The influence of task demands and experience on diagnostic accuracy: Investigating the assumptions of a default interventionist dual systems model

Abstract

<p>There are various dual process models of human cognition. While many models of cognitive control propose processes that are selected exclusively or in combination, a default-interventionist model of reasoning assumes that processing occurs in serial stages. System 1 processes are believed to recruit unconscious memory retrieval processes by default and precede System 2 processes (Evans & Stanovich, 2013; Kahneman, 2011). System 1 processes are also considered to be overly sensitive to the automatic influences of the environment and thereby also to various cognitive biases and errors; hence System 1 is inferior. On the other hand System 2, which represent conscious logic and normative reasoning processes, is not considered susceptible to such automatic influences and thereby capable of overriding errors made through System 1 reasoning; hence System 2 is superior. This default-interventionist model has become highly influential in theories about best practices in medical education (Croskerry, 2009; 2003; Klein, 2005; Redelmeier, 2005), and has encouraged a view that increased conscious processing and reflective thought will improve performance. Such a view is in stark contrast to models of human memory in psychology that suggest contextual or automatic influences of the environment are not only critical for learning, but also critical for adaptive processing and the development of expertise (Yonelinas, 2002; Larsen & Roediger, 2012). In this thesis I investigate and critique several assumptions of the default-interventionist model by testing the relationship between processing time, reflective thought, experience and accuracy. The results of two large studies do not support basic assumptions presented in the literature and instead demonstrate that experience and knowledge are better predictors of performance.</p>