This paper describes a dynamic model of backmixing in solvent extraction columns which allows for the relatively low speed of the back flow compared to the main forward flow. It shows how a procedure based on transfer function matching that fully exploits the symbolic manipulation capabilities of scientific computing software can be used to determine the parameters in this model. The performance of this procedure is demonstrated by applying it to the tracer pulse data for a plate solvent extraction column. The values of these parameters thus obtained are compared with published values and their significance is briefly discussed. (c) 2005 Elsevier Ltd. All rights reserved.