How to Build a Database Search Strategy for a University Assignment

A useful search begins before you type anything into a database. Translate the assignment question into concepts, decide what evidence would answer it, and record the terms you try. Searching only the full essay title often produces either nothing or a pile of unrelated results. A small search plan saves time later because it makes the inclusion decision visible. Use the library resources your university provides and treat search results as candidates to inspect, not ready-made proof.

Break the question into concepts

Underline the population, issue, setting and outcome in the brief. For a question about how flexible working affects employee retention in small firms, the concepts are flexible working, retention and small firms. Keep the concepts separate while searching. A phrase such as “effect of flexible working on retention in small firms” may be too restrictive because authors can use different words. Write one sentence describing what a relevant paper must actually study, then use that sentence to screen abstracts.

Build a synonym bank

For each concept, list plausible variants and discipline-specific terms. Flexible working may appear as remote work, hybrid work or telecommuting. Retention may be called turnover intention or employee attrition, though those outcomes are not identical. Put close alternatives together with OR, and connect different concepts with AND. Quotation marks can keep a fixed phrase together. Check the database help page for its particular truncation and field-search rules before using them.

Search in stages

Begin with a broad combination and examine the first useful records, especially their subject terms and keywords. Add or remove one concept at a time so you can tell which change improved relevance. If an excellent paper appears, examine its reference list and papers that cite it where the database allows. Search across more than one suitable index when the assignment spans disciplines. A single search engine cannot establish that you have seen every relevant study.

Record decisions as you go

Keep a compact search log containing database, date, exact query, useful filters and why you kept or rejected a source. Save the DOI or permanent record link, not only a browser tab. Separate a source that discusses your topic from one that directly supports your argument. Revisit your terms when new studies use unexpected vocabulary. If your lecturer requires a systematic method, follow that specific protocol rather than presenting an informal search as exhaustive.

Turn results into an evidence set

Read methods and limitations before building an argument from a promising abstract. Compare how studies define the same concept: retention measured by actual staff departures differs from stated intention to leave. Prefer sources that fit the question, not merely those with the highest citation counts. When evidence conflicts, record the difference in setting, sample or measurement rather than discarding an inconvenient result. The final paper should show how selected evidence answers the brief.

Worked example

Suppose your search returns many articles on remote work but few on small businesses. Try “hybrid work” OR “telecommuting” with “SME” OR “small business”, then inspect whether each study reports actual turnover or only employee attitudes. Keep both types if useful, but do not describe them as the same outcome. A two-column note—what the study measured and how it relates to the assignment—will make the later literature review much clearer.

A search log you can reuse

Create four columns: query, result count, promising records and reason for revision. On the first line, try your broad concepts. On the second, add one population term. On the third, replace an outcome with a recognised synonym. This sequence shows what narrowed the search and whether the narrower set lost an important study. When you find a relevant paper, capture its exact subject headings and test them as new terms. Stop when additional searches mostly reproduce sources you have already assessed, within the scope of the assignment. A search log is also a safeguard against a last-minute citation you cannot retrace. If a source has no accessible full text, record that access limit rather than relying on an abstract for a detailed methodological claim. In the final paper, describe the search method only to the degree the brief asks; do not call it systematic without a defined protocol.

Checklist before you submit

Extract three to five core concepts from the brief.
List alternatives without treating different outcomes as interchangeable.
Save exact searches and source identifiers.
Screen the methods and limitations before quoting a result.

Questions students ask

How many keywords should I use?

Use enough to represent the question, then change one concept at a time. Relevance matters more than a fixed keyword count.

Should I search only Google Scholar?

It can help you discover material, but also use relevant library databases and check access, coverage and source quality.