Abstract
The embedded and contingent nature of classroom-based formative assessment means that validity in the norm-referenced, summative tradition cannot be understood in exactly the same way for formative assessment. In fact, some scholars (e.g., Gipps, Beyond testing: towards a theory of educational assessment, 1994, Falmer Press, London, UK) have even contended for an entirely different paradigm with an independent set of criteria for its evaluation. Many others have conceptualized the validity of formative assessment in different ways (e.g., Nichols et al., 2009, 28 (3), 14–23; Stobart, Validity in formative assessment, 2012, SAGE Publications Ltd, London, UK; Pellegrino et al., Educ. Psychol., 2016, 51 (1), 59–81). This article outlines a framework for evaluating the argument-based validity of CBFA. In particular, I use Kane (J. Educ. Meas., 2013, 50 (1), 1–73) as a starting point to map out the types of inferences made in CBFA (interpretation and use argument) and the structure of arguments for the validity of the inferences (validity argument). It is posited that a coherent and practical framework, together with its suggested list of inferences, warrants and backings, will help researchers evaluate the usefulness of CBFA. Teachers may find the framework useful in validating their own CBFA as well.