Skip to main content
It looks like you're using Internet Explorer 11 or older. This website works best with modern browsers such as the latest versions of Chrome, Firefox, Safari, and Edge. If you continue with this browser, you may see unexpected results.

Best Practice for Literature Searching

For the sciences of food and health

Creating a search strategy

Once you have determined what your research question is and where you think you should search, you need to translate your question into a useable search. Doing so will:

  • Make it much more likely that you will find the relevant research and minimise false hits (irrelevant results)
  • Save you time in the long run
  • Help you to stay objective throughout your searching and stick to your plan
  • Help you replicate and update your results (where needed)
  • Help future researchers build on your research.

If you need to explore a topic first, your search strategy can initially be quite loose. You can then revisit search terms and update your search strategy accordingly. Record your search strategy as you develop it and capture the final version for each place that you search.

Remember that information retrieval in the area of food is complex because of the broadness of the field and the way in which content is indexed.  As a result, there is often a high level of ‘noise’ when searching food topics in a database not designed for food content. Creating successful search strategies involves knowledge of a database, its scope, indexing and structure.


Shaping a search strategy
A good search strategy will include:

  • Key concepts and meaningful terms
  • Keywords or subject headings
  • Alternative keywords
  • Care in linking concepts correctly
  • Regular evaluation of search results, to ensure that your search is focused
  • A detailed record of your final strategy. You will need to re-run your search at the end of the review process to catch any new literature published since you began.

Search matrix


 


Using a search matrix helps you brainstorm and collect words to include in your search. To populate a search matrix:

  • Identify the main concepts in your search
  • Use two parallel strategies to populate the other boxes:
    • Run initial searches with your terms, scanning abstract and subject terms (sometimes called descriptors, keywords, MeSH headings, or thesaurus terms, depending on which database you are using) of relevant results for words to add to the matrix.
    •  Explore a database thesaurus hierarchy for suitable broader and narrower terms.

Note: You don’t need to fill all of the boxes in a search matrix.

You will find that you need to do some searches as you experiment in running it and this will help you refine your search strategy. For the search on this example question:

  • Some of the broader terms turned out to be too broad, introducing a host of irrelevant results about pork and chicken
  • Some of the narrower terms were unnecessary, as any result containing “beef extract” is captured by just using the term beef.

See the revised matrix (after running searches) tab!


This revised matrix shows both adjustments made to terms, and how the terms are connected with Boolean operators.  Different forms of the same concept (the columns) are connected with OR, and each of the different concepts are connected with AND.   

Search tools

 

 

Boolean operators tell a database or search engine how the terms you type are related to each other.  

 

 

 

Use OR to connect variations representing the same concept. In many search interfaces you will want to put your OR components inside parentheses like this: (safe OR “food safety” OR decontamination OR contamination OR disinfect*). These are now lumped together into a single food safety concept for your search.

Use AND to link different concepts. By typing (safe OR “food safety” OR decontamination OR contamination OR disinfect*) AND (beef OR “cattle carcasses”)—you are directing the database to display results containing both concepts.

NOT eliminates all results containing a specific word.  Use NOT with caution. The term excluded might be used in a way you have not anticipated, and you will not know because you will not see the missing results.


Learn more about using Boolean operators: Research Basics: Using Boolean Operators to Build a Search (ifis.org)


The search in the matrix above would look like this in a database:

("food safety" OR safety OR decontamination OR contamination OR disinfection) AND (thaw* OR defrost* OR "thawing medium") AND ("sensory quality attributes" OR "sensory perception" OR quality OR aroma OR appearance OR "eating quality" OR juiciness OR mouthfeel OR texture OR "mechanical properties" OR "sensory analysis" OR "rheological properties") AND (beef OR "cattle carcasses")

Thesaurus terms will help you capture variations in words and spellings that researchers might use to refer to the same concept, but you can and should also use other mechanisms utilised by databases to do the same. This is especially important for searches in databases where the thesaurus is not specialised for food science.

