We describe a consulting project in which the statisticians found that
they needed to develop a new method of analysis in order to reach val
id conclusions about the effect of a herpes vaccine on the immune syst
ems of human subjects. This method estimates the frequency of blood ce
lls responding to a viral antigen at a single, carefully chosen diluti
on. The traditional analysis of such data uses a cutoff to separate ex
perimental sites (''wells'') which contain no responding cells from we
lls which contain at least one responding cell, whereas our method use
s the scintillation count to estimate the actual number of responding
cells for each well. We describe the experiment in some detail, as we
found that one of the major challenges facing the statisticians was th
e need to understand the biology in enough detail to provide a relevan
t model.