We analyze the security of an interactive identification scheme. The scheme is the obvious extension of the original square root scheme of Goldwasser, Micali, and Rackoff to 2 ^m th roots. This scheme is quite practical, especially in terms of storage and communication complexity. Although this scheme is certainly not new, its security was apparently not fully understood. We prove that this scheme is secure if factoring integers is hard, even against active attacks where the adversary is first allowed to pose as a verifier before attempting impersonation.