I bet this is one of those studies where it can turn out the dataset is entirely faked and it wouldn't change a thing in the field because it's such an obvious facet of well known established facts about humans empathising with random objects.