Consciousness research has been plagued by a lack of conceptual clarity and confusion between types of data and measures, which affects the effects found, so that it is not uncommon to find conflicting effects for the same phenomenon. Therefore, it is essential to develop methodologies that address these problems and experimental designs that allow for the collection of the necessary data to apply them. In this area, visual perception, and in particular perceptual grouping, has been in the spotlight, with the question of whether local forms group into global forms in the absence of awareness, with various effects being found. To investigate these grouping effects in the absence of awareness, the present study examines whether a masked stimulus is processed unconsciously by combining different designs of a masked priming paradigm and both objective and subjective measures of awareness collected in separate blocks (offline measures) or trial-by-trial (online measures). Two novel methods are used to investigate whether primes are processed unconsciously or not: a correction of Greenwald's regression using a Bayesian generative model, and sensitivity versus awareness curves based on general recognition theory. Both methods avoid the post hoc selection of subjects and trials and account for several methodological problems such as regression to the mean, which can lead to misinterpretation of the results. Our results indicate a certain dependence between perceptual processing and level of awareness. Interestingly, the use of different paradigms to assess unconscious processing and the resulting theoretical assumptions may lead to somewhat divergent results.