In the past I've presented many comparisons between quantised inertia (MiHsC, horizon mechanics) and the emdrive thrust data, and here is another one (see the graph below) updated with the new NASA 2016 data (White et al., 2016). The ten comparisons between the thrust predicted by MiHsC/QI (this way) along the x axis, and the thrust experimentally observed, along the y axis, are shown by the open squares. Ideally, they should all lie on the diagonal line, and they are close to it, but you'll notice that the NASA thrusts (NASA2014 and NASA2016) are overestimated (too far to the right ) by anything up to a factor of ten (the scales are logarithmic), and Shawyer's first test (Shawyer1) is too far to the left.
I wondered (see the NASA shift) if it could be because NASA put a thin dielectric at the narrow end of their emdrives (Paul March confirmed they did), and Shawyer put a dielectric at the wide end of his first emdrive (Shawyer1). So I have now calculated the change this makes to MiHsC/QI since a dielectric changes the speed of light. The amended thrusts are shown on the plot with the black diamonds, and they lie closer to the diagonal 'correct' line, which is encouraging and some evidence that MiHsC/QI is the right explanation (but let's see if the reviewers of my new paper, to be submitted, agree with the way I've calculated the effect of the dielectric). Here is the same data in table format:
McCulloch, M.E., 2015. Testing quantised inertia on the emdrive. EPL, 111, 60005. Preprint
White, H., P. March, J. Lawrence, J. Vera, A. Sylvester, D. Brady, P. Bailey, 2016. Measurement of impulsive thrust from a closed rf cavity in vacuum. AIAA J. of Propulsion and Power.