The deliberate or institutionalized practice within scientific research of selecting only hypotheses, experimental designs, data, or analyses that are likely to yield a preferred, publishable, or fundable result. This includes p-hacking, HARKing (Hypothesizing After Results are Known), and the file drawer problem (not publishing null results). It corrupts the scientific process by making the literature a curated museum of "successes," not an accurate map of reality.
Scientific Picking *Example: A pharmaceutical company runs 20 trials on a new drug. The two that show a mild positive effect (likely by chance) are published. The 18 showing no effect or harm are filed away. This Scientific Picking creates a public, peer-reviewed "fact" of the drug's efficacy that is a complete statistical mirage.*
by Abzugal Nammugal Enkigal February 4, 2026
Get the Scientific Picking mug.