[Falsely positive values in multi-channel analysis: An iniquiry into reference and patient groups (author's transl)].

The number of falsely positive values occurring in 12-channel analysis was determined in two groups of patients and reference individuals. It revealed that the portion of falsely positive values actually found was statistically significant beyond that calculated on the assumption of a binomial distribution. Partly distinct correlations of the parameters combined to a profile as well as clear deviations from the normal distribution have to be taken into consideration as reasons for this discrepancy between theory and reality. The results show that the application of the binomial distribution leads to statements which significantly differ from the conditions actually present.