UMCG 18 juni 2013

Mini-symposium Opslag en ontsluiting onderzoekdata RUG/UMCG 18 juni 2013

Programma 11:00 Ontvangst met koffie 11:10 Elmer Sterken (RM): Samenvatting uitkomst interne inventarisatie 11:40 Peter Doorn (DANS): Vertrouwen in data, vertrouwen in de wetenschap 12:10 serveren lunch 12:15 Vragen / discussie 12:30 Hans Hillege (GUIDE, epidemiologie) Optimizing Data Stewardship in Clinical Research 13:00 Daan Raemaekers (GIA): Digitale archeologie: wetten en wensen 13:30 Edwin Valentijn (Kapteyn): Data about Data 14:00 Theepauze 14:15 Tom Postmes (Heijmans): Zorgvuldige en integere omgang met data: pleidooi voor een bredere aanpak. 14:45 Slotwoord RM

21-6-2013 | 2

Opslag en ontsluiting OZ-data mini-symposium RUG 18 juni 2013

Elmer Sterken: • Aanleiding • Samenvatting inventarisatie SEP instituten

21-6-2013 | 3

Aanleiding Advies van de KNAW-Commissie OZ-gegevens (cie. Schuyt, sept. 2012) Opdracht: 1. Bestaande praktijken omgang OZ-gegevens in kaart brengen 2. Routines op de werkvloer aangeven die integer wetenschappelijk handelen bevorderen 3. Aanwijzen en toedelen verantwoordelijkheid overdracht normen in onderwijs en begeleiding (jonge) onderzoekers

21-6-2013 | 4

Citaat rapport Schuyt: “Er bestaat naar het oordeel van de commissie een relatie tussen het zorgvuldig omgaan met onderzoeksgegevens en integere wetenschapsbeoefening.”

21-6-2013 | 5

Aanbevelingen cie. Schuyt ad OZ-data: › Maatwerk per discipline › Vrije beschikbaarheid OZ-data is de standaard › Niet meer regels nodig, maar ‘vitalisering’ van bestaande

21-6-2013 | 6

Inventarisatie SEP instituten Respons: 29 SEP instituten: • 14 gereageerd (48%)

• 12 vragenlijst ingevuld (41%) • 7 (v/d 13) van FWN (Kapteyn, ZIAM, Stratingh, ITM, ALICE, GRIP, CEES) • Alle 3 van FLet (ICOG, CLCG, GIA) • 2 (v/d 3) van GMW (ICS, Heijmans)

• FWB/GRIPh: geen empirische OZ-data • GMW/Nieuwenhuis: NL gedragscode Vereniging voor OnderwijsResearch (VOR)

21-6-2013 | 7

40 aanmeldingen voor dit mini-symposium: › › › › › › ›

2 extern (DANS, Hanzehogeschool) 14 van centrale eenheden: UB (4); CIT, R&V (3); CvB, Archief (2); 8 van FWN (CEES , Kapteyn, ZIAM, Stratingh, ITM, ALICE) 6 van GMW (Heijmans, ICS, Nieuwenhuis) 4 van UMCG (CMB, GUIDE, BCN) 3 van FLet (CLCG, GIA) 1 elk van FRG, FRW, KVI

› Afwezig: FWB, FGG, FEB

21-6-2013 | 8

Inventarisatie SEP instituten Vraag 1) Afspraken eigenaarschap ?

Ja: Nee:

5 7

a) Schriftelijk vastgelegd?

Ja: Nee: Ja: Nee: Ja: Nee: Ja: Nee:

1 1 2 0 3 0 6 0

b) Bekend ? c) Ook voor ruwe data? d) Beleid financiers bekend?

21-6-2013 | 9

Inventarisatie SEP instituten Voorbeelden eigenaarschap OZ-data: • • • •

GRIP: vastgelegd in CAO (?) ICS: eigendom vakgroep ICOG: eigendom onderzoeker CEES: eigendom RUG

› Meestal: impliciete afspraken /aannames › Uitzondering: extern gefinancierd OZ (GS2 + GS3)

21-6-2013 | 10

Inventarisatie SEP instituten Vraag 2) Afspraken waarborging kwaliteit ?

Ja: Nee:

4 8

a) Schriftelijke afspraken ?

Ja: Nee: Ja: Nee: Ja: Nee:

1 1 1 0 3 1

b) Bekend ? c) Ook voor ruwe data?

Ja: CEES, ALICE, GIA (Malta), ICS

21-6-2013 | 11

Inventarisatie SEP instituten

Vraag 3) Waarborging privacy ?

Ja: Nee:

3 9

a) Schriftelijke afspraken ?

Ja: Nee: Ja: Nee: Ja: Nee:

2 1 2 0 2 0

b) Bekend ? c) Onderscheid ruwe vs def. data?

Ja: Stratingh, ALICE, ICS

21-6-2013 | 12


Vraag 4) Geheimhouding octrooi‐ Ja: eerbare data ? Nee:

4 8

a) Cf. RUG regels in "De waarde van kennis" Ja: Nee: b) Bekend ? Ja: Nee:

3 1 3 1

Ja: ZIAM, ITM, ALICE, GRIP

21-6-2013 | 13


Vraag 5) Gebruik digitale archieven?

Ja: Nee:

9 3

a) Ook voor ruwe data?

Ja: Nee: Eigen groep: Peers: Open Access:

5 2 6 3 4

c) Ontsluiting?

Nee: GRIP (DVD), ICOG, Heijmans Ja, bijv.: DANS, Target, Datawiki, UB repository, Dropbox, Google, Y-schijf

21-6-2013 | 14

Inventarisatie SEP instituten Vraag 5) < vervolg > (digitale archieven) d) Allen dezelfde?

