The Cloze Technique and ESL Proficiency.

The variety of fill-in-blank test known as the “cloze” procedure is discussed as a device for teaching and testing ESL proficiency. Research with native speakers and the sparse literature available concerning studies with non-native speakers are explored briefly. An experiment is conducted to attempt to partially determine the discriminative power of a cloze test (scored by the exact-word method) and its validity as a device for measuring ESL skills. Students in beginning, intermediate, and advanced ESL along with two control groups of native (ENL) speakers (freshmen and graduates, respectively) are tested. Differentiation of levels of proficiency among the ESL groups seems adequate, but ENL freshmen are not significantly distinct from advanced ESL students though they are significantly inferior to KNL graduate students. The cloze test correlates best with the dictation (.82) on the UCLA ESLPE 2C, and next best with the reading section (.80): multiple correlation with all sections is .88. It is concluded that the cloze method is a very promising device for measuring ESL proficiency.