I use Biggs' solo taxonomy for as many marking criteria as I can get away with. I find that writing a detailing marking scheme can never fairly reflect student performance. When you write it you always have in mind the way a student is going to answer the question but you never manage to cover the full diversity of responses. I found that when I have been marking the statistics assignments I wanted to change the marking scheme as some students had done well but not managed to address the specific points of the criteria. This is always a problem when you have large classes and the marking has to be shared out. You want to make the marking as transparent and objective as possible but in the end this often leads to less fair marks than a more generic and subjective set of criteria.
That is why I like Biggs as I align it to the grade boundaries:
Pre-structural - Level 4 (fail pass at level 6).
Unistructural - Level 5 (3rd at level 6)
Multi-structural - Level 6 2:2
Relational - Level 6 2:1
Extended abstract - 1st - MSc Distinction.
Then you can fill in the evidence as you read the work and you are not thinking there is 1 mark for this and 2 for that ...