Validating Argument-Based Opinion Dynamics with Survey Experiments

The empirical validation of models remains one of the most important challenges in opinion dynamics. In this contribution, we report on recent developments on combining data from a survey experiment with an argument-based computational model of opinion formation in which biased processing is the principle mechanism. We first review the development of argument-based models, and extend a model with confirmation bias by noise mimicking an external source of balanced information. We then study the behavior of this extended model to characterize the macroscopic opinion distributions that emerge from the process. A new method for the automated classification of model outcomes is presented. In the final part of the paper, we describe and apply a multi-level validation approach using the micro and the macro data gathered in the survey experiment. We revisit previous results on the micro-level calibration using data on argument-induced opinion change, and show that the extended model matches surveyed opinion distributions in a specific region in the parameter space. The estimated strength of biased processing given the macro data is highly compatible with those values that achieve high likelihood at the micro level. The model provides a solid bridge from the micro processes of individual attitude change to macro level opinion distributions.