This research project, funded by the Ethics and Values Studies component of the Science and Society Program, examines the role of values in computational modeling, how these values are shaped by professional and organizational cultures, and the effects of values on the success of computational models as products. Computational models are used every day in a wide variety of products and policies. Such models are often treated as pure mathematical abstractions with no values content or ethical implications. However, these models are not pure abstractions, but rather they are only as good as their inputs, including the values of computational modelers. The research builds on a field study of a corporate research laboratory funded by NSF. The first major finding of the exploratory study was that the design and use of computational models is affected by both embedded and contextual values. The second major finding was that organizational culture plays a key role in influencing the computational modeling process. The third major finding was that values help to shape the success of models as products. Further research at additional field sites with different organizational cultures is necessary to determine the general validity of the findings and to examine the relationship between organizational culture and the values of computational modeling. The first field study was successful in demonstrating the validity and effectiveness of the method, and generated new research questions that require additional research. Based on the research objectives and the findings of the first field study, three hypotheses guide the research: 1) embedded and contextual values play important roles in all of the stages of the modeling process; 2) values embedded in and contextual to the design and use of computational models are tied to organizational and professional cultures; and 3) values embedded in and contextual to the design and use of computational models influence the success not only of the computational modeling process but also of computational models as products. The two remaining field sites for this research include an academic laboratory and a government laboratory. The research method, including interviews, surveys, participant observation, and focus groups, has been validated and refined through the first field study. Data collection and analysis will include three field site visits over a one-year period for each site. This project will lead to important products, including conference presentations, manuscripts for peer-reviewed journals, and a book manuscript. This study demonstrates the importance of values in computational modeling. In particular, the finding that both embedded and contextual values play an important role in the design and use of computational models is a novel and important finding for the literatures on the ethics of modeling, the ethics of computing, and value-sensitive design that merits further investigation. The study also demonstrates the influence of organizational and professional cultures on values in computational modeling. This research generates a better understanding of how values impact the success of the modeling process and of models as products. The findings of this study will improve practice by computer scientists by demonstrating the importance of ethical choices related to values in computational modeling. The findings will inform managers of computational modelers about the important role that values play in the success of computational models. This study also demonstrates the importance of professional and organizational codes of ethics. Managers need to be responsible in making and enforcing ethical standards within organizations, while modelers should be encouraged to follow both professional and organizational codes of ethics. In addition, the research contributes to ethics education for computational modelers and their managers. The findings of this study can be used for general classes in computing and information ethics as well as specific courses and training in professional ethics for computational modelers.
