Crowdsourcing for Engineering Design: Theoretical and Experimental Studies

Crowdsourcing is a practice of getting ideas and solving problems using a large number of people on the Internet. The outcomes of crowdsourcing contests depend on the decisions and actions of the participants. For an effective use of crowdsourcing within engineering design, there is a need to understand participants\u27 decisions in crowdsourcing contests. Towards addressing this need, the objective of this thesis is {\it to understand the individual decision making in design processes, and the interactive decision making in crowdsourcing contests}. To achieve this objective, theoretical models are used in conjunction with behavioral experiments and analysis of field data. The first study of interactive decision making evaluates game-theoretic models of contests, and field data from GrabCAD challenges. Game-theoretic models provide a theoretical basis for predicting outcomes from rational interactive decisions, while field data allows testing of these predictions. Since existing game-theoretic models do not account for various aspects of design problem and process-related factors in contests, the second study analyzes individual decision making in design. The second study establishes theoretical models of information acquisition in design based on decision theory, and observes actual decisions of student subjects using a controlled experiment. These models capture learning and uncertainty of outcomes in information acquisition process, and predict rational decisions. Controlled experimentation allows observation of the effects of problem-related factors on individuals\u27 actual decisions. In results from the first study, game-theoretic models show limited applicability to the field data due to invalidity of certain assumptions and unobservability of factors such as effort and cost. Predictions from these models, that are tested to be true on the field data, suggest that following factors affect participation or quality outcomes positively: (i) higher prize amount, (ii) higher prize allocation to top prizes, (iii) ideation type problems over expertise-based problems. Additional insights, observed from the analysis of the field data but not modeled in game theoretic models, show that participation increases with larger number of prizes and lower task complexity. In results from the second study, problem-related factors, task complexity and design cost, are observed to affect individuals\u27 decisions in design. The results of the analysis show that higher design cost negatively influences information acquisition as individuals search less, while higher task complexity motivates information acquisition as individuals search more