Measurement Sensitivity: What, Why, and How?

Did students learn? How much did an athlete improve? How many more steps were taken after an intervention? To be able to answer these questions is essential in measurement practice and it is directly related to the sensitivity of the measures employed. Yet, little literature has covered sensitivity and related topics, and it is basically ignored in constructing measurement tools. As a result, many so called valid and reliable instruments are useless since they are not sensitive enough to detect the differences or changes of interest. What is measurement sensitivity then? It is the measure being able to detect a difference or change when there is a one. While sensitivity is related to measurement errors (systematic/random), validity (measures what it purports to measure), reliability (consistency of the measure), it is impacted mainly by the precision of a measure, which is the extent of a measure being free of random error. While some generic indexes of precision, e.g., Standard Error of Measurement and Root Mean Square Difference, have been developed and are being used in practice, they can only provide “global/generic” precision estimations. Fortunately, conditional, local and context-specific prevision indexes have been developed under both classical and new testing theories. After providing an overview on what sensitivity is and why it is important, how to determine and evaluate sensitivity will be described in detail. Related physical activity assessment examples will be illustrated, and advanced concepts (e.g., “tolerance for error”) and practices (e.g., jackknife procedure) will be introduced.