The Occupational Depression Inventory (ODI) was recently developed to assess depressive symptoms that individuals specifically attribute to their work. One purpose of the ODI is to respond to limitations of current assessments of job-related distress, most notably, assessments relying on the burnout construct. In this study, we conducted a thorough examination of the psychometric and structural properties of the ODI using exploratory structural equation modelling bifactor analysis and Mokken scale analysis. The study involved three samples of employed individuals, recruited in France (N = 3454), Switzerland (N = 1971), and Australia (N = 1485). Results were consistent across the three samples. The ODI exhibited essential unidimensionality and excellent total‐score reliability―as indexed by McDonald's omega, Cronbach's alpha, Guttman's lambda‐2, and the Molenaar‐Sijtsma statistic. We found evidence for measurement invariance across sexes, age groups, and samples. Mokken scale analysis revealed that the ODI's scalability was strong. No monotonicity violation was detected. Invariant item ordering showed sufficient accuracy. In all three samples, suicidal ideation was the least commonly endorsed item―thus acting as a sentinel item―and fatigue/loss of energy was the most commonly endorsed item. The ODI exhibits excellent psychometric and structural properties, suggesting that occupational health specialists can effectively employ the instrument.