- Rich Timpone
- Yongwei Yang
In past work, the criteria of Truth, Beauty, and Justice have been leveraged to evaluate models (Lave and March 1993, Taber and Timpone 1996). Earlier, while relevant, Justice was seen as the least important of modeling considerations, but that is no longer the case. As the nature of data and computing power have opened new opportunities for the application of data and algorithms from public policy decision-making to technological advances like self-driving cars, the ethical considerations have become far more important in the work that researchers are doing. While a growing literature has been highlighting ethical concerns of Big Data, algorithms and artificial intelligence, we take a practical approach of reviewing how decisions throughout the research process can result in unintended consequences in practice. Building off Gawande’s (2009) approach of using checklists to reduce risks, we have developed an initial framework and set of checklist questions for researchers to consider the ethical implications of their analytic endeavors explicitly. While many aspects are considered those tied to Truth and accuracy, through our examples it will be seen that considering research design through the lens of Justice may lead to different research choices. These checklists include questions on the collection of data (Big Data and Survey; including sources and measurement), how it is modeled and finally issues of transparency. These issues are of growing importance for practitioners from academia to industry to government and will allow us to advance the intended goals of our scientific and practical endeavors while avoiding potential risks and pitfalls.