Information filters : learn from personal and social experience, or how subjectivity explains "the" learning mechanism

People learn and make decisions by selectively gathering, processing and communicating information – a complex chain of actions prone to bias. In doing so, people do not only interact directly with the environment to gather information, but also with other individuals. Indeed, social context allow more efficient learning and decision-making by mitigating individual limitations. In this thesis, personal and social learning are viewed as two complementary levels of information filtering – a recursive process, the base case of which being individual (subjective) experience via sampling. Identity representing subjectivity in a learning process is a filtering parameter on both levels. A well-established phenomenon in risky decision-making – Description-Experience Gap – was used to explore the difference of personal and social learning. Firstly, newly developed decisions-from-observation and decisions-from-description experimental paradigms showed that decision-making by social and personal experience are similar learning processes, with communication tending to decrease personal experience bias by over-reporting it. Secondly, decisions-from-description paradigm with either present or absent social source showed that people assume social source even when it is absent. Lastly, identity (mis)alignment between social source and receiver showed to affect evaluation of the source, but not the evaluation of information they delivered. Overall, by comparing and combining personal and social learning strategies, and by performing Bayesian meta-analysis, this thesis shows that personal and social levels of information filtering are closely related with identity playing a mediating role. The experimental results also suggest that personal and social information filters are two different perspectives constituting a single complementary process necessary for efficient decision-making