Social-Behavioral Modeling for Complex Systems

40 Lessons on Decision Aiding for Social‐Behavioral Modeling

Paul K. Davis

RAND Corporation and Pardee RAND Graduate School, Santa Monica, CA, 90407, USA

One important function of social‐behavioral modeling (SBM) should be decision aiding (Davis et al. 2018; Davis and O'Mahony 2019), but for that function social‐behavioral (SB) models will need to have characteristics that are currently unusual. In this chapter, I discuss those characteristics by drawing on experience in other domains in which model‐based decision aiding has been important. The focus is primarily on informing strategic‐level planning.

Strategic Planning Is Not About Simply Predicting and Acting

A common image of decision aiding is that leaders have a well‐posed problem to which they need a solution. In this image, various model‐based analyses predict which option would provide the best results, i.e. which option would be optimal. Classic methods include decision analysis and linear programming. The models employed are assumed to be mathematically sound and the data accurate, although perhaps expressed in statistical terms. The methods may be so good that the solution found can be accepted with confidence. If uncertainty analysis is necessary, it can be simple sensitivity analysis. For example, models may predict heavy snow and ice tomorrow with 20% probability, thereby aiding a city's decision‐makers as they contemplate whether to have emergency storm groups on duty.

Strategic decision‐making ...