Answers are grounded in this paper and cite their sources.
Small-sample experiments (n = 3–10) dominate preclinical and pilot research, yet, to our knowledge, no closed-form decision-theoretic framework links experimental costs, base rates and error control for small-sample screening pipelines. We present a cost-calibrated framework for pipelines in which initial screening precedes downstream confirmation, unifying frequentist testing, Bayesian posterior thresholding and Bayes-factor analysis as special cases of a single cost-parameterised family. Under a conjugate normal model we derive an analytic false-positive bound, a Bayes risk decomposition with a closed-form crossover condition, and a pipeline optimality theorem with controlled false discovery rates. Simulations and analyses of 248 effects from four domains, three of which provide replication or independent-confirmation data, show improved expected utility for moderate-to-large effects, moderate-to-high base rates and available confirmatory follow-up, and no benefit in low-base-rate or single-stage settings. Open-source R and Python implementations enable direct use without decision-theoretic expertise.
No takes yet. Share an insight, caveat, or question.
Hackenberger et al. (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: