So here is a discussion I am having with a colleague. I put the references in alpha order and assigned the first Martin, 2007 "a" and the second "b" as you describe above. My colleague assigned "a" to the first Martin, 2007 that appeared in the text of the paper and "b" to the second. She told me that it didn't make sense to have (Martin, 2007b) come before (Martin, 2007a) in the text of the paper even though I cited b before I cited a. Your thoughts?