Tisane: Authoring Statistical Models via Formal Reasoning from Conceptual and Data Relationships

Policy makers rely on models to track disease, inform health recommendations, and allocate resources. Scientists use models to develop, evaluate, and compare theories. Journalists report on new findings in science, which individuals use to make decisions that impact their nutrition, finances, and other aspects of their lives. Faulty statistical models can lead to spurious estimations of disease spread, findings that do not generalize or reproduce, and a misinformed public. The challenge in developing accurate statistical models lies not in a lack of access to mathematical tools, of which there are many (e.g., R [63], Python [52], SPSS [58], and SAS [24]), but in accurately applying them in conjunction with domain theory, data collection, and statistical knowledge [26, 38]. There is a mismatch between the interfaces existing statistical tools provide and the needs of analysts, especially those who have domain knowledge but lack deep statistical expertise (e.g., many researchers). Current tools separate reasoning about domain theory, study design, and statistical models, but analysts need to reason about all three together in order to author accurate models [26].