In [37] we drew the tentative conclusion that designing a digital tutee with low self-efficacy would be a good choice since the results suggested that students with low self-efficacy benefitted from interacting with a digital tutee with low rather than high self-efficacy. At the same time, ...