| Abstract: | Many investigations concerned with human decision making have centered their attention on the issues of cue integration and cue importance. Unfortunately, an equally important issue—that of cue relevance—has been either neglected or equated with the importance issue. Two experiments tested the ability of information integration theory (N. H. Anderson, 1981, Foundations of information integration theory, New York: Academic Press) to disentangle cue relevance and cue importance formally as well as empirically. The decision environment was divided into two components: aspects of the information present during a decision (information context) and aspects of the situation surrounding a decision (situation context). It was hypothesized that variations in information context would affect cue importance while changes in situation context would affect cue relevance. Furthermore, it was hypothesized that the weight parameter (w) of information integration theory would reflect the information context manipulations but not situation context manipulations. Results supported previous findings showing the influence of information context on w. As predicted situation context did not affect w; however, it did affect the variability of the responses to a cue's values. The relationship between perceived variability and situation context suggests that the scale dispersion parameter σ of information integration theory may be used as a measure of relevance (K. L. Norman, 1980, Organizational Behavior and Human Performance, 25, 289–310). |
