There are well-known systematic disagreements in partial photoneutron reaction cross sections obtained using quasimonoenergetic annihilation photons in experiments based on neutron multiplicity sorting methods. Using newly proposed criteria we demonstrate that a large part of the systematic uncertainty comes from certain shortcomings of experimental methods of neutron multiplicity sorting. To develop methods of correction of data obtained in experiments a new approach to data evaluation was developed in which a combined model of photonuclear reactions is used to decompose experimental total neutron yield reaction cross sections into partial reaction contributions. Evaluated cross sections of partial photoneutron reactions obtained using this method show a good agreement with results of alternative experiments.