Being a good Bayesian would entail taking the evidence as-is, only adjusting your estimate of the "true" value after the fact taking other measurements into account.