Accuracy and relevancy of probit analysis for in vivo dose-response laser tissue experiments

The most common method of analysis for 'sin vivo laser tissue experiments" is the probit regression. The data gathered at these experiments are specific in that there are very few repetitions of the exact stimulus exposure; thus the response frequencies for most stimulus are either '0' or '1'. Though such type of data is acceptable in probit, it seems that such data might not produce robust estimates of the ED50 and the slope. The accuracy of the probit's estimation was investigated by the use of Monte-Carlo simulation. Preliminary results suggest that the accuracy of the probifs estimations is conditional and might be biased in a way that raise doubts about the validity ofthe conclusions based on probit's estimations.