Dilution experiment where at each dilution there are 3 samples and within these there are an additional 10 sub-samples. Measurements on these samples are taken, but for values 40 and above, the result is censored.
The main question of interest is to provide a probability the next observation from a particular sample will be censored given one observation from that sample.
Given the data we have, we can construct a curve for the relationships between the standard deviation and mean in the original data. From the consultant, it appears the relationship is positive.
Since we only have one observation for the main question of interest, the value for this observation is our best guess about the mean for that sample and the standard deviation will be determined by the mean-sd relationship in the previous paragraph. Given the mean and standard deviation, you can calculate a probability a normal with that mean and sd will be above the censoring threshold.
blog comments powered by Disqus