The present study examined the reliability of a dichotic consonant-vowels (CV) pairs task using an ABX procedure. Twenty-four right-handed subjects (12 females, 12 males) completed a dichotic CV pairs task twice in a test-retest approach. The conventional free-recall or directed-attention procedure was replaced by an ABX procedure, in which participants were required to determine whether a CV heard binaurally was present in a dichotic pair presented on the same trial. Results showed the expected right-ear advantage in dichotic listening for verbal material. However, reliability was not as high as has been reported with a directed-attention procedure. Methods for improving the reliability of the ABX discrimination task are discussed. In addition, the discussion emphasizes the usefulness of the ABX approach for testing clinical populations.