A Monte Carlo method is employed to estimate coefficient standard deviations and their confidence intervals for rate constant temperature equations. The method used, unlike parametric statistics, requires no assumptions about the distribution of errors or the covariances between pairs of parameters. in this way, biased estimates of coefficient confidence intervals are avoided, and therefore more reliable values for thermodynamic activation functions’ confidence intervals are obtained. The advantages and importance of using weighted regression in this context are also shown. The proposed procedure is applied to rate constants of the solvolysis of 2-iodo-2-methylpropane in methanol, between 0 and 55 degrees C. New rate constants are reported at 0.00 and 5.00 degrees C. A comparison is made with data derived from conventional weighted least squares regression.