Literature Review Search Log: A Reproducible Checklist for Researchers

First decide what kind of review you are doing

A narrative thesis chapter, scoping review and systematic review have different methods and reporting expectations. Ask your supervisor and read the programme or target-journal instructions before promising a “systematic” search. The search log below is useful for any review, but it is not itself a systematic-review protocol.

For systematic reviews, the PRISMA-S search-reporting extension explains what to report about information sources, full strategies, limits, supplementary searches and record management. The Cochrane Handbook’s searching chapter gives detailed methodological guidance for intervention reviews. Apply the guidance that fits your discipline and study design.

Record each search in a form someone else could rerun

1. Research question and version. Record the question, eligibility criteria and date they were agreed. Save later changes rather than silently rewriting the original plan.
2. Source and platform. Name the database or website and the platform used to reach it. “Google” or “library” is too vague for a reproducible record.
3. Date and coverage. Record the day searched and any database coverage or publication-date limits. A date filter should follow your eligibility criteria, not an automatic “last five years” rule.
4. Exact strategy. Save the full query as entered, including field tags, Boolean operators, phrase marks, subject headings and language or document-type filters. A screenshot alone is difficult to reuse.
5. Result count and export. Record the number returned before deduplication and the exported file name. If a platform caps exports, note that limit.
6. Selection trail. Document duplicates removed, title/abstract decisions, full-text decisions and exclusion reasons at the appropriate level for your review type.
7. Update and hand search. Log any later searches, citation chasing or reference-list checks separately, with dates and methods.

Store the log alongside the citation library and a read-only copy of each export. Your institution’s data handling rules still apply to any unpublished material or personal data.

One search-log entry, with illustrative values

Question: what barriers do UAE small businesses report when adopting digital payment tools? Source: the named bibliographic database and its institution-provided platform. Search date: 14 October 2026. Query: (“small business” OR SME) AND (“digital payment” OR “mobile payment”) AND (UAE OR “United Arab Emirates”). Limits: English-language journal articles, if those limits are justified in the protocol. Result count: enter the actual count shown by the platform; do not invent it. Export: record the RIS or other file name and the number successfully exported. Next step: test whether the query retrieves a few already-known relevant studies, then refine terms and log the revised version as a new entry.

This is a formatting example, not a validated strategy for that research question. Different databases use different field syntax and subject vocabularies. A librarian or information specialist can help test and translate a strategy when search completeness matters.

What makes a search record unusable?

• A final reference list with no record of excluded or duplicate results.
• A broad query described only in prose, so another researcher cannot reproduce it.
• Filters applied after viewing the results but not recorded.
• Mixing exploratory searches with the final planned strategy without labels.
• Reporting the same database twice through different platforms as if it were two independent sources.
• Using AI-generated citations without checking that each source exists and was actually retrieved.

Our UAE academic-source guide discusses where to search and how to assess individual sources. The search log serves a different purpose: it records the route from a question to a transparent source set.

Keep search, deduplication and screening separate

A reproducible search log records how records entered the project. A screening record explains which records were included or excluded against the eligibility criteria. Keep both. If 100 records are exported and 12 duplicates are removed, do not say the search returned 88. The source returned 100; 88 unique records remained for screening. Make the distinction in a flow diagram when your review type requires one.

For a narrative literature review, the exact level of screening detail should match the thesis brief and supervisor’s expectations. For a systematic review, the PRISMA-S guidance sets a higher search-reporting standard, and the Cochrane Handbook provides detailed methods for intervention reviews. Neither source means every ordinary thesis chapter should claim to be systematic.