In many practical applications, data about an unknown continuous distr
ibution arise in a grouped form. For these cases, estimation of popula
tion entropies and divergences must be done by means of their sample d
iscretized estimates. In this paper, the problem of loss of informatio
n due to the discretization of the data is studied for (h, phi)-entrop
ies and R(phi)h-divergences. Quadratic convergence theorems are given
and asymptotic distributions are obtained.