A Proposed Framework for Testing the Oral Language of Second/Foreign Language Learners

Within the proposed framework, oral tests consist of several different speech interactions, each of which encompasses a variety of contextual variables that affect its output. These interactions are performed as simulations in order to tap unmonitored language, which approximates the most stable discourse style, namely, the vernacular. The sample of speech interactions included on an oral test is randomly selected from specifications of those interactions most relevant to the specific context and purpose of the assessment. The language samples are then assessed with rating scales, whose criteria vary according to the intended use of the test results. The accuracy of the scores, the quality of the test as a whole, and the assumptions made about oral language are examined through the relevant types of reliability and validity as well as through extensive research.