Define the question and collect evidence
Set a manageable topic and scope. Keep a source table with each study's question, sample, design, measures, findings and limitations. Record citation details as you read.
Organize by the argument
Group studies by a shared question, theme or method. Compare their evidence within each section. Explain whether differing samples or measures could account for disagreement; avoid treating every study as equally informative.
Example
Invented studies for a synthesis exercise, not real citations or archive submissions.
- Topic
- How is background sound related to recall during a short reading task?
- Study A
- A fictional experiment randomly assigns 40 volunteers to quiet or instrumental music. Mean recall scores are 8 and 7 out of 10, respectively; no uncertainty estimates are supplied here.
- Study B
- A fictional survey of 60 students asks whether they prefer music while studying. It measures preference, not recall performance.
- Synthesis
- These studies address different outcomes. Preference for music cannot confirm or contradict a recall effect. Study A's mean difference alone does not establish statistical significance.
Review your draft
Use real sources you have read, cite their findings accurately and distinguish reported results from your interpretation. End with the question your review resolves or leaves open.
Further reading: Purdue OWL: writing a literature review.
Past literature-review requests
Real historical request types, described without student text or files.
- Psychology literature-review request — September 2021
- Psychology literature-review request — August 2021
- Psychology literature-review request — May 2021
- Psychology literature-review request — May 2021
Common questions
Is a literature review an annotated bibliography?
No. An annotated bibliography discusses sources individually. A literature review connects them to develop an overall account of the evidence.
Must studies agree to belong together?
No. Disagreement can be useful when you explain differences in design, participants, measures or context instead of simply counting positive findings.