skanowanie0036 (12)

literary, and cultural facts. Since the choice is of a two-way one, the probability of a chance success is much higher than in a multiple-choice test with four options. Even when a third choice “Don't know” is added, there is little guarantee that students wili not prefer to guess the answer.

—    Multiple-choice tests provide answers to ąuestions on lexical meanings of words or items of cultural interest. Of the choices Offered (usually 4) only 1 is appropriate, the others are based on probable errors.

—    The fill-in-the-blank tvoe of test may be multiple-choice or it may reąuire the student to write in an appropriate word. These tests are useful for assessing knowledge of grammatical structure, use of tenses, levels of language, and vocabulary. The teacher must ensure that the sentences are quite unambiguous and can be completed appropriately by only one word.

—    Matching tests are commonly used as vocabulary tests. Students are asked to match synonyms, antonyms, names of objects with oocupations, etc. Tests of this type can also be used for testing knowledge of facts about the foreign culture.

6.6. Cloze Tests

Cloze tests are a special and very freąuent type of tests. In a J cloze test words are removed from a reading passage at regular I intervals, leaving blanks. For example every fifth word may be | removed. The reader must then read the passage and try to guess the j missing words. The cloze procedurę is thus a techniąue for 1 measuring reading comprehension. But it can also be used to judge j the difficulty of reading materials. The reader is given a score m according to how well the words guessed matched the original j| words, or whether or not they madę sense. Two types of scoring ] procedurę are used:

1. The reader must guess the exact word that was used in the m original. This is called the exact word method.

2. The reader can guess any word that is appropriate or acceptable in the context. This is called the acceptable word method (also the appropriate word method).

6.7. Test Eyaluation

Piloting the test. If time and resources permit, a test should be piloted on a group of leamers similar to that for which it is designed. The pilot administration provides the test constructor feedback on the items and the appropriateness of predicted timing.

Item difficulty. Item difficulty refers to the proportion of correct responses to a test item. A test which aims to differentiate among respondents should have items which forty to sixty percent of the respondents answer correctly. If the purpose of the test is to determine whether nearly all leamers have achieved the objectives, then the item difficulty should be ninety percent or better.

Item discrimination. The item discrimination index tells how well an item performs in separating better leamers from the poorer ones. The index is intended to distinguish respondents who know the most or have the skills or abilities being tested from those who do not. The teacher can calculate item difficulty and discrimination for a test in class with the students' help.

Test reyision. If an item has a difficulty coefficient of lower than forty percent and higher than sixty percent, and if the discrimination coefficient is Iow, then the item should probably be eliminated.

Used sources

Brown, H.D. 1987. Principles of Language Leaming and Teaching. Prentice Hall k' Regents.

Oller, J. W. 1979. Language Tests at School: A Pragmatic approach. London: . Longman Group Limited.

Oller, J.W. 1983. Issues in Language Testing Research. Rowley, MA: Newbury House Publishers.

73