Searching via databases

Overview

Searching for information can feel a little like drinking from the proverbial firehose. While relying on search algorithms to interpret your quick and dirty search will usually do the trick for lower-stakes day-to-day searches, when it comes to in-depth searching for professional or academic purposes, you need to do a little bit more work constructing your search to track down the resources you need.

Developing

A search for information starts with a need for information. Sometimes it helps to have this need phrased in the form of a question if it isn’t already. This can help you identify the concepts within your need, and it’s the concepts that drive your search term selection.

Identifying concepts

Look at this research question:

How does diet affect the symptoms of people with depression?

What are the concepts the above question touches on? Which of them are actually useful in a search?

“Diet” and “depression” are the two major concepts that jump out immediately. They are the crucial elements of this question, and someone searching on this question would definitely be interested in papers that mentioned both of these concepts.

The concept of “affect” or something being affected by something else is certainly important to how you end up answering this question but is it actually helpful for a search aimed at gathering literature related to the question? For example, if you had this question would you still be interested in papers that discussed changes in diet for people with depression, even if the paper didn’t use the word "affect" in it? And wouldn’t using “improve” also lend bias to your search in that requiring it to appear in a paper would rule out papers that found dietary change made no difference? With this in mind, “affect” is less important and possibly detracts from the quality of the search.

Finally, “symptoms” and “people” are both too general and vague to be useful as search terms. Adding “symptoms” to the search may mean you miss papers where the word “symptoms” isn’t used; also, a search for “diet” and “depression” implies some relationship between the two, so the search doesn’t need to be so specific about what that relationship is (i.e. that it relates to symptoms).

Key point:

From your research question, consider which concepts are the most important and practical to search with.

With a database search, sometimes less is more!

Generating keywords

Once we have identified the concepts that are useful for our search, the next step is to generate some search terms for each. While each concept is a useful search term on its own, we need to consider other ways the concepts could be expressed and other words that are associated with them in order to capture as much of the relevant literature as possible.

Continuing with our example, we can generate these additional search terms by thinking about them ourselves and by looking at the words used in the relevant resources we find. Often it helps to put these into a table or some other kind of search planning document in order to keep track of what you’re finding.

Here’s an example of such a table for tracking purposes; note that below the concepts appear as search terms along with a couple of additional search terms related to each:

Concepts ->	Diet concept	Depression concept
Search terms ->	Diet	Depression
Search terms ->	Nutrition	Depressed
Search terms ->	Food	Mood

Key points

Your concepts are a starting point for search terms that help you find relevant records in a database. Possible search terms include synonyms, antonyms and other related terms for each concept.

Keep track of your search terms using a search planner.

Identifying where to search

Once you have a set of search terms generated, it’s time to start putting them to work. This makes choosing which database(s) to search in is the next step.

Depending on the nature of your question you may be looking at discipline-specific databases or multi-disciplinary databases. Generally, discipline-specific databases have very good depth of coverage on a narrow field whereas multi-disciplinary are more broad as they attempt to cover many fields. It may also be important to consider if the database focuses on a particular geographical, historical or other area, and how that fits with your search in terms of what you could expect to find within the database.

Key point

You will need to understand a little bit about the database in order decide whether to use it for your search.

Examine the database’s Help or About sections, or the database information or guides provided by the library that is faciliating access to the database.

Using database search interfaces

While some databases can infer the relationship between your search terms, it is best to be as direct as possible and tell the database what you want. A fundamental method of doing this is to use the words AND, OR and NOT. These are common search operators (AKA Boolean operators) which tell the database how to treat the search terms within a search:

• Putting AND between search terms tells the database that you want to see records that feature both of the terms; in other words, if only one term appears in a record, you don’t want to see it. Using AND narrows your search by putting extra requirements on it.

• Putting OR between search terms tells the database you’d be happy to see records that feature either of the search terms; in other words, if either term appears in a record, you want to see it, even if the other term(s) don’t appear. Using OR broadens your search as it gives the database more options for records to show you.

• Putting NOT before a search term tells the database that you don’t want to see any records containing that term. Using NOT narrows your search by removing irrelevant results. However, NOT should be used very cautiously as it could remove relevant results that also feature (perhaps inadvertently or as a comparison) the search term you have used NOT with. While it’s good to know about NOT, it’s unlikely you’ll use it very often in your searches.

It can also be helpful to use parentheses to group together your search terms. This isn’t always necessary but is a useful way to control the relationship between the search terms so that the database doesn’t do something unexpected.

Continuing with our diet and depression question, this example search uses AND, OR and parentheses to tell the database what you’re looking for:

(diet OR nutrition OR food) AND (depression OR depressed OR mood)

In the above, we’re asking the database to show us records that feature any one or more of the search terms within each set of parentheses (or each concept group). That is, if a record has any one of the search terms from the diet concept group, as well as any one of the search terms from the depression group, the database will show it to us. If a database record features a term from one group but not another, it will not be shown in the search results for the above search.

Key point

When needed, using AND, OR, NOT or parentheses to control the relationship between your search terms is a useful way to tell the database exactly what you’re looking for.

Skilled - Contribution needed

• Field codes

• Subject headings

• Trunctation, wildcards, phrase searching

Adept - Contribution needed

• Proximity searching

• Search filters

• Citation searching

Further reading

Search tips and planning from Deakin University Library