Validity and reliability

Assessment needs to accommodate accountability, learning, selection, quality control and cost control. The stakeholders include trainees, employers, regulators, courseware providers and advisors’ clients. An adequate assessment strategy needs to be fair to all concerned. A fair assessment instrument is valid and reliable.

Valid tests (or assignments or other assessment instruments) measure what they purport to measure. A reliable test involves the same standards, no matter when it is taken or who marks it. Reliable tests give consistent results.
Imagine you are doing a test which is supposed to measure competence in driving a car. The first question asks you when cars were invented, and by whom. This is an invalid question! You don’t need to know the history of cars to be a competent car driver.

Now, suppose you do an assignment and get an A+. Business commitments prevent you sitting the exam and you have to repeat the subject next year. This time you fail the assignment. This assignment is an unreliable assessment instrument.
A good assessment instrument is valid and reliable: it consistently gives the right results.