1.3 Validity and reliability
Any assessment should be both valid and reliable. Validity asks whether grades generated by the kind of statutory testing described earlier represent a student’s achievement in the whole of the subject. Can a series of timed, written tests at the end of a key stage assess all those things MFL teachers think are important for students to learn about?
An over-reliance on test results may lead teachers to make generalisations and judgements about a pupil’s capability in all aspects of a subject, based on the formal testing of a subset. For example, a National Curriculum level tells us nothing about a pupil’s problem-solving ability or creativity, nor about their ability to work in groups or engage in extended tasks. Perhaps all that can be said is that the tests simply tell the teacher about the capabilities of pupils to answer questions at a particular time and of a particular type (and in the conditions and circumstances of the test) – no more and no less.
Reliability asks whether pupil performance changes (or not) depending on the particular questions that are set. Ideally, assessments should give every pupil optimal opportunity to demonstrate what they know. In practice, however, tests have been found to be biased against pupils from particular backgrounds, socio-economic classes, ethnic groups or gender (Pullin, 1993). Equity issues are particularly important when assessment results are used to label pupils or deny them access to courses or careers in the future.