Data archiving in Canada: problems and prospects Presentation to NRF by Laine GM Ruus University of Toronto. Data Library Service 16/05/2002.

Data archiving in Canada: problems and prospects

Presentation to NRF

by Laine GM Ruus

University of Toronto. Data Library Service

16/05/2002

Outline

• Why archive data?

• Current problems in Canada (and a personal view of some solutions)

• Current trends, in problems and solutions

Why archive data?

• Promotes research ethics/research integrity

• Enables replication analysis

• Allows re-analysis with refined or new social theories

• Allows re-analysis with new techniques (eg. bootstrapping, etc.)

Why archive data (cont’d)

• Allows comparative analysis, similar data with new universes

• Enables analysis of change, similar data at different points in time

• Maximizes return on initial investment in data collection

• Enables training of policy makers with local data

Why archive data (cont’d)

• Promotes increased numeracy in the population

• Preservation for future generations of an aspect of our culture not measured by other means.

Current climate for data archiving in Canada

• 3 major data producers: government, academia, commercial sector

• Copyright Act:– Crown copyright: government produced data

and information products belong to the Queen

• Government information policy– set by Treasury Board for government, without

consultation with non-government sectors

Government sector as a data producer

• Statistics Canada is major data collector for socio-economic data

• Government has no data archive

• National Archives has not collected data over last 15 years

• Treasury Board policy since ca 1984 treats government data and information products as a commodity

Academia as a data producer

• SSHRCC has data deposit regulations, but no enforcement

• CIHR (formerly MRC), and NSERC have no data deposit regulations

• individual university research funds have no data deposit regulations

• no national data archive

Academia as a data producer (cont’d)

• no tradition in academic sector of depositing research data

• attitude of individual ownership vis-à-vis data files

• history of using US data/10

• no tradition of citing data files in publications

Academia as a data producer (cont’d)

• Few requirements among periodical editors requiring citation of data files

• no tradition among tenure boards to treat creation of a data file as equivalent to publication

• only commercial value in software and applications awakening universities to their rights under Copyright Act.

Commercial sector as a data producer

• Subject to Copyright Act and new Personal Information Acts

• No national data archive

• No uniformity in approach to archiving their data

• No national body with which to negotiate arrangements

Solutions to the Canadian problems (a very personal view)

• Canada needs a national information policy

• Canada needs a national data archive

• Government (all sectors) need government data archives

• Need to promote a culture of data deposit and data sharing in the academic sector (SSRC, CIHR, and NSERC, etc.)

Solutions to the Canadian problems (a very personal view) cont’d

• Need to educate hiring bodies and tenure boards that data file creation is a valuable academic activity

• Need to sell commercial sector on benefits of data archiving and data sharing

• Need to promote numeracy in the population.

Recent trends in the data archiving/data service sector

• Longitudinal data

• Research data centres

• DDI/DTD

• WWW data extractor interfaces

• GIS

• Proliferating formats

Recent trends: longitudinal data

• Data producers increasingly collecting longitudinal/panel data

• Enhanced capability to test theories re social change over time

• Increased problems of preserving privacy and confidentiality

• Requires more sophisticated research techniques

Recent trends: research data centres

• Secure access to more detailed or sensitive data

• Creates segregation of research capabilities (data haves vs data have nots)

• Data producers less likely to produce public use microdata files

Recent trends: DDI/DTD

• Data Documentation Initiative Data Type Definition

<http://www.icpsr.umich.edu/DDI/>

• A standard format for metadata describing microdata files

• Being expanded to encompass aggregate and time-series data

Recent trends: data extractors

• Data extractors provide access to data/analyses of data via Internet protocols

• Enabled by development of DDI/DTD

• Selected data extractors linked at:<http://www.chass.utoronto.ca/datalib/misc/dli/extracts.htm>

Recent trends: data extractors (cont’d)

• Two major data extractor developments:– NESSTAR

<http://www.nesstar.org>– Virtual Data Center project

<http://TheData.org>

Recent trends: GIS

• Geographic information systems

• New theoretical models based on spatial analysis

• New software capable of spatial analysis (ArcGIS)

• Increasing demand for geocoded aggregate and microdata

The nice thing about standards is that there are so many to

choose from!

Recent trends: proliferation of formats

• Data archiving becoming more difficult

• Many new proprietary formats and flavours to deal with

• Increasing number of formats for which we have not yet developed preservation formats, eg GIS shape files, relational databases, etc.

To finish...

• Without data archives, we will loose about 50 years of our culture

• Successful long-term preservation of our electronic culture will partly depend on bringing copyright legislation, internationally, into the 21st century

• Data archiving is not a national problem, nor a problem that is unique to any one country. The problems and solutions are similar in all countries. We can all learn which solutions are best and/or worst from each other.

• IASSIST (International Association for Social Science Information Service & Technology) is one of the venues in which we learn from each other.

<http://www.iassistdata.org/>

• This presentation is available at:<http://www.chass.utoronto.ca/~laine/misc/sada02.ppt>