Effective population size predicts the repeatability of selection response after accounting for additive variance

Pre-registration / analysis-plan freeze for an empirical reanalysis testing whether effective population size (Ne) predicts the reliability (across-replicate dispersion) of artificial-selection response after holding selection regime fixed, with additive variance (VA) shared within series. Lead analysis: Jones, Frankham & Barker (1968) long-term Canberra factorial — among-replicate SD of cumulative response proxy R0–50 ~ log(Ne) + selection intensity. Secondary: Frankham et al. (1968) short-term realized-h2 dispersion (same design). Confirmatory: Weber (1990) wing-tip and Weber & Diggins (1990) ethanol. Yoo (1980) same-Ne reference. Griffin (2017) parked. IMPORTANT: Interim looks at Frankham short-term occurred before Jones was designated lead. See ANALYSIS_HISTORY.md and DECISION.md. This is a frozen exploratory reanalysis package with a pre-specified lead analysis, not a blinded confirmatory trial. Independent confirmatory series require a new registration. Author: R. Craig Stillwell (Independent Scholar) Local freeze: 2026-08-01