De validiteit van luistertoetsen Moderne Vreemde Talen

For ten years now multiple choice tests have been used in the Dutch school system to measure listening comprehension of English, French and German. The tests were developed in a research program, conducted at the Insitute of Applied Linguistics by Dr. ? Groot. Now that the tests have been in use for 10 years we are confronted with the following questions. Are the tests still reliable, as they were 10 years ago? In how far does the multiple choice technique give a true picture of the listening comprehension of students? Does the multiple choice technique help studens to cope with language material that they could not have coped with otherwise, in other words, to what extent does the language material used in tests suggest a higher level of listening comprehension than the students actually have? An experiment has been carried out at C.I.T.O. (Central Institute for Test Development). Students had to answer both multiple choice questions and open ended questions concerning the same language material. The results suggested that the language material used in tests was verydifficult for students to handle in an open ended question test form. The results also suggested that various levels of difficulty of the langua material used within a single test was reflected in the open ended test results, but not in the results of the multiple choice tests. The multiple choice technique seems to obscure the relative difficulty of the various test components. It has been found that an appropriate use of the multiple choice technique can cover only a restricted range of language material. The measuring technique must not restrict the choice of language material, and thereby influence content validity. A possible solution to the problem would be the development of a new kind of test. In this test a great variety of language material should be tested with a great variety of testing techniques: a great variety of language material in order to improve the content validity of the test, a great variety of testing techniques in order to reduce, as much as possi ble, the disadvantages of every single testing technique by itself.