Yes, every screening and matching is done against the fulltext of each document in the collection.