The strength of the olivocochlear reflex has been assayed by comparing ipsilateral cochlear responses with and without contralateral sound. In humans, ipsilateral cochlear responses have usually been inferred by measuring otoacoustic emissions (OAEs), whereas, in animal work, they have been assessed by measuring compound action potentials (CAPs). Thus reports that the reflex strength is smaller in humans than in animals cannot be interpreted until the differences between the two tests are better understood. The present study directly compares reflex assays using distortion-product (DP) OAE and CAP measures in the same animals. For ipsilateral frequencies of 2-8 kHz and levels from 25 to 80 dB SPL, efferent reflex strength was computed from the CAP or DPOAE amplitude-versus-level curves measured with and without contralateral noise. The "effective attenuation" produced by efferent activation was, with few exceptions, greater when measured with the CAP than with the DPOAE assay. Differences between the two measures increased as frequency increased, with differences as large as 10 dB observed. These results, coupled with previous measurements on humans and animals, suggest that the efferent reflex is at least as strong in humans as has been shown in animal experiments.