The document defines key terms related to test item design such as stem, key, distracter, and normal item format. It discusses ideal characteristics of test items such as using a clear context and avoiding negative stems. Common problems in item design are outlined like non-homogeneous response options or difficulty stemming from instructions rather than the task. Guidance is provided on ensuring items have a single correct answer and assessing the intended language aspects. The importance of validity, reliability, practicality and backwash is covered.
Assessing Language Skills Effectively with Well-Designed Test Items
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14. Problem: In communicative tests, instructions should not be based on grammar, but on meaningful contexts. Wrong: Write sentences using make, let, and be allowed. Better: You got admitted in a university away from for hometown so you have to move to a student residence. Write sentences expressing the things that are prohibited or permitted in this context.
15.
16. Problem: It is never suggested to use made-up words or grammatical structures as distracters. Wrong: People say that my children are malraised/spoiled because I buy them whatever they ask for. Better: People say that my children are consented/spoiled because I buy them whatever they ask for.
17.
18. Problem: The difficulty in an item should lie in the task itself, not in the instructions or the stem. Wrong: When a police officer arrests someone, the officer must inform the person of a certain rights, including the right to remain silent and the right to an attorney. Why, according to the text, is this required? Better: According to the text, why must police officers inform suspects of their rights during an arrest?
19.
20.
21. Problem: The correct answer should be based on the content of a passage and not on previous knowledge or common sense. Wrong: Read the text and say if these sentences are TRUE or FALSE. a) The three basic principles to be environmentally friendly are reduce, reuse and recycle . ______ Better: Read the text and answer the following questions. a) What are the three basic principles to be environmentally friendly? _________________________________________
22.
23.
24. An item is VALID if it assesses authentic communication and interaction, and also reflects what students do in class. It has POSITIVE BACKWASH if students feel they have been tested fairly and become aware of the importance of language proficiency. Wrong: Dictation . Listen to the instructions to start a weblog and write the information on the space provided. Then translate the information into your native language. Better: You want to create a weblog and a friend of yours is giving you the information to set up an account. Listen to the instructions and fill in the missing information to learn how to start your own blog.
25. An item is RELIABLE and if it measures students’ performance consistently. It is PRACTICAL if it takes little time to design, organize and mark, and involves resources that are easily available. Wrong: In your opinion, how can computers help us? _______________________________________________________________________________________________________________________________________ Better: Mention three ways in which computers can help firefighters in case of an emergency? 1)____________________________________________ 2)____________________________________________ 3)____________________________________________
26. Integrative and open-ended test formats are usually very VALID and also have a POSITIVE BACKWASH, since they involve communication and interaction in real-like situations. Example: While reading today’s newspaper, you find two interesting articles. Express your opinion about them. However, integrative and open-ended test formats are often impractical (difficult to interpret) and unreliable (open and subjective).
27. On the other hand, discrete item test formats are usually very RELIABLE and PRACTICAL, BECAUSE they provide enough information, correct answers are limited and they are quick and easy to mark. Example: It is said that pit-bulls are a breed/race of dog that is particularly violent. Nevertheless, discrete item formats frequently lack validity (meaningful communication) and positive backwash (luck is as important as hard work).