Several methods for determining persistence length from size-exclusion chromatography data are evaluated for a stiff polymer, poly(n-hexyl isocyanate) (PHIC), and the more flexible poly[2,7-(9,9 di-n-hexylfluorene)]. The molar mass dependence of the root-mean-square radius obtained from light scattering detection is used to calculate the persistence length of PHIC, based on the Kratky-Porod wormlike chain model. The persistence length estimate is in reasonable agreement with literature values, and the results are relatively insensitive to the sample concentration. Directly solving for persistence length by non-linear regression is more suitable than linear approximations that use ratios of root-mean-square radius and molar mass. The persistence length calculated for the same polymer from viscometry detection, and the Yamakawa-Fujii-Yoshizaki hydrodynamic cylinder model for wormlike chains, is reasonable only when the viscosities measured by viscometry detection, and the molar masses measured by light scattering detection, are extrapolated to zero concentration values. The concentration effects are not significant for the more flexible PHF, and viscometry data are suitable for determining persistence lengths, even by linear approximation methods that use the ratio of intrinsic viscosity and molar mass. Molar masses calculated from a universal calibration curve and the viscometry detector are shown to be erroneous for the freely draining PHIC and should not be used for persistence length determination. Other methods that examine selected regions of viscosity conformation plots are shown to be of limited utility. (c) 2005 Elsevier Ltd. All rights reserved.