The US Environmental Protection Agency recommends on one of their websites that –if in doubt- experimental partition constants are to be preferred over predicted ones. Do you agree?
Download this page as a pdf