The driver's brake behavior has a direct effect on the deceleration of the train in terms of strength and duration, which can affect passenger comfort, and in the long run also efficiency and energy consumption. Increased knowledge of these effects may be used to improve passenger comfort, save energy, and as an important input for rising technology developments such as Connected-Driver Advisory System (C-DAS) and Automatic Train Operation (ATO). To examine the effect of various deceleration levels on passenger comfort, a field study was conducted with a Swedish EMU of Bombardier Regina X52 type in regular traffic in northern Sweden. The train driver was braking according to a 2X2 factorial study design, altering both the main deceleration (until 20 km/h) and the deceleration to stop (from 10 – 0 km/h). Deceleration to stop includes two parameters, the deceleration and jerk effect. A high-performing GPS was used to measure deceleration and passengers were asked to rate the level of comfort on a scale from 0 (only comfortable) to 100 (very uncomfortable). The hypothesis was that both decelerations respectively affect the passengers' experienced comfort.  In total, there were 23 stops, and 91 passengers took part, rating various numbers of stops according to their traveled distance. According to the 2X2 design, the train driver altered the brake level.  The measured average levels were -0,76 m/s2 (high) and -0,46 m/s2 (low) for main deceleration. For deceleration to stop the average levels were -0,55 m/s2 (high) and -0,37 m/s2 (low).  A 2X2 between-subject factorial ANOVA with ETA- square as effect size was carried out using SPSS. This revealed a significant main effect of main deceleration, such that higher deceleration force led to higher ratings of discomfort, M= 23.79 (SD = 19.63) compared to M = 9.38 (SD = 13.00), F (1, 652) = 116.20, p < .01, 

η2p𝜂p2 = .151. There was also a significant main effect of the deceleration to stop, such that higher deceleration force led to higher ratings of discomfort (M = 17.81 (SD = 18.25) compared to M = 13.79 (SD = 17.22), F (1, 652) = 5.99, p = .015, 

η2p𝜂p2 = .009. There was no interaction effect.  From this study, and a somewhat unexpected, it can be concluded that the experienced comfort at all these various levels of deceleration was generally high. Further, both the main deceleration and the deceleration to stop, respectively, affect the passenger comfort. These results can be used as input for train-driver training and railway technology developments such as C-DAS and ATO.