1. When Search Becomes Research – Part I
Query design
Anat Ben-David
Science, Technology and Society, Bar-Ilan University
anatbd@gmail.com
Erik Borra
Digital Methods Initiative, University of Amsterdam
erik@digitalmethods.net
2. When Search Becomes Research
Turning Google into a research tool
“We look at Google search results and see society, instead of
Google” (Rogers, Stevenson, Weltevrede 2009)
3. Words/keywords
When words become “issue language”
Actors and their terminology: Keywords as positioning efforts
Note program, anti-program as well as efforts at net neutrality (Cf.
Akrich & Latour, 1992)
4. Keywords and source sets
“side by sidedness” offical / non-official … different kinds of actors
(living side by side) in an issue map. (Rogers, 2004)
5.
6.
7.
8.
9.
10. example:
Program / anti-program
How is the issue of “google street view” and privacy is being treated when google-
related sites are excluded from the search?
1. “Google street view” +privacy 33,100,000 results
2. “Google street view” +privacy site:google.* 1,840 results
3. “Google street view” +privacy –site:google.* 36,000,000 results
11. search/research
Search operators and syntax
Use, for example: +, ~, OR, NOT, SITE:, “”
See, for example: http://www.googleguide.com/category/query-
input/
12. example:
Program / anti-program
A query design example using advanced operators:
• “~Cellular Phone” + “brain tumor” + “not associated”
• “~Cellular Phone” + “brain tumor” + “270%”
• Compare two queries across different actors. Add “site:.edu”,
“site:.com”, .”site:.org”, etc.
13. search/research
Research protocol For using Google
Google Settings:
• For the “universal Google” go to http://google.com/ncr or http://
google.com/intl/en
• Log out of your Gmail account
• Google preferences:
Set interface and search language
SafeSearch: Off
Google Instant: Off
Nr of Results: 100 per page
14. search/research
Research protocol For using Google
Clean browser
• Log out
• Clear cookies and the browser’s search history
• Or: create a “research browser” (i.e. install a new one)
“Turning off search history personalization”
15. example:
Program / anti-program
A query design example using advanced operators:
• “~Cellular Phone” + “brain tumor” + “not associated”
• “~Cellular Phone” + “brain tumor” + “270%”
• Compare two queries across different actors. Add “site:.edu”,
“site:.com”, .”site:.org”, etc.
16. Example:
nationality of issues: Rights types
Can the search engine be repurposed to show which rights are specific per country?
Method
1. Query the term "rights" in national terminology per different Google
country (e.g. ‘droits’ in .fr, ‘rechten’ in .nl)
2. Fetch the top 10 unique rights types.
3. Visualize top 10 issues per country and mark unique issues.
https://wiki.digitalmethods.net/Dmi/NationalityofIssues
18. search/research
Research protocol
Saving results for verification and retrieval
• “Save page as” in the browser, name files and folder consistently
• Collect right types in spreadsheet (incl. translation)
• Merge results and collect saved files in one place
19.
20.
21.
22.
23. search/research
Questions and related tools
Using Lippmannian Device aka Google scraper:
Resonance of controversial terms
• What are the relevant issues in the controversy?
• Where do controversial terms resonate?
24. EXAMPLE:
CLIMATE CHANGE SKEPTICS
Where do the skeptics get “air time”? Where are their audiences?
BBC cancels ‘Planet Relief’ program about environmental issues
“The only reason why this became an issue is that there is a small but
vociferous group of climate ‘skeptics’ lobbying agains taking action”
- BBC News, 5 september 2007
https://wiki.digitalmethods.net/Dmi/ClimateChangeSkeptics
25. example:
climate change skeptics
Query design: What are the sources?
Top 100 results for the query “climate change”
http://www.google.com/search?q="climate+change"&num=100
26. example:
climate change skeptics
Query design: What are the issues?
Derive list of climate change skeptics
Sources: motherjones.com, wikipedia.org, heartland.org
Compare the three lists and retain the skeptics that are mentioned in
at least two of the lists
27. example:
climate change skeptics
Skeptics
S. Fred Singer
Robert Balling
Sallie Baliunas
Patrick Michaels
Richard Lindzen
Steven Milloy
Timothy Ball
Paul Driessen
Willie Soon
Sherwood B. Idso
Frederick Seitz
28. example:
climate change skeptics
Google Scraper: Batch query Google
http://tools.issuecrawler.net/beta/scrapeGoogle
Enter sources in the top box
Enter keywords in the bottom box (mind the quotes)
Click “scrape Google”
29. Warning: excessive usage will bring this tool down
Make sure to pay attention to query design
Body Text
Body text