Subaperture stitching interferometry is widely used in large-aperture plane surface measurements. The reference surface error forms a primary reason leading to erroneous stitching surface figures. The methods to remove this error include the normal stitching algorithm with the absolute test and a special algorithm with interlocked compensators without the absolute test. Here, we perform a series of subaperture stitching experiments to demonstrate that the reference surface error can easily be calculated wrongly using the algorithm with interlocked compensators. Consequently, we propose the use of a sensitivity coefficient to describe the sensitivity of the stitching algorithms of the two methods. Next, we perform simulations of pairwise subaperture stitching to illustrate the variations in the sensitivity coefficient with different overlapping area ratios. Finally, two plane surfaces with size of 200 mm × 200 mm are stitched with the reference surface error as reduced by the two methods. Comparisons of the residual surface errors of the two methods show that the algorithm using interlocked compensators is less accurate and less stable, which could be monitored by the sensitivity coefficient. Thus, for reducing the reference surface error in plane subaperture stitching interferometry, the absolute test method should be considered first.