The difference of 10% itself isn't much of an issue. What is important is making sure your test is repeatable no matter what system you are using. This means good clean impedance peaks with no bumps, blips, dips, etc. You should be able to put your woofer in the jig and run the same test over and over again without more than maybe 1% difference in any parameters. I did a post explaining some things about T/S measurement awhile ago that ou can find here.
I have heard talk about inaccurate measurements with older versions of the woofer tester, but I don't remember what parameter(s) were off. That said, I doubt its impedance measurements are off, I bet you could run impedance plots and calculate Vas yourself to be sure.
Vas (Really Cms) is sensitive to temperature, and also the suspension is nonlinear so measurement at different signal levels and /or temperatures can give you different values.