A note on Dean's overdispersion test