Self-report mental health assessments are the primary tool for measuring mental health in research contexts. These assessments have many limitations, in large part because there is still much we have not adequately formalized about the basic nature of mental health and mental illness, and how to improve the former and prevent and treat the latter. This paper outlines a new approach for developing measurement tools in mental health research. We describe principles that are essential for the process of authentically trying to understand the nature of mental health problems suffered by humans. In this novel framework, psychological measurement is considered a meta-psychology problem: how can we use our understanding of the mind to facilitate access (i.e. measure) the inner world of human beings? We propose three principles that are fundamental: (1) reverence for the data, (2) openness to discovery, and (3) compassionate listening (ROC). These principles are necessary components for creating a safe psychological context, which is analogous to having a “clean petri dish”, for psychological data to be captured. Without these principles in play, the research context will contain psychological contaminants that inhibit authentic sharing of information in assessment tools. These psychological contaminants, including shame, avoidance/denial of painful information, fear of honesty, and mistrust of research, are sources of noise that directly corrupt the integrity of psychological assessment data. The proposed three principles are a way of creating safe containers that can receive information that wants to be known. Implementing these principles in research contexts has the potential to generate cleaner psychological data in mental health research, which can globally increase observed effect sizes across the field of psychology and generate scientific insights that alleviate psychological suffering.