Understanding, Predicting and Preventing bias (through cognitive and individual differences psychology)
Dr Matthew Welsh, Australian School of Petroleum and Energy Resources, University of Adelaide
June 30, 2021 (SDP Invited Talk)
Abstract: Decision biases are common and robust, affecting the majority of people in many situations and resisting debiasing. While biases can be limited with focussed interventions, general education and knowledge have little effect on susceptibility. In fact, despite seven decades of research, analyses of industry data show unchanged levels of bias and our ability to pre-select good decision makers is limited.
This suggests a failure in identifying root causes of bias, which may be aided by understanding a key division in scientific psychology – between cognitive psychology and psychometrics. The former examines how people, in general, think and includes most work on heuristics and biases and bounded rationality. The latter looks at individual differences between people on various traits and how these predict behaviour or performance. These streams are often separate, meaning individual differences work in cognitive fields (e.g., decision making) is rare and reliant on outdated or unreliable measures. For example, some researchers claim intelligence does not predict decision-making ability – focussing instead on ‘decision style’ measures. Examining the evidence, though, reveals these ‘intelligence’ measures to include, e.g., self-reported college entrance exam scores.
There is a need to bridge this gap, bringing appropriate psychometric measures to bear on bias research. In addition to improving our ability to predict bias susceptibility, linking biases to established psychometric traits can clarify the cognitive processes underlying bias – moving away from a ‘one-heuristic-per-bias’ model and towards a cogent taxonomy of decision making. As one example, hindsight bias and overconfidence are typically treated as unrelated. Consideration of underlying cognitive processes, however, suggests a common cause and, more specifically, that individual differences in memory abilities will predict susceptibility to both biases.
Click on the file below to hear a sample of the presentation.