Project pathway · question to evidence

Find evidence without drowning in it.

Turn one focused question into a search you can explain, refine and repeat. Use the annotated example, then keep your real searches in the downloadable log.

Bring: one focused research question.
Leave with: a concept plan, exact search record and a shortlist ready to appraise, then save in EndNote or Zotero.
Boundary: a documented student search is not automatically a systematic review.

The central habit

Search in passes, not for one perfect query.

Start broad enough to learn the language of the topic. Inspect useful results, add better terms and repeat. Keep enough detail for another person, including future you, to understand what you did.

Before searching: complete the research-question guide. A database cannot rescue a vague question.

By the end you can

  • choose a proportionate search rather than promise a systematic review;
  • turn question concepts into keywords and subject headings;
  • use AND, OR, quotation marks, truncation and brackets deliberately;
  • record, screen and save useful evidence without losing the trail.

1 · Scope the task

Decide how exhaustive the search needs to be.

Orientation

Learn the field.

Use a reading list, recent review and Library Search to discover key authors, terms and debates. Follow promising reference lists.

Student project

Search transparently.

Use a planned set of sources and concepts, record exact searches, screen consistently and explain sensible limits.

Systematic review

Follow the protocol.

Use the required review methods, databases, reporting standard, duplicate processes and information-specialist input.

Use the label your method earns. “I searched three databases and kept a log” describes good practice. It does not, by itself, make the work a systematic review.

2 · Plan the concepts

Break the question into ideas a database can match.

Worked question

Among older adults, is social isolation associated with mortality?

Concept 1 · population

Older adults

older adult* · ageing · aging · elderly

Concept 2 · exposure

Social isolation

“social isolation” · loneliness · “living alone”

Concept 3 · outcome

Mortality

mortality · death · survival

  1. Write one concept per column. Do not paste the full research question into the database and hope.
  2. Add synonyms, spelling variants and abbreviations. Borrow useful language from relevant papers.
  3. Check the database thesaurus. A subject heading such as MeSH may retrieve records that use different wording.
  4. Keep only concepts that help. Adding every PECO/PICO element can make a search unnecessarily narrow.

3 · Build the search

Combine alternatives with OR; combine concepts with AND.

Annotated example

A first-pass keyword search

(older adult* OR ageing OR aging OR elderly) AND ("social isolation" OR loneliness OR "living alone") AND (mortality OR death OR survival)

In plain language: a result must match at least one term from each coloured concept group.

  1. OR widens within a concept. The record can use any listed alternative.
  2. AND narrows between concepts. The record must represent every connected concept.
  3. Quotation marks hold a phrase together. Use them only when word order matters.
  4. The asterisk asks for word endings. adult* may retrieve adult and adults.
  5. Brackets keep the logic together. They stop the database mixing your OR and AND groups incorrectly.

Database syntax varies. Check its help page and inspect how it translated your query. In PubMed, use Advanced Search and Search Details to see what was actually run.

4 · Search, inspect, refine

Use the right place for the question, then learn from the results.

01

Orientate

Begin with the module reading list, a recent review or a known useful paper.

02

Search broadly

Use Newcastle Library Search to find books, articles and material across collections.

03

Search by subject

Use the relevant subject guide and databases. PubMed is useful for biomedical literature, but no single database covers every question.

04

Chase citations

Look backwards through reference lists and forwards to newer papers that cite a useful source.

Too many results?
  • check that every concept is present;
  • search important terms in title or abstract fields;
  • replace vague terms with more specific ones;
  • apply only limits you can justify.
Too few results?
  • add synonyms and spelling variants;
  • remove a weak or over-specific concept;
  • check subject headings and indexing;
  • remove accidental date, language or article-type limits.
Results look wrong?
  • open two or three records and inspect their language;
  • check brackets, phrases and field codes;
  • look at the database’s translated query;
  • record the change and why you made it.

5 · Keep the trail

Record the search while it is still on screen.

Downloadable workbook

Plan, log and screen in one place.

The workbook includes a start page, concept planner, 50-row search log and 100-row screening log.

Where

Database or platform, plus the specific collection if relevant.

When

Date searched, because databases and records change.

Exactly what

The complete query as run, including fields, phrases and brackets.

Limits

Dates, languages, study types or filters, plus the reason for each.

What happened

Results found, records exported and what you changed next.

A screenshot can help, but searchable text in a log is much easier to reuse, compare and report.

6 · Screen and save

Make one consistent decision, then keep the useful source.

  1. Export the records. Move results into your chosen manager or screening workflow and remove clear duplicates.
  2. Screen title and abstract. Use the same question and inclusion rules for every record; be cautious about excluding uncertain records early.
  3. Check the full text. Record a clear reason when an apparently relevant paper is excluded.
  4. Link reports from the same study. One study may produce several papers; do not automatically count each report as independent evidence.
  5. Appraise before relying on it. Map the study, inspect the results and record what each useful paper can and cannot support.
  6. Save what you use as you read. Verify the record, attach or link the full text where permitted and keep the appraisal note with it.
Continue the recommended route Read and appraise each useful paper → Turn a shortlist into structured judgements before the evidence enters your argument.
Go deeper · official searching guidance

Use these authoritative guides when you need database-specific detail or are planning a more comprehensive review.

Newcastle University · Planning

Plan how you will find information

Move from a question and concepts to keywords, sources and an iterative search.

Open Finding Information
Newcastle University · Techniques

Use search operators deliberately

Extend the worked example with truncation, wildcards, controlled vocabulary and citation searching.

Open Search Techniques
Newcastle University · Discovery

Use Library Search well

Search the catalogue, refine results and use Advanced Search for more control.

Open Library Search guidance
PubMed · Official help

Translate the logic into PubMed

Use Advanced Search, field tags, filters, search history and Search Details.

Open the PubMed User Guide
Cochrane · Comprehensive searching

Understand review-level searching

Read the fuller standard for sources, strategies, deduplication, study records and reporting.

Open Cochrane Handbook Chapter 4