Essays on the wisdom of crowd

Abstract

This dissertation consists of three essays. These essays together attempt to investigate on how to use different decision rules and aggregation mechanisms to improve collective judgment accuracy in wisdom of crowd applications. The first chapter focusses on the application area of open innovation and compares the performance of commonly used decision rules (scoring and ranking rules) for idea evaluation & selection. The results suggest that the crowd wisdom extracted using the scoring rule outperforms the crowd wisdom derived using ranking rule in terms of the likelihood of selecting the highest-quality ideas. From a managerial perspective, this study provides guidance for the choice of idea evaluation process and thereby the design of open innovation initiatives. The second and the third chapters focus on the application area of judgmental forecasting and investigate an important context, that of forecasting environments characterized by structural breaks. The second chapter demonstrates the presence of systematic biases (under and over forecasting) in forecast judgments under the presence of structural breaks. Further, this chapter proposes a novel and effective asymmetric trimming rule of forecast aggregation and prescribes two forecast ensemble methods to leverage such aggregated forecasts and improve forecast performance. Chapter 3 further tests the performance of the proposed forecast ensemble methods under extended conditions and suggests improvements to one of the proposed ensemble methods. Overall, the dissertation aims to unfold decision rules and aggregation heuristics by demonstrating potential cognitive limitations and biases of the decision maker and further, prescribes aggregation mechanisms to improve decision quality.