A Human-Computer Interaction Approach for Integrity in Economics

Emerging data science platforms using simplified and automated user interfaces can help research become significantly more transparent and ethical. By depending on standard human-generated code, many statistical software programs commonly used in economics and the social sciences inadvertently rely on the human willpower of scientists, and inspite of an assumed invincibility, such individuals are nearly necessarily prone to errors and research integrity compromises, as is increasingly clear. Removing the vast majority of arbitrary and subjective data judgments, including the generation of code, from researcher control would free behavioural and social scientists from human limitations. Automating the text annotations that accompany data visualizations in figures and diagrams using emerging natural language processing tools can also free scientists from overconfidence or the temptation to embellish findings. Scientific communities across economics as well as other social science fields should embrace such systems to enhance the integrity and transparency of the next-generation of research.