Approach-induced biases in human information sampling

Abstract

2Information sampling is often biased towards seeking evidence that confirms one's prior beliefs. Despite such biases being a pervasive feature of human behavior, their underlying causes remain unclear. Many accounts of these biases appeal to limitations of human hypothesis testing and cognition, de facto evoking notions of bounded rationality, but neglect more basic aspects of behavioral control. Here we demonstrate involvement of Pavlovian approach biases in determining which information humans will choose to sample. We collected a large novel dataset from 32,445 human subjects, making over 3 million decisions, who played a gambling task designed to measure the latent causes and extent of information-sampling biases. We identified three novel approachrelated biases, formalized by comparing subject behavior to a dynamic programming model of optimal information gathering. These biases reflected the amount of information sampled ('positive evidence approach'), the selection of which information to sample ('sampling the favorite'), and the interaction between information sampling and subsequent choices ('rejecting unsampled options'). The prevalence of all three biases was related to a Pavlovian approachavoid parameter quantified within an entirely independent economic decision task. Our large dataset also revealed that individual differences in information seeking are a stable trait across multiple gameplays, and can be related to demographic measures including age and educational attainment. As well as revealing limitations in cognitive processing, our findings suggest information sampling biases reflect the expression of primitive, yet potentially ecologically adaptive, behavioral repertoires. One such behavior is sampling from options that will eventually be chosen, even when other sources of information are more pertinent for guiding future action.

Approach-Induced Biases in Human Information Sampling

“…This confirms our previous observations concerning the validity of behavioral data acquired via smartphone [25]. We make the large behavioral dataset freely available for download [31], providing an empirical testing ground for models of human information seeking.…”

“…Moreover, subjects’ approach-induced biases in information sampling can be readily parameterized within this framework. We anticipate that further, more refined models will be subsequently tested by downloading the raw behavioral data from Dryad [31]. …”

“…Raw data, along with MATLAB scripts for reproducing all figures shown in the paper and code for the dynamic programming model, are freely available for download from http://dx.doi.org/10.5061/dryad.nb41c [31]. …”

Learning how to behave: Cognitive learning processes account for asymmetries in adaptation to social norms

“…This model demonstrated two ways in which the competitive norm was learned faster and was harder to forget, compared to other norms. First, learning rates associated with active behaviours were consistently higher than those for avoidance behaviour, in line with non-social learning findings(42,47,51). In addition, behaviours that carried a competitive intent, e.g., omission of reward or inflicting a punishment, were more readily transferred to other players.…”

“…Punishments and rewards are processed by different neural mechanisms (45), and can have different effects on learning (46). In a similar manner, active and passive actions are perceived and processed differently (42,47). Such asymmetries can give rise to different biases that shape human decision making (48), and can lead to different psychiatric conditions (49).…”