Given the influence of the human factor on decision-making processes by aviation operators of the “leading edge”, therefore, — of the flight safety, individual and group preference systems as formalized representations of air traffic controllers about a number of spectrum ordered by danger have been investigated from n=21 of characteristic errors. That has a pronounced positive proactive character. Explicating their opinions, the air traffic controllers simultaneously form discriminating, memorizing skills, therefore, — avoidance of mistakes in a professional activity: the subjects made one third fewer errors in comparison with other air traffic controllers in the process of simulator training. Group preference systems allow identifying the features of the functioning of individual societies — dispatch shifts, possible group deformations, as well as the impact on their members of the presentation features of the performing technological procedures specifics by instructors. m=37 individual preference systems of the air traffic controllers, which were involved in research, were built by pairwise comparison of the danger of errors and the application of a differential method of distributing the indicator of their total danger. That contributed to conducting of 420 pairwise error hazard comparisons. The implementation of a multi-step procedure for identifying and screening out 10 marginal opinions has led to a statistically consistent group system of preferences: Kendall’s concordance coefficient equals is W=0,700 and it became statistically significant at a high level of significance of a = 1 %. The decision matrix was formed from mА=27 individual preference systems of the air traffic controllers, which, by the definition, is a “cost matrix” and for the solution of which a methodology for the correct application of the classical decision criteria by Wald, Savage, Bayes-Laplace, Hurwitz has been implemented. It revealed the identity of the group systems of preferences obtained by Wald and Savage criterion, as well as the Bayes-Laplace criterion and such a strategy of group decisions as summation and averaging of ranks. The empirical preferences are generally the same: Spearman’s rank correlation coefficients are unusually high (RB–L–W/S=0,8922, RB–L–HW=0,9263, RW/S–HW=0,9477) and statistically gullible at a high level of significance for human factor studies . The following values of the normative indicator of the not distinguishing dangers of error risk in the group preference systems are obtained: R*BL=0 , R*HW=0,19·10-2, R*W/S=5,58·10-2. The value of this indicator for the group is R*g=0,52·10-2.