the application of a comprehensive model of communicative language ability (cla) to language teaching and testing has always been an imperative in l2 education since hymess proposal of communicative competence in the 1970s. recent l2 research has clearly underscored the importance of sufficient pragmatics representation as an essential component of cla in pedagogical and testing practices in l2 classrooms. motivated by this theoretical orientation, this study primarily drew on bachman’s (1990) and bachman and palmer’s (1996, 2010) cla model to explore the status of pragmatics representation in efl high school classrooms and the national university entrance test (nuet) in iran. further, the study, inspired by messick’s (1996) conceptualization of consequential validity, set out to scrutinize the (probable) impact of the high-stakes nuet on pragmatics construct-representation in efl classrooms and on high school students’ interlanguage pragmatic development in iran. the participants comprised a random sample of 100 efl teachers, 220 (pre-nuet and 3rd grade) high school students, and 50 undergraduate students. a triangular approach was employed using 3 researcher-made likert-type questionnaires, a classroom observation checklist to observe 28 high schools and nuet-preparatory classrooms, semi-structured interviews with the efl teachers and the high school students, and the nuet language test item examinations over the past 5 years. the results of descriptive statistics and chi-square tests indicated that pragmatics was seriously underrepresented in efl curriculum and instruction in iranian high schools, meaning that the efl classrooms did not fulfill their critical function of raising and developing learners’ pragmatic awareness and competence. further, pragmatic competence was underrepresented in the nuet, and that the nuet use negatively impacted pragmatics representation in the instructional materials and pedagogical activities in iranian efl classrooms. the findings suggest that the impact and consequences of high-stakes testing practice would not be beneficial to the stakeholders, the educational system, and the society unless the construct-underrepresentation and construct-irrelevant factors are removed or minimized in the tests.