That´s an interesting topic, mainly because lot of people in my field (neurodegenerative science) are performing maze and do not always check visual acuity of animal.
In our lab we perform always SHIRPA test before starting the first run of behavior (http://empress.har.mrc.ac.uk/browser/?sop_id=10_002_0). By the way, they have a nice article on sensory system and vision on it.
We also use visual cliff, that you can find on this article
This isn't my area of expertise, but a few things: 1. reaching reflex (part of a SHIRPA or Irwin screen; very gross measure) - hold animal by tail and lower toward the edge of a cage or countertop - watch vibrissae carefully and determine if animal reaches for edge prior to vibrissae touching. 2. Visual cliff test. 3. I've also seen people measure visually evoked ERPs in response to light flashes at different angles relative to eyes.