We analyse the causes of a discrepancy between the earlier obtained samples of the mass distributions of exoplanets detected by the transit method and the radial velocity (RV) one and corrected for some observational selection effects. It is found that this discrepancy can be removed by introducing the following restrictions into the procedures forming the samples: (i) to consider, among transit exoplanets, only those which masses were determined by the RV method (i.e. excluding the TTV method); (ii) to take into account exoplanets with orbital periods P ∈ [1, 100] days and masses M ∈ [0.02, 13]MJ (Jupiter masses). In addition, we compare here the distributions by projective masses (which is Msin i, where i is the orbital inclination of an exoplanet). For this, the mass distribution of transit exoplanets is transformed into the projective mass distribution. Due to these three changes in the procedure, the obtained RV and transit distributions exhibit a similar behaviour in an interval of M ∈ [0.02, Mmid]MJ and coincide at M ∈ [Mmid, 13]MJ, where Mmid ≈ 0.17MJ.