Speaking module

For the first time, this study combined models and principles of authentic assessment from two parallel fields of applied linguistics as well as general education to investigate the authenticity of the TOEFL iBT speaking module. The study consisted of two major parts, namely task analysis and task survey. Utilizing Bachman and Palmer’s (1996) definition of authenticity, the task analysis examined the degree of the correspondence between the characteristics of the speaking module tasks in the TOEFL iBT test and those of target language use (TLU) tasks. In the task survey, a Likert Scale questionnaire of authenticity was developed by the researcher based on Herrington and Herrington’s (1998; 2006) four criteria of authentic assessment. The questionnaire was sent through email to 120 subjects who had already taken the test in order to elicit their attitudes towards the degree of the authenticity of the speaking section tasks. The results of the task analysis revealed a limited correspondence between the characteristics of the test tasks and those of the TLU tasks. However, the results of the task survey indicated that except for one factor (indicators), most of the test takers had a positive view toward the authenticity of the speaking module tasks in terms of the three other factors (context, student factor, task factor).

This study represents the inaugural attempt at assessing the authenticity of the tasks encompassed in the IELTS Speaking Module. The evaluation is conducted from the vantage points of applied linguistics and general education, and serves to enhance comprehension of authenticity and authentic assessment. In order to achieve this objective, an analysis was conducted on the Speaking Module tasks using Bachman and Palmer's (1996; 2010) model of test usefulness from the discipline of applied linguistics, as well as Herrington and Herrington's (1998; 2006) inventory of the fundamental attributes of genuine assessment from the realm of general education. The results of both task analyses revealed low indices of authenticity of the tasks. The high degree of the convergence of the results from the analyses could open new horizons to experts in applied linguistics to exchange ideas about authentic language assessment with those in general education. Furthermore, such characteristics as problem-solving skills, higher-order thinking skills, integrated assessment, and multiple indicators of learning, are not included in Bachman and Palmer’s framework.