ERROR ANALYSIS IN ENGLISH SPEAKING SKILLS OF PRIMARY SCHOOL 4TH GRADE STUDENTS

Abstract

An error analysis study was conducted in order to investigate the errors of fourth grade Turkish students with regard to English speaking skill. Demirel (2012) stated that among the skills in foreign language teaching, the speaking skill might be the most difficult skill to acquire. There are various reasons behind the scene such as learning difficulties in speaking a foreign language, the continuum of these learning problems in the following years, hesitation of students to speak English and attend English courses, exam-oriented learning and teaching environment, and not putting the information into the long-term memory. Considering these reasons, the present study showed an attempt to examine errors in the speaking skill of primary school students. Error analysis is employed to find the sources of errors by making an in-depth investigation (Corder, 1974). In the current study, via a quasi-experimental research design, a 9-week implimentation was carried out to improve the speaking skills of the participants. Descriptive analysis was used to analyze of the data. The results of speaking pre-tests and post-tests in the experimental (n=17) and control groups n=(15) were utilized. "English Speaking Skills Rubric for Primary School Students" was used as a measurement tool. The outcomes were analyzed by employing Man Whitney U test. In the error analysis focusing on the development of students' speaking skills based on the difference scores, the type of sub-dimension that the students were found to be more successful, respectively, as follows: Task achievement 10.7 difference points (out of 20 points) 53.5%, Comprehension and expression 4.7 difference points (out of 10 points) 47%, Fluency 7 differences points (out of 15 points) 46.2%, Vocabulary 8 differences points (out of 20 points) 40%, Grammar and structure 8 differences points (out of 20 points) ) 40%, Intelligibility 6 points (out of 15 points) 39.6%. In this situation, while the most errors were detected in the intelligibility sub-dimension, the least errors were observed in the task achievement.