typically, a 'method comparison' approach is sufficient to determine if a one site is equivalent - or not - to a second site. Check out my paper on measurement system verification and validation in the resources section of this site. (when viewing 'new posts' the resources appear on the right side of the screen...)
Without more details of what you are measuring and what the measurement system is, I would recommend selecting material at the low and high end of the specification (both in spec and out), probably 15 parts at each of the limits for a total of 30 parts. have the old site measure them twice and have the new site measure the same parts twice.
Without more details of what you are measuring and what the measurement system is, I would recommend selecting material at the low and high end of the specification (both in spec and out), probably 15 parts at each of the limits for a total of 30 parts. have the old site measure them twice and have the new site measure the same parts twice.