Sources of and effects of interexpert correlation: an empirical study

Abstract

Expert estimates are relied on as sources of data whenever experimental data are lacking, such as in risk analysis. Correlation between experts poses problems in the aggregation of multiple estimates. The sources and structure of interexpert correlation are identified, and some of the ramifications are discussed. Expert estimates were found to correlate to the experts' problem-solving techniques and not to any features of their professional backgrounds. The paths that experts used to reach solutions included algorithms and assumptions. Both assumptions and algorithms were found to correlate to the answers because they are integral parts of the path. In addition, correctness of assumptions was found to correlate to the accuracy of the answers.< <ETX xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">&gt;</ETX>