Peer-reviewed research

Research papers are useful for tested methods, prior definitions and evidence about retrieval behavior. The abstract alone is rarely enough; readers should check the sample, setting, measurements and limitations.

• Methods and limitations are usually explicit.
• Findings can be compared with related work and later citations.
• Supports nuanced claims when the study conditions are reported.

What it does not establish

• Publication can lag behind commercial search product changes.
• A controlled study may not reflect a user's market or prompt mix.
• Access to full methods and data is sometimes restricted.

Why this record exists

Preview record for academic evidence. The production record must link to a specific paper.