Rating elements via partial orders

Question

I would like to rate a set of $n$ elements, with each element assigned a rating from ${0,\dots, 10}$.

The way in which I want to rate is by repeatedly selecting subsets of $k$ elements and querying a user to rank them relative to each other. I would like a means of minimizing the number of necessary queries to assign ratings (i.e. "how do I pick which elements I should ask about?"), and a way of aggregating my partial orders into the appropriate buckets ("when do I stop, and how do I map the partial order into my range?").

Are there any decent references I should be looking at?