Social participation is the main outcome of physical rehabilitation programs. The aim of this study is to improve the measurement of social participation, using an instrument called the Assessment of Life Habits Scale and the Rasch model. The interval level measurement, the dimensionality and the generalizability of the item hierarchy were verified. The data from a large sample of people with spinal cord injury was analyzed and specific results were compared with expert opinions. The main properties of the instrument were satisfactory and the agreement with expert opinion was high. Principal component analysis showed multidimensionality. The item difficulty hierarchy obtained with spinal cord injury experts was different from the one obtained with traumatic brain injury experts, indicating a different difficulty level of items in relation to each population characteristics. We conclude that the instrument is appropriate for the measurement of social participation and suggest ways to improve the instrument.