Searching the keyword search engines Internet Research Skills Niall O Dochartaigh, Dept' of Politica

1
Searching the keyword search engines Internet
Research Skills Niall O Dochartaigh, Dept. of
Political Science and Sociology
2

• Get to know the literature first
• Identify key sources
• Develop a knowledge of your subject area
• Building on this, devise specialised queries.
• Search engines can reinforce offline inequalities
  in power and visibility
• Google and Chinese government censorship

3
The big four Google, Yahoo, Ask.com, Windows Live

• Google the largest index
• None are comprehensive
• Huge overlap, but much that is unique
• Use all four for a needle in the haystack
  search
• Diferent ranking criteria results ordered
  differently
• Google and Yahoo Word documents, PDF files,
  PowerPoint and Excel files

4
Grouping results

• Windows Live more results from
• Yahoo more from this site.
• Google and Ask.com group results from the same
  site together

5
(No Transcript)
6
(No Transcript)
7
(No Transcript)
8
(No Transcript)
9
(No Transcript)
10

• Default is an AND query
• Google gives higher ranking to exact phrases
• Word order is significant
• Case is not eg Snow White snow white
• Limited stemming eg child children

11

• Search Engine Watch (searchenginewatch.com)
• Search Engine Showdown (searchengineshowdown.com)

12
Google (www.google.com)

• The ranking system
• similar pages
• Ten word maximum

13
(No Transcript)
14
(No Transcript)
15
Yahoo (www.yahoo.com)

• Googles largest competitor
• Ranking popularity and relevance
• Links to the Directory

16
(No Transcript)
17
Ask.com

• Ranking Subject-Specific Popularity.

18
Windows Live

• Great ambitions

19
Advanced searching

• Windows Live option after you have run an
  initial search
• Ask.com option after you have run an initial
  search

20
(No Transcript)
21
(No Transcript)
22
(No Transcript)
23
(No Transcript)
24
Search for the exact phrase

• outrageous assault
• Restrict by language

25
Restrict by date

• Yahoo and Google last three months, six months
  or year
• Ask.com sophisticated options
• Displaying results

26
Restrict by location

• Windows Live geographical tags
• Windows Live Adjusting the ranking criteria
• File formats

27
(No Transcript)
28
Restrict by site / domain

• siteedu
• siteharvard.edu
• sitewww.ksg.harvard.edu
• Cant include the file path

29
Searching URLs

• anthro
• www.un.org
• www.un.org/documents
• inurlwww.un.org/documents

30
Searching titles

• intitleanthology

31
Searching for links

• Google and Windows Live linken.wikipedia.org
• linken.wikipedia.org/wiki/Quaero
• Alexa.com

32
(No Transcript)
33
Search tip An advanced search using link and
domain

• linkboingboing.net/archive.html
• -siteboingboing.net

34
Meta Search Engines

• Simplified queries
• Specialised databases
• Beaucoup (www.beaucoup.com)
• Dogpile (www.dogpile.com)
• Metacrawler (www.metacrawler.com)
• Search Allinone (www.searchallinone.com/)

35
Comparing search engines with meta search engines

• A needle in the haystack search
• "CO 730" Archives Iraq mandate
• Amazon.co.uk, Google Book Search and Google
  Scholar

36
(No Transcript)
37
(No Transcript)
38
Boolean searching

• Israel AND (Palestine OR Palestinian),
• revolution AND (communist OR socialist)

39
(No Transcript)
40
Math Searching
41
How Search Engines Search

• They dont search the Web
• Searching an index
• Robots
• Theyre not up-to-date

42

• Priority to major sites and to rapidly changing
  sites
• Search engines often bring up hits which dont
  exist
• Cached Web pages

43

• Missing out on the databases
• robots.txt
• Sampling big sites
• Indexing part of a document
• Many pages have scarcely been indexed at all

44
Devising search queries

• Know your subject
• Know what youre looking for
• Develop specialised queries based on this
  knowledge
• Start your search with a reality check
• Flexibility

45

• If your query does not turn up the required
  results in the first ten or twenty hits devise a
  new query and search again, and then devise
  another, and another

46

• If the item youre looking for is in the fifth or
  sixth screen of hits then effectively your search
  hasnt found it at all
• Learn how to quickly assess the potential
  relevance of links

47

• Aim for a unique combination of words and phrases
  which is certain to appear in a document directly
  related to your topic but which could not
  possibly appear in other documents
• An unusual combination of just a few words can be
  unique

48
Search terms
49
Quotes

• The quote does not have to be meaningful
• Workers of the World Unite
• "tale of the spectre"

50
Names

• Lenchen Demuth

51
Place names

• Even the most placeless and theoretical topics
  will often be associated with particular places
• Common place names can very effectively modify a
  search that is already quite specific

52
Subject phrases
53
Look before you leap

• Identifying common genres
• College course descriptions
• www.anycollege.edu/econ/Taylor/macro.html
• www.anycollege.edu/psych/courses/py101.html
• Bibliographical references
• Link pages

54
Technical difficulties

• Contents lists
• frisbees.htmlred
• When back doesnt work
• Parasite frames

55
Understanding error messages

• Error 400 bad request
• Error 401 Unauthorised
• Error 403 Forbidden
• Error 404 Not Found
• Error 500 Internal Server Error
• Too many connections Network connection
  refused by server Host not found