ONLINE SEARCHING

Online Searches are used in:

• The Library Catalog
• Indexes of Journal Articles
• Internet Search Engines
• Other Databases

The basic concepts and techniques presented here can help you understand how to do online searches using any of these information tools.

Most online search formats have examples or tutorials that will help you apply the basic search concepts and use options that are available.

About Databases

A database is an organized collection of information.  We usually think of an electronic database, but print resources and other formats may also be considered databases.  A database is a way of linking one type of information to another: A phone book links names to phone numbers; a class schedule links class names to times and rooms; a journal index links subject terms to citations of related articles.

A database is made up of records, items with associated information linked to them.

Records are made up of fields containing specific information (the data.)

Fields that can be searched are access points for getting to the information—using one field to get to the record to see the rest of the linked information. 

A phone book is an example of a database with a print format and alphabetical searching.  The residential listings are by name, one access point.  The business pages are listed by name and the yellow pages list them by business type, a second access point.  Each record is linked to a phone number.  The fields are name, address, and phone number.  The business type is an added field in the yellow pages. 

Electronic databases usually have an advantage of more access points.  For example, card catalogs for libraries had listings by the author, title, and subject.  An online library catalog can also be searched for any word in the title or subject, the publisher, or other words used in the record.

Examples of Databases and what is in them:

Elements of a Search

An online index, the library catalog, or an Internet search engine will have a search interface as a way to use the online search tool.  The search interface will include a place to enter your search terms (query), something to start the search, and a way to see the results (retrieve, select, and capture.)

Query: Tell the search tool what to look for.  Set up the search terms typed in a box or form.  Options to target and limit the search vary among databases.  (More on this in the next sections of this guide.)

Search: Start the process by clicking on Submit, Search, Go, or something similar.  Often pressing Enter is an option.  This sets the matching process into motion.  The search terms are checked against all the records in the database, and any matches located.  Online searches have become so fast that we are not aware of the time taken to compare information and find matches. Imagine having to compare all the records by hand, or even early computer searches matching punch-card holes. 

Retrieve: A list of results (“hits,”) usually with brief information for each matched record.  Options to sort and refine the results vary.  Sometimes there is a limit to the number of matches that will be retrieved.

Select:  Mark an item of interest or open a complete record with additional information fields showing.   Many databases allow creation of a list.  Some databases have an option to select a link to full text.  Selecting an item from the results of an Internet search will open a web page. 

Capture: A way to keep chosen items.  The items can be a single record, a results list, a marked list, or full text.  The methods to capture them may be printing, e-mail, downloading, or saving to an account within the database.  

Basic Types of Searches

Browse searches only look at the beginning of the field to be searched.  (In alphabetical lists, A, An, or The at the beginning is almost always skipped.)  The result will show a list of records in alphabetical or numerical order, starting at a position nearest to being a match to the search term.   From there you can move up or down the list. 
Compare to phone book search: Opening to the right page and looking at the list.

Keyword searches look for a match anywhere in the fields to be searched.  The search may be for words used in specific fields or a search of all the fields of a record at the same time.  A match must be exact, so differences of spelling and punctuation can be important.  Keyword searches are often used in combinations.  The results list will show all items that include a match.
Compare to phone book search: Getting a list of everyone with the same street name or the same first name.

Specialized vocabulary searches bring up all the items that have been linked to a subject heading, whether or not the same words appear in the rest of the record.  Subject headings may be organized in a hierarchy of more specific terms.  Identification numbers are another type of controlled terms.
Compare to phone book search: Finding all the businesses under the same heading in the Yellow Pages.  For example, car dealerships are all found under “Auto Dealers” whether or not “auto” is in the name. 
Specialized or controlled vocabulary search setups often have available thesaurus or index of terms linked to the database to help identify subject headings.
Compare to phone book search: Using the index at the beginning of the Yellow Pages that helps locate subjects. 

Boolean searches combine the results of more than one search.  A Boolean search may involve two or more keywords or a keyword and a controlled vocabulary term, etc.  (More on this in the next section.)

Relevance searches rank items by how many terms match or how many times the chosen terms are used.  (More on this with sorting of results.)

Boolean Searches

Combining Search Terms

Boolean Operators

The basic concepts of connectors for different combinations were developed by the mathematician George Boole in the 1850’s.  The same relationships are used in math, logic, and computer operations, not just in setting up searches.

• Additional information: More about George Boole from a website about the Pioneers who made computers possible.

A AND B        Both are present                               A         B         AB     

A OR B           Either is present                                A         B         AB

A NOT B         A is present, B is excluded                A         B         AB

AND, OR, and NOT are called the Boolean operators.

These are often represented with circles called Venn Diagrams. 

