The "finite-size" effects in the isobaric yield ratio (IYR),which are shown in the standard grand-canonical and canonical statistical ensembles (SGC/CSE) method,are claimed to prevent the actual values of physical parameters from being obtained.The conclusion of SGC/CSE may be questionable for neutron-rich nucleus-induced reactions.To investigate whether the IYR has finite-size effects,we re-examine the IYR for the mirror nuclei [IYR(m)] using a modified statistical abrasion-ablation (SAA) model.It is found that when the projectile is not so neutron-rich,the IYR(m) depends on the isospin of the projectile,but size dependence can not be excluded.In reactions induced by the very neutron-rich projectiles,contrary results to those of the SGC/CSE models are obtained,i.e.,the dependence of the IYR(m) on the size and the isospin of the projectile is weakened and disappears in both the SAA and experimental results.