This paper presents two studies examining the dimensionality and construct validity of the union commitment scale (Gordon, Philpot, Burt, Thompson & Spiller, 1980). In the first study, analysis of the 20-item (Friedman & Harvey, 1986) and 30-item (Cordon et al., 1980) versions of the scale in two large samples (N = 229 and N = 551) suggests that (a) both versions of the scale share a common factor structure and (b) the interpretation of the belief in unionism dimension is confounded with the use of negatively worded items. We propose a shorter 13-item scale measuring union loyality, responsibility to the union and willingness to work for the union. In the second study (N = 847), the dimensionality and construct validity of the shorter scale is examined and supported.