Reliabilty Test on Speaking Proficeiency for Engineering Students of Polines

Abstract

English language skills (listening, speaking, reading and writing) according to the result of survey on stake holders about the need analysis of company towards active aabilty of English for the graduates of State Polytechnic of Semarang conducted by UPT Bahasa (language Centre) in 2004, must be taught and tested. For the speaking test, UPT Bahasa uses the analytic approach that is the one which considers that any aspect attached to the English speaking ability basically develops at different rate so it should be evaluated  separately. This research tries to uncover wht=ether the approach has higher relibiality and efficientcy than the other approach i.e. holistic approach which considers speaking proficiency as general performance  and that it is represnted by a single score. The data used are taken from the result of dprsking assessment of 60 participants from 2^nd and 3^rd grade of Civil, Electronics and Mechanical Engineering. Based on the level of reliability coefficient 0.70 – 0.79 (Hughes, 1989) the result of the research shows the analytic approach has a correlation mark of 0.803 at level of significance 0.000, whereas holistic approach points at 0.653 correlation at the significance level of 0.000 which means that the analytic approach is reliable; but the holistic is not since the correlation points at less than 0.70.

Keywords: speaking, test, reliablit