Andrej Košir ^* Dejan Zagar ^* Matija Svetina ^* Marko Perkovic Franc Dimc Tanja Brcko Satler

• University of Ljubljana, Ljubljana, Slovenia

This article examines data derived from physiological responses measured during simulation exercises involving marine pilots during port approach procedures, but focuses on faults in statistical reliance rather than the particulars of what the studies set out to determine. While it is worth noting that pilots make one of the best sets of subjects for simulation of maritime challenges, they are also the most difficult to gather for such a study given the nature of their work, its stress, responsibilities and difficult conditions. The study focuses on interpreting the results arising from postprocessing errors during conventional statistical signal processing, particularly emphasizing the mean.The study explores unintentional misinterpretations while analysing dominant characteristic curves in complex maritime environments. As ships grow, tasks such as port entry become more intricate, amplifying the cognitive load on crew and pilots responsible for safe navigation. The study scrutinizes the impact of conventional statistical signal processing, with a specific emphasis on meancentred approaches, on the interpretation of outcomes during port approach procedures. The paper critically analyses responses in a group of experienced marine pilots based on simulated data and an anomaly likelihood approach. Highlighting the limitations of standardization in summarizing data, we question the generalizability of the S curve during a typical conventional signal processing.We generate a large dataset of signals, and one signal with a dominant characteristic suggests an intentional imbalance to simulate scenarios where most signals lack a characteristic shape. The deliberate imbalance reflects the distribution of real-world data, where specific patterns or shapes may be rare compared to more common patterns. A study revealed that while the average response of experienced pilots exhibited an approximate S shape, indicating a psychophysiological reaction preceding stress, detailed post hoc analysis showed that one participant primarily influenced this pattern. This challenges the generalization of the S curve as a typical psychophysiological response to stress, prompting a discussion on uncritical data interpretation. The research contributes to the broader discourse on data interpretation, urging a re-evaluation of conventional statistical signal processing methods to enhance the reliability of conclusions drawn from dominant characteristic curves.