The psychological construct of control is increasingly thought to be an important variable in psychosomatic medicine, but there are limitations with how it has been measured by previous instruments. The current study details four limitations of previous instruments as a rationale for developing a new multi-faceted, multi-dimensional instrument to measure control, the Shapiro Control Inventory (SCI). Discriminant, divergent, and incremental validity of the SCI is compared with the Rotter's Internal/External Locus of Control Scale and Wallstons' Multidimensional Health Locus of Control Scales on five groups, one normal and four clinical groups (depression, generalized anxiety disorder, borderline personality, and panic disorder). Analysis of variance showed significant differences among populations on each of the nine SCI scales. The majority of these results were highly significant, whereas for Rotter's scale the results were barely significant, and for Wallstons' three scales, not at all significant. Correlations of the SCI scales with Rotter's and Wallstons' scales were small to moderate, but in no case explained more than 15% of the variance. The SCI also provided incremental validity over Rotter's and Wallstons' scales both for sensitivity (clinical versus normal) and specificity (between clinical groups). The discussion section provides suggestions for future research.