The purpose of this study was to evaluate the psychometric properties of the Diabetes Self-Care Scale (DSCS). A convenience sample of 175 adults with diabetes who met the inclusion criteria from a local hospital in southern Louisiana participated in this study. Data from a pilot study with 50 respondents were also used to calibrate the instrument. The analysis tested the hypothesis that the difference between any respondent's agreement with an item and the difficulty presented by that item is equal to the natural logarithm of the odds of a response. Results indicated that respondent separation reliability is acceptable (.80) and item separation reliability is high (.99). All 35 items measure a single construct of diabetes self-care. Construct validity was supported by the meaningfulness of the item endorsability order and by the consistency of that order across respondents. Seven recommendations for modification of the instrument, future research, and practice are proposed.