CombiStats v7 will automatically invoke the Spearman-Kaerber method in cases where it is impossible to estimate the slope from the data. Here is an example where the options are set to use the probit model, but because the slope cannot be estimated, the Spearman-Kaerber method is used instead. The output shows a message to inform the user of this fact.

Note that the Spearman-Kaerber method requires that the doses are equidistant. It is the responsibility of the user to make sure that this is indeed the case. If the doses are not equidistant, CombiStats will use the smallest distance (d) between adjacent doses giving unequal responses. The common slope of the rectangular curves is then fixed at 1/ln(d).