Originally posted by Lloyd_Christmas decentering best evaluated with max aperture which for this lens is f2.8???
Misalignments are in my opinion easiest spotted when shooting wide open, yes.
This can be done with a brickwall test (or a similar subject) as have already been suggested, the difficulty thereby is to perfectly align the camera with the subject though.
Because of that I find it more convenient to use the horizon (preferably with some trees or other details in it)for a quick test; place it in the centre of the image -> focus manually (with LV; on a tripod) to exclude focus errors -> take an image -> rearrange the horizon (but leave the focus untouched) to the top and bottom of the image and take 2 more images.
Compare the image quality of the edges and borders of the 3 images -> I don't know this lens from first hand experience, but there might be some field curvature visible; slightly blurry edges might be normal wide open, therefore you are looking for edges that are more blurry (and have more chromatic aberrations) than the others.
Decentering is also a bigger problem with wide angle lenses, therefore I'd suggest doing this at least for the 16mm focal range, but additional tests for 35mm and 50mm might also be of interest.