We describe the construction of a Job in General (JIG) scale, a global scale to accompany the facet scales of the Job Descriptive Index. We applied both traditional and item response theory procedures for item analysis to data from three large heterogeneous samples (N = 1,149, 3,566, and 4,490). Alpha was .91 and above for the resulting 18-item scale in successive samples. Convergent and discriminant validity and differential response to treatments were demonstrated. Global scales are contrasted with composite and with facet scales in psychological measurement. We show that global scales are not equivalent to summated facet scales. Both facet and global scales were useful in another organization (N = 648). Some principles are suggested for choosing specific (facet), composite, or global measures for practical and theoretical problems. The correlations between global and facet scales suggest that work may be the most important facet in relation to general job satisfaction.