Ja: Nee: e) Afspraken kwaliteit en metadata? Ja: Nee: e i) Schriftelijke afspraken ? Ja: Nee: e ii) Bekend ? Ja: Nee: f) Ervaring met Data Management Plan? Ja: Nee:

2 7 5 2 5 0 5 0 1 6

21-6-2013 | 15


Vraag 6a) Regels voor overdracht Ja: na vertrek tijdelijk WP ? Nee:

5 7

b) Idem na vertrek vaste staf?

4 7 3 8

c) Idem voor data studenten?

Ja: ZIAM, CEES, ITM, ALICE, ICS

Ja: Nee: Ja: Nee:

21-6-2013 | 16

Inventarisatie SEP instituten Vraag 7) Noodzaak aanscherping Ja: beleid ? Nee:

9 3

a) Suggesties hoe ?

7 3 3 1 3 1 0 2

b) Waar te beleggen

Nee: Kapteyn, GRIP, ICOG

Ja: Nee: OZI: Faculteit: CvB: Journals: Discipline: VSNU / SEP / KNAW:

21-6-2013 | 17

Inventarisatie SEP instituten 8) Gebruik faciliteiten UB of CIT ?

Ja: Nee:

6 6

a) Zo ja, naar tevredenheid?

Ja: Nee: Ja: Nee:

3 0 3 1

Ja: Nee:

4 8

b) Zo nee: wel wenselijk geacht?

9) Overige suggesties ?

Nog geen facilitering CIT of UB maar wel gewenst: ITM, Heijmans; Kapteyn: enkel voor Big Data.

Data Archiving and Networked Services

Vertrouwen op data, Vertrouwen in de wetenschap (Trust in data, trust in science) Peter Doorn Directeur DANS

Ingrid Dillo Hoofd Beleid & Communicatie

Mini-symposium Opslag en ontsluiting onderzoekdata RUG/UMCG dinsdag 18 juni 2013, 11:00-15:00 Van Swinderenhuys, Groningen DANS is an institute of KNAW en NWO

Contents • About DANS • Data fraud and data quality • Trust and responsible data management – – – –

About trust and certification How to build trust European framework for certification DSA, DIN, ISO

• The role of journals – Data Availability Policies – Data reviews

• Roles and responsibilities: Front and Back-office functions

What is DANS? Institute of Dutch Academy and Research Funding Organisation (KNAW & NWO) since 2005

First predecessor dates back to 1964 (Steinmetz Foundation), Historical Data Archive 1989

Mission: promote permanent access to digital research information

Data Seal of Approval EASY: Electronic Archiving System for self‐deposit

Persistent Identifier URN:NBN resolver

NARCIS: Gateway to scholarly information In the Netherlands

Niederlande Renommierter Psychologe gesteht Fälschungen

The KNAW “Schuyt report” on data practices • 94 interviews with researchers • Variation across and within disciplines • Pattern: data management in small-scale research more risky than in big science • Risk: missing checks and balances, especially in phase after granting research proposal and before publication • Peer pressure is an important control mechanism

http://www.knaw.nl/Content/ Internet_KNAW/publicaties/p df/20131009.pdf

Data sharing cultures across scientific domains On a central storage facility outside my department or institute 9%

Other 1%

On a network disk of my department or institute 31%

Locally: on my own computer(s), or on computer(s) of my department or laboratory 36% On external hard disks or backup media (CD, DVD, tape, etc.) 23%

Psychology: a closed data community • Psychologists hardly have a tradition of sharing data. • Differences between sub-disciplines within psychology in terms of their view on data sharing and data archiving.

Dutch Academy (KNAW) endorses Schuyt recommendations • Hans Clevers, President KNAW: “Maximum access to data promotes the pre-eminently scientific approach whereby researchers check one another’s findings and build critically on one another’s work” • The Academy supports the free movement of data and results. Taking into account variations across and within scientific disciplines, free availability of data should be the default. • We do not need additional rules or codes of conduct, but should focus on revitalising existing rules and making these better known. • Examination of data management practices should become an integral part of official research evaluations.

Trust: What do we mean? What do we rely on? RANG IS ALLEEN RANG ALS ER RANG OP STAAT

You can rely on us

Can you?

What is a trusted digital repository?

Things are not always what they say they are. Things do not always state what they are.

Trust in research data • Trust is at the very heart of creating, storing and sharing data: – – – –

Data creators Data users Data repositories Funders

• Data quality – – – – – –

valid accurate consistent integrity timely complete

Trust by definition implies uncertainty

What is trust built on (in trusted archives)? • Dedicate yourself (mission statement) • Do what you promise (stable, sincere and competent reputation) • Be transparent (peer review, get certified)

Standards of trust

ESFRI Research Infrastructures and Trust

Requirements for CLARIN Centres “Centres need to have a proper and clearly specified repository system and participate in a quality assessment procedure as proposed by the Data Seal of Approval or MOIMS‐RAC approaches” Building Trust: CESSDA Self‐Assessment Project Participants from fifteen CESSDA member organisations discussed the CESSDA‐ERIC requirements and agreed upon using the Data Seal of Approval (DSA) guidelines as a tool to gain information on the level of their conformance with the DSA and the CESSDA‐ERIC requirements.

Certification of digital repositories Formal Certification

Extended

• International framework • 3 standards • 3 levels (basic, extended, formal)

Certification

Basic Certification

EUROPEAN FRAMEWORK FOR AUDIT AND CERTIFICATION OF DIGITAL REPOSITORIES to be promoted by the EU

Certification Standards: Data Seal of Approval (DSA)

19!

