This study was motivated by Higuchi (2007), in which many advanced learners of Chinese made incorrect responses to seemingly easy items in a two-choice cloze-type test of the Chinese numeral classifier sequence yi ge. In Experiment 1, we tested native Chinese speakers as subjects, using the same material but allowing three possible choices for each item: (1) yi ge, (2) zero, or (3) yi ge/zero (i.e., either acceptable). The results revealed that there were several items which elicited a yi ge/zero response, thereby indicating that the validity of a two-choice test is questionable.In Experiment 2, both a two-choice and a three-choice test were administered to Japanese learners of Chinese. The results showed that the subjects made more correct...