What gets tested is what gets taught. Ever since Frederick Kelly, Dean of the College of Education at the University of Kansas, introduced the multiplechoice test in 1914, we’ve increasingly tested what’s easy to measure, not necessarily what matters. While a fine approach for many basic skills, it falls far short of facilitating the deeper learning demanded in the 21st century. When high stakes are added to these tests, they further narrow the focus of teaching and learning, place unproductive stress on educators, and diminish student engagement and motivation. These days, a growing chorus of parents, educators, and policymakers is voicing frustration and anger with top-down accountability and high-stakes testing. As members of two not-for-profit education organizations— one focused on assessment and the other on research and instructional practices—we find nothing wrong with testing itself; indeed, we believe evidence of what students know and can do should be at the heart of schooling. We are concerned, however, about what seems to be an almost myopic focus on high-stakes accountability based on tests of basic knowledge and skills to drive improvements in educational outcomes, to the exclusion of using formative and performance assessment to facilitate student growth and deeper learning (i.e., the ability to apply knowledge and skills in novel situations) (Pelligrino & Hilton, 2012).