Evaluating the reliability of measurement instruments is common practice in organizational psychology. When measurement proceeds using composites of multiple indicators, reliability is often examined with internal consistency estimates. This is common practice for both measurement instruments with a reflective measurement model, where a latent psychological construct is assumed to cause the item scores, and with a formative measurement model, where the composite score defines the construct. In the latter case, however, internal consistency indices such as Coefficient Alpha or Omega are not informative about reliability. In this paper, we review the assumption of reflective measurement and discuss why internal consistency estimates assume this model. We then provide an illustration of a checklist (originally proposed by Jarvis et al., 2003) intended to aid researchers, reviewers and editors in recognizing reflective measurement. Finally, the paper describes how researchers can estimate the reliability of instruments that use formative items. With these contributions, the paper aims to promote awareness of measurement assumptions when estimating reliability, and to provide tailored alternatives for assessing the reliability of formative instruments.