In this example, the left-side circles represent all of the items in a database that use the word “cats” and the right-side circles represent all of the items that use the word “dogs.”

•  Additional information: An animated illustration from Creighton University gives another example of  Boolean Logic. 

Sample search terms using Boolean connectors:

AND narrows a search.  AND corresponds to All of these.

OR broadens a search.  OR corresponds to Any of these.

A memory phrase: “AND narrows the band—OR brings in more.”

NOT narrows a search.  Use NOT with caution, since it can eliminate useful items that use both terms.

In many databases, search terms can be linked to particular fields. 
For example the first sample search above might be made more specific:
(History in Subject) AND (Wisconsin in Title) AND (Forest in Keyword)

In some databases, there are more options using proximity—specifying that term A must be in the same field as term B, or that term A must be within a certain number of words of term B.   There may also be other variations on connectors.  See the examples or tutorial in your chosen database for more specifics.

• Additional information: Boolean Searching on the Internet from the libraries of State Universities of New York at Albany.

Advanced Search—check it out!

In many search formats, “Advanced Search” is the setup with boxes for multiple search terms that make it easy to combine concepts.  Sometimes the same type of setup is called “Guided Search” or has another name.   These can actually be easier to use than typing the terms and the combining operator into a single-line “Basic Search."  

“Command line” searches may also be available.  (In some databases this is what “Advanced Search” means.)  These are the searches that do require advanced skills.  There will be a specific syntax for setting up the search, such as an order of nesting parentheses or use of specific abbreviations to designate fields.  This is a way to craft a very precise search.  For a complex search or one that will be repeated, it may be worthwhile to go through the examples or tutorials for the database to learn more.

• Additional information: Four Nets for Better Searching, an introduction to using Advanced Search features in Google, from San Diego State University.

Search Terms and Limits

Choosing Search Terms

Explore possible terms

• Start by writing down what you are looking for and picking out the major descriptive words.
• Think of synonyms or other related words.
• Try to use distinctive terms.  Pick the words most likely to lead directly to your topic with few other meanings.  If there are multiple authors, search on the one with an unusual name.
• Avoid general terms such as “science” or “women” unless combined with other terms by AND

“Mining” for more terms

• Look at entries for your topic in dictionaries, encyclopedias, or other reference sources to find additional words or phrases to try as search terms.
• When you find some useful items, look at the descriptors used and try them as search terms.  In many databases, you can click on subject terms to perform a new search, or copy them into a search to combine them with other terms.  This can be a very effective way to search.

Phrases

Searching on the phrase “Lake Superior” is different than searching for all records that use the two words “lake” and “superior.”  Phrases are often enclosed in quotation marks as the signal to the search process to consider the contents as a single unit.  Some databases have other ways of designating phrases.

Truncation and Wildcards

Truncation is “cutting off” the end of a word to search for various forms of the word at the same time.

For example, child* will also bring in childhood, childish, children, and other related words.

Think about what else will be brought in by a truncation.  For example, import* will add importation, imports, importing and other related terms, but it will also add every item using the word important.  In that case, it may be better to put in the related terms linked by OR.

A truncation symbol is one type of wildcard, a symbol for a placeholder that may represent a single character, any number of characters, or even “one or zero” characters.  See the examples of the database for the types of wildcard searches available.   The asterisk (*) is the most common wildcard symbol.  Others used include question mark (?), number sign (#), and exclamation point (!).

Some search tools will automatically include plurals and sometimes other related terms.

• Additional Information:  Truncation and Wildcard Symbols used in a list of databases, from Chalmer Davee Library at UW River Falls.  Most of the databases used at UW Superior are listed.  (Caution: it is possible that a different vendor is used for access to a database, changing the search features.  If the suggested symbol doesn't seem to work, use the information in the database's own help and examples to find out what truncation and wildcard uses are supported.)

Stopwords

Most search tools will not do a keyword search on the small words that appear in many records. 

Examples: about, as, by, from, how, of, so, the, where, with.  The list varies somewhat by database.

A phrase may include stopwords; this is usually the best way to search for a title with small words.

A browse search may have stopwords, except that “a,” “an,” or “the” at the beginning are dropped.

For example, some possible searches for the book The Wind in the Willows:
Keyword search: “wind AND willows”
Phrase used as a search term: “the wind in the willows” or “in the willows”
Browse search: “wind in the” or “wind in the willows”

Other words that appear frequently may be difficult to search.  For example, the word “books” appears in so many records in a library catalog that it is best to specify a title or subject search 

Check the Spelling!

Because the search only brings in exact matches, any difference will affect the results.

