TWU logo

Advanced Research Skills Workshop for Graduate Students

Stepping up Search

So many of us are so used to a single search box and a keyword or two that we fail to recognize that this pattern of searching is inefficient and likely to produce less than adequate results. Keywords are notoriously poor signals to meaning, so that what we think we are searching for often produces something quite different. Our more sophisticated academic databases incorporate metadata to enable faceted searching. This multi-step process works much better than the single search box - single search option.

The Structure Behind Academic Databases

1. Metadata

Question: Why does Google have a single search box and very few other options? Answer: Because Google is working with almost no metadata. This is important when it comes to searching for complex academic information.

So what is metadata? Think of a phone call, which has two parts to it: the content of the call (what was said) and the metadata (the details: who called who? when? how long? where was each party located?). Data is actual content. Metadata constitutes the descriptors that come alongside of (meta) the data.

Academic databases are based on metadata, constructed records detailing the main features of books and articles. The record includes things like title, author, publisher, date, page numbers, and so on. 

Here is the metadata for a journal article:

From whom was humanity saved?: the ransom soteriology of Origen's Commentary on the Epistle to the Romans
AuthorsWaldow, Daniel
SourceInternational Journal of Systematic Theology, 21 no 3 Jul 2019, p 265-289
Peer Reviewed: Yes
ISSN: 1463-1652
Publication Year: 2019
Language: English
SubjectsOrigenes , ca 184-ca 253

The items in red are metadata tags that can be searched with a database search engine, so that, for example, if you wanted all the articles in the database by Daniel Waldow, you could ask it to search for Author: Waldow, Daniel. A search engine like Google can't do this, because it lacks the metadata, which is created by librarians or database providers. 

For more on metadata, see

2. Controlled Vocabularies

Keywords in search are handy. They are also treacherous. A keyword lacks context. For example, the word "rock" could mean a hard substance, a type of music, or a type of motion, among other things. Thus keyword searches bring in multiple contexts, resulting in lots of irrelevant results. What is more, you need to think of all the right keywords. Most concepts have more than one way of describing them. Thus a keyword search is bound to miss relevant resources whose titles use different terminology.

Because book and article records provide metadata for searches, it is possible to add controlled vocabularies to them. What are controlled vocabularies? They are terminology that standardizes the way we describe things. For example, if we can provide a standardized way of providing author names, we can find everything by the same author in the database.

One significant form of controlled vocabulary is the subject heading. Think of a subject heading as a standardized tag that pulls together all the keyword terminology to describe a concept and makes them searchable under one term.  For example, is it Communion, The Lord's Supper, or Eucharist? A controlled vocabulary would standardize all three under the term "Lord's Supper." Why and who determines what subject word is used? (The why is fairly obvious - If you search on any one of the keyword terms, you won't get results that use one of the other two.  By searching on the standardized controlled vocabulary, you get everything, whether it has the keyword Eucharist or Communion or Lord's Supper.)

So who is responsible for the choice of subject headings? Most often it is the Library of Congress in Washington, DC, simply because they started and have maintained this controlled vocabulary system. When the metadata is created for a book or article, specialized add subject headings based on what the book or article is about, not just what terminology it uses in a title. Here is a tutorial:

Maximizing Search

1. Though a good beginning, keywords are treacherous.

Keywords are necessary if you have limited or no metadata (e.g. Google, Google Scholar). But they are treacherous in that you have to think of all ways to describe a topic or you will miss things, and because keywords lack a context, so you will get results that use the right word(s) but belong in a different context with a different meaning.

2. Use terminology from your research question.

If you are using keywords, draw your terminology from your research question or thesis. Like this:

(Open to full screen)

3. Search the problem, not the topic.

 Many searchers have a tendency to research a topic - global warming, Kingdom of God rather than focusing a search on the specific problem at hand. In research, you are not searching a topic. You are finding resources that will address the problem at the heart of your project. This means you need to:

 a. Search narrowly, incorporating the problem into your search by using search terminology from your research problem statement.
b. Never split up research problem concepts into separate searches. For example, if you are searching on atonement in the Book of Romans, search on atonement AND Romans rather than doing a search on atonement and then a second search on Romans. The latter will get you more search results, but relevance will drop dramatically.
c. Recognize that books deal with topics more broadly than journal articles, so you may, with book searches, have to search more broadly than your question demands.

4. Find and use the metadata.

 Academic databases have metadata that should allow you to search more intelligently than you can with Internet search engines:

a. One way to identify metadata is to use the database's advanced search. For example:

The subjects and title tags to  the right, above, are metadata enabling a focus on articles about the subject of atonement that have the word Romans in their titles.

b. Another way to find metadata to focus your search results is in the list of filters, often in the column to the left of your results:


c. You can also find metadata in the record for one of your results. Just click on its title:


Paul , Saint, Apostle -- Theology -- Soteriology
Greek language -- Terms -- Hilastērion
Atonement -- Biblical teaching

These metadata terms will be hyperlinked to allow you to click on them and get results.

5. Aim for highly relevant and tightly bounded results.

 I often see bibliographies that deal with issues at the edge of the research problem. They are thus diverse in their content with few of them actually addressing the problem directly. For Romans AND atonement, you may find books or articles on atonement in the New Testament or salvation in Romans but few directly on atonement AND Romans. Keep the content as tightly around the research problem as possible so that the bibliography looks uniform rather than diverse.

6. Tutorials



The following are tasks you can perform to build your skills. 

Task #1 - Read Research Strategies: Finding your Way through the Information Fog, chapter 4-6.

Task #2 - Do a search in EBSCOhost eBook Academic Collection (library home page - e-book collection), starting with a keyword basic or advanced search, based on your research problem. As much as possible, draw your keywords from terminology in your research problem. If you get too few results, you may have to broaden your focus. Then, on the results page, use subject headings to narrow your results. List 8 book citations in your institution's format (note that the Cite function, to the right of a search result when you click on a title, can help get you started).

Task #3 - Do a search on your research problem in Atla Religion Database (library home page - databases), starting with keywords and then using subject headings. List 8 article citations in your institution's format (note that the Cite function, to the right of a search result when you click on a title, can help get you started).