Recommendation or evaluation? Task sensitivity in information source selection

doi:10.1086/323730

Return

Recommendation or evaluation? Task sensitivity in information source selection

2001-12-01

PRE

DOI：10.1086/323730

Original request for help

Save

Abstract

En 中文

Ironically, when consumers turn to an information source to assist in decision making, they are faced with the added responsibility of having to make a decision about the information source itself. A normative model is presented that shows that consumer assessment of an information source should include both the task for which an information source is sought and the appropriate probability of success for the task. A key distinction between tasks is whether or not the consumer knows the agents' (or prospective agents') ratings of the alternative in question at the time of his or her assessment of the information source. A series of experiments is presented that examines consumer assessment of the diagnosticity of one type of information source, consumer agents (e.g., movie critics and stock analysts) across the three common tasks of seeking recommendations, seeking evaluations, and choosing between agents who have provided conflicting advice. Results show that consumers frequently select inferior agents for providing recommendations and choose product alternatives that should be avoided because of a failure to recognize when a task calls for a conditional rather than overall assessment of agent prior performance. A final study attempts to isolate the reasons for these shortcomings by examining the process underlying consumer diagnosticity assessment of information sources. Implications and future research to address and improve consumer assessment of information sources across tasks are discussed.

Keywords:

NATURAL SAMPLE SPACES

BASE RATES

CONDITIONAL PROBABILITIES

PREFERENCE HETEROGENEITY

IRRELEVANT INFORMATION

COVARIATION ASSESSMENT

BAYESIAN-INFERENCE

CONSUMER-BEHAVIOR

DECISION-MAKING

JUDGMENTS

AI Summary

Key information extracted from the uploaded paper, including a brief overview, abstract, background, key highlights, visual analysis, and future outlook.