In this paper we present preliminary results on a study on the effect of instructional method (observational learning and learning by doing) and reflection (yes or no) on academic text quality and self-efficacy beliefs. 56 undergraduate students were assigned to either an observational learning or learning-by-doing condition, with or without reflection. The participants were tested on academic text quality, self-efficacy beliefs and satisfaction with the instructional method. Our preliminary results suggest that there is no main effect of instructional method on academic text quality and self-efficacy beliefs. Observational learning and learning by doing seem equally effective for learning how to write a large and complex writing assignment, even though participants who learned by doing were more satisfied with the method than participants who learned by observing models. We found no effect of reflection. However, there appears to be an interaction between instructional method and reflection: in observational learning reflection seems to have a negative effect on academic text quality while in learning by doing it appears to have a positive effect. We will repeat the experiment in September 2016. [For full proceedings, see ED571332.]