Improving numerical measures of human feelings: The case of pain

Numerical self-report scales are extensively used in economics, psychology, and medicine to quantify subjective feelings, ranging from life satisfaction to the experience of pain. These scales are often criticized for lacking an objective foundation, and defended on the grounds of empirical performance. We focus on the case of pain measurement, where existing self-reported measures are the workhorse but known to be inaccurate and difficult to compare across individuals. We provide a new measure, inspired by standard economic elicitation methods, that quantifies the negative value of acute pain in monetary terms, making it comparable across individuals. In three preregistered studies (completed between June 7th to Sept 26th, 2022), 330 healthy participants were randomly allocated to receive either only a high- or only a low-pain stimulus or a high-pain stimulus after having double-blindly received a topical analgesic or a placebo. In all three studies, the new measure greatly outperformed the existing self-report scales at distinguishing whether participants were in the more or the less painful condition, as confirmed by effect sizes, Bayesian factor analyses, and regression-based predictions. We conclude that standard economic methods can greatly improve the measurement of experienced pain across individuals. The new measure can be used in experimental studies and random parallel-assignment clinical trials, and opens the door to potential improvements in pain management.