• • • • • •

DANS initiative International Board 16 guidelines Self assessment Transparency 14 seals awarded

Data producers are responsible for the quality of research data, repositories for storage and long‐term access, and users for correct use of data

The research data: • can be found on the Internet • are accessible (clear rights and licenses) • are in a usable format • are reliable • can be referred to (persistent identifier)

Data quality and the role of journals • Quality control of scientific output concentrates on publications: peer review • Few journals have a data availability policy (DAP): – When a paper is submitted for peer-reviewed publication, data access for reviewers is required – When the paper is published, readers should be able to access the data

• Data repositories: marginal checking (metadata, format, privacy) • Community reviews by data users

The federated data infrastructure: a collaborative framework Research

Gov’t. agencies

Private sector

Research

University Libraries / local data facilities

Data‐Research

E‐Science / E‐Humanities

Data Providers

Gov’t. agencies

Private sector

Data Users

Research Infrastructures

NWO Gebieden

Front office

Acquisition, services, support, training, consultation Data curation/stewardship, management, archiving Systems and infrastructure

Storage Cloud/Grid, processing, backup facilities

Back office Basic Infra‐ structure

Based on: Riding the wave: How Europe can gain from the rising tide of scientific data ‐ Final report of the High Level Expert Group on Scientific Data. A submission to the European Commission, October 2010

Front Office – Back Office Model Front office – Universities (libraries, local data centers) – Disciplinary research infrastructures (ESFRI/National Roadmap) – Possibly: NWO areas (data contracts for funded projects)

Back office – DANS (humanities, social sciences) – 3TU.Datacentrum (technical sciences) – …

Basic Technical Infrastructure – SURFsara – Target

Roles and responsibilities: Front Office Focus supporting researchers at the home institution • Research Data Management: awareness raising, information, training • Offer Virtual Research Environments (research tools, data storage during research; Sharepoint, Dataverse; transfer for long‐term archiving in trusted digital repository) • Liaising with back offices • Data acquisition

What is Dataverse? Dataverse is a ‘virtual web archive’ • contains ‘data studies’ • customized and managed by its administrator or owner • collaboration between the Harvard-MIT Data Center (now part of Institute for Quantitative Social Science) and Harvard University Library • open-source application for publishing, referencing, extracting and analyzing research data

Dutch Data Verse Network (DVN)

Roles and responsibilities: Back Office Focus on long term storage, standards and expertise: • Long‐term preservation of and access to data in a trusted digital repository • Providing expertise to the front office: training courses for data librarians, consultancy, contact persons • Expertise, standards and innovation in the area of permanent storage, data management and re‐use of data • In collaboration with Front Office: expertise/training courses for the research community

Research Data Netherlands • Mission: the promotion of sustained access and responsible re‐ use of digital research data • Building on existing cooperation DANS‐3TU.Datacentrum (D4L training, Dutch Data Prize) • Expanding cooperation of back‐office function • Stepping stone to federated data infrastructure: open to other trusted digital repositories

Conclusions • Sharing data increases the transparency of research, therefore reducing the risk of fraud • Funders/Universities should require that project proposals contain a data management plan, and that such plans contain a section on accessibility of data after publication of the results • DANS (and 3TU.DC) offer back-office services complementary to Research Data Management at Universities (Libraries)

Data Archiving and Networked Services

Thank you for your attention http://www.dans.knaw.nl http:// www.narcis.nl http://dataintelligence.3tu.nl/ [email protected] [email protected]

DANS is an institute of KNAW en NWO

Optimizing Data Stewardship In Clinical Research Hans Hillege Head Trial Coordination Centre Dept. Epidemiology UMCG Opslag en ontsluiting onderzoekdata RUG/UMCG dinsdag 18 juni 2013, 11:00-15:00 in het Van Swinderenhuys

Thank you for your attention !

Data integrity, reliability and fraud in medical research n=3247

All %

Mid career Early career

Falsifying data

0.3

0.2

0.5

Using anothers’s ideas

1.4

1.7

1.0

Overlooking other’s use of flawed data

12.5

12.2

12.8

Changing the design, methodology or results in response to pressure from an funding source

15.5

20.6

9.5

Inadequate record keeping related to research projects

27.5

27.7

27.3

Inappropriately assigning authorship credit

10.0

12.3

7.4

Using inappropriate research Designs

13.5

14.6

12.2

Nature, vol. 435, 9 Juni 2005

Data integrity, reliability and fraud in medical research • 21% had discovered incorrect data after publication in previous manuscripts they had co‐authored. • Fraudulent data was discovered by 4% of respondents in their previous work. • 4% also noted ‘smudged’ data. • 21 % were involved in a paper that was submitted despite disagreement about the interpretation of the results, although the disagreeing author commonly withdrew from authorship. Mark Otto Baerlocher, Jeremy O'Brien, Marshall Newton, Tina Gautam, Jason Noble European Journal of Internal Medicine 21 (2010) 40–45

CCMO Annual Report 2011 • Every year approximately 18,000 applications for medical research • 57% interventional research, 43% observational research • 30% of the applications in medicinal research – 59% commercial, – 41% non-commercial (i.e. investigator initiated)

• In total over 307,000 subjects; – over 37,000 (12 %) involved in medicinal research – 100,000 (33 %) in the interventional research – 170,000 (55 %) in observational research

Trial Coordination Center Scope: facilitating (clinical) researchers with the design, conduction, analyses and reporting of biomedical research Project management, protocol writing, data entry, data management, monitoring, statistics, reporting and software development

Competence in Biomedical Research

Mission statement • The TCC assists researchers in performing state of the art clinical research involving advanced medical and non-medical therapeutics to optimize the care and cure of patients. • The TCC is committed to improving the quality of biomedical research by maintaining a state of the art research infrastructure and by focusing on the design, initiation, implementation and reporting of clinical research. • The TCC aims to achieve and maintain a stable group of well-trained biomedical research work force.

