Revisiting multiple profiles of learner compositions: A comparison of highly rated NS and NNS essays

This study explored multiple profiles of highly rated learner compositions following the model proposed by Jarvis et al. (2003). The composition of multiple profiles focused on the co-occurrence of linguistic features identified by Cluster Analysis and functionally interpreted across essays written by native speakers (NS) and non-native speakers (NNS) of English. Data were obtained from a corpus of essays in response to a prompt that was similar to those from the Test of English as a Foreign Language (TOEFL) iBT Independent Writing Tasks. The essays were written by international graduate and undergraduate students enrolled in eight universities in the United States (U.S.) and NS undergraduate students from a university in the southeastern U.S. Each essay was rated by two trained raters from a rubric developed for the study. Only essays that received a score of five (5) (on a scale of 0-5) were included in the final dataset (NNS, N=24; NS, N=51). Results show that highly rated essays could be classified into six significantly different linguistic clusters which also vary across NS and NNS groups. Some implications for writing assessment, the teaching of academic writing, and materials production both for NS and NNS students are discussed.