Boolean Searching

Latest revision as of 14:33, 23 August 2010

[edit] LexisNexis Academic: Advanced Boolean Search Strategies

Note: it is best to use these advanced strategies with the Power Search form. Many of the other search forms include an "Edit in Power Search" link in the Tips that will convert a search you have built on another form to a Power Search.

Boolean search logic give you the tools to construct sophisticated queries that have high recall (return almost all relevant documents in the database) and high precision (do not return many irrelevant documents). When combined with segment search techniques described in the Academic Document Sections these tools become even more powerful.

[edit] Universal or Wildcard Characters

Often there are variations of your search terms that you would like to retrieve. For example, searching the word CHILD will not find the word CHILDREN, although this term might be relevant to your search.

To make your search more effective, you can search for word variations using the asterisk (*) as a wildcard symbol. An asterisk (*) replaces one letter, can be used more than once in a word, and can be used anywhere EXCEPT as the first letter of a word. For example, searching on the term wom*n will locate records containing both woman and women.

Use an exclamation mark (!) as a truncation to replace more than one letter at the end of a search term. You can only use this symbol once in any word. For example, searching on the term immigra! will locate records containing the terms immigrant and immigration.

Note: LexisNexis Academic automatically includes plural forms of most words when performing a search. It is not necessary to use wildcard characters to search for variations such as "boat" and "boats"; for more on this see the section on Plurals below.

[edit] Using the AND Connector

Because AND can connect words that are far apart from one another or in different segments, searches using AND usually find more documents than searches using the W/n connector. As a general rule, use AND when it doesn't matter where your search words appear in a document. Use the W/n connector when there is a connection between your search terms and you need to find the terms near each other.

If your search terms are fairly unique, the AND connector can find documents that are related to your research. Using the AND connector can also help you get started on your research, until you begin to find more specific concepts and terms for your search. For example, if you want information about how land can be preserved in ohio using a land trust, you could use this search:

land trust AND Ohio

However, to find documents that are relevant when your search terms are less specific, you may need to use the W/n connector. For example, the following search will find more relevant documents than if the AND connector were used:

business loss w/10 tax deduction

If you're looking for a document in which the same word occurs twice, such as a court case with Marvin v. Marvin as respondents, do not use the AND connector. The following search would find many unrelated documents:

marvin AND marvin

Instead, use the W/n or W/seg connector, such as:

marvin W/5 marvin
marvin W/SEG marvin

[edit] Finding phrases containing the word "and"

If you're searching for a phrase that contains the word "and," omit the word "and" from your search request and use W/1 as a connector. Otherwise, "and" is interpreted as a connector instead of a literal word. For example, if you're looking for "profit and loss" statements, your search request should be

profit W/1 loss

Using the W/1 connector ensures that the words "profit" and "loss" appear near each other in the document.

[edit] Using AND when searching abstracts

Because abstracts are usually short, the AND connector is a good choice when searching for multiple terms. The following search in an abstract file finds a large number of documents, but the chances of document relevancy are great:

satellite AND launch!

Using the OR Connector

Use the OR connector to find documents that contain either or both of the words or phrases linked by OR. Use the OR connector to link search words that are synonyms, antonyms, alternative spellings, or abbreviations.

Examples:

Synonyms

   lawyer OR counsel OR attorney

Antonyms

   regulated OR deregulated

Alternatives

   takeover or take over

Abbreviations

   international business machines

  OR i.b.m. OR ibm

The search looks for the words or phrases linked by the OR connector, not the word "or" itself. The words or phrases linked by OR can be in any part of a document.

[edit] Using the W/n Connector

Use the W/n connector to find documents with search words that appear within "n" words of each other. The value of "n" can be any number up to 255. Use W/n to join words and phrases that express parts of a single idea or to join closely-associated ideas.

Words or phrases linked by W/n must be in the same segment (a specific part of a document). Either word may appear first.

Note: W/n connectors cannot be used in combination with W/s or W/p connectors.

For example, the following search request tells the research software to find documents in which both words appear in the same segment, within three or fewer words of one another.

william w/3 hearst

It retrieves documents containing the words William Randolph Hearst; William R. Hearst; and Hearst, William R.

