When parents need not have children — Cognitive biases in information modeling

Abstract

Cognitive biases associated with human judgment and choice are widely studied, recognized, and documented in behavioral decision research. It is also well accepted that an understanding and acknowledgment of these biases are vital to mitigate their effects. However, research into cognitive biases in information modeling is virtually lacking. Lest one assumes that research on cognitive biases is irrelevant to the field, information modeling is a cognitively intensive activity and is, thus, highly susceptible to such biases. There is a pressing need then, to identify and understand these human biases in order to lessen their effects. This paper describes an experiment designed to investigate the use of syntactic and semantic information by modeling experts. The experimental results indicate that when interpreting information models, modeling experts tend to focus on the syntactic aspects of information and totally ignore the semantic information, even in situations where the semantic information is clearly more representative of the real world situation. These biases exhibited by modeling experts are explained using the learning paradigm in cognitive psychology.