffield. Key interests include crisis communication, the emerging field of data journalism and innovating methods for the study of social media. She is a co-author of The Data Journalism Handbook (2012) and the forthcoming Researching Social Media (Sage), with computer scientist Mike Thelwall. She blogs at www.reseachingsocialmedia.org and tweets as @flygirltwo.
Social practice, information and social informatics
Dr Andrew Cox, Information School
Thursday 13th December at 4pm in the Information School lecture room (Regent Court, RC-204) with refreshments afterwards
This talk explores the value and limits of taking a practice approach (Schatzki 2002) to the study of "information behaviour" and "social informatics" (socio-technical studies). The case is illustrated by examples from digital photography and food blogging.
The role of curiosity and serendipity in how people browse digital information systems.
Thursday June 7th; Professor Elaine Toms
Information Retrieval Research Group
Libraries and librarians in the network world: defining our value proposition
Thursday May 31st; Professor Sheila Corrall
Libraries and Information Society Research Group
Elaborations of grounded theory and of arenas theory in information management research.
Thursday May 24th; Dr Ana Cristina Vasconcelos, Ms Barbara Sen, Ms Ana Guedes Pereira Rosa
Knowledge and Information Management Research Group
Use of information to support health and wellbeing in older people.
Thursday May 17th; Dr Peter Bath, Health Informatics Research Group
Drill-and-practice is not necessarily a behaviourist pejorative tool: an example of its successful application as a self-learning component of a constructivist phonetics blended learning environment.
Thursday May 10th; Dr Miguel Baptista Nunes, Information Systems Research Group
Exploring university students' experiences of inquiry and research, and implications for pedagogy.
Thursday March 29th
Professor Philippa Levy, Educational Informatics Research Group
Enriching gazetteers by detecting the informal vernacular place names from the web
Monday 26th March
Basheer Al-Farwan, PhD student, Information Retrieval Research Group
In our study we describe how to enrich the existing geographic gazetteers by gathering the geographical references that still have no representation or not stored yet in the gazetteers. Such references are called “Vernacular References”, which are references that refer to geographical places; they may refer to a common place or places that have no precise boundaries and coordinates or places that locally have different names.Vernacular place names may differ sometimes from the administrative definition or the official name of the same place name; this will make it difficult to capture such names or even to use them in the geographic services. Gathering vernacular references and storing them in the existing gazetteers is a big demand due to the limitations in the existing gazetteers, as they have a lack in storing and representing such references.
Our study is only concerned with the vernacular references that are still vague and unknown due to their informal or historical names. Finding these vernacular place names and assigning them to their official names will help to enrich the geographic databases and then lead to benefit several geographic services. One of these services, the Geographic Information Retrieval (GIR) systems, as enriching gazetteers will lead to have more flexible GIR systems and a better performance. Our work will discuss the problems we could face when detecting the vernacular place names from web sources and our method to detect and define them and then assign each vernacular name to its official name.
Use of data fusion methods to find biologically active molecules in chemical databases
Thursday March 15th
Professor Peter Willett, Chemoinformatics Research Group
Participatory Models in Networks, Crowds and Communities
Monday 12th December at 15:30, ICOSS Conference Room (see map).
Seminar from 15:30 until 16:30 with refreshments afterwards. No need to book.
Professor Caroline Haythornthwaite
University of British Columbia
Professor Haythornthwaite is Director, School of Library, Archival and Information Studies, University of British Columbia. She joined UBC in 2010 after 14 years at the University of Illinois at Urbana-Champaign, where she was Professor in the Graduate School of Library and Information Science. In 2009-10, she was Leverhulme Trust Visiting Professor at the Institute of Education, University of Londonpresenting and writing on learning networks; and in summer 2009 she was a visiting researcher lecturing on distributed knowledge, social networks, and e-learning at the Brazilian Institute for Information in Science and Technology (IBICT), Rio de Janeiro, Brazil. She has an international reputation in research on information and knowledge sharing through social networks, and the impact of computer media and the Internet on work, learning and social interaction.
This research presents a Social Network perspective on online peer production, identifying "crowds" and "communities" as two ends of acontinuum of contributory behaviour. Peer production, it is argued,seems to operate on two distinct models – a crowdsourcing modelbased on micro-participation from many, unconnected individuals, and a virtual community model, based on strong connections among a set ofstrongly committed members. At one end of the scale, `lightweight´ collaboration is characterized by low interpersonal commitment, yetstrong investment in a common interest or purpose. At the `heavyweight´ end of the scale, strong interpersonal connectionsoperate with strong-ties with other community members and community purpose. Building on the literature and cases of crowds and virtualcommunities, the presentation defines a set of dimensions that distinguish these two forms of organizing, based on such factors ascontribution type and group size, power structures, and recognition/reward systems.
