There are numerous test-specimen configurations for determining the static shear modulus of composites. Selection of specimen configuration depends primarily on product component configuration. Two-ring specimens in use are the split-ring twist specimen and the four-point-loading ring-twist specimen. The latter is more advantageous, since it uses an intact ring. However, reported validation tests on aluminum rings showed a large error in predicted shear modulus. The present paper provides an improved analysis that takes into account the tilting of the cross section on the loading bar.