Spin-orbit splitting energy may be proportional to |dV/dr| at the electron's position. The potential V of an atom is different among different valency although it is very small. We must note that another interaction between (2p spin) and (3d spin), which is smaller but still observable. If we observe 2p3/2 and 2p1/2 peaks of Na3VO4(s=0), VOSO4(s=1/2), V2O3(s=1), then we expect Na3VO4 should be zero spin-spin interaction, but experimentally 2p1/2 is broadest in spite that 2p3/2 is narrowest amongthe three vanadium compounds as mentioned in my book: "X-Ray Spectroecopy for Chemical State Analysis", DOI: 10.1007/978-981-19-7361-1_3
The statement is correct, but the nature is deeper than this statement.