actually said:

The accuracy of the level application on my one phone is out bt over
7 degrees. That is a simple "level" app..

This is the first indication of what's possible out of a mobile device in
this thread, so I thank you for figuring out that your level app has an
accuracy of plus or minus 7 degrees (if I understood you correctly).

How did you find that out though?

The "rigid" level
application on my Blackberry PlayBook is very accurate - How you
reference it to the wheel will be the biggest variable that can
through your accuracy off. A trammel type setup made from a straight
bar of metal (or "straight" hardwood) with 2 screws protruding to
reach the edge of the rim, adjusted to be identical in protrusion,
will transfer the wheel angle accurately to the "level". You can
determine if the rim is true to the spindle by checking the level with
the bar upright with the wheel turned 180 degrees to make sure the
reading is the same with the wheel turned.

I agree that the jig attached to the wheel has to be exactly on target
(within the stated accuracies, all of which add up).

Here's an example of a camber jig for home use:
http://i.cubeupload.com/XocXQ9.jpg

Here's an even better camber jig setup for home use:
http://i.cubeupload.com/J0UuYd.png

WITH CARE you can check your camber to a reasonably high level of
accuracy. To get the camber "normalized" you need to roll the car
back and forth a few feet so the car "settles" on it's suspension.

This is good practical advice that you need to both roll the car back and
forth to let it settle on the suspension, and you need to add slip plates
under the wheels so that they slip nicely when adjusted.

A
professional setup uses a "slip plate" that allows the wheels to slide
in and out with little resistance. Normal procedure is to bounce the
car on the slip plates to "normalize" the suspension.

This is also good advice to bounce the car and to use slip plates for
measuring and adjusting toe so that the wheels move freely.
http://www.gnttype.org/techarea/susp...ics/align3.jpg

I've done it long pre-smart-phone using a simple bubble level to
verify the alignment was "close enough" afterr an accident in central
Africa severely damaged the front of my Peugeot.

The really good news is that, like in your case, a simple bubble level
might suffice simply because a decent rear camber spec is zero degrees
anyway, which is the easiest angle to measure.
http://i.cubeupload.com/J0UuYd.png

In summary, what I've learned in the past day are a few things:

1. A practical value for rear camber is 0 degrees to a smidge negative
2. A practical value for toe-in is 0 inches to a smidge positive (inward)

Both those are so close to zero that I can check that they are zero, and
then I can tweak them to a "smidge" inward.

But that is a different problem from checking them, which seems to be
easily doable using a few common tools based on my googling today:
http://www.tomhoppe.com/index.php/20...-camber-gauge/