Statistics on the Web: VIVA and Beyond

1
Statistics on the Web VIVA and Beyond

• Donna Tolson
• Scholars Lab
• University of Virginia Library

Virginia State University Radford
University June 6 7, 2007
2
Todays topics

• General stuff about statistical information

Finding using domestic statistics
Finding using international statistics
3
Texts vs. Statistics

• Qualitative
• Database collection of texts or citations
• Key word
• Rarely need metadata
• Majority published privately

• Quantitative
• Database collection of observations
• Variable
• Usually need metadata
• Majority published publicly

4
Statistics vs. Data

• Usually aggregated
• Count, , µ, xy, S, ranking
• The quick answer
• WYSIWYG
• For use as context, evidence, or argument

• Usually microdata
• Matrix or spreadsheet of values
• Never a quick answer
• Exists to be analyzed
• For research, testing hypotheses

5
A few things to know.

• Statistics come from data about populations
  (enumeration) or samples (survey).
• The bigger the sample, the more reliable the
  statistics.
• Aggregated statistics generally cant be
  dis-aggregated, or filtered. This is what
  microdata is for.
• Confidentiality is strictly enforced affects
  access to statistics about small populations and
  microdata.

6
The statistical reference interview
How will the information be used?Context
statistics vs. Analysis dataset
Statistics about what or who?Person, household,
country, etc. unit of analysis
When?Most current vs. time series
Where?Geographic level as well as precise
location, Single vs. multiple locations
Beware the data fallacyJust because someone
asks the question doesnt mean there is an answer!
7
Locating statistics

• In library catalogs, search for specific variable
  (e.g. income) AND statistics
• Web sources may not be cataloged time to Google
  
• Work backwards from data sources cited in
  articles
• Consult comprehensive sources
• University web sites Michigan, Virginia
• Government web sites FedStats
• Published statistical databases LexisNexis

8
Domestic Statistics Government sources

• Statistics available on most anything regulated
  or funded by government
• Shrinking number of pre-analyzed statistics
• Growing number of interactive data sites
• Rarely, government withdraws data ex.
  immigration statistics
• Data you wont find in government sources
• Religion data
• Market-based information (who buys what)
• Polling data or opinion surveys

9
Interactive data sites

• PROs
• Increased access to large datasets
• Allows selection of variables
• Customizable queries
• Automatically handles weighting, etc.
• Can omit need for statistical software

• CONs
• No standard system
• Multi-step process
• May use short, confusing variable names
• Rarely includes analysis
• Not for statistical novices

10
Locating U.S. statistical info

• Begin with StatUSA or FedStats or U.S.
  Statistical Abstract
• Is geography or subject more important? Search
  accordingly.
• If youre in search of a few statistics, reports
  are fine. If youre in search of many statistics
  or data, skip pdfs and look for online or
  downloadable data (unless youre looking for
  citations of data sources).

11
Federal sites worth getting to know

• American Factfinder for demographics, economics
  housing
• CDC for health data and vital statistics
• BJS for crime law enforcement
• NCES for education
• BLS for labor force and employment

12
Locating historical statistics

• Generally, online data only available from
  mid-1990s
• Selected sources for historical data
• FedStats look for historical statistics links
• Census
• Selected historical data and publications
• Historical Census Browser
• Be ready to help patrons navigate off the web and
  into print Documents (oh no!!)

13
International Statistics

• Several comprehensive sources
• Most countries have statistical agencies, but few
  have as much free information as U.S.
• Global statistics are part of the data fallacy
  generally an aggregate of as many countries as
  possible is the best you can do
• Use IE

14
Comprehensive Sources

• UN Common Database (274 countries areas)
• Organization for Economic Cooperation and
  Development (30 member countries)
• Eurostat (EU member countries)
• World Bank (185 member countries)
• Nationmaster

15
Country-specific Sources

• Most other countries have one or two agencies in
  charge of producing statistical information
• BLS Statistical Sites on the Worldwide Web
• Univ. of Auckland Official Statistics on the Web
  
• Remember to
• use IE
• look for English translation (can be tricky!)

16
Locating NGO statistics

• Public opinion surveys polls
• World Public Opinion
• Some polling archives offer some data for free
• The Gallup Poll exit polls
• Roper Center for Public Opinion presidential
  approval ratings
• Statistics data on religion
• Association of Religion Data Archives
• Adherents.com
• Market-based data

17
Thank you!

• Donna Tolson dtolson_at_virginia.edu