Generating search terms: Find out more

Before you dive into specialist literature it always helps to get some basic information first and contextualize your topic. You can use reference works for that:

• specialised encyclopedias targeting a certain topic or discipline: these are often substantial books with dozens, even up to thousands of short pieces on all aspects of a topic or on all issues in a discipline.
• handbooks: introductory works with dozens somewhat longer pieces describing the current state of knowledge, often arranged thematically or by scientific perspective.
• guides: books that are mostly more practically oriented or describing how to do research in a certain field.

Although there are thousands of these type of reference works there is no guarantee that they will be available for each and every topic. Also the delineations between the different types mentioned above may be quite vague.

Finding these reference works is not that hard. Using the advanced search in WorldCat. Search the title field with handbook OR encyclop*. On a second line of the same search again select the title field and type in a search term describing your subject (e.g. energy, transport*, horse*, renaissance, copyright, drug abuse or gravitational waves) .Try various (broader and narrower) terms that describe your topic. An increasing share of the search results will be available as e-books.

You can also try other terms describing reference works: dictionary, compendium, gazetteer, manual, textbook or truncate some of the terms, e.g. encyclop*. If you need to find non-English reference works, you need to work with non-English terms.

The most important reference works are also mentioned in the list of search systems on the library website.

The most important advice we can give you on generating search terms is try to think of the terms likely to be used in the document. This so-called 'thinking in terms of the document you are looking for'  helps you to avoid looking for information with the help of terms that you would use, but encourages you to use terms the author has probably used.

This means, rather paradoxically, that you try to imagine beforehand the content of the document you are hoping to find. Try to consider:

• The jargon the author could have used: very scholarly, journalistic, business jargon, very formal language, street language etc.
• By playing with these different styles, you may influence the nature of your results: the flu is the same as influenza, but searching on these terms generates information from different sides. W
• What phrases/concepts you expect in the document. Put these multi-term phrases/concepts between double quotes
• Please be careful because by using brackets you decimate the number of hits. It would be better to use this method once you have found out that the same phrase is often repeated ('continental drift') or when something is a fixed concept ('Utrecht University')
• How the answers to your question are formulated.
• Especially if you are looking for facts with the plain Google web search engine it is somtimes useful to enter a sentence which already includes the answer, but then without the answer of course; e.g. "Tesla is the inventor of" or "the director of Pulp Fiction is"

Thinking in terms of the document is easier once you have seen more relevant articles, thus know more about the subject and are more familiar with the jargon used.

This advice applies to classical documents such as books and articles, but also to web pages, reports, blogs and tweets.

You don't have to think of every search term all by yourself. Especially in the case of  foreign languages, your active vocabulary is smaller than your receptive vocabulary: you do know or recognize the words, but you cannot think of them yourself. So make good use of anything that suggests good search terms:

Subject related sources:

• Wikipedia usually provides many terms in an article that are related to your subject
• subject encyclopedias provide much context and accepted scientific terms
• in subject dictionaries you can check the exact meaning of a particular concept
• Search engines/online bibliographies often give suggestions and sometimes show the the key words which authors and editors have added to the articles 
• Reading through documents already found will help thinking of searhc terms. Remember them by highlighting or noting them down. If you use RefWorks as your literature management tool, you can often read again the key words/descriptors of earlier found literature.

Language thesauruses: lists of synonyms and related concepts in a particular language which you know receptively but not actively:

• Roget's Thesaurus for English (online version)
• Roget's Thesaurus for English (print version)
• Visuwords for various types of relations between words (English)
• Van Dale Thesuarus for Dutch (print version)
• Het Juiste Woord for Dutch (print version)

Translation dictionaries

• translation dictionaries are used to find the translations of the term
• Van Dale online dictionaries
• Wiktionary (provides also synonyms and antonyms)
• Google Translate (also provides synonyms etc. for some languages)

more about dictionaries in the LibGuide woordenboeken (in Dutch)

To many economic publications the subject codes of the Journal of Economic Literature (JEL) Classification System have been added. In the RePEc Ideas search engine you can browse by JEL-codes and find related publications with the help of JEL-codes. Despite the fact that no separate search field is available, you can also search by these codes, but then you will only find publications to which the codes have been added as keywords. In Econlit you can also search by JEL-codes. It is possible to indicate that the codes must occur as keywords and not as words in the text.

• MeSH (Medical Subject Headings)

For scientific terms:

• Thesauruses:
• Earth sciences: Georef Thesaurus (print book): relations between thousands of Earth Science subject terms.
• Subject dictionairies:
• Human Geography: Dictionary of human geography: definitions, concepts and basic literature references
• execute a search on your topic using Scopus and then check the list of keywords in the refine menu.