• Typos and words you might need to check for the correct spelling.
• Variations—especially for very old or very new terms and terms translated from other languages.
• British spellings, such as behaviour or centre.
  • More information: American and British Spelling Variations, a support page for XPDNC, a directory of worldwide labour (labor) information.
• Even presence or absence of punctuation can make a difference; try it both ways.
• Most search tools look for both lower case and capital letters, so it is usually better to enter search terms in lower case.

Limits

When a limit is applied, the matching process does not look at all possible items.  Parts of the database are eliminated from consideration so the search is faster and more specific.  In some databases, the limit options only appear in the Advanced search.

Some common types of limits:

• Date—Limit to recent materials for topics requiring current information, choose a range of dates for a historical perspective, or break a search into sections by date
• Language—Limit to English only, or bring in documents in other languages.
• Publication Type / Document Type—Limits available vary among databases.  Some examples:
  • Limiting to “articles” in JSTOR will eliminate opinion pieces and book reviews
  • In MEDLINE and PsychInfo, “review” articles go over the state of knowledge and previous research on a subject; you may want to search just for review articles that could help you understand the, or you may need to eliminate them if assignment requires using only original research articles
• Limiting to Scholarly (Peer Reviewed) articles is available in Academic Search, and similarly a limit to Refereed publications is available in Expanded Academic.   See Scholarly Journals and Popular Press for more information.
• Specific Journals—Limit to just the major journals in your field.  This can be useful when many different meanings of the same terms bring in many unrelated hits.
• Limit to articles with full text available in the same database.

Each search format has particular limits that can be applied.  Some examples:

• JSTOR allows choosing one or more subject groupings of journals to be searched. 
• Academic Search allows a limit by number of pages in the article—eliminate brief or too-long items. 
• PsychInfo has limits by the age group being studied; gender of subjects; and human or animal studies. 
• American History and Life has choices of time periods.
• Google has options to search only images, or only particular types of domain names (.org, .edu).

There are many other possibilities.  Explore the available limits as you set up your search.

Some databases allow searching only the latest group of records added.  Running the same search in each new update would help you keep informed on a topic.

Limiting to “scholarly” or “peer-reviewed” publications is available in some of the large general journal indexes.  See “Scholarly Journals and Popular Press” for more information.

The Results Set

Out of all of the possible items in the database, your query has delivered a subset that match the search terms you entered.  The records brought in by your search are displayed as a list of results.

Looking at Results: Too Much or Too Little?

Online search results are a balance between Recall and Precision.

The goal for most searches is to be in the middle between high recall and high precision.  Expect to look through some items that are not on target to find the ones you want, and expect to do multiple searches as you refine what works. 

Higher recall is necessary for an in-depth project or as the background before doing original research. 

Higher precision is useful when only a few sources are needed.

In most cases, consider narrowing the search if you are scrolling through more than 100 to 150 hits. 

Measurements of search effectiveness often make mathematical calculations of recall (completeness of results) and precision (purity of results.)

Too Many Hits—Ways to Narrow the Search:

• Additional terms linked with AND
• Limit the fields to be searched—Title, Subject, or other specific fields
• Use the thesaurus or list of subject terms to find specific subject headings to search on
• Combine subject searches with additional terms
• Limit by Date, Language, Format, Publication Type, or other options
• Limit to scholarly or peer-reviewed publications (an option in some databases)
• If there are too many items that are relevant and it is difficult to choose among them, consider narrowing the topic to one aspect of the subject.

Not Enough Hits—Ways to Expand the Search

• Additional terms linked with OR
• Look for possible terms used in the items that are closest to what you need, especially subject terms.
• Look for synonyms or related terms to use—look first in the material you have found, then look in a thesaurus.  Try Credo Reference (in the Databases list--formerly XRefer Plus) or specialized encyclopedias and dictionaries in the Reference area.
• Use truncation to bring in other forms of the words used as search terms
• Search in full text where possible
• Try different databases
• Search off-line: Use the reference list of encyclopedia entries and other print sources to lead to other items
• Try a broader subject to find items that may have a chapter or section on your subject

Working with Results

Sorting

Depending on the search tool, results may be listed in various ways:

• Alphabetical order by author or by title
• Chronological—newest items first (or sometimes oldest items first)
  In the Library Catalog, sort by “Publishing Date Descending” to see the newest items first.
• Strongest matches first (Relevance ranking—more below)
• Numerical order, sometimes by an internal control number
• Items with most previous uses first (especially on the Internet)
• Clustering of like items (such as used by the search engine Vivisimo)
• Other criteria, or random

Many databases have options for changing the order of sorting, sometimes before the search and sometimes while using the results set.  There is often a maximum number of items that can be sorted, so it may be necessary to narrow a search to use the sorting.

