Abstract
Randomized errors to varying extents have been introduced in a constructed data set stimulating receptor binding saturation analyses. The parameters Bmax and Kd were then determined using six different evaluation techniques, four linear and two non-linear. It is concluded that a non-linear evaluation gives more reliable results, especially when errors are large. Furthermore, an inevitable correlation between Bmax and Kd regardless of evaluation technique results in possible misinterpretations of simultaneous changes in both parameters.