For the past 70 years, test development has been dominated by what is called classical test theory (CTT). However, there are many problems associated with CTT, including: the resulting scales tend to be long; their interpretation is highly dependent on the normative sample; the assumption that each item contributes equally to the total score is often wrong, as is calculating a single index of measurement error for all possible scores; and it is difficult to equate different tests developed using CTT. Recently, a new approach to scale development has appeared, called item response theory (IRT), which overcomes all of these problems and, in certain cases, results in a scale with true interval-level properties. This article is an introduction to IRT. It concludes by discussing why IRT hasn't been adopted more widely, and some of its limitations.