Thirty-five dogs watched forty second videos of three human faces, one of them their own owner’s, and blinked at a higher rate when the faces on screen were blinking, an increase that showed up for owners and not detectably for strangers, with no significant blink-for-blink timing between dog and human

Two different comparisons sit inside the result, and a third prediction failed. The blink-video effect appeared only when the face on screen was the dog's owner; the direct owner-versus-stranger comparison separated only for female owners; and the prediction that dogs would time their blinks to a human's was not confirmed.