Rules and regulations clinical research

• National – Health related laws • WMO, GW, WBIG, KWZ, WBP, WGBO

• European – Directives • Clinical Trial Directive (implemented in NL in de WMO)

• International – ICH Guidelines (Good Clinical Practice en 70 andere guidelines) – ISO standards (ISO 14155 medical devices)


Source Herman Pieterse

Medical research and participation subjects?

N

Subject is subject to acts or conduct imposed?

N

Treatments compared via randomization?

Y

N

N

Non-WMO

Research with nonregistered drug or medical device ?

Review according nWMO-classified research

Y

Y N

The acts fall within standard treatment?

N

Y

Research with product registered in other indication, dose?

Y

N

Y

Review according to WMO

Y

In doubt?

Y N

Physical or mental Integrity affected by Investigative measures?

The Medical Research Involving Human Subjects Act (WMO) • Proper protocol • Approval of a recognized METC • Written informed consent for the operation and if data be traceable to the natural person also consent for the processing of the data (WBP) • Independent physician • Life insurance

The Medical Research Involving Human Subjects Act (WMO) • The member states of the European Union have the 'Clinical Trials Directive' (CTD) and the 'Good Clinical Practice' (GCP) guidelines in their legislation. • With these directives the international ICH -GCP principles have been addressed

The Medical Research Involving Human Subjects Act (WMO) • For the research with medicinal products the national WMO and European 'Clinical Trials Directive' (CTD) should be respected and the ICH-GCP guidelines regarding the design, implementation, reporting and archiving of clinical research with medicinal products, should be addressed

Benchmark nWMO research • Wet Geneeskundige Behandelingsovereenkomst • Wet Bescherming Persoonsgegevens • Wet Beroepen Individuele Gezondheidszorg • Kwaliteitswet Zorginstellingen • Het Privacyreglement van de zorginstelling • Code of Proper Use of Human Tissue • (Wet Zeggenschap Lichaamsmaterialen in concept aanwezig)

Content UMCG Research Code • • • • •

Good mentoring Scientific integrity Respect for participants in medical research Relationship external sponsor ……

Principles (Clinical) Research must meet • Diligence • Reliability • Accountability • Impartiality • Independence

Code of Proper Use of Human Tissue • Observational versus Experimental scientific research with humans • Fills vacuum in current laws & regulations • Self regulation of/for relevant stakeholders • Human Tissue Act (WZL) • Applies to observational research on human tissue • All Human Tissues, except: – Fetal tissue – Organs meant for donation – Tissues of deceased persons


Preparation

Initiation

Execution Management Data mgnt Monitoring

Data mgnt Closure and analyses

• Data management starts when completed CRF becomes available • Perform data validation (queries) • Arrange Data Review Meeting • Close database

Preparation

Initiation

Execution Management Data mgnt Monitoring

Data mgnt Closure and analyses

• Write statistical report • Write integrated study report • Send summary of report to investigators, CCMO and aEC

Kwaliteitsborging mensgebonden onderzoek 2.0 Nederlandse Federatie van Universitair Medische Centra • Training • Risk classification as an instrument for optimal quality assurance • Monitoring • Data and Safety Monitoring Board • Auditing • Reporting to the sponsor • Research codes • Archiving

NFU advisory board ‘Data infrastructuur voor de biomedische wetenschappen’ • Framework: 'Data stewardship' • Theme 1: Processes and architecture research data • Theme 2: Required IT infrastructure • Theme 3: maintenance of expertise and programming R&D

TCC en Data steward ship • Data stewardship is the management and oversight of an organization's data assets to provide business users with high quality data that is easily accessible in a consistent manner. • While data governance generally focuses on highlevel policies and procedures, data stewardship focuses on tactical coordination and implementation. • Data stewards can also be responsible for carrying out data usage and security policies as determined through enterprise data governance initiatives.

Volgorde en interacties tussen de processen van het kwaliteitsmanagementsysteem van het Trial Coordination Center (TCC)

DIRECTIE VERANTWOORDELIJKHEID Activiteiten: P A.2 Verklaring uitgifte en distributielijst P A.4 Kwaliteitsbeleid en doelstellingen P A.5 Organisatieschema P A.6 Directievertegenwoordiger P D.1 Vergaderstructuur ICS D.1.1 TCC bespreking ICS D.1.2 PGL overleg

ANALYSE EN VERBETERING

Resultaten: # Voorwoord # Organisatieschema # Directiebeoordelingsverslag # Notulen TCC bespreking # Notulen Productgroepleiders overleg

Resultaten: # Meta analyse: - overzicht testen reviewverslagen - overzicht (klanttevredenheids)evaluaties + lessons learned - overzicht klachten - overzicht voortgang verbeterpunten en effecten

Activiteiten: ICS A.7 Meta analyse ICS B.5 Het werken aan continue verbetering ICS B.7 Beoordeling van middelen en leveranciers

MANAGEMENT VAN MIDDELEN Activiteiten: ICS B.10.1 Security ICS B.10.2 Management of hard- and software ICS B.10.3 Backup ICS B.10.4 Software patches ICS D.4 Aanstellingsgesprekformulier ICS D.5 Rechtspositieformulier ICS D.6 Jaargesprekformulier ICS D.7 Opleidingsplan ICS D.8 Opleidingsregistratie

BEWAKING EN METING

Resultaten: # Kwaliteitsplan # BHS (Beheer Hard & Software) registratie en archivering # Dagelijkse back-up pc files en databases door ICT UMCG # Software & Patches Installation Registration # Aanstellingsbesluit # Rechtspositioneel besluit # Opleidingsplan

