Assessment is an ongoing instructional process aiming at understanding and improving student learning. Bias reporting occurs when there is a mismatch between students’ perceived knowledge and the actual knowledge. Learning occurs when students realize their bias reporting and take steps to realign their knowledge structure. Unfortunately, providing feedback on bias reporting is not feasible under normal educational setting. Therefore, we propose a new feedback type, feedback on knowledge state, as an alternative. This study develops a Computer Adaptive Assessment Software (CAAS) for multiple-choice items using Number Right Elimination Testing (NRET) method to provide feedback on knowledge state. This study also gauges teacher and students’ perceptions toward feedback on knowledge state and explores how CAAS can help 449 students and 19 teachers to learn. The results show that teachers and students perceived feedback on knowledge state as useful. It helped students learn by giving positive and negative reinforcement. In addition, it also provoked thinking leading students to check their own thinking.