Ways of Knowing Compassion: How Do We Come to Know, Understand, and Measure Compassion When We See It?

Abstract

Over the last decade, empirical research on compassion has burgeoned in the biomedical, clinical, translational, and foundational sciences. Increasingly sophisticated understandings and measures of compassion continue to emerge from the abundance of multidisciplinary and cross-disciplinary studies. Naturally, the diversity of research methods and theoretical frameworks employed presents a significant challenge to consensus and synthesis of this knowledge. To bring the empirical findings of separate and sometimes siloed disciplines into conversation with one another requires an examination of their disparate assumptions about what compassion is and how it can be known. Here, we present an integrated theoretical review of methodologies used in the empirical study of compassion. Our goal is to highlight the distinguishing features of each of these ways of knowing compassion, as well as the strengths and limitations of applying them to specific research questions. We hope this will provide useful tools for selecting methods that are tailored to explicit objectives (methods matching), taking advantage of methodological complementarity across disciplines (methods mixing), and incorporating the empirical study of compassion into fields in which it may be missing.

Keywords: altruism; compassion; compassion meditation; empathy; methods; phenomenology.

FIGURE 1

Mapping the ways of knowing compassion. This figure maps the major methodologies reviewed here into theoretical spaces. The shape of the methodology denotes frame of reference. Color represents the extent to which that method has ecological validity. Positioning on the x-axis corresponds to the extent to which a method measures internal versus external aspects of compassion. Positioning on the y-axis corresponds to whether the methodology is generally used to measure state or trait compassion or is used to measure both. Methods on the line between state and trait can be used to measure both, depending on the specifics of the methodology.