A cross-cultural study on individual, team, organisational and institutional factors affecting knowledge sharing willingness in the IT services industry
Thursday 8th December at 13:00, Information School lecture room (RC-204)
No need to book.
PhD student, Information School
What factors affect knowledge sharing willingness? Prior studies explored individual, team, organisational and institutional factors. However, most research investigated only one or two levels. The presenter argues that none have examined all four levels in a single research setting, which this present study aims to address. Secondly, almost all enquiries focused on one country and a select few on two countries. This study aims to analyse an organisational setting that spans across five countries and across three continents. The presentation on 8 December will introduce knowledge sharing willingness and provide an insight into the research setting, pilot data collection and analysis. Initial results, namely the identification of 98 factors, are discussed. Details of the next phases and anticipated timetable conclude the presentation.
Evaluating the Intellectual assets of the Scholarship & Collections Directorate at the British Library
Wednesday 30th November at 13:00, Information School lecture room (RC-204)
No need to book.
PhD student, Information School
This presentation will give an overview of the progress made thus far in the evaluation of the intellectual assets in the Scholarship and Collections directorate at the British Library. A summary of the background to the project will be made, including a definition of intellectual assets, key themes from the literature, and an introduction to the British Library. Next, the researcher will explain the methods being used for the data collection and analysis process, and how they were employed during the pilot study earlier in the year. The researcher will then discuss the data which have already been collected, and the conclusions drawn. Finally, the anticipated course of the project will be explained, including the eventual results that the researcher hopes to produce.
Measurement and Determinants of Innovation in European ‘Low-Tech’ Sectors
Professor Elsa Fontainha, Lisbon School of Economics (ISEG), University of Lisbon
Tuesday 1st October 2013
The research aims to identify the role and determinants of process and product innovation in European low-technology sectors. Two types of outcomes are presented. First, using data from e-Business Survey of the European Commission (2,654 Construction firms) and applying econometric models, the results for the sources of process and product innovation are : sector of construction innovates, and the factors that contribute more to this innovation are suppliers and growth of business; firm size is more relevant for process innovation than for product innovation; companies that are guided by international markets innovate more than those that focus on local and regional markets. Second, a panel micro database collected from the firms’ accounts (11,123 companies) with information for eight years (2004-2011) is built and the preliminary results obtained from Probit models show that for firms in ‘low tech’ sectors, being a small firm and having high levels of investment increase countercyclical behavior. Past performance also has a role. Innovation presents mixed results.
Is Relevance Hard Work? Evaluating the Effort of Making
Relevant Assessments. Dr Robert Villa, IR research group, Information School, University of Sheffield, Tuesday 8th October 2013
The judging of relevance has been a subject of study in information retrieval for a long time, especially in the creation of relevance judgments for test collections. While the criteria by which assessors’ judge relevance has been intensively studied, little work has investigated the process individual assessors go through to judge the relevance of a document. In this paper, we focus on the process by which relevance is judged, and in particular, the degree of effort a user must expend to judge relevance. By better understanding this effort in isolation, we may provide data which can be used to create better models of search. We present the results of an empirical evaluation of the effort users must exert to judge the relevance of document, investigating the effect of relevance level and document size. Results suggest that “relevant” documents require more effort to judge when compared to highly relevant and not relevant documents, and that effort increases as document size increases.
A Taxonomy of Enterprise Search
Monday 28th November at 14:00, ICOSS Conference Room (see map).
Seminar from 14:00 until 15:00 with refreshments afterwards. No need to book.
Dr. Tony Russell-Rose
Dr. Russell-Rose is currently director of UXLabs, a UX research anddesign consultancy specialising in complex search and informationaccess applications. He began his career investigating the use ofstatistical and knowledge-based language modelling techniques toimprove the accuracy of handwriting recognition and informationretrieval systems. With a Royal Academy of Engineering fellowship, heworked at HP Labs on speech recognition interfaces to mobile devicesand then at BT Labs on intelligent agents for information retrieval.He moved on to lead the information retrieval group at Canon, and thento a group at Reuters to develop advanced search user interfaces andprototypes, and a software development team Advanced Computation Labof Cancer Research UK. He was headhunted by Endeca where he createdand lead their user experience consulting practice, before assuminghis current role. He has a PhD in HCI and is currently writing a bookon search experience design.
