Many calibrations of conventional time domain reflectometry (TDR) have been done with the apparent electrical permittivity as the independent variable and water content as the dependent variable (or vice versa). Since the apparent permittivity is calculated as a function of the TDR pulse two-way travel time (t_t) along a waveguide of length, L, this is practically the same as calibrating in terms of (t_t/2L). Recent studies have confirmed that the travel time increases not only with water content, but also with increasing bulk electrical conductivity (BEC). Also, the apparent permittivity (and thus travel time) increases as the effective frequency, F_e, of the signal decreases (eg., as cable length increases). We conducted calibrations for TDR derived water content in three soils (17%, 30%, and 48% by mass of smectitic and montmorillonitic clays) in terms of travel time, BEC, and F_e. Slopes on the travel time term varied significantly when only t_t was considered, and calibrations in the two more clayey soils did not account for temperature induced variation in reported water content. Slopes varied even more when both t_t and BEC were included. When all three properties were considered, slope differences on the travel time term were minimized and temperature effects on reported water contents were practically eliminated. All three values (t_t, BEC, and F_e) may be measured from TDR waveforms, allowing correction for temperature effects without separate measurements of temperature.