Massive neutrinos leave peculiar imprints in the different cosmological data sets allowing cosmology to test their absolute mass scale. The cosmic microwave background (CMB) anisotropies has always been considered a powerful probe in this sense and the recent full sky measurement from the Planck satellite provide a unique tool. Planck has a wide frequency coverage designed to provide accurate discrimination of the Galactic emission from primordial anisotropies and it measures a wide range of angular scales. This unprecedented precision has triggered the search for fine effects in CMB due to the slightly different expansion history and structures formation in the presence of massive neutrinos. A careful statistical analysis is required for obtaining robust results. For both the standard cosmological model and its extensions that include massive neutrinos, we compare the traditional Bayesian approach based on Markov chains (MCMC) with a frequentist approach that uses profile likelihood. The comparison is particularly interesting due to the asymmetry induced by the physical constraint for the masses to be positive. We show that results from the two methods are remarkably consistent and that this study allows a better understanding of some important subtleties in the limits that are set.