Resultaten: # Intern audit verslag # Klachtenregistratie + analyse # Test- en reviewverslagen # Herstelverslagen # Evaluation form # Lessons learned

Activiteiten: ICS B.4 Intern audit verslag ICS B.6 Inkoop van middelen ICS B.8 Klachten en interne meldingen ICS B.9 Review en testen ICS C.1.7 Evaluation Form

Primair proces / producten: Klant verwachtingen

C.1 Project management C.2 Protocol

C.4 CRF

C.7 Database

C.10 Data validation

C.5 Randomisation

C.8 Monitoring

C.11 Statistics

C.3 EC-submission

C.6 Software develop

C.9 Data entry

C.12 Reporting

SUPPORT PROCESSEN Kwaliteitsmanagementsysteem (KMS) P A.1 Inhoudsopgave P A.3 Matrix ISO-paragrafen P A.5 Organisatieschema P B.1 Opzet van het kwaliteitssysteem en beheer van procedures en ICSen ICS B.1.1 Information Carrier Specification ICS B.1.2 Procedure

Klant tevredenheid

SUPPORT PROCESSEN ICS B.1.3 Form ICS B.1.4 Form Authorisation ICS B.1.5 Instructions P B.2 Aanpassen van het kwaliteitssysteem P B.3 Beheersing van kwaliteitsregistraties ICS D.1.3 PG overleg ICS D.1.4 Project team overleg ICS D.1.5 Projectoverleg

ICS D.1.6 Standaard notulen met actielijst ICS D.2 Huisregels ICS D.3 Sollicitatiemap ICS D.9 Declaratieformulier ICS D.10 Betaalbewijs creditcard

TCC - ICT Land scape

Detailed Clinical Models (DCM’s)

TCC en Data steward ship • Data warehouse – – – – – – – – –

Manageability and integrity of research data sets 3-level (database) architecture ( Stage-Generic –Publish Revision management on data sets (Stage) Semantic data integration (Generic) Improvement of data quality Data Sets in a controlled manner publish (Publish) Use of international standards (HL7, DCM) Match data from other information systems State of the art pseudonimisation tooling

Research Workspace / Logon

Research Workspace

TCC en Data steward ship Research workspace Manageability and security of research data sets • Virtual research work (VDI technology) • No and/or limited data up- and download • Accessibility int-, (UMCG) and external • Standard office and analysis tooling technology available • Sharing of information (data sets, protocols etc. ) • Own High performance compute cluster part of infrastructure • First 25 workstations available, soon enlargement

Conclusions We think that the current architecture of the TCC/UMCG • facilitates appropriate utilization of state of the art technical resources by promoting and supporting careful design and planning, initiation, execution and analyses of clinical research • provides a framework to assist researchers in adhering to good clinical/epidemiological research principles;

With the aim to improve the acceptance of clinical research using sound scientific methods.

Preserving Public Trust

faculteit der letteren

groninger instituut voor archeologie

18-06-2013 | 1

Digitale archeologie: wetten en wensen Mini-symposium opslag en ontsluiting onderzoeksdate RUG/UMCG Daan Raemaekers, 18 juni 2013



18-06-2013 | 2



18-06-2013 | 3

Formulieren • Vastleggingen in het veld • Beschrijvingen van vondsten: aardewerk, bot, steen enzovoorts enzovoorts



18-06-2013 | 4

1. Wetten › Verdrag van Malta (1992)  Marktwerking  Zorgen om kwaliteit - Opgravingsvergunning (ook RUG) - Register - ARCHIS - Kwaliteitsnorm Nederlandse Archeologie KNA: archeologie als ambacht - Onder andere: ‘deponeren doe je zo’



18-06-2013 | 5



18-06-2013 | 6



18-06-2013 | 7

Goed geregeld?



18-06-2013 | 8

2. Wensen › KNA richt zich alleen op Nederland, maar GIA werkt ook in Italië, Griekenland, Egypte, Irak, Spitsbergen…. › KNA werkt niet met terugwerkende kracht….



18-06-2013 | 9

2. Wensen › GIA is in 1920 begonnen als BiologischArchaeologisch Instituut en voerde tot ‘Malta’ alle archeologische projecten in NoordNederland uit. -> -> -> ->

11.000 veldtekeningen 8.000 glasnegatieven 5.00 grootbeelddia’s enzovoorts

Geen depot voor documentatie



18-06-2013 | 10

Pilot met Bibliotheek: beeldbank opgravingsfoto’s Ezinge



18-06-2013 | 11



18-06-2013 | 12

2. Wensen › Beeldbank  Valorisatie: wekelijks verzoeken om beeldmateriaal. Nu erg bewerkelijk  Nieuw onderzoek: falsificeren oude interpretaties en nieuwe interpretaties



18-06-2013 | 13

2. Wensen › Digitale archivering standaard via DANS Nadelen: • Bedrijven willen data ‘over de schutting gooien’. RUG-GIA niet! • Kosten bedragen 5 € per GB per jaar



18-06-2013 | 14

2. Wensen › Digitale archivering standaard via DANS › GIA ontwikkelt met Bibliotheek repository om:  Zelf beheer te voeren  Greep op kosten te houden  RUG-onderzoek zichtbaar te maken en te houden



18-06-2013 | 15

Conclusies › Voor KNA-projecten is digitale opslag en ontsluiting goed geregeld  Voor lopende projecten ontwikkelen we RUG-GIA repository



18-06-2013 | 16

Conclusies › Oude projecten behouden hun waarde  Valorisatie  Falsificeren oud onderzoek en nieuwe interpretaties   

Ontwikkelen beeldbank foto’s Bibliotheek-GIA Ontsluiten dagrapporten via ARCHIS Noodzaak tot digitale ontsluiting overige onderzoeksgegevens zoals veldtekeningen



18-06-2013 | 17

Conclusies › Digitale archeologie: › Wetten: goed geregeld › Wensen: steun vanuit RUG-centraal  Op de kaart zetten van digitaal eigendom  Ontsluiten analoge onderzoeksgegevens

Data about data Edwin A. Valentijn Prof Astronomical Information Technology

Target University of Groningen

18 Juni 2013 Target

Dit project wordt medegefinancierd door het Europees Fonds voor Regionale Ontwikkeling en door het ministerie van EZ, Pieken in de Delta. Het project wordt mede mogelijk gemaakt door de gemeente Groningen, de provincie Groningen, de provincie Drenthe en SNN en staat onder auspiciën van Sensor Universe.

