[en] The approximations of the interacting boson approximation of Iachello and Feshbach and the ability of this approach to describe the ¹⁶O nucleus are tested. In the interacting boson approximation some ''exchange'' parts of many-boson matrix elements and violation of the Pauli principle in many-boson states are ignored. Matrix elements, two-boson wave functions and energies resulting from the interacting boson approximation are compared with exact two-hole two-particle calculations for the lowest 2⁺, 0 and the lowest excited 0⁺, 0 states in ¹⁶O. It was found that neither of the above two approximations of the interacting boson approximation are valid. Calculations without these two approximations improve the approximate energies but not the wave functions