Step 1: Eliminate the "filler" words in your research topic that will not help you retrieve relevant articles. What you have left will be the keywords you'll search:

Step 2: Think about whether your main concepts have any synonyms or related terms that you should search (you want to try to predict what other terms authors might have used for your topic.)

Note: If you can't think of any synonyms, that's okay -- some concepts don't have any relevant synonyms or related terms. To help you keep track of the keywords and synonyms, build a keyword table:

* An asterisk at the end of a word tells the database to find any possible ending to that word, so dens* will find dense, density, or densities.

Step 3: Combine your keywords using connectors like like "AND" and "OR" when you enter your search into a database -- the keyword table can help you understand how to connect your keywords.

Quotation marks can help you narrow down your search:

"soft drinks"

This search will exclude articles where search terms appear individually, such as "soft pillow" or "sugary drinks."

But be careful that you don't inadvertently narrow your search too far:

soda AND "bone density"

This search would miss articles where authors use the phrases "bone mineral density" or "bone densities."

In many databases, an asterisk (*) tells the database to search for different word endings so that you don't have to enter and search all variations of a word.

Instead of searching (simulate OR simulates OR simulation) just search simulat* --it's easier!

Getting too many results?

To narrow searches and get fewer results:

• Use database search refinements such as limiting by language, a span of years, document type, source, etc.
• Try searching some or all of your search terms in more restrictive fields, such as the title field.

Getting zero results (or a very small number)?

To broaden searches and get more results:

• Use more synonyms connected with OR, as in:
  • (costs OR ecomonics OR financing OR funding).
• Use broader terms. For example:
  • Underwater tunnels or railway tunnels are broader terms than a specific underwater tunnel like the Channel Tunnel.

For each new database you search, look to see how your results are being sorted.

Databases usually default to either:

• Listing the most recently published articles first, or:
• Listing the most relevant articles first (the database determines relevance by counting how many times your search term appears in each article).

*If you have too many results, try sorting by relevance to see if some good articles rise to the top

But remember: when you sort by relevance you run the risk of older articles rising to the top of your results list.

Some databases, such as Web of Knowledge and Google Scholar, provide metrics showing the number of times an article has been cited by other articles. Looking at the articles that cite an article is a way of seeing how the research has progressed over time.

In Web of Knowledge, you can sort your results by times cited- highest to lowest to see which articles have been cited the most (the most highly-cited articles might be particularly important if lots of other authors have cited them).

*Citation metrics can be confusing -- if you'd like more information, remember you can always contact a librarian.

Some databases allow you to manipulate and analyze a result set in a number of ways.

If you'd like more information, remember you can always contact a librarian.

Most databases have a link to a help page that will show you what search features are supported.

For example, in Web of Science, you can:

• Show which authors have written the most papers on a topic.
• Show what years have the most number of papers on a topic.