Target

Data about data

Target

data public en private • Astronomie: – grote surveys – optisch ruw+ resultaat – Grote surveys – infrarood ruw+ resultaat – MUSE ruw + resultaat – Radio surveys ‐ Lofar resultaat – Ruimtevaart ‐ Euclid ruw + resultaat

• Tekst – Scans

• Lifelines Target

ruw + resultaat

ruw

Astronomy International organizations public • ESO – all raw data; some processed data • ESA – all processed data; some raw data • Strasbourg ‐ the International Virtual Observatory all processed data 1000’s of databases

Target

Zelf regulering • Ontdekking ‐ publicatie (referee) – verificatie • Fraude, statische bias, systematische fouten – Het verschil is vaak flinterdun – de drive, ambitie, verwachtingen, funding van wetenschappers

• “preprint”archives‐ self refereeing community Target

Absolute wetenschap paradox • D. Bohm Reality would mean something that would have an existance independently of being known • River, falsified

Target

Target

Peta ‐100 Peta

Data about

Target

Big Data

WISE information system– fully datacentric All data beyond pixel data is Metadata BIG Data all pixel data <–>data servers Data about data all Metadata <–> database compute clusters / GRIDs all I/O to db • all components scalable • all components distributed Target

www.astro-wise.org

Target

AstroWise ‐> Wise • 64 bit identifyer • Terabytes of pointers • Extreme datalineage ++ • Modeling dependencies • Distributed

Target

Experimental Astronomy Volume 35 januari 2013 Topical issue om AstroWise 19 papers 389 pages

Target

Extreme data lineage

Target

Quality view (a)

Target

Quality view

Target

KiDS: data quality KIDS_129.0_-0.5

r e = 0.04

PSF stable over FOV

Target

Most distant quasars in the universe KiDS QSO at z~5.8 in 5 months

i

Z

Y

Ly-break+alpha NTT/EFOSC2, 30jun12

Target PhD 2012 Johnson Mwebaze / Venemans

J

H

K

communities • • • •

Disciplines VO Communities VO‐ AstroWise Big Projects AstroWise Small projects WISE Databases on demand

Target

VO standards Table access protocol

Target

Target

KiDS: first results 7 high-z QSOs to date

Target

Quality control • Distributed • web based IJKDijk

Target

• Lifelines

Big Data answering questions WISE technology

• Monk • Target Holding • Crowdy news, Lipperhay‐DataProvider • FEI Target

10 Peta byte Target testbed with IBM GPFS • • • •

Tiered storage system – 4 levels Data about Big Data ‐ GPFS 400 Million files – big and small Research collaboration ‐> Upgrades – Performance: hardware, software, design

Target

Euclid Mission Archive Target ‐ ESA

Target

Astro‐WISE – LOFAR LTA

IBM- Blue Gene/P Target

5 Petabyte/year

Collaboration • Data about Data – Abstractions‐ implementations – Standards ‐protocols – Project management

• Custom made – Projects – small – big – branches – Disciplines, communities ‐ world Target

1 bit 0 1 2 states

No physics Electron Qm only act of measurement by observer

Abstractions – agreements – models - standards Target

7 bit ASCII

7

Target

128 ASCII

8 bit

Target

16 bit ‐> kilo

Target

16 ‐24 ‐> Mega

IBM 5 Mbyte harddisk 1956

Target

Giga ‐ Tera

Target

Target

The Universe

Target

21-6-13 |

Zorgvuldige en integere omgang met data: pleidooi voor een bredere aanpak Tom Postmes Sociale Psychologie / Interpersoonlijk gedrag

Context • Schuyt • Stapel • Consequenties voor het vakgebied • Consequenties voor de RuG • Consequenties voor Nederlandse wetenschap • Lopen we inmiddels voorop?

Wat is het probleem • Fabricatie, Falsificatie, Plagiaat (FFP) • Fang et al. (2013) PNAS • Questionable Research Practices • Levelt et al. (2012) • Rapportage resultaten • Fanelli (2011)

Wat is het probleem Time-to-Retraction. The time interval

retraction varied according to the caus cles retracted because of fraud takin retract (Table 2). A gradual trend to retraction over time was detected ( factor did not correlate with time-to-r retracted because of error, plagiarism, but did exhibit a modest correlation because of fraud in high-impact journa a shorter time-to-retraction (Fig. 4B). A were responsible for multiple retractio groups with greater or equal to five 43.9% (n = 390) of retractions for fraud modern biomedical literature (Fig. S articles by authors with 10 or more r because of fraud (Table S2). This find discovery of multiple fraudulent artic investigation of a single instance of fra traction of a 2010 Blood article by Sa lowed in rapid succession by the ret articles originating from the laboratory

• Fabricatie, Falsificatie, Plagiaat (FFP) • Fang et al. (2013) PNAS • Questionable research practices • Levelt et al. (2012) • Rapportage resultaten • Fanelli (2011)

Fig. 1. (A) Number of retracted articles for specific causes by year of retraction. (B) Percentage of published articles retracted for fraud or suspected fraud by year of publication.

because of fraud or error differed significantly from that of jour-

Citation of Retracted Articles. Previous that many retracted articles continue t work (15, 16), but others have docume of retraction on citation frequency (1 examine this question comprehensivel variation among the most frequently (Table 3). Some retracted articles ex tained decline in citations following re continued to be cited (Fig. S3).

