The ''informative'' approach, an optimum design methodology, has been
applied to two empirical models of solvents, that of Koppel and Palm a
nd that of Kamlet, Abboud and Taft with the purpose to determine if th
ey are identical, similar or very different. The experimental results
necessary to establish the corresponding models are the keto/enol equi
librium constants of 2,4-pentanedione carefully determined by H-1-NMR
spectroscopy. The conclusion is that both models are only similar, bei
ng one more precise (KP) and the other more robust (KAT).