This is a paper presented at the Education University of Hong Kong in December 2017. Published abstractAssessment in the form of testing and examination has long dominated educational practice with substantial consequences attached to high quality performance (e.g.,graduation,promotion,certification,scholarship,etc.). It has also been long presumed that assessments are accurate indicators of quality in a student’s learning, a teacher’s instruction, and an institution’s education.However, increasing emphasis has been put on the use of additional and alternative assessment methods and on using assessments formatively to improve student learning, teacher instruction, and institutional outcomes.Yet, evidence shows us that these goals are difficult to attain largely because of human and social factors and conditions that interact with our best efforts at high-quality assessment. Although good assessment can offer some protection from charlatans, it turns out our testing and our interactive assessments are unsurprisingly error-prone. Nonetheless, an open society persists by pursuing insights from assessments carefully and with attention to the plausibility that we could be wrong about educational processes and outcomes.