Why laboratory values of G scatter beyond their stated uncertainties
In plain words
About a dozen top experiments, each claiming an error of a few parts in 100000, give values of G that spread by about 5 parts in 10000. Something in how G is measured is not understood.
Precise statement
Laboratory determinations of G (torsion balances in time-of-swing, angular-acceleration and servo modes, beam balances, simple pendulums, atom interferometry) spread over about $5e-4$ relative, roughly ten times their typical quoted uncertainties of $1e-5 \text{ to } 5e-5$, near $G = 6.674e-8\,\mathrm{cm}^{3}\,\mathrm{g}^{-1}\,\mathrm{s}^{-2}$. Identify the unaccounted systematic effects (or establish any genuine dependence of the measured G on method, material or geometry) that explain the excess scatter.
What would settle it
A consensus set of independent measurements with different methods agreeing within quoted uncertainties at the 1e-5 level, together with an identified correction to the outliers.
Status in the literature
Unverified note
NIST published a 10-year blinded torsion-balance result in Metrologia in 2026 that lies about 235 ppm below the BIPM value obtained with the same apparatus design.
Related problems
- More general than Atom interferometry determination of G at 10 ppm
- More general than Same apparatus design, different G: the NIST versus BIPM discrepancy