  • Phrase searching, putting two or more words inside quotation marks like “food safety” will ensure that those words appear in a single field (i.e. title or abstract or subject heading) together as the phrase. Phrase searching can eliminate false hits where the words used separately do not represent the needed concept.
  • Some databases allow you to use proximity searching to specify that words need to be near each other. For instance, if you type ripening N5 cheese you will get results with a maximum of five words between ripening and cheese.  You would get results containing cheese ripening as well as results containing ripening of semi-hard goat cheese.

Learn how to test if a phrase search or a proximity search is the better choice for your search: Proximity searching, phrase searching, and Boolean AND: 3 techniques to focus your literature search (ifis.org)


Note: Proximity symbols vary from database to database. Some use N plus a number, while others use NEAR, ADJ or W. Always check the database help section to be sure that you are using the right symbols for that database.

Truncating a word mean typing the start of a word, followed by a symbol, usually an asterisk (*).  This symbol tells the database to return the letters you have typed followed either by no letters (if appropriate) or letters.  It is an easy way to capture a concept that might be expressed with a variety of endings. 

Sometimes you need to adjust where you truncate to avoid irrelevant results.  See the difference between results for nutri* or nutrit*

 

nutri* nutrit*

nutria     

nutritious     

nutrition       

nutritional       

nutritionally  

nutricereals    

nutrikinetics

nutritionists  

nutrative  

nutrient 

nutrients

nutrigenetics

nutrigenomic

nutrigenomics

and more. 

nutrition 

nutritious           

nutritional     

nutritionally  

nutritionists  

nutritive  

 

Inserting wildcard symbols into words covers spelling variations.  In some databases, typing organi?ation would return results with organisation or organization, and flavo#r would bring back results with flavor or flavour.  


Note: While the truncation symbol is often *, it can also be $ or !.   Wildcard symbols also vary from database to database. $ or ? are sometimes used. Always check the database help section to be sure that you are using the right symbols for that database.


In building a search you can combine all the tools available to you.   “Brewer* yeast” , which uses both phrase searching and truncation, will bring back results for brewer yeastbrewer’s yeast and brewers yeast, three variations which are all used in the literature.

Best Practice!

BEST PRACTICE RECOMMENDATION: Always check a database's help section to be sure that you are using the correct  proximity, truncation or wildcard symbols for that database. 

Handsearching

It is good practice to supplement your database searches with handsearching. This is the process of manually looking through the table of contents of journals and conferences to find studies that your database searches missed. A related activity is looking through the reference lists of relevant articles found through database searches. There are three reasons why doing both these things is a good idea:

  • If, through handsearching, you identify additional articles which are in the database you used but weren’t included in the results from your searches, you can look at the article records to consider if you need to adjust your search strategy. You may have omitted a useful variation of a concept from your search string.
  • Even when your search string is excellent, some abstracts and records don’t contain terms that allow them to be easily identified in a search, but are relevant to your research.
  • References might point to research published before the indexing began for the databases you are using.

For handsearching, target journals or conference proceedings that are clearly in the area of your topic and look through tables of contents. Sometimes valuable information within supplements or letters is not indexed within databases.

Academic libraries might subscribe to tools which can speed the process such as Zetoc (which includes conference and journal contents) or Browzine (which only covers journals).  You can also see past and current issues’ tables of contents on a journal’s webpage.

Handsearching is a valuable but labour-intensive activity, so think carefully about where to invest your time.

Best practice!

BEST PRACTICE RECOMMENDATION:  Ask a colleague, lecturer, or librarian to review your search strategy. This can be very helpful, especially if you are new to a topic. It adds credibility to your literature search and will help ensure that you are running the best search possible.


BEST PRACTICE RECOMMENDATION:  Remember to save a detailed record of your searches so that you can run them shortly before you are ready to submit your project to see if any new relevant research has been published since you embarked on your project. A good way to do this is to document:

  • Where the search was run
  • The exact search
  • The date it was run
  • The number of results

Keeping all this information will make it easy to see if your search picks up new results when you run it again.


BEST PRACTICE RECOMMENDATION: If you are publishing your research, take note of journals appearing frequently in your search results for an indication of where to publish a research topic for good impact.