Specifying the value of "n" There is no magic formula for choosing the value of n, but these guidelines may prove useful:

Choose from these options:

W/3 - W/5: words will appear in approximately the same phrase

W/15: words will appear in approximately the same sentence

W/50: words will appear in approximately the same paragraph

CAUTION: Although you may choose a number for n as large as 255, you may wish to choose a number less than 100. Choosing a number greater than 100 is likely to retrieve documents in which your search words are used in unrelated contexts.

Using Multiple W/n Connectors
If W/n connectors have the same number, they operate from left to right. If they have different numbers, the smaller number operates first. The following search finds "airport" within five words of "noise" and then some form of the word "abate" within five words of either "airport" or "noise":

airport W/5 noise W/5 abat!

The following search first looks for documents in which forms of the word "assign" appear within five words of "collateral," then finds occurrences of "lease" within 8 words of the forms of "assign":

lease W/8 assign! W/5 collateral

The search above finds documents that contain all three search terms, and the search terms would be within 13 searchable words of one another.

[edit] Using the AND NOT Connector

Use the AND NOT connector to find documents in which a search word or phrase is to be excluded.

For example, the following search finds documents where the word "trust" occurs but the word "charitable" does not.

trust AND NOT charitable

Because the exclusion covers the entire document, a document would be excluded if the word "charitable" appears anywhere in the document. Therefore, even if "charitable" is used as a term of distinction in a document, the document would not be included in the search results. For example, a document that includes the phrase "this is not a charitable trust" would not returned, even though that is the type of trust you want information about.

[edit] Using AND NOT with segment searches

When documents have information in some consistent part or segment, you may use AND NOT with less risk. For example, if you want to find court cases that mention asbestos, but do not have the Manville Corporation as a plaintiff or defendant, you may place the AND NOT connector at the end of your search, as in the following example:

AND NOT name (manville)

This restricts the operation of AND NOT to the NAME segment. If you do not use a segment search, but instead end your search with

AND NOT manville

you eliminate undesired cases, but you also eliminate any cases that mention the word "manville," even as a reference, and any case that used the word "manville" unrelated to the Manville Corporation.

Putting AND NOT last If you include AND NOT in your search, it should be the last connector you use. Otherwise it may produce undesired results.

If you want to exclude court cases with the Manville Corporation, as either a plaintiff or defendant, the last part of your search should look like this:

AND NOT name (manville)

If you put another search word after that part of your search, such as:

AND NOT name (manville) AND bhopal

you would not eliminate all documents with "manville" in the NAME segment. In fact, you might not eliminate any. By linking "bhopal" to NAME (manville) with the AND connector, only documents that have both "manville" in the NAME segment and the word "bhopal" somewhere in the text would be eliminated.

[edit] Using the PRE/n Connector

Use the PRE/n connector to find documents in which the first search word precedes the second by not more than the stated number of words. As with W/n, both words must be in the same segment.

For example, t he following search finds documents in which "pay" precedes "television" by three or fewer words:

pay PRE/3 television

If you use the following search request instead, you find documents that refer to pay television, as well as documents that discussed what television networks pay their employees and how often television news teams pay attention to local events.

pay W/3 television

PRE/n is primarily useful in situations where a different word order significantly alters meaning. For example, "summary judgment" is significantly different from "judgment summary."

If you are searching for individuals whose names might produce undesired results if used in reverse order, such as McGeorge Bundy, consider using the PRE/n connector.

[edit] Using the W/p (Within Paragraph) Connector

Use the W/p connector to find documents with search words that appear within the same paragraph. You may also use W/p when you want your search words to have a general relationship to each other.

For example, the following example finds "rule" within the same paragraph as "sanction":

rule 11 W/p sanction

The following example finds "take over" or "takeover" within the same paragraph as "poison pill:"

take over OR takeover W/p poison pill

Note: W/p connectors cannot be used in combination with W/n connectors.

[edit] Using the W/seg Connector

The W/seg connector tells the research software to find documents in which both of your search words appear within the same segment. Words joined with W/seg can occur together in any segment.

For example, the search below finds documents in which "opec" and "gasoline" are in the same segment:

opec W/SEG gasoline

W/seg is slightly more restrictive than the AND connector. In the example above, a document in which "opec" is in the headline and "gasoline" is in the text would not be found by this search, because the words are not in the same segment.

W/seg is primarily useful for searching files of highly segmented information, such as annual reports. If you want to find annual report footnotes that mention unitary taxation, your search might look like this:

tax! W/SEG unitary

[edit] Using the W/s (Within Sentence) Connector

Use the W/s connector to find documents with search words that appear within the same sentence. You may also use W/s when you want a close relationship between words without specifying an exact proximity.

