A new method is presented for analyzing reaction rate measurements to obtain cross sections. In the usual approach, a complete forward or slowing down calculation is required for each beam energy at which the reaction rate is to be obtained. The approach here uses an adjoint formulation, yielding reaction rate vs energy in a single pass, making the analysis easier to perform and the physical process more transparent. The accuracy of the approximations involved in the present approach is tested in two cases by comparison with rigorous Monte Carlo results. For certain conditions of sample thickness and cross section as shown in this paper, the usual trial and error procedure for finding cross sections that fit the measured reaction rates can be avoided. It is then possible to invert the reaction rates directly into cross sections. A test case is described in which this direct inversion process proved to be stable and accurate.