Discussion In addition to confirming a recent ris tractions, this study provides a numb Perhaps most significantly, we find th

Wat is het probleem • Fabricatie, Falsificatie, Plagiaat (FFP) • Fang et al. (2013) PNAS • Questionable research practices • Levelt et al. (2012)

Steeds meer moesten de Commissies tot de conclusie komen dat er, ook zonder strikte fraude, in algemene zin sprake was van een cultuur waarin op slordige, selectieve en niet-kritische wijze met onderzoek en data werd omgegaan. Het ging niet om kleine ‘normale’ onvolkomenheden in de statistische bewerkingen of

• Rapportage resultaten

opzet en uitvoering van het experiment, maar om schendingen van fundamentele regels van ordentelijk

• Fanelli (2011)

wetenschappelijk onderzoek die de conclusies van het onderzoek zeer ernstig kunnen beïnvloeden. (p. 47)

Wat is het probleem • Fabricatie, Falsificatie, Plagiaat (FFP) • Fang et al. (2013) PNAS • Questionable research practices • Levelt et al. (2012) • Rapportage resultaten • Fanelli (2011)

Omvang probleem • Onbekend • Groeiend bewustzijn • Angst dat het om enorme aantallen zou gaan • Risico voor beleid

Wie is verantwoordelijk voor oplossingen "Vitalisering bestaande regels" door wie? • De werkgever • Directe collega's (rol whistleblowers) • "Het vak", tijdschriften, etc.

Wie is verantwoordelijk voor oplossingen "Vitalisering bestaande regels" door wie? • De werkgever

Het is aan de onderzoeksinstellingen om een klimaat te scheppen en te onderhouden ... Colleges en

• Directe collega's (rol whistleblowers)

besturen van onderzoeksinstellingen dienen erop toe te zien dat de zorgvuldigheid .. werkelijk aandacht krijgt.

• "Het vak", tijdschriften, etc.

.. met nieuwe lokale toezichtmaatregelen in de vorm van kwaliteitsprotocollen in universiteiten en universitaire medische centra dient terughoudendheid te worden betracht. .. Er ligt .. een zware verantwoordelijkheid op de schouders van de leden van wetenschappelijke gemeenschappen om voortdurend kritisch te blijven op de normen en waarden voor de omgang met onderzoeksgegevens in het eigen vakgebied. Commissie Schuyt, 2012

Wie is verantwoordelijk voor oplossingen "Vitalisering bestaande regels" door wie? • De werkgever • Directe collega's (rol whistleblowers) • "Het vak", tijdschriften, etc.

De commissie is tegen het van buiten af of van boven opleggen van regelgeving en toezicht. Toch is dat de enige oplossing. .. We hebben allang niet meer van doen met amateurwetenschappers met een roeping. Wetenschap is geïnstitutionaliseerd, waardoor de rol van de wetenschappelijke verenigingen is geminimaliseerd. Onderzoekers communiceren met elkaar, soms via wetenschappelijke verenigingen, maar ze zijn niet aangesteld in of door verenigingen. (Miedema, CvB UMC, 2012)

Wie is verantwoordelijk voor oplossingen "Vitalisering bestaande regels" door wie? • De werkgever

Hoe worden fraudeurs ontdekt? (Stroebe et al., 2012)

• Directe collega's (rol whistleblowers) • "Het vak", tijdschriften, etc.

5% 8% 50%

28% 10%

Whistleblowers Audits Outsiders Non-Replicatie Peer review

Beleidsmogelijkheden • Zelfregulatie • Preventie • Toezicht • Opsporing

Beleidsmogelijkheden • Preventie

Leren van justitie? Pakkans misdrijven

(CBS, 2011)

• Zelfregulatie • Toezicht • Opsporing

85% op heterdaad waarvan 87% door burgers Bron: NPA (2007)

Concrete maatregelen tegen FFP en QRP • Extern delen van (gepubliceerde) data, materialen, syntax, etc • Audits • Verdenkingen en klachten • Vertrouwenspersonen, procedures voor onderzoek, strafregeling, behandeling en resocialisatie, • Intern delen van data, materialen, syntax gedurende onderzoeksproces • Procedures en protocollen voor onderzoek, analyse, opslag, rapportage • Ethisch klimaat, wetenschappelijk debat

Verwachtte effectiviteit?

Fraude

Questionable Res Pract

Extern delen data Audits Regeling verdenkingen en klachten Intern delen data Procedures en protocollen Ethisch klimaat, wetenschappelijk debat

Effectiviteit extern delen data •Onbekend •Verwacht geen wonderen ! gebrekkige peer-review •Kosten en risico's? •Voordeel: Klinkt als "hard" beleid

Rapportage

Effectiviteit audits •Voordeel: klinkt als "hard" beleid •Symbolische waarde •bijdrage verandering cultuur •Door beperkte rijkweidte geringe effectiviteit

Effectiviteit verdenkingen en klachten • Vertrouwenspersonen: ? • Integriteitsonderzoek: • kostbaar • Met name effectief als verdachte bekent • Gebreken in procedures • Onderzoek door instelling ! problemen vraagstelling • Samenstelling integriteitscommissies • Gebrek regie LOWI, geen standaarden

Effectiviteit intern delen data •Toezicht door collega-experts: bewezen effectief •Geïntegreerd beleid: klimaat, debat, & toezicht

