I found this new study soon to be published via this tweet and it is more than interesting:

So Michael Schneider & Franzis Preckel did a great job analyzing 91 meta-analyses and I think their conclusion is a very clear summary:

This study is the first systematic and comprehensive review of the meta-analyses on the variables associated with achievement in higher education published in the international research literature. The results are generally compatible with those of similar reviews for school learning or for learning in general. They offer evidence-based arguments for current debates in higher education. Instructional methods and the way they are implemented on the micro-level are substantially associated with achievement in higher education. This emphasizes the importance of teacher training in higher education. Among the different approaches to teaching, social interaction had the highest frequency of high positive effect sizes. Lectures, small-group learning, and project-based learning all have positive associations with achievement provided they balance teacher-centered with student-centered instructional elements. As yet, instruction and communication technology have comparably weak effect sizes, which did not increase over the past decades. High-achieving students in higher education are characterized by qualities that, in part, are affected by prior school education, for example, prior achievement, self-efficacy, intelligence, and the goal-directed use of learning strategies. Thus, universities indirectly benefit from a high instructional quality in K– 12 schools. The application of our findings into practice is a complex process that requires the collaboration of practitioners, educational researchers, and policy-makers. The effect sizes of the meta-analyses included in this systematic review indicate that such an evidence-based approach has great potential for increasing achievement in higher education.