We describe here a simple non-parametric method for the evaluation of data obtained from primed lymphocyte typing (PLT) experiments. The method is designed to compensate for both differences in the PLT responsiveness and stimulatory capacity by a double normalization procedure based on a representative negative reaction (the median) for each PLT-cell and for each secondary stimulator. The index of response is called the Normalized Median Response (NMR). The NMR-method was developed from data from one of a series of PLT experiments aimed at HLA-D typing. The NMR-method was applied to eight other PLT experiments and was compared to two other calculation methods based on the maximal response of each PLT-cell: (i) a method without compensation for stimulator variation, and (ii) a method including non-parametric compensation for stimulator variation. The eight experiments involved 37 different PLT cells tested against 74 different stimulating cells yielding a total of 1,334 secondary combinations. The NMR-method descriminated significantly better than the other methods when the results were compared to the HLA-D types obtained with homozygous typing cell technique. In particular, the number of extra positive reactions was diminished by the NMR-method. The reproducibility and the sources of variation were studied in the eight experiments and in five additional experiments. The NMR-method reduced the combined priming-to-priming variation and day-to-day variation to a level comparable to the variation between duplicate testings of the same PLT cell tested towards the same set of secondary stimulators in the same secondary PLT experiment.