They set up 80 identical nest boxes in woodland areas in Madingley (Cambridgeshire) during the pre-breeding season (February-April), when Jackdaws prospect potential breeding sites. The nest boxes had two perches, one at the base and the other just by the entrance. They fitted the inside of the entrances of the nest boxes with a circular printout, visible from the outside. Each nest box was randomly assigned one of four types of printout: a dark circle (control) a dark circle with white 'eyes' of the same dimensions of a Jackdaw (eyes only), a printout of a face-on photo of a Jackdaw with eyes retouched to increase the brightness (bright eyes), and the same printout but with dark eyes (dark eyes). They used remote video cameras to record Jackdaw visits to the nests, and the time they spent sitting on the nest itself and the perches.
The results were striking. Jackdaws alighted on the nest box significantly less when the bright eye print was in the nest entrance. In addition, they spent very little proportion of their visit time on the perch closest to the entrance both in the eye only and bright eye treatments. The visiting Jackdaws actually entered the nest box in two occasions, one in the control and another one in the dark eye treatment.
Davidson GL, Clayton NS, & Thornton A (2014). Salient eyes deter conspecific nest intruders in wild jackdaws (Corvus monedula). Biology letters, 10 (2) PMID: 24501271