Classical Test Theory

Classical test theory (CTT) is an approach to measurement that considers the relationship between the expected score (or “true” score) and observed score on any given assessment. The word classical is used in the sense that the theory is considered to be the first practical application of mathematics to describe this relationship. CTT offers a relatively parsimonious, elegant, and intuitive way to scale individuals according to some theorized latent construct. This entry further describes CTT and its basic principles and estimation procedures, then discusses its framework for determining a measure’s proportion of true score variance, standard error of measurement, item analysis, and validity. Finally, it looks at the limitations to the theory.

Although more contemporary, model-based approaches to measurement, such as item response theory (IRT), have garnered more focus, CTT retains its relevance and importance for several reasons. First, CTT offers a relatively simple and intuitive analysis of response characteristics for an assessment. Even if the goal is to utilize more contemporary methods of measurement, CTT provides an initial framework of analyses to explore data; its relatively simple approach augments data diagnostic efforts. Second, CTT follows a less rigorous set of assumptions than the more complex IRT approach to measurement. It can be easily be applied to a wide variety of testing situations. Third, CTT requires fewer data demands for scaling procedures. Fourth, CTT extends from a framework of computations that are simpler in nature; variance, covariance, and correlation statistics lay the groundwork for CTT. Thus, almost any statistical software or data management program can be employed for most CTT analyses.