1
Key Words: database, online test, English proficiency test, language competence Originally, Chung Hua University (CHU) has planned to group students according to their Chinese language proficiency and this project was brought up to serve as a prerequisite for it. Since some students at CHU do not have enough Chinese language competence, Chinese remedial courses will be offered. This project is focused on grouping participants 2
modern Chinese literacy and does not get involved culture knowledge that most Chinese language tests will cover. Instead, the project on placement test similar to English proficiency test adopts easily promoted online test module and only gets involved language competence. The target participants of this project are non-language freshmen. While reading skills are strongly emphasized in this online placement test, writing skills are also incorporated into it. The online placement test served as a prerequisite for grouping students according to their Chinese language proficiency is a pioneering experiment: We are trying to build up a test bank that covers all aspects of Chinese language competence and provides both incorrect and correct examples illustrating mistakes that Chinese students tend to make. Therefore, this test bank is divided into three categories content-based, type-based, and level-based. This test bank closely combined with database and online test system can randomly select test questions from different categories but with similar content genre and level of difficulty. Chinese language teachers should collaborate closely on the project with program developers: Chinese language teachers should concentrate on the collection and revision of test questions for a long time. Then, the program will be executed to test the data offered by Chinese language teachers. Finally, freshmen will serve as subjects to try this test bank and test system so that Chinese language teachers can receive feedback and program developers can debug the program. Of course, this experimental project can not achieve immediate results. What matters is that we will design a set of expandable test bank and modules that can be developed in the future. 3
1 a b c a b a b c a b A0 A1 A2 B1 C1 C2 D1 D2 E2 4
F1 G1 5
This online evaluation system for Chinese language proficiency will be introduced with the following two major parts (1) Chinese word and phrase test bank, and (2) web-based test system. 1. Chinese word and phrase test bank A. Testing System With accumulated experience of test writing and revision, we have further achieved two goals this year: classification and organization of test bank and design of various test types. Referring to some theories about psychological analysis and language test, we have made modifications to our test bank. Roughly speaking, tests can be classified according to the following different features or purposes: (1) goal-based: achievement test, competence test, and potential assessment (2) test genre: fragmented test and general test (3) scoring type: (a) objective test, subjective test, and semi-subjective test 6
(4) channel type: paper and pen test and computer-aided test Since this project is defined as an online evaluation system for Chinese as a tool, it is close to features of competence test. Although we are aware of the fact that subjective or semi-subjective tests can not be totally replaced by online and objective tests, we still adopted fragmented, objective, and computer-aided testing model considering the purpose of this project and the limitation of our working conditions. B. Construction of Test Bank The content of test bank was roughly divided into five categories: words (including phoneme), collocation, idioms, sentence patterns and word order, and semantic transition. However, this classification has been changed into four categories: words, idioms, syntax, and paragraphs. C. Design of Test Patterns and Rules of Test Question Revision (1) If reference books are available, test items are selectively chosen and carefully revised from them. (For example, the test items of idiom test were taken from idiom dictionary.) (2) Test patterns are simplified to adhere to the principle of fragmented test. All test items are multiple-choice questions with one correct answer. The priorities of test items are easy to edit and clear to understand so that the combination of test items in one test pattern can match the function of randomly chosen test items. The main consideration of test item choice is that we only test the smallest unit of a language one time. That is, we only test one test point in one question. For instance, if we try to test pronunciation, we just test one word in one question. While we try to test students idiom ability, we won t use wrong words to mislead them we only test their idiom ability. Since our test consists of test items randomly chosen by online test system, how the online test system can avoid choosing some certain repeated test items should be considered. Another problem that we have to face is the repeated occurrence of some test points. Right now, we are using key words to avoid these problems. We are using of the first character (Hanyu Pinyin) of each target words, expressions, or idioms as a key word. Whenever the online test system is randomly choosing test items, it will match Hanyu Pinyin of key words and select those test items with differently pronounced key words. 2. Web-based test system The computer center of Chung Hua University is responsible for the design and 7
development of web-based test system. It has achieved the following goals at this stage: (1) The purchase and installation of operating system and database software. (2) The establishment of development environment and analysis of system flow. (3) The function of user s right management. (4) The system management of test bank and design of user interface. (5) The automation of test item choice and management and setup of test questions. (6) The configuration and management of examinees. In addition to the design of online test system, we have also conducted a preliminary, hand-on test and obtained some test results for further analysis and improvement. We can also use the test results as a benchmark for the adjustment of difficulty level in the future. 8
9 4. 1
10
11 1
12
13
14
A0 A1 A2 B1 C1 C2 D1 D2 F1 G1 E2 15
A0 A1 16
A2 17
B1 C1 18
C2 19
D1 誔 20
D2 21
E2 22
F1 23
卧 踪 G1 羣 4. 24
1 2 3 25
26
1 2 3 4 D20001 D 2 0001 27
3 28
A0 161 A1 103 A2 213 B1 498 C1 417 C2 199 D1 026 D2 060 F1 024 G1 044 E2 124 1869 29
/ / 30
巻 31
32
33
34
35
36
37
38
(1) (2) 39
40
41
42
43
44
45
47
48