On the meaning of formative measurement and how it differs from reflective measurement: Comment on Howell, Breivik, and Wilcox (2007).

D. Howell, E. Breivik, and J. B. Wilcox (2007) have presented an important and interesting analysis of formative measurement and have recommended that researchers abandon such an approach in favor of reflective measurement. The author agrees with their recommendations but disagrees with some of the bases for their conclusions. He suggests that although latent variables refer to mental states or mental events that have objective reality, to gain knowledge of the existence of these states or events requires that emphasis be placed on the nature and interpretation of the relationship between latent and manifest variables. This relationship is not a causal one but rather a kind of correspondence rule that contains theoretical, empirical, operational, and logical meanings as part of its content and structure. Implications of the above views are discussed for formative and reflective measurement.