cherrypicking

Cherrypicking, or cherry-picking, is the selective gathering or presentation of items to support a preferred conclusion while ignoring others. In everyday language, it refers to picking the ripest cherries, but in metaphor it denotes choosing only favorable evidence. The expression likely originates from the practice of selecting prime fruit during harvest.

In research and data analysis, cherry-picking is a form of bias and manipulation. It occurs when data are selectively included or excluded, when analyses are conducted in ways that yield significant results, or when only studies with positive outcomes are reported. Related terms include selective reporting, data dredging, and p-hacking. Cherry-picking can also appear in journalism, policy debates, and marketing, where favorable anecdotes, quotes, or statistics are highlighted while contradictory information is omitted.

Consequences include distorted conclusions, reduced reproducibility, and diminished credibility of researchers or publications. It can exaggerate effects, obscure uncertainty, and mislead audiences.

To mitigate cherry-picking, researchers use preregistered protocols, disclose all measured outcomes, share data, perform intention-to-treat analyses, and conduct sensitivity analyses. Systematic reviews and meta-analyses aim to minimize selective inclusion, and independent replication helps verify findings.