Theresa Jiin-ling Tseng
Tunghai University, Taiwan
The construct of communicative competence has been greatly emphasized in second language teaching at different levels of education. The purpose of this study is to examine the construct validity of the assessment practice of a Freshman EFL for Non-Majors (FENM) program that stresses communicative competence. The methods of this study include an inspection of both the assessment practice recommended in the Teachers’ Handbook, Freshman English for Non-Majors 2002-2003 (“FENM Teachers’ Handook,” 2002) and samples of the FENM program-wide exams, and interviews with three experienced teachers on their assessment practice. The findings indicate that the test format (multiple-choice questions) used in the program-wide exams of reading and listening does not match the characteristics of communicative language testing. Consequently, students’ reading and listening skills that are measured do not permit teachers to predict the students’ ability to use the skills to negotiate with others successfully in a naturalistic situation. The oral assessments often times do not allow authentic interaction to take place since most teachers let students write and memorize the scripts before the oral assessment. As a result, it is difficult to measure an important attribute of communicative competence—students’ ability to process unpredictable data in real time. Overall, the construct validity of the assessment program appears to be low. The results of this study may spur teachers of CC-oriented programs to reflect upon whether their test practice measures what they intend to measure, and what can be done to increase construct validity of their tests. Recommendations on using authentic tasks and guidelines to increase construct validity of communicative language testing are provided at the end of this article.
Key Words: communicative competence, communicative language testing, construct validity, language assessment, authentic tasks, measuring communicative competence