Effects of information sequence and irrelevant distractor information when using a computer-based decision aid

Decision Sciences, Winter 2001 by Reneau, J Hal, Blanthorne, Cindy

H1: Judgment accuracy will be influenced by the sequence of information presented. Accuracy will increase when more diagnostic (predictive) information is presented at the end rather than at the beginning of the information sequence.

Irrelevant Distractor Information

Statistical theory would suggest that more information is always better. That is, relevant information can only increase the accuracy of judgments. In the subsequent discussion we define relevance in an objective, statistical sense. Information cue x is said to be relevant if it is predictive of event z, and irrelevant if unrelated to z. Further, x is said to be more relevant than information cue y if x is more predictive of z than is y. When both x and y are available to a decision maker and they are correlated, we label x relevant and label y irrelevant, with the understanding that in natural settings relevance or irrelevance of an information cue is relative, not absolute. Finally, we use the terms relevant and diagnostic synonymously.

Consider a baseball fan whose favorite team has won all of its games in the first week of the season; the same fan subsequently observes the team lose all of its games in the second week of the season. According to statistical theory, both the accuracy and confidence the fan has in his judgment of the team's ability should be higher based on the combined data from both weeks than from either week alone. Yet, this may not be the case in practice where humans do not process information as intuitive statisticians.

A substantial portion of the judgment literature is concerned with issues of confirmation and disconfirmation in information processing. A "confirmation bias" in both the acquisition and interpretation of information is well documented (see Fischhoff & Beyth-Marom, 1983, for a review). Information acquisition is controlled by the structure of the previously described decision aid, however, information interpretation, remains in the human domain. Thus, interpretation of information received remains a subjective process.

The influence of normatively irrelevant information on interpretation may be explained by the belief perseverance hypothesis. To illustrate, Lord, Ross, and Lepper (1979) examined the effects of irrelevant information in experiments where subjects' initial beliefs were documented. Subjects differentially evaluated the quality and persuasiveness of the information and interpreted the irrelevant information as being supportive of their initial beliefs. Similarly, Ajzen, Dalto, and Blyth (1979) found that the effect of ambiguous information held whether the initial position was assigned by researchers or chosen by subjects. More recently, Russo, Medvec, and Meloy (1996) investigated the distortion of new information in the absence of an initial preference. They found that while developing a preference for one alternative, which they describe as the "weakest form of preference," subjects distorted new information so as to favor the leading alternative.