Comments on 'Bayesian estimate of the Newtonian constant of gravitation' with an alternative analysis

This paper provides critical comments on the methodology employed in 'Bayesian estimate of the Newtonian constant of gravitation' ( Dose 2007 Meas. Sci. Technol. 18 176 - 82), where many independent estimates of G and their quoted uncertainties are combined to form a single estimate and standard uncertainty. It is argued that the use of the non-Gaussian likelihood functions is unwarranted and that some of the prior distributions used are inappropriate. Another criticism is that the uncertainty of the final estimate is interpreted differently from the uncertainties of the contributing estimates. A non-Bayesian means of analysis is presented for problems of this type. The corresponding estimate is G = 6.67408(12) x 10(-11) m(3) kg(-1) s(-2), with the figures in parentheses indicating standard uncertainty.