Cognitive processes in self-report responses: tests of item context effects in work attitude measures

Much applied research relies on multi-item, self-report instruments. Drawing from recent cognitive theories, it was hypothesized that the items preceding a self-report item, its item context, can generate cognitive carryover and prompt context-consistent responses. These hypotheses were tested in 2 investigations: a field experiment involving 431 employees of a nonprofit urban hospital and a laboratory replication involving 245 undergraduate business students who held full- or part-time jobs. In both studies, evaluatively neutral items were placed in specially arranged blocks of uniformly positive, uniformly negative, or randomly mixed items on 3 modified Job Descriptive Index scales. Responses to the neutral items differed across the 3 forms, but scale-level psychometric properties remained unchanged. The implications of these item- and scale-level results for a variety of self-report measures in organizations are discussed.