1.5 Keywords: The Language of Databases

Let’s talk about the language of searching. We know you can type almost anything in Google’s search box or other web search engines and find something. But when you’re doing research, it is important to think about your search terms more carefully. You’ll get better and more relevant results in most finding tools if you choose your search terms wisely. Here is an illustration of what we mean.

In his book on effective online research, William B. Badke explains what would happen if a trip to the grocery store operated the same way a typical web search engine works.^[1] He uses the example of looking for a particular brand of chicken soup, but unlike a typical grocery store, this store has no signs that list what can be found in each aisle, and all the brand names and labels of products have been taken away. Instead, you have to search for your chicken soup through all the aisles of the store.

Badke uses this example to explain the difference between unstructured web searches that look for your terms almost anywhere they might appear (the kind of web search most of us do every day), versus the kind of highly-focused searching that is possible in most scholarly indexes. We’ll take things a few steps further in our version of Badke’s supermarket analogy. Let’s say you go to the grocery store with no signs and no labels looking for corn. Perhaps you ask a literal-minded employee to help you find “corn” in the huge store. You know what you mean by corn, but the employee has other ideas, and shows you popcorn, candy corn, canned corn, corn starch, high fructose corn syrup as an additive in many foods, corn pads for your grandma’s feet, fresh corn on the cob, and corn as an ingredient added to dog food. You’re forced to walk through all the aisles, looking through countless products that weren’t at all what you had intended.

We’re sure you have seen this in action when you’ve searched for information on the web. Part of this is due to the way that most web search engines work, and part is due to how we often search the web without much thought. Basic searches in most search engines will retrieve anything that includes your words just about anywhere. In our simple supermarket example, our search would have benefited from more description or detail. For example, specifying what kind of corn you were looking for would have been helpful.

The main point of Badke’s analogy is that article indexes label segments of the item records they contain, for the same reasons that supermarket aisles are labeled – to help you find what you’re looking for in a quick and organized way. In other words, scholarly indexes let you focus your search to find your terms in very specific areas – for example, in the title, author, or subject areas – the same way that you could restrict a grocery store search to one specific aisle like the frozen vegetables section. The keywords and search strategies you use have a huge impact on your search results, so consider them carefully.

Natural Language versus Controlled Vocabulary

There’s another important difference between most web search engines and many scholarly indexes. Web search engines are more likely to support what’s called natural language searches, which are search terms and phrases using everyday language and even complete sentences. You can easily search Google or other web search engines using flexible natural language searches like the following:

• examples of 20th century art from the United States
• Forever 21 clothing sweatshops scandal
• what is the future of electric cars
• Twitter audience backchannel during presentations
• what should I not say during job interviews

With natural language searches, you can guess any words that might work. Web search engines are typically designed to work well with natural language searches.

In contrast, scholarly indexes and databases are less likely to support this type of casual searching and rely instead on highly-defined subject headings, or what’s called controlled vocabulary. You’ll probably have more success when searching scholarly indexes if you take time to discover the subject headings used by the index you’re searching. Use those terms instead of the natural language searches you might typically use with web search engines.

Here are some examples of controlled vocabulary you might use in scholarly indexes to find materials similar to those described above in the natural language searches:

• Art, modern–20th century–United States–Exhibitions
• Clothing trade–corrupt practices
• Hybrid electric cars
• Multimedia systems in presentations
• Employment interviewing

These specific phrases might seem clunky and awkward, but they are examples of controlled vocabulary selected to avoid confusion between similar terms. With controlled vocabulary, you’ll need to discover and use the subject headings or descriptors used in the index you’re searching.

The search terms and strategies you use will depend on context. For example:

• What type of finding tool are you using?
• What type of search vocabulary does that tool accept?
• Are some terms or strategies more successful in the finding tool you’ve chosen?

Search boxes may all look alike, but they may not handle your search terms or phrases in the same way. When you search multiple tools for a project, you will need to use different search terms or strategies in each of the different tools. Take the time to explore!

You’ll learn more about keyword and subject searching in specific finding tools in later chapters.