For geographic names:

• International:
  
• Geonames.org Open database containing > 8 million place names
• Getty Thesaurus of geographic names online: >1,1 million geographic names describing > 900,000 places
• Buitenlandse Aardrijkskundige namen (in Dutch): foreign toponyms in Dutch for all countries, by the Taalunie
• Vertaalhulp elementen uit buitenlandse toponiemen (in Dutch): a translation tool for parts of place names in all kinds of languages. By the Taalunie.
• UNGEGN World Geographical names via a map
• Netherlands:
• Plaatsenregister of Plaatsengids.nl

many more Geo gazetteers and databases of place names via UNGEGN

• Psychological Indexing terms, included in the search engine PsycInfo
• Sociological Indexing terms, included in the search engine Sociological Abstracts

Spelling variations are a major problem in search actions, because many search engines are not able to  suggest spelling corrrections, let alone search for different correct versions.

Problems mainly arise in the case of:

• different spelling of terms in the different languages and language variants (British English versus American English)
• diffenence in transcription  from languages with a non-Latin script
• different spellings over the years, for instance due to official spelling reforms

Ideally the preferred spelling is used within a language region which is often recorded in spelling guides or authoratative dictionaries. However, they do not contain proper names of persons, organisations or places.

The many shapes of Qaddafi...

NB: Results from 20120328. Due to the otherwise unreliable numbers the search term was "Ghadaffi" in Google Books. In Google Scholar the search was limited to articles having the name in the title.

The example of the former Libian colonel clearly shows how search results may differ due to spelling variants. Moreover, in this example it is hardly possible to use the * to truncate on the root of the word, apart from the fact that Google and Google Scholar do not support this search method.

For that matter, really reliable databases have their own conventions relating to preferred terms and preferred spelling of the keywords added to an article. So with the help of the preferred term, articles can be found in which that term does not necessarily occur.

There are plenty of buzz words going round to indicate all kinds of (search) terms. They do not always mean the same. Usually they are used in the following sense:

• author keywords = keywords the author himself adds when submitting his article
• controlled vocabulary = a fixed keywords list or thesaurus used by a publisher or journal or in an entire discipline
• index terms/subject headings/descriptors = preferred keywords usually used by makers of databases to describe a text
• keyword = a term relating to the subject used to characterise a text; does not necessarily have to occur in the text; sometimes interpreted as keywords given by the editors or authors, sometimes broadly interpreted as 'search term'
• query = the search term(s) and extra syntax (operators, wildcards, parentheses) you give as input to a search engine
• search term = a word that you use to search for something in a database
• thesaurus = a coherent set of words, usually in a particular discipline, including broader, narrower, related and preferred terms

With some search methods you do not have to generate search terms; with the snowball method you just follow links starting from a publication you have at hand. That way you get to cited publications, citing publications, more of the same author or more in the same journal. Some databases contain clickable keywords in article descriptions taking you to related information. Others have direct links 'related' or 'similar'.

For books another option is browsing shelves in the library: not very targeted but it may give suprising results. Note that books are increasingly acquired in e-book format, so this option will become less valuable, especially in science, life sicences and social sciences.

You search strategy defines what you search, where you search and how you perform your search. In the course of your search process you take many decisions that affect the quality of search results and the time needed to get those results.

The main decisions in your search strategy relate to:

1. What: think through in advance what information you really need: subject, type of information (analysis, news, statistics, opinion, overview etc.), level and recency. The exact terms you are going to use in your search are of crucial importance.
2. Where: What you are looking for determines where you should go to find it: unfortunately there is not one search engine or database that has it all.
3. How: There are various methods of searching. The systematic/bibliographic method (using search terms in scholarly databases) and the snow ball method (finding new information related to what you already have) are the most important. The exact application of these methodes depends on the options available in the database or search engine.

The special LibGuide search strategy (in Dutch) has more on building succesful search strategies.

It is a good thing to think of as many possible alternatives as possible. Which ones you use in the end depends on:

1. The content of the database/search engine. Keep your search terms in line with the language and text level of the documents catalogued in the database. Many scientific databases only contain documents in the English language, so entering Dutch search terms is not very useful. In the case of search engines covering several languages you could:
1. perform a seperate searches in different languages
2. Combine several languages in an OR-search, for instance innovatie OR innovation
3. (almost never) by limiting on the root/stem of the word and so find different languages in one go:  innovat*
2. The nature of the information you are looking for. There are always small differences in meaning of terms, and especially in the kind of publications in which these terms are used. When you are doing research on soft drugs, you could enter: cannabis, marijuana, marihuana, hash, pot, stuff, weed, dope, shit and search in different languages, cities and subcultures and many more words for cannabis. If you know enough of your subject, you could influence the kind of information you get by the exact choice of your terms: scientific, popular or coming from a specific culture.

It is wise to check beforehand how many and what kind of results a certain search term yields before you combine it with other terms. But please note that the numbers of results in Google are totally unreliable, even relatively speaking.

Many search engines save your searches during the browser session. You can use this information to:

• Keep track of your searches, so you won't perform searches twice
• Retrieve an earlier search, for instance when after trying alternatives you decide that the earlier search was the best after all
• Quickly compare the results of several searches
• Work systematically with the search history option. First search on single terms and then look at your search history. It shows numbers of hits. Use that information to select terms you want to combine in new searches.

Some search systems offer the option to combine search terms directly from the browser history. For instance by ticking previous searches and combine them with AND or OR.

If you want to keep your search history also after closing your browser session you should create an account in the search system. Google offers this option as well as most catalogues and databases provided by library.