After reading the critique, I updated some of the arrows in my concept map on the article.
Two other critiques deal with the lack of a quantification of the relationships and a lack of inclusion of time in the relations. I found these critiques slightly less poignant. The article never claims to quantify relations and although I estimate that a time factor is important (learners may change their conceptions of learning over time), within the framework of one module (a few months) it looks fairly constant to me.
Finally, Richardson points out that the relationship between learning approach and attainment is not so straightforward as one might intuitively expect. Learners who apply surface learning tend to do poorly on exams (strong relation), but there is only a weak and variable relation between deep learning and doing well on exams. The relationship may be clouded by the role and type of assessment that is used.
In Cambodia we assume that student-centered learning will lead to better learning and better exam results. Although nation-wide assessment procedures are outside our scope, they may influence how students learn. The article and the subsequent discussion call for caution for both relationships.