In order to assess general job satisfaction, researchers often use composites formed by summing scores from multiple satisfaction facets. The appropriateness of composites, however, is a matter of contention: some researchers have argued that composites provide an effective means of assessing general job satisfaction (e.g., Spector, American Journal of Community Psychology, 13, 693-713. 1985), whereas others have argued that composites are inappropriate and should be replaced with global job satisfaction scales (e.g., Ironson et al., Journal of Applied Psychology, 74, 193-200. 1989). To address this debate, we compared the construct validities of composite job satisfaction scales with those of global job satisfaction scales. We first created a nomological network that specifies a hypothetical pattern of relationships that exists between general job satisfaction and several external variables. Using this network as a guide, we conducted two studies (total N = 676) that examined the construct validities of four global scales and five composite scales. Within both studies, we found that composite satisfaction scales and global satisfaction scales generally yielded similar relationships with the external variables included in the nomological network. This suggests that the two types of scales are equally effective at assessing general job satisfaction. Researchers, therefore, should not summarily disregard previous studies that have used composite measures. These findings also suggest that it is appropriate to use composite measures in instances where global satisfaction data are unavailable, a situation that can occur when working with archival or organizational datasets.

为了评估总体工作满意度，研究人员经常使用通过对多个满意度方面的分数求和而形成的综合体。然而，复合材料的适当性是一个有争议的问题：一些研究人员认为，复合材料提供了一种评估总体工作满意度的有效手段（例如，Spector，《美国社区心理学期刊》，第13卷，第693-713页，1985年）。认为复合材料是不合适的，应该用全球工作满意度量表来代替（例如，Ironson等人，Journal of Applied Psychology，74，193-200。1989）。为了解决这一争论，我们将综合工作满意度量表的构造效度与全球工作满意度量表的构造效度进行了比较。我们首先创建了一个法理网络，该网络指定了一般工作满意度和几个外部变量之间存在的假设关系模型。以此网络为指南，我们进行了两项研究（总计N= 676），考察了四个全球规模和五个综合规模的建构效度。在两项研究中，我们发现综合满意度量表和总体满意度量表通常与法理网络中包含的外部变量产生相似的关系。这表明两种类型的量表在评估总体工作满意度方面同样有效。因此，研究人员不应总忽略先前使用综合措施的研究。这些发现还表明，在无法获得全球满意度数据的情况下（当使用档案数据或组织数据集时可能会发生这种情况），可以使用综合度量。