Research Forum at the 75th Annual Meeting of the Society of American Archivists August 22, 2011 - August 27, 2011 Chicago, IL Does trust matter? Ixchel M. Faniel, Ph.D. OCLC Elizabeth Yakel, Ph.D. Nancy McGovern, Ph.D. Kathleen Fear Morgan Daniels Adam Kriesberg Universit y of Michigan This project is possible with funding from the Institute of Museum and Library Services.
Agenda Dissemination • Mot ivat ion Information Packages • Research Quest ions for Information Reuse • Research Met hods (DIPIR) Proj ect Inter-university • Mot ivat ion Consortium of Political • Research Quest ions and S ocial Research • Research Met hods (ICPS R) S urvey • Overview of At t ribut es Audit and Certification • Example Mapping of Trustworthy Digital • Example Hypot heses Repositories (IS O/ TRAC) Does Trust Matter? 2
THE DIPIR PROJECT Does Trust Matter? 3
Research Team Nancy McGovern ICPS R Elizabeth Ixchel Y akel Faniel UM S chool of OCLC DIPIR Information Proj ect William Eric Kansa Fink Open UM Museum Context of Zoology Does Trust Matter? 4
Research Motivation Our interest is in this overlap. Two Maj or Goals Data reuse 1. Bridge gap between research data reuse and digital curation research 2. Determine whether Disciplines Digital reuse and curation curating curation and reusing practices can be research data generalized across disciplines Does Trust Matter? 5
Research Questions 1. What are the significant properties of data that facilitate reuse by the designated communities at the three sites? 2. How can these significant properties be expressed as representation information to ensure the preservation of meaning and enable data reuse? Does Trust Matter? 6
Research Methodology Sep 2012 – Sep 2013 Phase 3: May 2011 – Apr 2013 Mapping S ignificant Phase 2: Properties as Representation Collecting & Information Analyzing User Data across the Oct 2010 – Jun 2011 Three S ites Phase 1: Proj ect S tart up Does Trust Matter? 7
ICPSR SURVEY Does Trust Matter? 8
Findings from digital curation literature • S uggest significant properties that provide “ reliable, long-term access to managed digital resources” (Trusted Digital Repositories: Attributes and Responsibilities 2002 ) • Funct ionalit y, relat ionships, appearance (Coyne et al. 2007) • Look and feel (Hedstrom et al., 2006; Matthews et al. 2009) • Comput ing environment and usage (Morrissey 2010) • Purpose and use (Ashley et al. 2008) • Interest in determining the range of significant properties a trusted repository might have to accommodate Does Trust Matter? 9
Findings from data reuse literature • S uggest significant properties that help users understand the data and whom and what to trust • Dat a cleaning & manipulat ion (Carlson & Anderson 2007) • Dat a collect ion met hods (Faniel & Jacobsen 2010) • Ident it y of dat a collect or (Knorr Cet ina 1999; Van House 1998; 2002) • S elect ion and calibrat ion of dat a collect ion inst rument s (Wallis et al. 2007) • Qualit y checks (Carlson & Anderson 2007; Zimmerman 2003) • Interest in determining trustworthiness of data producers and an understanding their actions Does Trust Matter? 10
Research Questions • What are the significant properties of data that facilitate reuse by the designated communities at ICPS R? • What differences do the attributes of a trusted digital repository (TDR) make to researchers using data from that repository? • What at t ribut es of TDRs as out lined in TRAC do researchers care about ? • How do t heir percept ions about reposit ories influence t heir propensit y for dat a reuse? Does Trust Matter? 11
Research Methods • Outline concepts in reuse & curation literatures • Map TDR attributes in IS O TRAC to concepts • Develop hypotheses – In progress • Operationalize concepts – In progress • Administer survey to ICPS R dataset users – Fall 2011 Does Trust Matter? 12
Comparison of Literatures Curation Literature Reuse Literature Concepts Trusted data repository Trusted data producer Repository context perceptions of trustworthiness of source Primarily data-centered Primarily user-centered User context characteristics of user Identifying & Identifying & Data context maintaining data maintaining data significant properties of data context that allows context that allows data to be rendered data to be interpreted over the long term over the long term consistently accessible accessed and supported Delivery context and supported at producer discretion ways data are accessed and supported Both literatures agree that decisions to reuse are “ not by metadata alone… ” Does Trust Matter? 13
