Various studies have shown that experiments on nest defense and enemy recognition (e.g. recognition of adult brood parasites) can be confounded by many factors. However, no study has described a confounding effect of control dummy type. Here, I show experimentally that the choice of control dummy may influence the results of an experiment and lead to erroneous conclusions. I tested recognition abilities of the Blackcap (Sylvia atricapilla), currently a host rarely used by the Common Cuckoo (Cuculus canorus). Blackcaps responded very differently to two kinds of control dummies: they ignored the Eurasian Blackbird (Turdus merula) dummy, but attacked the Rock Pigeon (Columba livia) dummy as frequently as they attacked the Common Cuckoo. The differing results may be explained by the fact that the Rock Pigeon is more similar to the Common Cuckoo than the Eurasian Blackbird is, and consequently elicited more aggressive behavior than the latter. Thus, absence of discrimination in enemy-recognition studies may reflect a methodological artifact resulting from varying abilities of particular hosts to discriminate along a continuum of recognition cues. This result has serious methodological implications for further research on enemy recognition and aggression in general: a control dummy should not be too similar to the dummy brood parasite; otherwise, the chance of detecting existing recognition abilities is low. Further, I argue that coevolution only increases pre-existing aggression in the particular host species. Therefore, increment analysis (assessing changes in host antiparasitic responses during the nesting cycle while controlling for background aggression to control dummies) provides a more accurate picture of hosts' recognition abilities than the traditional approach (when the total level of antiparasitic response is analyzed).