Recent trends in language assessment emphasize moving away from traditional testing to learning oriented assessment approach. This implies that teaching, learning, and assessment are intertwined processes. That is, teachers would assume more responsibility towards assessment than before, which in turn, would require a good command of language assessment knowledge on the part of teachers. That is probably why teachers’ language assessment knowledge (LAK) has been the focus of research in the last few years (Taylor, 2009, 2013; Fulcher 2012, Inbar 2016). However, despite its importance, research findings reveal that EFL teachers do not receive adequate training on language assessment in teacher education programs (Inbar-Lourie, 2008; Malone, 2008; Farhady &Tavassoli 2017, 2018).
The purpose of this paper is to report on the findings of LAK research in Iran and Turkey and make a couple of arguments. First, I would like to argue that consistent and cross contextual findings of research on insufficiency of EFL teachers’ LAK may not reflect the reality of teachers’ LAK. All might have been due to the framework suggested some thirty years ago (AFT, NCME, & NEA, 1990), and upon which most LAK tests have been developed, no longer reflects the reality of assessment today. Second, following Brookhart (2011) and Gotch and French (2014), I would like to argue that the present definition of LAK, which is based on summative normative principles, may not be construct valid anymore. We may need to ask the classic question of what it mean to have an adequate LAK. Following these arguments, suggestions for redefining the construct and setting standards for LAK will be made.