In academic education peer review is seen as an effective educational approach to support learning. However, in practice peer reviews often remain superficial and lack critical depth. Therefore in many courses it is common practice to grade peer review assignments. Since grading based on quality of the peer review is time-consuming from a teachers perspective, grading is often based on timely delivering and the number of reviews. In this paper we describe an innovative approach to increase both the participation and quality of peer review and online discussions. In short, this approach involves grading only the 'best contribution' students' provide to the peer review. These 'best contributions' are self-selected by the student from his/her peer review(s). 'Best Contribution Grading' (BCG) aims to 1) Increase the quality of peer review, 2) Increase participation, 3) Offer a safe learning environment, 4) Deliver a scalable teaching method. The principle of BCG is evaluated in four academic courses at Wageningen University using a cross-over design. The effect of BCG is expressed in terms of student satisfaction, quality of peer review, participation grade, and textual analytics of peer reviews. During the academic year 2018/2019 the evaluation experiments are conducted. Results are expected in summer 2019.