In micellar liquid chromatography, MLC, the hydrophobicity of a compound is the predominant effect on its retention and interaction with micelles. The capacity factors at zero micellar concentration, k(m), and the solute-micelle association constants, K-AM, have recently been used as the hydrophobicity index of compounds and are important in QSAR studies. These parameters could be estimated (by regression) from the (k,[M]) data, where k is the capacity factor and [M] the surfactant concentration minus the critical micelle concentration. k(m) and K-AM are usually obtained from the intercept and slope, respectively, of the plot 1/k vs, [M]. In spite of the general use of this equation, the reliability of the estimates has not received special attention. Two new regression models have been proposed: a nonlinear model (plot k vs. [M]) and a new linear model (k vs. k . [M]). The reliability of the k(m) and K-AM estimates is studied with the aid of simulated data, and the conclusions obtained from the theoretical study are contrasted with experimental data, The influence of variables like the hydrophobicity of compounds, K-AM/k(m) ratio, the uncertainty of experimental k data and the experimental design have been considered, Some critical aspects of the models and recommendations for their use are provided. (C) 1997 Elsevier Science B.V.