Estimation of flow or transport parameters often involves observations such as flow or concentrations that may be significant over multiple orders of magnitude, typically dictating the use of weights inversely proportional to the product of a coefficient of variation multiplied by the observation. This allows observations of considerably different magnitude to have similar importance in terms of guiding the parameter-estimation process. However, if a simulated value is of smaller absolute magnitude than the observed, the weighted residual will be limited to the inverse of the coefficient of variation, while weighted residuals of simulated values greater than the observed are unbounded. This produces asymmetry in the potential contribution to the objective function from simulated values whose magnitudes are less-than and greater-than the observed. This work uses a simple transformation to create a supplemental observation set. The combined observation set is used to demonstrate an objective function with improved characteristics: including the transformed observations results in a balanced set of objective-function contributions from simulated values smaller-than and greater-than the observation magnitude. The potential for improving the process of parameter estimation, generating sensitivities, as well as precautions for interpreting final statistics, are considered.