Several mechanisms, dependent on the laser radiation and tissue parameters, affect laser tissue interaction. A common in vivo dose response experiment contains, for every parameter value, a few subjects, which are exposed to several doses. These experiments are commonly analyzed using probit models where for every parameter value, the measured data from all subjects are unified to be a single statistical batch. The analysis is then performed under the assumption that the batch responses are statistically independent and that they all originate from the same distribution. However, it is well known that even individuals, that have many common characteristics, may respond differently to equal drug dose or identical medical treatment. Thus, it is reasonable to assume that unaccounted subjects' dissimilarity may affect and bias the results of the statistical analysis. As a part of our study, on the stochastic characteristics of laser (ocular) tissue interactions, we reanalyzed the raw data of the U.S. Army Medical Research Detachment Walter Reed Army Institute of Research (USAMRD-WRAIR) spectral ED50's experiments. Some effects of specimen dissimilarity on the results of the experiment analysis will be presented and discussed.