•Novice meditators can be trained to report their phenomenological experience.•Self-reports are associated with history of practice, not with desirable responding.•Balance between focused attention (FA) and open monitoring (OM) practices predicts the ability to dissociate them.•Daily practice of FA predicts the felt stability of long sessions.•Total practice of OM predicts phenomenological proficiency. Empirical descriptions of the phenomenology of meditation states rely on practitioners’ ability to provide accurate information on their experience. We present a meditation training protocol that was designed to equip naive participants with a theoretical background and experiential knowledge that would enable them to share their experience. Subsequently, novices carried on with daily practice during several weeks before participating in experiments. Using a neurophenomenological experiment designed to explore two different meditation states (focused attention and open monitoring), we found that self-reported phenomenological ratings (i) were sensitive to meditation states, (ii) reflected meditation dose and fatigue effects, and (iii) correlated with behavioral measures (variability of response time). Each of these effects was better predicted by features of participants’ daily practice than by desirable responding. Our results provide evidence that novice practitioners can reliably report their experience along phenomenological dimensions and warrant the future investigation of this training protocol with a longitudinal design.