ISO TRAC MAPPING Does Trust Matter? 14
Overview of ISO TRAC Areas of Focus 1 • Organizational Infrastructure • “ charact erist ics of t he reposit ory organizat ion t hat affect performance, account abilit y, and sust ainabilit y.” (p. 9) • Digital Obj ect Management • “ reposit ory funct ions, processes, and procedures needed t o ingest , manage, and provide access t o digit al obj ect s for t he long t erm.” (p. 21) • Infrastructure & S ecurity Risk Management • “ adequacy of t he reposit ory’ s t echnical infrast ruct ure and it s abilit y t o meet obj ect management and securit y demands of t he reposit ory and it s digit al obj ect s.” (p. 43) 1 Definitions from Trustworthy Repository Audit and Certification: Criteria & Checklist, 2007 Does Trust Matter? 15
Summary of ISO TRAC Mapping • 63 attributes were mapped • 43 related to repository context • 20 related to data context • 11 related to delivery context • 20 related to user context Does Trust Matter? 16
S ignificant properties of data DATA CONTEXT Does Trust Matter? 17
Example of Data Context ISO TRAC 4.2.5.2 ISO TRAC p. 4-11 Quote from CBU07 “ [… ], like t he quest ions on “ The repository shall have abort ion, t hey have changed tools or methods to over t ime. And t hey're doing determine what experiment s t o figure out Representation Information bet t er wordings, so like half is necessary to make each of t he sample got one Data Obj ect quest ion, and t he ot her half understandable to the of t he sample got a different Designated Community.” quest ion. S o t he codebook is like your guide t o all of t hat [… ]” Does Trust Matter? 18
Ways data are accessed and supported DELIVERY CONTEXT Does Trust Matter? 19
Example of Delivery Context ISO TRAC 4.5.1 ISO TRAC p. 4-23 Quote from CBU08 “ The repository shall specify “ [… ] like wit h t he religious affiliat ion dat a, I've looked minimum information and looked [… ] but [… ] I don't requirements to enable the even t hink I found a good Designated Community to t opical subj ect cat egory [… ] discover and identify S o it 's definit ely helpful when material of interest.” t hings are organized t hrough different cat egories t hat you can search t hrough for t opics." Does Trust Matter? 20
Perceptions about the trustworthiness of source REPOSITORY CONTEXT Does Trust Matter? 21
Example of Repository Context ISO TRAC 4.2.9 ISO TRAC p. 4-15 Quote from CBU03 “ S o t here’ s like a [reposit ory “ The repository shall provide t hat ] has a bad [… ] hist ory an independent mechanism basically. It was very biased for verifying the integrity on how t hey developed t heir of the repository met hodology and such. And collection/ content.” t hen t hey said t hey correct ed what t hey did like recent ly or what ever, but now t hat hist ory creat es a hist ory of bad reput at ion. I don’ t know anyone t hat uses [t he reposit ory].” Does Trust Matter? 22
Characteristics of user USER CONTEXT Does Trust Matter? 23
Example of User Context ISO TRAC 3.3.1 ISO TRAC p. 3-5 Quote from CBU09 “ The repository shall have “ Because I am so novice in defined its Designated these areas, I would Community and associated heavily value the opinions knowledge base(s) and of like professors that shall have these definitions knew more than me [… ]” appropriately accessible.” Does Trust Matter? 24
NEXT STEPS Does Trust Matter? 25
What contributes to decisions to reuse? • Trust in the data • Reposit ory cont ext – perceived t rust wort hiness of t he source • Relevance of the data • User cont ext - charact erist ics of t he user environment (e.g. t ask, experience level) • Quality of the data • Dat a cont ext - significant propert ies of t he dat a t hat make it fit for use • Ease of use of the data • Delivery cont ext – perceived effort needed t o access and get support for dat a Does Trust Matter? 26
Focusing on Trust – 3 Sources • Data producer – reputation of the person who originally collected the data • Repository – reputation of the institution providing access to the data • 3 rd party – reputation of an independent entity endorsing reuse of the data (e.g. faculty advisor, Data S eal of Approval Board) Does Trust Matter? 27
Recommend
More recommend
