An Empirical Evaluation On An In-depth Measure Of L2 Vocabulary Knowledge

Vocabulary knowledge has no longer been viewed as a single dimension but a multidimensional construct. And approaches to testing vocabulary knowledge have generally been divided into tests of vocabulary size or breadth and tests of vocabulary quality or depth. This paper intends to conduct an empirical study on an in-depth measure of L2 vocabulary knowledge by using a DVK (Depth of Vocabulary Knowledge) test, which was coined and developed by Qian (1999; 2002).The paper applied three research instruments: MC (Multiple-choice Vocabulary Test), DVK and MT (Mid-term Vocabulary Test). MC consisted of a pool of 30 discrete multiple-choice vocabulary items from CET-4 Grammar and Vocabulary Section from the year of 1996 to the year of 2005. Then these 30 correct choices in MC were transformed into DVK headwords. The MT test was then utilized as an outside criterion for comparison. The study proposes to address the following research questions:1. Do MC vocabulary tests change in difficulty level when they are remade as DVK items?2. Does the DVK measure correlate with MC vocabulary tests?3. How is the concurrent validity of the DVK measure?4. Are the instructions and format of the DVK comprehensible to test takers?5. What is the appropriate time needed for completing the DVK?Altogether 52 students in BLCU participated in this study. Their scores on these three measures were collected and their teachers' appraisal was also utilized by the author to provide validity evidence for MT as an outside criterion. The descriptive and reliability statistics were obtained for these measures. The computed results by KR 20 indicated that DVK was quite reliable, with a reliability coefficient of 0.92, 0.20 higher than that of MC. Proportion Correct (P+) was chosen as the means for reporting the difficulty levels for all 30 items in DVK and MC. The results of the data analysis have...