Natural assessment of oral language proficiency of limited English proficient students in paired reciprocal learning

Abstract

This study describes a naturalistic oral language assessment instrument used to determine language growth of limited English proficient (LEP) students in paired and non-paired learning environments. The purpose of this study was: (a) to assess the relationship between paired learning and oral language proficiency, and (b) to develop a direct observation language assessment instrument that observed English language use in its natural context that was both grammatical and functional. Based on the data presented in this study, it can be concluded that: 1. The English language growth of paired learners was either not effectively detected by the assessment instrument over the span of the study, or there was no growth at all. The non-paired learners demonstrated greater English language growth than the paired learners, although gains were also small. 2. The Naturalistic Assessment of Oral English Proficiency (NAOEP) instrument developed in this study: (a) exhibited good inter-rater reliability on all language variables, but still requires further refinement of specific language item descriptors, (b) demonstrated a general lack of score stability over time for both groups affecting the trustworthiness of the instrument, and (c) effectively measured non-language behaviors, but lacked the sensitivity to measure language growth in peer interaction due to the variety of communication forms involved. The NAOEP instrument can effectively be used by teachers in informal oral language assessment of individual students as they interact in the classroom. However, the instrument is impractical for use with large numbers of students. Various factors have possibly influenced the results of this study. The lack of stability of scores, short number of weeks, number and length of observations, and lack of sensitivity all may have contributed to the results. Caution should be taken with the interpretation of findings.