Background: Obligate brood parasites exert strong selective pressure on target hosts. In response, hosts typically evolve anti-parasitism strategies, of which egg recognition is one of the most efficient. Generally, host egg-recognition capacity is determined using model eggs. Previous studies have shown that some host species, which are capable of detecting parasite eggs, do not reject model eggs. However, it is unknown that whether the reaction to model eggs varies among distinct populations of the same host in relation to the degree of parasitism pressure. Results: Here, we compared the rejection frequencies of model eggs and real eggs between mainland and island populations of the plain prinia (Prinia inornata), which are respectively sympatric and allopatric with their brood parasite, the common cuckoo (Cuculus canorus). Our results indicated that the mainland and island populations rejected real eggs at similar rates, but rejected model eggs, which were similar in size to real eggs but heavier, at significantly different rates: the island population rejected fewer model eggs, possibly because the rejection motivation of this population was lower due to absence of parasitism. Conclusions: Our results indicated that some factors affecting the decision to reject, such as rejection motivation, varied according to the degree of parasitism pressure, and thus influenced the frequency of egg rejection. Furthermore, our results suggested that model eggs should be used with caution in comparative studies of egg recognition abilities among species or populations subjected to different intensities of brood parasitism. That is, model eggs may fail to accurately detect egg recognition in host populations with little to no risk of parasitism.