Breaking continuous flash suppression (b-CFS; Stein, 2019) is a reaction-time-based measure that has been established as an important tool in the study of consciousness and stimulus detection. In recent years, the method gained popularity and insights in the domain of person perception. Still, there are only few studies investigating the theoretical and statistical underpinnings of the procedure. Past re-analyses of b-CFS studies indicate that the results and conclusions vary if you change even small aspects of the data pre-processing and analysis pipeline. In the present talk, we will draw on 101 studies employing b-CFS to (1) highlight between-study variability in pre-processing and analysis of reaction time data and (2) show that data transformation and outlier trimming massively influence results and conclusions across multiple datasets and for Frequentist, Bayesian and meta-analytic analysis approaches. Finally, we will discuss several implications for researchers employing b-CFS and other reaction-time-based measures to promote theoretically grounded, replicable, and open data pre-processing and analysis.