Measuring measurement error

The chapter first discusses where and how one can find different observations for a phenomenon at a certain place and time. Examples from such diverse fields as GDP, schooling, consumer spending, greenhouse gas emissions, the investment-saving balance and public debt illustrate the availability of alternative observations. Against the background of earlier developed practical measuring rods for measurement error, the chapter sets out the basic instrument that will be used in this book to assess measurement error, the implicit minimal measurement error (IMME). IMME is a very simple statistic with a clear intuitive interpretation - the measurement error is ‘at least as big’ as the percentage reported for IMME. Measurement errors by definition cannot be known with more accuracy than the original data. Case studies deploying IMME report significant inaccuracies. Finally, the chapter sketches a procedure to investigate measurement error by means of seven steps that focus on the concept to be measured, the motivation of sources and periods to be considered, an ex ante error prognosis, data collection, cleaning of the data, calculation and, finally, reporting.