Skip to Main Content

Research Workshop

Lecture: Searching by Keyword

What is a keyword search?

A keyword search looks for results by matching the words you put into a search engine. This is different from a subject heading search: in a keyword search, you are using your own search terms, not predetermined, standardized terms. When you type keywords into a search bar, you are asking the search engine to find that particular string of characters anywhere in its records, regardless of where it may show up. This approach has pros and cons.

Pros: Keyword searches, because they are not standardized, are more flexible. You can combine terms, change terms, change their order, and use a lot of tricks to get the search engine to show you the results you want.

Cons: Keyword searches are "flat," meaning they have no context. The search engine cannot tell, from a simple keyword like "Iraq," whether you are looking for resources on the history of Iraq, the people of Iraq, the wars in Iraq, the geography of Iraq, the religions of Iraq, etc.; all it knows is that you want anything that has the string of letters "Iraq" in it.

When should you use a keyword search?

There are several instances in which it is appropriate to use a keyword. They include:

(1) When you don't know or can't find a subject heading.
(2) When the search engine you are using doesn't allow you to search by subject (Google, for instance, only allows keyword searches).
(3) When you are trying to relate/integrate multiple subjects in your paper (if you are comparing Chaucer and Spenser, for example, a keyword search including the search terms "Chaucer" and "Spenser" can be more efficient than doing multiple subject heading searches).

How do you use a keyword search?

Here are some different types of keyword searches:

Suppose you are researching a topic that is made up of multiple words, like "Spice Girls." (Because someone somewhere would actually research that.) If you go to JSTOR and type in the two words "spice" and "girls" by themselves, you are telling the search engine that you want it to pull up any results with either or both of those words. As you can imagine, this would grievously increase your research time by forcing you to slog through totally irrelevant results. However, if you put quotation marks around the phrase "Spice Girls," now you are telling the search engine that you only want results that have those two words in that exact order. This significantly reduces the number of results you have to wade through. (Some databases want you to use parentheses instead of quotation marks, while others will accept either symbol.)

Truncation: Type the beginning of a keyword, and then add an asterisk (*) to the end. (Some databases use a different symbol instead of an asterisk, but most will have a help page that lets you know which symbols to use.) This will allow you to search for variations of words with the same root but different endings. For example, search "interact*" instead of "interaction"; in addition to "interaction," "interact*" would search for the keywords "interact," "interactive," and "interactivity" as well.

Wildcard: Put an asterisk in the middle of a keyword to search for variations of words with the same root but different spellings. For example, searching “wom*n” would yield search results for both the singular “woman” and the plural “women.” Both truncated and wildcard searches save you time and get you more results.

George Boole was a smart guy who applied mathematical principles to enhance the research process. He did this by introducing three "Boolean operators" (as they are now known): AND, OR, and NOT. Each of these operators has different effects on a search.

AND searches only find information that is common to two distinct search terms. If you think of your search terms as different Venn diagram circles, AND limits your results to the area where those two circles overlap. Searching "Islam AND Iraq," for instance, will eliminate all results that only have the word "Islam" or only have the word "Iraq," leaving you only with results that include both these words. This is a great way to narrow down your search results, which can be helpful if your topic has many different elements.

The trick, however, is to refrain from overloading on AND terms. For example, if you are looking for material on Osama Bin Laden’s influence during the Iraq War of 2003, some keywords you might come up with include "Osama Bin Laden," "Iraq War," "2003," "influence," and so on. But how likely is it that you will find all those terms in one record, like an article entitled “Osama Bin Laden’s Influence on the Iraq War of 2003"? Not super likely. Therefore, don’t search "(Osama Bin Laden) AND (Iraq War) AND 2003 AND influence," which is too narrow and specific, but try instead "(Bin Laden) AND (Iraq War)," which is broad enough that it won't exclude potentially relevant material, but narrow enough that it won't overload you with results. Remember, with AND searches, you are always looking for the smallest number of terms necessary to get relevant data.

OR searches find information contained within two separate concepts. Again, if you envision your search terms as a Venn diagram, OR searches find all the data contained within two separate circles. This search method is great for synonyms, like "cars OR automobiles," "paedobaptism OR infant baptism OR child baptism," or for closely related concepts; it anticipates different ways a particular concept might be described, which saves you the trouble of running multiple searches. However, note that OR searches will expand your search results, since you are looking at multiple sets of data.

What if you want to relate two concepts, but both of those concepts have synonyms? You can combine AND and OR searches by nesting terms together. For instance, if you are researching the effects of homelessness on the education of young people, you can try a search like this: "homeless AND (youth OR adolescen*) AND (education OR school*)." Does this contradict my previous warning about overly complicated searches? Not at all! This search only includes three AND phrases, I just included synonyms and truncations to save myself the time of running multiple searches. (Also note that putting parentheses around the OR terms keeps the search engine from getting confused.)

NOT searches, like AND searches, narrow down your search by excluding data that you are not interested in. Unlike AND searches, however, NOT searches achieve this by eliminating all the information that two distinct search terms have in common. Basically, NOT searches ignore all the material contained within the area where the two Venn diagram circles overlap. For example, if you are searching for material on the symbolism of virgins and unicorns but are not interested in the mythical or legendary aspects of these two topics, you could try a search like this: "(unicorn AND (maiden OR virgin)) NOT (myth* OR legend*)."

Nota Bene: Crafting a good keyword search takes time, effort, creativity, and intelligence. Your first few searches may not come up with anything. Don't be discouraged! Keep refining your search terms and limiters and try again. Even librarians come up against regular dead ends in their searches! Just be persistent and learn from your failed attempts.

Exercise 4

YOUR HOMEWORK FOR THIS LESSON

1. Locate a journal article on JSTOR that is relevant to one of your papers this term.

  • You can access JSTOR through the "Links" page under the "Library" tab on Populi.

2. Submit a PDF document containing the following information under the "Exercise 4" assignment on the Populi class page:

  • A link to the article you found
  • The exact contents of your search bar (must include at least one phrase search, truncation/wildcard, and/or Boolean operator)
    • E.g., (unicorn AND (maiden OR virgin)) NOT (myth* OR legend*) 
  • The total number of results your search yielded (must yield less than 1,000 results)