A method for determining the porosity of a subsurface geological formation traversed by a borehole is provided. The method generally comprises obtaining a means which relates ranges of apparent formation porosity (.phi..sub.a) as determined by a neutron-neutron log, formation fluid salinities, formation total capture cross sections (.SIGMA.), formation matrix and fluid constituents, and true formation porosity (.phi..sub.T) according to a predetermined equation. The predetermined equation relates the apparent porosity to a function of a modified migration length which is obtained in a semi-empirical manner (i.e. physics modifed by data). The modified migration length includes a slowing down length and a diffusion length, but causes at least the diffusion length to be a function of the slowing down length. The means which relates the apparent porosity to salinity, .SIGMA., matrix and fluid constituents, and .phi..sub.T solves a forward problem. Thus, the inverse problem of determining true porosity may be solved by obtaining determinations of salinity, formation matrix and fluid constituents, .SIGMA., and .phi..sub.a, and using the relating means to provide the true porosity. The provided method properly and separately accounts for the effects of capture cross section and salinity, and can be used to correct the apparent porosity determination of a neutron-neutron tool, regardless of the formation matrix or formation fluid constituents.