#Finding scientific information: Designing a search string and Boolean operators

Before actually performing searches, you have to turn a search problem into a concrete search string. The topic must be divided into parts, and the concepts are described with alternative terms, parallel concepts and synonyms. The search terms and their truncation symbols, the classification codes and the Boolean operators connecting them form a search string. Before performing a search you should, for example, think about the following points:

• central concepts in the problem
• synonyms and foreign language equivalents for the concept
• parallel, broader and narrower equivalents
• classification codes
• various limiters (e.g. language, year)
• how to combine search terms e.g. which terms must include in a reference, which terms are optional expressions, the proximity of search terms

A search for the topic "tsunami warning systems" is started by dividing the topic into its two central concepts. After that, broader, narrower and parallel terms, and different spelling forms are thought about. The more specific the term is, the smaller the font for writing it; broader terms are written in upper case. Here the font size describes the search term's position in the concept hierarchy, and not its importance or frequency.

Appropriate search terms are selected from word groups and they are built into a search string. The words are connected with suitable operators, truncation symbols and substitution symbols, in keeping with the conventions of the database being used. Also the way of expressing phrases should be checked from the instructions given by the database. In this picture, words chosen from the tag cloud have been built into a search string:
(tsunami* OR "seismic sea wave*") AND (tsunameter* OR "warning system*" OR "alarm system*" OR alarmsystem*)

Boolean operators OR, AND and NOT are commonly used for combining search terms and search strings. Operational principals are:

A or B looks for the references which contain either search term A or B or both of them. The operator OR is used, for example, when connecting synonyms and combining broader and narrower terms.

A and B looks for the references which contain both search terms A and B. The AND operator is used when, for example, you want to combine two subject areas.

A not B looks for the references which contain A, but it leaves out the references containing B. It also omits the references with both A and B. With the NOT operator you can exclude a term in the search results. You should, however, be very careful when using NOT because you might deprive yourself of some good references in the process.

You can determine the execution order by using parentheses to group search strings, search terms in parentheses are always combined first. Without parentheses the execution order in a search string depends on the database being used. A database processes search terms according to default precedence rules used in that database.

• (A OR B) AND C results in references which must contain C plus either A or B, so the search is always working correctly
   
• Instead A OR B AND C  - depends on the database being used:
  • if AND operator is executed first in the database (like e.g. in the Web of Science database) the search is not working as intended but finds references which must contain only A or both B and C
  • if OR operator is executed first in the database (like e.g. in the Scopus database) the search works as intended and finds references which must contain C plus either A or B

In databases there are search forms containing several search boxes connected with Boolean operators.  The execution order between these boxes is database specific. E.g. in Advanced search in Oula database boxes connected with AND operator are always combined first, and e.g. in IEEE XPlore database the search boxes are processed from top downward.