The paper discusses the features of the previously published method of the Bayesian statistical evaluation of simultaneous satellite measurements of the minor species OH, HO2, and O3 at the mesospheric altitudes. These features are due to the introduction of a priori constraints on true concentration values (masked by measurement noise), which are determined by the condition of photochemical equilibrium of the species. The method is based on the probabilistic view of the satellite measurement process where the true concentrations of OH, HO2, and O3 are considered as random variables. In such a technique, we construct the a posteriori probability density of these variables and compare its statistical characteristics with the initial measurement data. It is shown that there is ambiguity in the construction of the a posteriori probability density of OH, HO2, and O3, which is due to the different ways of limiting transition from the three-dimensional probability distribution to the surface one. The ambiguity significantly affects the statistical means and leads to an inevitable systematic error. We present the main options for choosing the probability density, depending on the type of the transition. To estimate the systematic error, we tested the method by using artificial noisy model data on OH, HO2, and O3 that simulate perfect (unbiased) measurements. It is shown that choosing a patch transition leads to the least systematic error. Applying the method to MLS/Aura data of July 2005 confirmed the conclusion made earlier that the satellite measurements of the HO2 concentration have a significant bias greatly exceeding the systematic error of the method. This leads, in particular, to a significant error in the localization of the concentration maximum of this component at the mesospheric altitudes.