RASCH MEASUREMENT

Tesio (2014) responded with a very thoughtful article in RMT 28:1 to two articles by Stenner, Burdick and Stone that appeared in RMT, 22:1 (2008) and 22:4 (2009), which quite aptly, he referred to as “enlightening”. The articles by these authors were concerned with distinguishing between what they referred to as index variables (also referred to as formative variables), and causal variables (also referred to as reflective). Essentially, as summarized by these authors, index variables are defined by their indices, whereas causal variables define or generate indices. Stenner, Burdick and Stone stressed that, although responses to a collection of items may fit the unidimensional Rasch model, their fit does not tell whether or not the variables are index or causal and that this distinction can only come fro