Improbable Research brings us a
paper by someone who ran the experiment.
"Animal results showed that primate latencies correlated significantly with the human [scariness] ratings, while non-primate latencies did not."
(I also feel a desperate logical need to add another question mark to my subject line.)