Intern toezicht Stapel netwerk normaal

Aantal auteurs op artikel 1

2

12%2%

34%

3

4 en meer

6%2%

14% 52%

netwerk Stapel 78%

Niet frauduleus

Frauduleus

Effectiviteit procedures, protocollen, klimaat, debat • Nadeel: klinkt als "soft" beleid • Maar hier is een evidence base: • redelijk harde effecten

Journal of Applied Psychology 2010, Vol. 95, No. 1, 1–31

© 2010 American Psychological Association 0021-9010/10/$12.00 DOI: 10.1037/a0017103

Bad Apples, Bad Cases, and Bad Barrels: Meta-Analytic Evidence About Sources of Unethical Decisions at Work Jennifer J. Kish-Gephart, David A. Harrison, and Linda Klebe Trevinõ Pennsylvania State University, University Park Campus

As corporate scandals proliferate, practitioners and researchers alike need a cumulative, quantitative understanding of the antecedents associated with unethical decisions in organizations. In this metaanalysis, the authors draw from over 30 years of research and multiple literatures to examine individual (“bad apple”), moral issue (“bad case”), and organizational environment (“bad barrel”) antecedents of unethical choice. Findings provide empirical support for several foundational theories and paint a clearer picture of relationships characterized by mixed results. Structural equation modeling revealed the complexity (multidetermined nature) of unethical choice, as well as a need for research that simultaneously examines different sets of antecedents. Moderator analyses unexpectedly uncovered better prediction of unethical behavior than of intention for several variables. This suggests a need to more strongly consider a new “ethical impulse” perspective in addition to the traditional “ethical calculus” perspective. Results serve as a data-based foundation and guide for future theoretical and empirical development in the domain of behavioral ethics. Keywords: unethical behavior, intuition, decision making, intention

For over 30 years, researchers have attempted to determine why individuals behave unethically in the workplace. Once viewed as the province of philosophers—a “‘Sunday school’ subject not worthy of serious investigation”— behavioral ethics has become a legitimate and necessary field of social scientific inquiry (Trevinõ, 1986, p. 601). Indeed, as ethical scandals have garnered attention across multiple sectors of society (e.g., business, government, sports, religion, education), research examining the determinants of individual-level unethical choices at work has grown dramatically (for reviews, see O’Fallon & Butterfield, 2005; Tenbrunsel & Smith-Crowe, 2008; Trevinõ, Weaver, & Reynolds, 2006). Between 1996 and 2005, over 170 investigations were published (O’Fallon & Butterfield, 2005). Yet, despite this increased attention, much remains to be understood about how and under what circumstances individuals make unethical choices. Recent qualitative reviews of the behavioral ethics literature (O’Fallon & But-

quantitative summaries to “derive statistically valid conclusions” about the proposed antecedents (O’Fallon & Butterfield, 2005, p. 405; see also Robertson, 1993). In this paper, we attempt to provide a clearer empirical and theoretical picture of what we know (and don’t know) about multiple sources of influence on unethical behavior at work. We begin by using two meta-analytic techniques to summarize evidence of the influence of individual characteristics (cognitive moral development, locus of control, Machiavellianism, moral philosophy, and demographics), moral issue characteristics (i.e., moral intensity; T. M. Jones, 1991), and organizational environment characteristics (ethical climate, ethical culture, and codes of conduct) on unethical choices. We then investigate the potential moderating effect of using intention as a proxy for behavior—a practice commonly employed by behavioral ethics researchers (see O’Fallon & Butterfield, 2005; Weber, 1992; Weber & Gillespie,

B

MORAL ISSUE CHARACTERISTICS

Effects of Moral Issue Characteristics

Concentration of Effect

-.07*

Magnitude of Consequences

-.07

Probability of Effect

-.17*

Proximity

-.10*

Social Consensus

-.23*

Temporal Immediacy

-.05

C

ORGANIZATIONAL ENVIRONMENT CHARACTERISTICS Egoistic Ethical Climate

.04; .09*

Benevolent Ethical Climate

-.17*; -.07*

Principled Ethical Climate

-.19*; -.21*

Ethical Culture Code Existence Code Enforcement

UNETHICAL INTENTION

Effects of Organizational Environment Characteristics

.13; -.08 .06; .10*

UNETHICAL INTENTION

UNETHICAL BEHAVIOR

-.08*; -.33*

Figure 2. Simultaneous unique effects of proposed antecedents on unethical intention and behavior. First entry is effect on unethical intention; second entry (after semicolon) is effect on unethical behavior. ! p ! .05.

Verwachtte effectiviteit, samengevat

Extern delen data Audits Regeling verdenkingen en klachten Intern delen data Procedures en protocollen Ethisch klimaat, wetenschappelijk debat

Fraude

Questionable Res Pract

Rapportage

~/+ ~/+ ~/+ + + +

~/+ ~/+ ~/+ + + +

nvt nvt nvt nvt

+ +

Implementatie tot op heden • Extern delen van (gepubliceerde) data, materialen, syntax, etc • Audits • Verdenkingen en klachten • Vertrouwenspersonen, procedures voor onderzoek, behandeling en resocialisatie, strafregeling • Intern delen van data, materialen, syntax gedurende onderzoeksproces • Procedures en protocollen voor onderzoek, analyse, opslag, rapportage • Ethisch klimaat, wetenschappelijk debat

Conclusie • Verantwoordelijkheid voor aanpak ligt bij werkgever in samenwerking met medewerkers • Verantwoordelijkheid leggen bij "het vakgebied" is uitstel van oplossing • Open data: doen, maar voor experimenteel onderzoek geringe opbrengst • "Softe" factoren hebben harde effecten • Balans tussen aandacht voor "output" en aandacht voor proces

UMCG 18 juni 2013

Recommend Documents