For example, the following search finds "sanction" within the same sentence as "frivolous":

sanction W/s frivolous

The following example finds "circumstances" within the same sentence as "mitigat!":

circumstances W/s mitigat!

Note: W/s connectors cannot be used in combination with W/n connectors.

[edit] Using the NOT W/n Connector

The NOT W/n connector tells the research software to find documents in which the first search word is found. The second word need not be in the document, but if it is, it cannot be within 'n' words of the first word.

For example, the following search finds documents in which the word "rico" occurs but the word "puerto" is not within two words of it:

rico NOT W/2 puerto

This finds documents that have "rico" (possibly standing for Racketeer Influenced and Corrupt Organizations Act) further than two words away from "puerto." This avoids retrieving documents solely on the basis of their mention of Puerto Rico.

[edit] Using the NOT W/seg Connector

The NOT W/seg connector tells the research software to find documents that have at least one segment in which the first search word appears, but not the other search word.

For example, the following search finds documents in which there is at least one segment with the word "tank" but not the word "m1":

tank NOT W/SEG m1

This does not mean the same document could not have both words in some other segment.

NOT W/seg is useful in searching highly structured and segmented files, such as annual reports or the CODE file in the INSRLW (insurance law) library.

Index terms are located in the INDEX segment of documents in INSRLW files. They are surrounded by spaces and # symbols. They are added to the documents that deal with those concepts, whether or not the words are actually used. Index terms are not "fenced off" with # symbols in any library except INSRLW.

For example, if you want to find a document in the code file of the INSRLW library that had been indexed with the index term # investment limitations # but not with the index term # securities #. Your search might look like this:

1. investment limitations # NOT W/SEG # securities #

[edit] Using the NOT W/sent Connector

The NOT w/sent connector tells the research service to find documents in which the terms you specify appear, but not within the same sentence.

Example: market NOT W/SENT share

finds documents in which the word market appears at least once, but not within the same sentence as the word share, which also appears at least once.

[edit] Using the NOT W/para Connector

The NOT w/para connector tells the research service to find documents in which the terms you specify appear, but not within the same paragraph.

Example: cable NOT W/PARA television

finds documents in which the word cable appears at least once, but not within the same paragraph as the word television, which also appears at least once.

[edit] Using the ALLCAPS Command

Using the ALLCAPS command, you may restrict your Boolean search to find words in which all letters are capitalized.

Example: allcaps (era)

finds documents referencing the Equal Rights Amendment (ERA). The documents this search request finds may also contain occurrences of the time period (era) or laundry detergent (Era), but only incidentally.

[edit] Using the ATLEAST Command

Use ATLEAST to require that a word or words appear ‘at least’ so many times in a document. Use ATLEAST when you want only documents that contain an in-depth discussion on a topic rather than just a mention.

For example, to find documents that contain an in-depth discussion of CERCLA, the Comprehensive Environmental Response, Compensation, and Liability Act, enter:

atleast10(cercla)

This search requires CERCLA to appear in every document at least 10 times. You can use any number from 1 to 255 with the ATLEAST command.

[edit] Using the CAPS Command

Using the CAPS command, you may restrict your Boolean search to find words in which capital letters appear anywhere in the word.

Example: caps (supervga)

This search finds documents containing superVga, SuperVga, Supervga, SUPERVGA, and so on.

[edit] Using the NOCAPS Command

Using the NOCAPS command, you may restrict your Boolean search to find words in which none of the letters are capitalized.

For example, to search on documents about foreign aid, but not the disease AIDS, use

nocaps (aid)

The documents this search request finds may also include references to AIDS, but only incidentally.

[edit] Using the PLURAL Command

The research software automatically finds singular, regular plural, and possessive endings for search words. However, there may be times when you want only the plural form of a word. For example, to search for documents discussing Steven Jobs, the founder of Apple Computer, enter:

steven W/3 plural (job)

You can further refine your search request by combining the CAPS and PLURAL commands:

steven W/3 caps (plural (jobs))

[edit] Using the SINGULAR Command

The research software automatically finds singular, regular plural, and possessive endings for search words. However, there may be times when you want only the singular form of a word. For example, to search for documents discussing job discrimination, enter:

singular (job) W/5 discrimination