More about Relevance Ranking

A complex mathematical weighting (an algorithm) assigns results to a higher or lower position in the list based on how strong the relationship is to the search terms.  Sometimes this is shown as a percentage.

Factors usually included: frequency (number of times the terms are used); proximity (use as a phrase or found near each other); all the terms found in the same field (one in the abstract and one in the author’s address gets a lower ranking); and specific fields (use in the title gets a higher ranking.)

A relevance search has a very high number of hits because items with lower ranking are still included, such as ones matching only one of the search terms, listed after the stronger matches.

Strategic placement of keywords has become common for web pages, so look past the first page of results when using a search engine such as Google.  The number of previous hits also affects ranking.

Combining Searches

Some databases have a feature called “Combine Searches” or “Search History.”  This can be used to build more precise searches by combining results of previous searches.  It can also be useful to try various combinations—for example, running a search on “tribal government” and then combining it with the name of a tribe, then a different tribe, and so on. 

Scanning and Selecting

There are several levels of consideration.  A hypothetical example:

• Scanning brief records of all the results (maybe 100 hits)
• Looking at a smaller number in more detail by opening the record; reading the abstracts
  (maybe 40 of the 100)
• Getting the full text of the ones that seem most useful (maybe 5 of the 40) 
  See “Finding Journal Articles” for more about locating full text when it is not directly linked.
• Choosing the ones that will actually be used for your project (maybe 2 of the 5)

Saving a List

In many databases, there is an option to “Mark” or “Save” items to create a list or folder of possible selections.  In some search formats, the list must be downloaded or printed before going to another group or another search; others allow adding to the list while trying various searches in the same database.  It can be very efficient to quickly mark “maybe” items, and then consider choices from the marked list.

Searching Takes Practice

It is normal to run several (and sometimes many) searches to find the information you want on a subject.  A search will “evolve” as you try out different search terms, learn more ways to use a particular database, and refine what you are looking for.

Try different databases.  The most useful search terms may be different from one database to another.

Sometimes preliminary searches can be useful in selecting a topic by getting an idea how much information there will be to work with.

Some things are simply not available in online formats.

Even if you find some useful material right away, do not assume you have found the best sources without further checking.  Use more than one database to find sources.

See the other How-to Guides for additional information, especially Information Resources Summary, Finding Books and E-Books, and Finding Journal Articles.

Reference Librarians can help you with a search.  It helps to make an appointment.

The Right Database 

You wouldn’t look in the dictionary for phone numbers, and you wouldn’t look in Biological Abstracts for articles about psychology.  For many topics, you will want to search multiple databases.  Use “More Information” next to each title in the List of Databases list for descriptions. 

Some of the categories of databases: 

• General journal indexes—Expanded Academic or Academic Search or other multi-subject indexes.
• Specialized journal indexes—Use the “By Subject” option on the databases list to see possibilities.
• News sources—Newspaper index and full-text collections or Lexis-Nexis News.
• Statistical sources—Lexis-Nexis Statistical, government sources, or databases on the Internet.
• Primary document sources—Collections of writings or images, usually with a special focus.  Some are in the databases list and others are found on the Internet.
• Reference sources--Credo Reference (formerly XReferplus) or other online encyclopedias, dictionaries, and almanacs.  Some are linked in Quick Reference.  Other reference sources in the Databases list have a particular subject, such as Columbia Grangers World of Poetry or Mental Measurements Index.
• Directories—Sources listing associations, colleges, and company information are linked to the databases list.  Other online directories are available from Internet sources.
• E-Books—Full text books available online in NetLibrary or ACLS Humanities E-Book (formerly History E-Book Project) in the databases; or Project Gutenberg or other online sources. (See additional sources in Finding Books and E-Books.)
• Government Sources—See Government Publications for many links to searchable Federal, State, and International sources.
• Internet—web sites searched by way of search engines (Google, Yahoo, Vivisimo, or others) or by linking from one site to another.

For many projects involving information research, you will conduct searches in more than one online database.  And don't forget the print sources!  See Information Resources Summary for more ideas.

• Additional information:  

  TILT - Texas Information Literacy Tutorial  An opportunity to learn more about using online searching plus selecting appropriate sources and evaluating and citing information.  From the University of Texas System.

  Choose the Best Search for Your Information Needs  and Choosing Invisible Web Databases   Internet search guides from Noodletools, a teaching resources web site.

  Internet Tutorials  Explanations of concepts related to using the Internet.  See especially How to Choose a Search Engine or Directory and Searching the Internet: Recommended Sites and Search Techniques

Close this Window

The above information is also available on handouts in the library.

---

Last updated on November 29, 2007.

---

The Jim Dan Hill Library is a Government Depository Library Belknap & Catlin, Box 2000, Superior, WI 54880 (715) 394-8343