Classic IR (information retrieval) is predicated on the notion ofusers searching for information in order to satisfy a particular"information need". However, it is now accepted that much of what werecognize as search behaviour is often not informational per se.Broder (2002) has shown that the need underlying a given web searchcould in fact be navigational (e.g. to find a particular site) ortransactional (e.g. through online shopping, social media, etc.).Similarly, Rose & Levinson (2004) have identified the consumption ofonline resources as a further common category of search behaviour.In this talk, we extend this work to the enterprise context, examiningthe needs and behaviours of individuals across a range of search anddiscovery scenarios within various types of enterprise. We present aninitial taxonomy of "search modes", and discuss some initialimplications for the design of more effective search and discoveryplatforms and tools.
Conceptualising the Library Collection for the Digital World: A Case Study of Social Enterprise
Tuesday 22nd November at 13:00, Information School lecture room (RC-204)
No need to book.
PhD student, Information School
This research uses a case study of library collections for social enterprise to explore broader issues relating to the concept of the collection in the modern library.
Social enterprise is a relatively new interdisciplinary field and can be summarised as business with a social purpose. The communities interested in information about this subject are highly networked and geographically distributed, sharing valuable information resources using informal online publishing tools and social media. This creates challenges for traditional collection development and management processes from community analysis to collection evaluation.
It is a field of interest to a wide range of stakeholders and relevant material is held in many different types of library, providing an opportunity to take a snapshot of library collection issues across library sectors.
This presentation will describe the progress of the research so far and discuss some initial findings.
Factors shaping Scholars´ Personal Information Management (PIM) in the research process: a study of PAAET, Kuwait
Friday 18th November at 13:00, Information School lecture room (RC-204)
No need to book.
PhD student, Information School
Within their research, scholars often build and then have to manage significant Personal Information Collections (PIC). This consists of both publications that they use in their research and the data they themselves collect. From an information science perspective, there is interest in this, not only from how this relates to their information seeking behaviour, but also directly from their personal information management practices and the factors that shape these. The research presented in this talk develops an in-depth understanding of what scholars´ PICs are like; how they are created and used within the research process; and explores the factors that shape such collections, such as discipline and seniority.
Information, Technology and Research:
Adoption, Adaptation, and Innovation Across the Disciplines
Friday 4th November at 3:00, ICOSS Conference Room
No need to book. Seminar from 3:00-4:00pm with refreshments afterward.
Dr. Eric Meyer
Oxford Internet Institute
University of Oxford
Dr Eric T. Meyer is a Research Fellow at the Oxford Internet Institute of the University of Oxford since 2007. His main area of research has been understanding, from a social informatics perspective, how e-research is enabling innovation in practices and in the kinds of research questions that can be pursued.
Dr Meyer will discuss how technology is being adopted and adapted by different scholarly disciplines in both social sciences and sciences, and how research innovation and technology advances are inter-linked. The talk will demonstrate using empirical evidence that while disciplinary differences are important, not all the stereotypes of disciplines stand up to scrutiny. In addition, he will report on data from a brand-new, soon to be released study on information practices in the physical sciences.
What can we learn from Web Search?
Monday 7th November at 1:00pm, Room 204, Regent Court
No need to book. Seminar from 1:00-2:00pm with refreshments afterward in the iSchool Common Room.
Dr. Helen Ashman
Director, Security Lab and Associate Professor
School of Computer Science and Information Science
University of South Australia
Dr. Helen Ashman leads the WebTech and Security Lab at the University of South Australia, and prior to that the WebTech Lab at the University of Nottingham. Her current research interests focus on Web search, hypertext modelling, personalisation and behavioural intrusion detection. She was founding co-Editor-in-Chief of the ACM Transactions on the Web.
Observing Web search and interactions with Web search shows up some interesting artifacts. When searchers make selections from search results, they give implicit relevance feedback on those results. However when they make multiple selections from a set of search results, they create an implied semantic relevance between the selections. This can be used to create semantically-similar, non-ambiguous aggregations of Web resources, with a number of potential uses, and some preliminary results will be shown in this talk.
However the aggregation at times seems to suffer from sparsity of multiple-selection data, and we postulated that this might be due to the volatility of search engines' top ten results. So we commenced a longitudinal study of search engine outputs to determine the stability of the top N search results, and whether it was different for different types. While result stability did vary according to query type, the top 10 was found to be more stable than lower-ranking results. Other interesting questions we considered with the collected data were whether there was significant overlap between search engines or between the search engines' different interfaces, and whether there was any apparent duplication of results. This talk will overview some of the findings from the first year or more of this experiment and discuss some ongoing work.