Start from the claim, not the title
Break the claim into its concepts. Who or what it is about, what is being done or examined, and what outcome or property matters. Usually two or three concepts, and rarely more, because each additional one narrows the result set sharply.
A title contains words chosen to interest a reader. A claim contains the things you need evidence about, which is a different and much shorter list.
Find the words the literature uses
Your working vocabulary is frequently not the published one. The phrase common on your unit may appear in almost no indexed articles, while a term you would never say is the standard heading.
Look at two or three papers you already know are relevant and read their keywords and subject headings. That is the fastest available route to the right vocabulary, and it takes minutes.
Synonyms widen, concepts narrow
Within a concept, join the alternatives so any of them will match. Between concepts, require all of them. That structure is what lets you widen one idea without disturbing the others.
It also makes the search legible to somebody else, including you in three months. A string built this way can be read as a sentence about what you were looking for.
Use the database's own headings
Most academic databases index articles against a controlled vocabulary, and searching those headings finds papers regardless of the words their authors happened to use. That is a substantially different search from a free-text one.
Combining both is usually best: headings for reliability, free text for recent material that has not been indexed yet.
Truncation and phrases
A truncation symbol catches word endings so one term covers a family of them. Quotation marks hold a phrase together so its words are not matched separately across a record.
Both are simple and both are frequently the difference between four hundred results and four thousand. The symbols differ between databases, which is why a string is translated rather than pasted.
Test and adjust deliberately
Run it, look at the first twenty results, and change one thing. If the results are relevant but too few, widen a concept. If they are numerous and off-target, the vocabulary is probably wrong rather than the structure.
Changing several things at once is how people end up with a search that works and cannot be explained, which matters if anybody asks you to justify it.
Write it down as you go
Record the database, the string, the date and any limits, every time. It costs seconds and it is what lets you re-run the search later, translate it for another database, or report it in a methods section.
For any systematic work this is not optional, and for ordinary coursework it is still the difference between a search you can repeat and one you would have to invent again.
What good looks like
A few hundred results you can screen, most of them plausibly relevant, produced by a string you could explain to somebody else in a sentence. That is a working search.
Thousands of results is a vocabulary problem or a missing concept. A handful is usually a term the literature does not use, not an empty field.
Questions this raises.
How many concepts should a search have?
Usually two or three. Each additional concept requires every result to match one more thing, which cuts the set quickly and starts excluding relevant work that happened not to mention your third idea in its title or abstract. Where four seem necessary, it is often better to search three and screen for the fourth.
Why do my results look nothing like what I wanted?
Almost always vocabulary rather than logic. The terms used in your workplace and the terms used in publication diverge more than people expect. Take two papers you know are relevant, read their subject headings, and rebuild the search from those. It usually transforms the results in a single attempt.
Should I use the database's headings or free text?
Both, combined into a single search. Headings find papers reliably regardless of the words an author happened to use, and free text catches recent material that has not been indexed yet. Relying only on headings misses the newest work, and relying only on free text misses anything phrased differently from your own vocabulary, which is a great deal more than most people expect.
Do I need to record my search for a normal essay?
It is not usually required and it is still worth doing. A recorded string can be re-run when you need more recent work, adjusted when a concept proves too broad, and reused for the next assignment on a related topic. For any systematic review it stops being optional entirely.
What if a database uses different symbols?
Then the string is translated rather than pasted, and that translation is part of the work. Truncation characters, phrase marks and the way headings are entered all vary. Every database publishes its own guide, and checking it once for the two or three you use regularly is time repaid many times over.
Dr. Marisol Quintanilla
PhD, Public Health, on the mph and research side of the hall. Writes for The Footnote and coaches the students who bring this work to the desk. The rest of the specialists.
Elsewhere in the hall.
Send the assignment and the deadline.
A specialist credentialed in your own field reads what you actually have and replies within two hours with a plan and a price. The reading costs nothing, and if the honest answer is that we are not the right hall for it, you will get that instead.
