| Measurement and evaluation are distinct but logically related Deep Throat Love processes. Measuring the learner's knowledge is a means of evaluating, not only the learner himself, but also the teacher and the teaching materials. The test results are neutral; they are merely the data from which to deduce something about any factor in the teaching situation. Typically, experiments whose aim is to evaluate the contribution of applied linguistics to language teaching take the teaching materials as the dependent variable, whilst the teacher and the learner are the independent variables. This is a familiar problem in all scientific experimentation, but especially so in the educational field. It is expressed in the adage, 'You cannot jump into the same river twice'. All sorts of procedures have been devised to overcome it, and they all hinge on the notion of'same'. Whilst we evidently cannot use the 'identical' teacher, materials or pupils for comparison, we can make an effort to find equivalents or matches. Instead of teaching the identical pupils a second time, we get a group which matches them in respect of as many independent pornpros variables (age, sex, intelligence, etc.) as we can. Or we find a set of materials which aim to teach a different part of the language, but one which is equivalent or matches it in difficulty. Or we look for teachers with equivalent skill in using different methods of teaching. But matching can never be perfect. For example, because we do not have any direct way of measuring teaching skill, or assessing the difficulty of sets of teaching materials, matching these can only be subjectively valid; or, in technical language, unreliable. I shall return to the whole question of reliability again in a later section. |
