
Use keywords and filters to search databases effectively. That is the plain answer, and it is still the best place to begin. A database search works best when the words are chosen with care, then narrowed with the right limits.
I start with the words a topic can actually live under. A database does not guess as a person does. It looks for the terms I give it, and it may also look for subject terms, which are the database’s own labels for a topic. Those labels can be more exact than plain words, but they are not the same in every database.
That difference matters more than many readers first expect. One database may search broad keyword fields. Another may search only titles, abstracts, or subject headings. A third may have a thesaurus, which is a list of approved subject terms. If I do not know which field a database searches, I do not know what the result list really means.
Keywords help at the start because they are flexible. They catch natural language, alternate spellings, and related terms. But they can also bring in noise. A broad keyword search may return many items that are only partly on topic. That is not a failure. It is a sign that the search still needs shape.
Filters do that shaping. They narrow results by date, language, source type, peer review status, age group, or other stored labels. In some databases, filters are called limiters. They are useful because they cut away records that do not fit the need at hand. They are also limited because they only work well when the database has tagged the records in the first place.
That is the part I trust least when the documentation is thin. A filter can look precise and still hide relevant items. A recent article may not be fully indexed yet. A subject filter may miss records that were tagged in a different way. A source may say it offers a filter, but not explain how complete that filter is across the whole collection.
So the method is simple, but not simple-minded. Use keywords to cast the first net. Use subject terms when the database gives them. Then apply filters with care, one by one, so the result set stays readable. If the database has a search help page or thesaurus, it is worth treating that page as part of the source itself, not as decoration.
I also think it helps to read the results as evidence about the database, not only about the topic. A thin result list may mean the topic is narrow. It may also mean the database is narrow, or that its search tools are weak, or that its coverage is uneven. The database does not tell the whole truth unless its coverage and limits are stated plainly.
This is the main habit I return to. Search in stages. Start broad enough to see the field, then narrow with filters that match the question. Keep one eye on the record labels and one eye on the database’s own limits. That is how a search becomes useful instead of merely busy.
I still want one honest warning on the page. Filters are not interchangeable across databases, and controlled terms are not the same from one system to the next. A search that works well in one place may fail in another, not because the topic changed, but because the database did.
That is why I value plain documentation so much. A source is only useful once its coverage, search tools, and known limits are named clearly. Without that, keyword searching can feel richer than it is, and filters can look more exact than they are.
The Source List keeps faith with that same standard: one digital source worth knowing, one search tip, and one honest limitation. That is the right scale for database searching too, because the best search is not the biggest one. It is the one that states what it can and cannot do.