“…Rigorous discussion or application of validation procedures is not uncommon, 36,54,55,68,106 but has not led to consistency of practice or development of methodology. In the models we reviewed, five main approaches to model validation or evaluation were apparent: checking modeled behavior against actors' stated or revealed behavior 36,50,53,54,57,67,68,87 (comparisons of the stated and revealed behavior 25 were not used); extensive sensitivity analyses; 36,39 comparison (or tuning to ensure agreement) of some modeled and measured outcomes; 37,39,50,54,56,60,70,77,84,85 comparison to the results of other models; 88 and reliance on previous validation of the same or a similar model 64,71,74,82,84 (Figure 2). Justification of the (usually single) selected approach is rare, as is discussion of implications for model usage and interpretation.…”