3 Steps to Mastering Literature Searches

Google and Yahoo make it easy to search the web, whatever you’re looking for you’ll probably find something out there! Thorough searching of bibliographic databases however requires a little more effort, although this is repaid by the quality of the material retrieved.  In this post we will be looking at how to search a bibliographic database. All bibliographic databases are different, but many of the same general principles apply. Here, we will look at PubMed which is a freely available and covers journals in life sciences and medicine.

.

How to Search

1. Get your search question clear

Is sounds obvious but this can be the trickiest part of setting up a search, basically the question is ‘what am I actually looking for?’. You need to be clear about what search terms you are going to use to ask your question. Take the topic below as an example –

In what ways does the regulation of bacterial chemotaxis differ between Gram negative and Gram positive organisms?

Just trying to type this as a search string into a bibliographic database won’t get you very far. One approach is to try splitting the question down into concepts and then associating relevant keywords with each concept. If you are unfamiliar with a topic this might require a little background research to identify suitable keywords and terms. For example, a flick through a microbiology textbook would help you find out that chemotactic regulation is based on a two-component regulatory system incorporating pairs of Histidine protein kinases and response regulator elements. One way you might build up you search terms and concepts is like this –

.

Notice how we now have some specific terms to work with. Rather than simply looking for Gram negative/positive organisms in general we’ve identified some specific examples of model organisms in which chemotaxis has been studied.

.

2. Putting your terms together

Start by considering possible term variations. Generally, bibliographic databases will only locate exactly what you tell them to find. Here are some sources of term variation you should consider –

• Plurals – searching for ‘bacteria’ (plural) won’t find ‘bacterium’ (singular).
• UK/US spellings – searching for ‘analyze’ won’t find ‘analyse’.

• Numbers – 2 or two or II
• Synonyms – The same hormone may be known as adrenaline or epinephrine.
• Abbreviations – PCR or Polymerase Chain Reaction?

Some databases have built in thesauri which help them find some term variations automatically. For example, searching for ‘Woman’ on PubMed will search for ‘Women’ as well.

Most databases allow you to specify wildcard characters. Wildcards are special symbols which you can include in a search term to substitute for a letter or group of letters. In Pubmed the asterisk ‘*’ will truncate words –

Bacteri*

Finds: Bacteria, bacterial, bacterium, bactericidal etc.

In some databases like Scopus and Web of Science you can allow substitution of single characters to help deal with differences in spelling –

Analy?e

Finds: Analyze and Analyse

Unfortunately this feature is not available in PubMed at the moment.

Once you’ve thought about term variations, you need to glue your search terms together using Boolean logic.

Imagine we’re doing some research into vaccine development for HIV. If you only want to find articles which contain reference to HIV along with “vaccine development”, then you need to connect these terms using the logical AND operator. In PubMed you’d use the following syntax to enter this search –

HIV AND “vaccine development”

This find 1016 records – a lot of results! Let’s add another term – ‘recombinant’ (say we’re interested in the use of recombinant virus and viral proteins as vaccine candidates).

(HIV AND “vaccine development”) AND recombinant*

This narrows the field down to 238 results. Notice how we used brackets tell PubMed to search for HIV and “vaccine development” occurring together and then exclude anything that didn’t mention recombinants of some kind.

We also have an abbreviation to deal with – HIV. To allow for the full name, we use another Boolean operator OR. The OR operator expands the search by allowing HIV, Human Immunodeficiency Virus or both terms to be found.

((HIV OR (“Human Immunodeficiency Virus”)) AND “vaccine development”) AND recombinant*

We’ve now added some nested brackets and the OR operator to the search. This search brings back 242 results, adding four papers we might have missed.

The final common Boolean logic operator is the NOT operator. This operator functions to exclude all records which contain a specified word or phrase. If you wanted to find everything in PubMed about vaccine development, but which wasn’t about HIV, you could try this –

“vaccine development” NOT (HIV OR “Human Immunodeficiency Virus”)

In summary,

• AND restricts
• OR expands
• NOT excludes

.

3. Set your limits

You can refine your results down further by applying limits to your search. PubMed lets you limit your search by many criteria including date of publication, document type and language. If you were only interested in reviewing recent clinical trials for recombinant vaccines for HIV that have been published in English, you could set the following limits using the drop down menus and selection boxes in PubMed.

• Search: ((HIV OR (“Human Immunodeficiency Virus”)) AND “vaccine development”) AND recombinant*

• Published in the last: 5 years.
• Type of Article: Clinical Trial.
• Languages: English

By applying these limits the results narrow down to just 4 reports.

.

Summary

By using a combination of advanced searching techniques and result limits, it’s possible to pinpoint very precise information in a bibliographic database. In the next and final post of this series we will look at how to manage your results efficiently.

.

I’m currently the Assistant Subject Librarian for Biochemistry at the Radcliffe Science Library, Oxford University. My job is to provide library support services to staff and students in the Biochemistry Department at Oxford University. Most of the information presented in these blog posts is based on materials I have used to help train students at the University in literature searching skills and upon my own personal experience as a researcher in microbial genetics. To find out more about me and my activities, please visit my blog and follow me on Twitter.

.

.

In case you missed it, check out Part I of the series:

Choose the Right Literature Research Tools (or Else…)

.

.