| In this study, the author developed a lexical tool in a form of software. This study focused its investigation of the tool as a computer-based vocabulary size test. The vocabulary test was in a form of an item-bank and was named Vocabulary Item Bank of English (VIBE). The target words in the VIBE test were taken the General Service List (West 1953) and the Academic Word List (Coxhead 2000) and were classified into three word frequency levels: (1) 1000-word, (2) 2000-word, and (3) academic-word level. The VIBE test was validated In two stages. In the first stage, three different gap-fillings formats were compared: Format 1 (one letter gap-filling), Format 2 (two-letter gap-filling), and Format 3 (one-letter gap-filling and ?ecognizing word meaning). The results of the format comparison indicated that Format 3 was better than the other formats in terms of its central tendency stability, reliability, variability and criterion-related validity. For those reasons, Format 3 was selected as the most promising format for the computer-based final version. In the second stage, selected format was validated further. The validation focused on four aspects: (1) the different scoring systems, (2) construct validity, (3) content validity, and (4) computer-based test equivalence. The first validation aspect was the different scoring systems of the VIBE test. Because the VIBE test had a double task: (i) supplying the missing letter, and (ii) selecting the correct meaning of the target word, three different scoring systems were possible. The first scoring system was based on (i) only, the second scoring system was based on (ii) only, and the third scoring system was based on both (i) and (ii). The second validation aspect - construct validity - was evidenced by the presence of an implicational scale. The scale showed that the mastery of the academic-word level implied the mastery of higher frequency word levels (i.e. the 2000- and 1000-word level). The third validation aspect - content validity – was supported by the fact that a large number of test items could be administered within a manageable testing time. The last validation aspect dealt with the equivalence of computer-based VIBE administration. The VIBE test administrations were found to be equivalent by analyzing the central tendency, variability, Standard Error of Measurement, and parallel-forms reliability. The end result of this study was a computer-based vocabulary-size test that was both reliable and valid as well as practical. Additionally, the findings related the different scoring systems of the VIBE test enabled the author to propose a new Strength Hierarchy of Word Knowledge Hypothesis to complement the existing one. This study concerns the development of a computer-based lexical test called Vocabulary Item-Bank of English VIBE test. This dissertation summary focuses on five topics: (1) background of the study, (2) the development of the VIBE test, (3) the validation of the VIBE test, (4) the practicality or advantages of the test, and (5) atheoretical implication that can be derived from the findings of this study. |