A new approach to the analysis of three body decays is presented. Model-independent results are obtained for the swave $Kpi$ amplitude as a function of $Kpi$ invariant mass. These are compared with results from $Kmpip$ elastic scattering, and the prediction of the Watson theorem, that the phase behavour be the same below $Keta^{prime}$ threshold, is tested. Contributions from $I=half$ and $I={3over 2}$ are not resolved in this study. If $I=half$ dominates, however, the Watson theorem does not describe these data well.}