17:610:530 Principles of Searching

Tefko Saracevic

EXERCISE 3

The exercise consists of three parts. Part 1. is a Dialog Refresher, designed for you to test yourself on the basic aspects and commands of Dialog searching. You do not have to turn in anything for the Refresher, but you are expected to master these things. So test yourself and make sure that you are at ease with these topics.

The second part consists of a number of questions for searching. The third part deals with question analysis. Turn in the results as explained below.

Part 1. Dialog Refresher

Each student should review whether s/he is mastering the following topics in Dialog searching. These are reflected in exercises provided, but each student should also do own exercises to master the topics. The topics are covered in Walker and Jones textbook and in Dialog tutorials.

Topics

E.g. Bluesheets can be found with each file; in addition File 415: Dialog Bluesheets, contains all the Bluesheets. File 415 can be searched to identify what files cover a given topic. If paying, use of file 415 is free. Where else can you find them?

e.g. help field 61 -to display a list of searchable fields for file 61; help file n - to display description of file n; help fmt n - to display format options for file n, etc.

e.g.begin 1 or b 1; b 1,61,202,438

e.g. select (or s ) library AND automation; select steps (or ss) library AND automation; s librar?; s cat?; s cat? ?; s catalog? ??

e.g. ss library(w)automation; ss librar?(3N)automation

e.g E library; E Au=Saracevic; s e3

e.g. ss library AND automation; ss library OR automation; ss library NOT automation; ss (librar? or school? or universities) AND (automation OR computers)

e.g. how to search for titles or descriptors only: select library/TI, select library/DE; how to do a search for documents from a given author, or a given journal - select AU=Saracevic, T?; s JN=Library Journal

e.g. s s3 AND New Jersey; s e3, e4, e6

e.g. if b 1, 61, 202 438, and s library AND automation) after the set retrieved in command line enter rd (or remove duplicates), to remove duplicate documents retrieved from different files.

e.g. t (or type) s1/5/1-12; can also be expressed as t 2/5/1-12; t 2/5/2,5; t 2/5, KWIC/1-12. KWIC highlights the windows of text that contain your search terms.

e.g ds (or display sets) - tells you the history of all sets formed since the last begin command. You can also display a range os sets - ds 10-22.

e.g. b 411; sf (select files) all social; select library AND automation. To then search given files you must enter the begin command - b 1, 7, 34,....

e.g command logoff hold disconnects you, but also holds your existing sets for 30 minutes so you can re-connect and continue your search.

e.g. help rates 1; to check cost during a session enter command costs. Also costs are displayed at logoff or at changing files.

e.g. command save temp saves temporarily the entire search strategy since the last begin command. The saved search is stored for seven days.. The saved search will get a number beginning with T like TC013. To execute the saved search after you began a file or files enter command execute steps (or exs) Tnnnn..

e.g logoff, also: log, logout, off, bye, quit, stop disc

Reminder: You should know about these things. Test yourself. You do not have to turn in anything on this refresher.

Part 2. Dialog exercises

Deliverables: Turn in the results as answers to given questions and as printouts where required to do a search.

1. What is the difference if you search in ERIC for various forms of the term 'information retrieval.' E.g.

    1. Describe the difference. Can any other combinations be searched?
    2. Which statements give the smallest, next smallest etc. to largest? (This is not a Dialog command - you should simply list them by hand).
    3. What happens when you expand on ' information' and 'information retrieval'? Can you find any related terms? Hint: "information retrieval" is a descriptor and next to it is a number under R. Expand the e number for "information retrieval' e.g. e e7 (if "information retrieval" was under e7).

2. Find citations to articles about searching of various web engines. Use ERIC, Library Literature, and Information Science Abstracts. Use OneSearch to search databases together. Eliminate duplicates. Type five most recent articles in short and full formats.

3. Use Dialindex to find the databases that have the most documents covering the subject of Internet searching in public libraries. Save the statement temporarily. Use your original search statement to search the two databases with the greatest number of citations. Examine a few of the records in each database to find which descriptors, if any, are used to cover this topic. Print four documents that you found with relevant descriptors.

4. Find newspaper reviews about the movie "The Pianist." Possible source are PAPERS files, or a subset from these. Type five references in format 2. 

Part 3. Expansion of terms in a query

Background

When a query term says 'player', it really means that you should treat it as a concept that may be expressed in other ways, i.e. also to search for terms such as 'contestant, athlete, participant, performer', and other related terms as warranted. A shorthand for this is called query expansion or question analysis. It means listing (as warranted) synonyms, near synonyms, and related, narrower and broader terms to be connected by a logical OR.

One method for doing this is to

Here is a general table for a question with three concepts that have a number of related terms connected by OR and then the three are to be connected by AND's:

 

 

Concept A

AND

Concept B

AND

Concept C

..OR..

A1

A2

A3

A4

.

.

.

B1

B2

B3

.

.

.

C1

C2

.

.

.

 

Exercise for part 3:

Here is a topic (question): Alcohol abuse among gifted teenagers.

  1. Identify main concepts. For each concept enumerate related terms as in the grid shown above. You can find a file that has a thesaurus (such as ERIC or PsychINFO) and add the terms, if any, from the thesaurus. You can do it from the top of your head. You can find some relevant documents and add terms from them. You can look in a textbook. You can talk to colleagues. You can use your imagination…
  2. Make the grid, as above, and hand it in.
  3. Select files (from Dialog or any other sources of your choice) that are most likely to contain answers for this question.
  4. Hand in the selection and of the files, with a brief explanation/evaluation as to why did you select them.
  5. You can do a search, but do NOT hand in the results. You can evaluate the results for your own use.

Thus the 'deliverables' are: